diff options
Diffstat (limited to 'editors')
-rw-r--r-- | editors/Config.in | 196 | ||||
-rw-r--r-- | editors/Kbuild | 14 | ||||
-rw-r--r-- | editors/awk.c | 2919 | ||||
-rw-r--r-- | editors/cmp.c | 135 | ||||
-rw-r--r-- | editors/diff.c | 1344 | ||||
-rw-r--r-- | editors/ed.c | 1049 | ||||
-rw-r--r-- | editors/patch.c | 254 | ||||
-rw-r--r-- | editors/sed.c | 1349 | ||||
-rw-r--r-- | editors/sed1line.txt | 425 | ||||
-rw-r--r-- | editors/sed_summary.htm | 223 | ||||
-rw-r--r-- | editors/vi.c | 3954 |
11 files changed, 11862 insertions, 0 deletions
diff --git a/editors/Config.in b/editors/Config.in new file mode 100644 index 0000000..7dbc9b6 --- /dev/null +++ b/editors/Config.in @@ -0,0 +1,196 @@ +# +# For a description of the syntax of this configuration file, +# see scripts/kbuild/config-language.txt. +# + +menu "Editors" + +config AWK + bool "awk" + default n + help + Awk is used as a pattern scanning and processing language. This is + the BusyBox implementation of that programming language. + +config FEATURE_AWK_LIBM + bool "Enable math functions (requires libm)" + default n + depends on AWK + help + Enable math functions of the Awk programming language. + NOTE: This will require libm to be present for linking. + +config CMP + bool "cmp" + default n + help + cmp is used to compare two files and returns the result + to standard output. + +config DIFF + bool "diff" + default n + help + diff compares two files or directories and outputs the + differences between them in a form that can be given to + the patch command. + +config FEATURE_DIFF_BINARY + bool "Enable checks for binary files" + default y + depends on DIFF + help + This option enables support for checking for binary files + before a comparison is carried out. + +config FEATURE_DIFF_DIR + bool "Enable directory support" + default y + depends on DIFF + help + This option enables support for directory and subdirectory + comparison. + +config FEATURE_DIFF_MINIMAL + bool "Enable -d option to find smaller sets of changes" + default n + depends on DIFF + help + Enabling this option allows the use of -d to make diff + try hard to find the smallest possible set of changes. + +config ED + bool "ed" + default n + help + The original 1970's Unix text editor, from the days of teletypes. + Small, simple, evil. Part of SUSv3. If you're not already using + this, you don't need it. + +config PATCH + bool "patch" + default n + help + Apply a unified diff formatted patch. + +config SED + bool "sed" + default n + help + sed is used to perform text transformations on a file + or input from a pipeline. + +config VI + bool "vi" + default n + help + 'vi' is a text editor. More specifically, it is the One True + text editor <grin>. It does, however, have a rather steep + learning curve. If you are not already comfortable with 'vi' + you may wish to use something else. + +config FEATURE_VI_MAX_LEN + int "Maximum screen width in vi" + range 256 16384 + default 4096 + depends on VI + help + Contrary to what you may think, this is not eating much. + Make it smaller than 4k only if you are very limited on memory. + +config FEATURE_VI_8BIT + bool "Allow vi to display 8-bit chars (otherwise shows dots)" + default y + depends on VI + help + If your terminal can display characters with high bit set, + you may want to enable this. Note: vi is not Unicode-capable. + If your terminal combines several 8-bit bytes into one character + (as in Unicode mode), this will not work properly. + +config FEATURE_VI_COLON + bool "Enable \":\" colon commands (no \"ex\" mode)" + default y + depends on VI + help + Enable a limited set of colon commands for vi. This does not + provide an "ex" mode. + +config FEATURE_VI_YANKMARK + bool "Enable yank/put commands and mark cmds" + default y + depends on VI + help + This will enable you to use yank and put, as well as mark in + busybox vi. + +config FEATURE_VI_SEARCH + bool "Enable search and replace cmds" + default y + depends on VI + help + Select this if you wish to be able to do search and replace in + busybox vi. + +config FEATURE_VI_USE_SIGNALS + bool "Catch signals" + default y + depends on VI + help + Selecting this option will make busybox vi signal aware. This will + make busybox vi support SIGWINCH to deal with Window Changes, catch + Ctrl-Z and Ctrl-C and alarms. + +config FEATURE_VI_DOT_CMD + bool "Remember previous cmd and \".\" cmd" + default y + depends on VI + help + Make busybox vi remember the last command and be able to repeat it. + +config FEATURE_VI_READONLY + bool "Enable -R option and \"view\" mode" + default y + depends on VI + help + Enable the read-only command line option, which allows the user to + open a file in read-only mode. + +config FEATURE_VI_SETOPTS + bool "Enable set-able options, ai ic showmatch" + default y + depends on VI + help + Enable the editor to set some (ai, ic, showmatch) options. + +config FEATURE_VI_SET + bool "Support for :set" + default y + depends on VI + help + Support for ":set". + +config FEATURE_VI_WIN_RESIZE + bool "Handle window resize" + default y + depends on VI + help + Make busybox vi behave nicely with terminals that get resized. + +config FEATURE_VI_OPTIMIZE_CURSOR + bool "Optimize cursor movement" + default y + depends on VI + help + This will make the cursor movement faster, but requires more memory + and it makes the applet a tiny bit larger. + +config FEATURE_ALLOW_EXEC + bool "Allow vi and awk to execute shell commands" + default y + depends on VI || AWK + help + Enables vi and awk features which allows user to execute + shell commands (using system() C call). + +endmenu diff --git a/editors/Kbuild b/editors/Kbuild new file mode 100644 index 0000000..76302aa --- /dev/null +++ b/editors/Kbuild @@ -0,0 +1,14 @@ +# Makefile for busybox +# +# Copyright (C) 1999-2005 by Erik Andersen <andersen@codepoet.org> +# +# Licensed under the GPL v2, see the file LICENSE in this tarball. + +lib-y:= +lib-$(CONFIG_AWK) += awk.o +lib-$(CONFIG_CMP) += cmp.o +lib-$(CONFIG_DIFF) += diff.o +lib-$(CONFIG_ED) += ed.o +lib-$(CONFIG_PATCH) += patch.o +lib-$(CONFIG_SED) += sed.o +lib-$(CONFIG_VI) += vi.o diff --git a/editors/awk.c b/editors/awk.c new file mode 100644 index 0000000..64371f0 --- /dev/null +++ b/editors/awk.c @@ -0,0 +1,2919 @@ +/* vi: set sw=4 ts=4: */ +/* + * awk implementation for busybox + * + * Copyright (C) 2002 by Dmitry Zakharov <dmit@crp.bank.gov.ua> + * + * Licensed under the GPL v2 or later, see the file LICENSE in this tarball. + */ + +#include "libbb.h" +#include "xregex.h" +#include <math.h> + +/* This is a NOEXEC applet. Be very careful! */ + + +#define MAXVARFMT 240 +#define MINNVBLOCK 64 + +/* variable flags */ +#define VF_NUMBER 0x0001 /* 1 = primary type is number */ +#define VF_ARRAY 0x0002 /* 1 = it's an array */ + +#define VF_CACHED 0x0100 /* 1 = num/str value has cached str/num eq */ +#define VF_USER 0x0200 /* 1 = user input (may be numeric string) */ +#define VF_SPECIAL 0x0400 /* 1 = requires extra handling when changed */ +#define VF_WALK 0x0800 /* 1 = variable has alloc'd x.walker list */ +#define VF_FSTR 0x1000 /* 1 = var::string points to fstring buffer */ +#define VF_CHILD 0x2000 /* 1 = function arg; x.parent points to source */ +#define VF_DIRTY 0x4000 /* 1 = variable was set explicitly */ + +/* these flags are static, don't change them when value is changed */ +#define VF_DONTTOUCH (VF_ARRAY | VF_SPECIAL | VF_WALK | VF_CHILD | VF_DIRTY) + +/* Variable */ +typedef struct var_s { + unsigned type; /* flags */ + double number; + char *string; + union { + int aidx; /* func arg idx (for compilation stage) */ + struct xhash_s *array; /* array ptr */ + struct var_s *parent; /* for func args, ptr to actual parameter */ + char **walker; /* list of array elements (for..in) */ + } x; +} var; + +/* Node chain (pattern-action chain, BEGIN, END, function bodies) */ +typedef struct chain_s { + struct node_s *first; + struct node_s *last; + const char *programname; +} chain; + +/* Function */ +typedef struct func_s { + unsigned nargs; + struct chain_s body; +} func; + +/* I/O stream */ +typedef struct rstream_s { + FILE *F; + char *buffer; + int adv; + int size; + int pos; + smallint is_pipe; +} rstream; + +typedef struct hash_item_s { + union { + struct var_s v; /* variable/array hash */ + struct rstream_s rs; /* redirect streams hash */ + struct func_s f; /* functions hash */ + } data; + struct hash_item_s *next; /* next in chain */ + char name[1]; /* really it's longer */ +} hash_item; + +typedef struct xhash_s { + unsigned nel; /* num of elements */ + unsigned csize; /* current hash size */ + unsigned nprime; /* next hash size in PRIMES[] */ + unsigned glen; /* summary length of item names */ + struct hash_item_s **items; +} xhash; + +/* Tree node */ +typedef struct node_s { + uint32_t info; + unsigned lineno; + union { + struct node_s *n; + var *v; + int i; + char *s; + regex_t *re; + } l; + union { + struct node_s *n; + regex_t *ire; + func *f; + int argno; + } r; + union { + struct node_s *n; + } a; +} node; + +/* Block of temporary variables */ +typedef struct nvblock_s { + int size; + var *pos; + struct nvblock_s *prev; + struct nvblock_s *next; + var nv[0]; +} nvblock; + +typedef struct tsplitter_s { + node n; + regex_t re[2]; +} tsplitter; + +/* simple token classes */ +/* Order and hex values are very important!!! See next_token() */ +#define TC_SEQSTART 1 /* ( */ +#define TC_SEQTERM (1 << 1) /* ) */ +#define TC_REGEXP (1 << 2) /* /.../ */ +#define TC_OUTRDR (1 << 3) /* | > >> */ +#define TC_UOPPOST (1 << 4) /* unary postfix operator */ +#define TC_UOPPRE1 (1 << 5) /* unary prefix operator */ +#define TC_BINOPX (1 << 6) /* two-opnd operator */ +#define TC_IN (1 << 7) +#define TC_COMMA (1 << 8) +#define TC_PIPE (1 << 9) /* input redirection pipe */ +#define TC_UOPPRE2 (1 << 10) /* unary prefix operator */ +#define TC_ARRTERM (1 << 11) /* ] */ +#define TC_GRPSTART (1 << 12) /* { */ +#define TC_GRPTERM (1 << 13) /* } */ +#define TC_SEMICOL (1 << 14) +#define TC_NEWLINE (1 << 15) +#define TC_STATX (1 << 16) /* ctl statement (for, next...) */ +#define TC_WHILE (1 << 17) +#define TC_ELSE (1 << 18) +#define TC_BUILTIN (1 << 19) +#define TC_GETLINE (1 << 20) +#define TC_FUNCDECL (1 << 21) /* `function' `func' */ +#define TC_BEGIN (1 << 22) +#define TC_END (1 << 23) +#define TC_EOF (1 << 24) +#define TC_VARIABLE (1 << 25) +#define TC_ARRAY (1 << 26) +#define TC_FUNCTION (1 << 27) +#define TC_STRING (1 << 28) +#define TC_NUMBER (1 << 29) + +#define TC_UOPPRE (TC_UOPPRE1 | TC_UOPPRE2) + +/* combined token classes */ +#define TC_BINOP (TC_BINOPX | TC_COMMA | TC_PIPE | TC_IN) +#define TC_UNARYOP (TC_UOPPRE | TC_UOPPOST) +#define TC_OPERAND (TC_VARIABLE | TC_ARRAY | TC_FUNCTION \ + | TC_BUILTIN | TC_GETLINE | TC_SEQSTART | TC_STRING | TC_NUMBER) + +#define TC_STATEMNT (TC_STATX | TC_WHILE) +#define TC_OPTERM (TC_SEMICOL | TC_NEWLINE) + +/* word tokens, cannot mean something else if not expected */ +#define TC_WORD (TC_IN | TC_STATEMNT | TC_ELSE | TC_BUILTIN \ + | TC_GETLINE | TC_FUNCDECL | TC_BEGIN | TC_END) + +/* discard newlines after these */ +#define TC_NOTERM (TC_COMMA | TC_GRPSTART | TC_GRPTERM \ + | TC_BINOP | TC_OPTERM) + +/* what can expression begin with */ +#define TC_OPSEQ (TC_OPERAND | TC_UOPPRE | TC_REGEXP) +/* what can group begin with */ +#define TC_GRPSEQ (TC_OPSEQ | TC_OPTERM | TC_STATEMNT | TC_GRPSTART) + +/* if previous token class is CONCAT1 and next is CONCAT2, concatenation */ +/* operator is inserted between them */ +#define TC_CONCAT1 (TC_VARIABLE | TC_ARRTERM | TC_SEQTERM \ + | TC_STRING | TC_NUMBER | TC_UOPPOST) +#define TC_CONCAT2 (TC_OPERAND | TC_UOPPRE) + +#define OF_RES1 0x010000 +#define OF_RES2 0x020000 +#define OF_STR1 0x040000 +#define OF_STR2 0x080000 +#define OF_NUM1 0x100000 +#define OF_CHECKED 0x200000 + +/* combined operator flags */ +#define xx 0 +#define xV OF_RES2 +#define xS (OF_RES2 | OF_STR2) +#define Vx OF_RES1 +#define VV (OF_RES1 | OF_RES2) +#define Nx (OF_RES1 | OF_NUM1) +#define NV (OF_RES1 | OF_NUM1 | OF_RES2) +#define Sx (OF_RES1 | OF_STR1) +#define SV (OF_RES1 | OF_STR1 | OF_RES2) +#define SS (OF_RES1 | OF_STR1 | OF_RES2 | OF_STR2) + +#define OPCLSMASK 0xFF00 +#define OPNMASK 0x007F + +/* operator priority is a highest byte (even: r->l, odd: l->r grouping) + * For builtins it has different meaning: n n s3 s2 s1 v3 v2 v1, + * n - min. number of args, vN - resolve Nth arg to var, sN - resolve to string + */ +#define P(x) (x << 24) +#define PRIMASK 0x7F000000 +#define PRIMASK2 0x7E000000 + +/* Operation classes */ + +#define SHIFT_TIL_THIS 0x0600 +#define RECUR_FROM_THIS 0x1000 + +enum { + OC_DELETE = 0x0100, OC_EXEC = 0x0200, OC_NEWSOURCE = 0x0300, + OC_PRINT = 0x0400, OC_PRINTF = 0x0500, OC_WALKINIT = 0x0600, + + OC_BR = 0x0700, OC_BREAK = 0x0800, OC_CONTINUE = 0x0900, + OC_EXIT = 0x0a00, OC_NEXT = 0x0b00, OC_NEXTFILE = 0x0c00, + OC_TEST = 0x0d00, OC_WALKNEXT = 0x0e00, + + OC_BINARY = 0x1000, OC_BUILTIN = 0x1100, OC_COLON = 0x1200, + OC_COMMA = 0x1300, OC_COMPARE = 0x1400, OC_CONCAT = 0x1500, + OC_FBLTIN = 0x1600, OC_FIELD = 0x1700, OC_FNARG = 0x1800, + OC_FUNC = 0x1900, OC_GETLINE = 0x1a00, OC_IN = 0x1b00, + OC_LAND = 0x1c00, OC_LOR = 0x1d00, OC_MATCH = 0x1e00, + OC_MOVE = 0x1f00, OC_PGETLINE = 0x2000, OC_REGEXP = 0x2100, + OC_REPLACE = 0x2200, OC_RETURN = 0x2300, OC_SPRINTF = 0x2400, + OC_TERNARY = 0x2500, OC_UNARY = 0x2600, OC_VAR = 0x2700, + OC_DONE = 0x2800, + + ST_IF = 0x3000, ST_DO = 0x3100, ST_FOR = 0x3200, + ST_WHILE = 0x3300 +}; + +/* simple builtins */ +enum { + F_in, F_rn, F_co, F_ex, F_lg, F_si, F_sq, F_sr, + F_ti, F_le, F_sy, F_ff, F_cl +}; + +/* builtins */ +enum { + B_a2, B_ix, B_ma, B_sp, B_ss, B_ti, B_lo, B_up, + B_ge, B_gs, B_su, + B_an, B_co, B_ls, B_or, B_rs, B_xo, +}; + +/* tokens and their corresponding info values */ + +#define NTC "\377" /* switch to next token class (tc<<1) */ +#define NTCC '\377' + +#define OC_B OC_BUILTIN + +static const char tokenlist[] ALIGN1 = + "\1(" NTC + "\1)" NTC + "\1/" NTC /* REGEXP */ + "\2>>" "\1>" "\1|" NTC /* OUTRDR */ + "\2++" "\2--" NTC /* UOPPOST */ + "\2++" "\2--" "\1$" NTC /* UOPPRE1 */ + "\2==" "\1=" "\2+=" "\2-=" /* BINOPX */ + "\2*=" "\2/=" "\2%=" "\2^=" + "\1+" "\1-" "\3**=" "\2**" + "\1/" "\1%" "\1^" "\1*" + "\2!=" "\2>=" "\2<=" "\1>" + "\1<" "\2!~" "\1~" "\2&&" + "\2||" "\1?" "\1:" NTC + "\2in" NTC + "\1," NTC + "\1|" NTC + "\1+" "\1-" "\1!" NTC /* UOPPRE2 */ + "\1]" NTC + "\1{" NTC + "\1}" NTC + "\1;" NTC + "\1\n" NTC + "\2if" "\2do" "\3for" "\5break" /* STATX */ + "\10continue" "\6delete" "\5print" + "\6printf" "\4next" "\10nextfile" + "\6return" "\4exit" NTC + "\5while" NTC + "\4else" NTC + + "\3and" "\5compl" "\6lshift" "\2or" + "\6rshift" "\3xor" + "\5close" "\6system" "\6fflush" "\5atan2" /* BUILTIN */ + "\3cos" "\3exp" "\3int" "\3log" + "\4rand" "\3sin" "\4sqrt" "\5srand" + "\6gensub" "\4gsub" "\5index" "\6length" + "\5match" "\5split" "\7sprintf" "\3sub" + "\6substr" "\7systime" "\10strftime" + "\7tolower" "\7toupper" NTC + "\7getline" NTC + "\4func" "\10function" NTC + "\5BEGIN" NTC + "\3END" "\0" + ; + +static const uint32_t tokeninfo[] = { + 0, + 0, + OC_REGEXP, + xS|'a', xS|'w', xS|'|', + OC_UNARY|xV|P(9)|'p', OC_UNARY|xV|P(9)|'m', + OC_UNARY|xV|P(9)|'P', OC_UNARY|xV|P(9)|'M', + OC_FIELD|xV|P(5), + OC_COMPARE|VV|P(39)|5, OC_MOVE|VV|P(74), + OC_REPLACE|NV|P(74)|'+', OC_REPLACE|NV|P(74)|'-', + OC_REPLACE|NV|P(74)|'*', OC_REPLACE|NV|P(74)|'/', + OC_REPLACE|NV|P(74)|'%', OC_REPLACE|NV|P(74)|'&', + OC_BINARY|NV|P(29)|'+', OC_BINARY|NV|P(29)|'-', + OC_REPLACE|NV|P(74)|'&', OC_BINARY|NV|P(15)|'&', + OC_BINARY|NV|P(25)|'/', OC_BINARY|NV|P(25)|'%', + OC_BINARY|NV|P(15)|'&', OC_BINARY|NV|P(25)|'*', + OC_COMPARE|VV|P(39)|4, OC_COMPARE|VV|P(39)|3, + OC_COMPARE|VV|P(39)|0, OC_COMPARE|VV|P(39)|1, + OC_COMPARE|VV|P(39)|2, OC_MATCH|Sx|P(45)|'!', + OC_MATCH|Sx|P(45)|'~', OC_LAND|Vx|P(55), + OC_LOR|Vx|P(59), OC_TERNARY|Vx|P(64)|'?', + OC_COLON|xx|P(67)|':', + OC_IN|SV|P(49), + OC_COMMA|SS|P(80), + OC_PGETLINE|SV|P(37), + OC_UNARY|xV|P(19)|'+', OC_UNARY|xV|P(19)|'-', + OC_UNARY|xV|P(19)|'!', + 0, + 0, + 0, + 0, + 0, + ST_IF, ST_DO, ST_FOR, OC_BREAK, + OC_CONTINUE, OC_DELETE|Vx, OC_PRINT, + OC_PRINTF, OC_NEXT, OC_NEXTFILE, + OC_RETURN|Vx, OC_EXIT|Nx, + ST_WHILE, + 0, + + OC_B|B_an|P(0x83), OC_B|B_co|P(0x41), OC_B|B_ls|P(0x83), OC_B|B_or|P(0x83), + OC_B|B_rs|P(0x83), OC_B|B_xo|P(0x83), + OC_FBLTIN|Sx|F_cl, OC_FBLTIN|Sx|F_sy, OC_FBLTIN|Sx|F_ff, OC_B|B_a2|P(0x83), + OC_FBLTIN|Nx|F_co, OC_FBLTIN|Nx|F_ex, OC_FBLTIN|Nx|F_in, OC_FBLTIN|Nx|F_lg, + OC_FBLTIN|F_rn, OC_FBLTIN|Nx|F_si, OC_FBLTIN|Nx|F_sq, OC_FBLTIN|Nx|F_sr, + OC_B|B_ge|P(0xd6), OC_B|B_gs|P(0xb6), OC_B|B_ix|P(0x9b), OC_FBLTIN|Sx|F_le, + OC_B|B_ma|P(0x89), OC_B|B_sp|P(0x8b), OC_SPRINTF, OC_B|B_su|P(0xb6), + OC_B|B_ss|P(0x8f), OC_FBLTIN|F_ti, OC_B|B_ti|P(0x0b), + OC_B|B_lo|P(0x49), OC_B|B_up|P(0x49), + OC_GETLINE|SV|P(0), + 0, 0, + 0, + 0 +}; + +/* internal variable names and their initial values */ +/* asterisk marks SPECIAL vars; $ is just no-named Field0 */ +enum { + CONVFMT, OFMT, FS, OFS, + ORS, RS, RT, FILENAME, + SUBSEP, ARGIND, ARGC, ARGV, + ERRNO, FNR, + NR, NF, IGNORECASE, + ENVIRON, F0, NUM_INTERNAL_VARS +}; + +static const char vNames[] ALIGN1 = + "CONVFMT\0" "OFMT\0" "FS\0*" "OFS\0" + "ORS\0" "RS\0*" "RT\0" "FILENAME\0" + "SUBSEP\0" "ARGIND\0" "ARGC\0" "ARGV\0" + "ERRNO\0" "FNR\0" + "NR\0" "NF\0*" "IGNORECASE\0*" + "ENVIRON\0" "$\0*" "\0"; + +static const char vValues[] ALIGN1 = + "%.6g\0" "%.6g\0" " \0" " \0" + "\n\0" "\n\0" "\0" "\0" + "\034\0" + "\377"; + +/* hash size may grow to these values */ +#define FIRST_PRIME 61 +static const uint16_t PRIMES[] ALIGN2 = { 251, 1021, 4093, 16381, 65521 }; + + +/* Globals. Split in two parts so that first one is addressed + * with (mostly short) negative offsets. + * NB: it's unsafe to put members of type "double" + * into globals2 (gcc may fail to align them). + */ +struct globals { + double t_double; + chain beginseq, mainseq, endseq; + chain *seq; + node *break_ptr, *continue_ptr; + rstream *iF; + xhash *vhash, *ahash, *fdhash, *fnhash; + const char *g_progname; + int g_lineno; + int nfields; + int maxfields; /* used in fsrealloc() only */ + var *Fields; + nvblock *g_cb; + char *g_pos; + char *g_buf; + smallint icase; + smallint exiting; + smallint nextrec; + smallint nextfile; + smallint is_f0_split; +}; +struct globals2 { + uint32_t t_info; /* often used */ + uint32_t t_tclass; + char *t_string; + int t_lineno; + int t_rollback; + + var *intvar[NUM_INTERNAL_VARS]; /* often used */ + + /* former statics from various functions */ + char *split_f0__fstrings; + + uint32_t next_token__save_tclass; + uint32_t next_token__save_info; + uint32_t next_token__ltclass; + smallint next_token__concat_inserted; + + smallint next_input_file__files_happen; + rstream next_input_file__rsm; + + var *evaluate__fnargs; + unsigned evaluate__seed; + regex_t evaluate__sreg; + + var ptest__v; + + tsplitter exec_builtin__tspl; + + /* biggest and least used members go last */ + tsplitter fsplitter, rsplitter; +}; +#define G1 (ptr_to_globals[-1]) +#define G (*(struct globals2 *)ptr_to_globals) +/* For debug. nm --size-sort awk.o | grep -vi ' [tr] ' */ +/*char G1size[sizeof(G1)]; - 0x74 */ +/*char Gsize[sizeof(G)]; - 0x1c4 */ +/* Trying to keep most of members accessible with short offsets: */ +/*char Gofs_seed[offsetof(struct globals2, evaluate__seed)]; - 0x90 */ +#define t_double (G1.t_double ) +#define beginseq (G1.beginseq ) +#define mainseq (G1.mainseq ) +#define endseq (G1.endseq ) +#define seq (G1.seq ) +#define break_ptr (G1.break_ptr ) +#define continue_ptr (G1.continue_ptr) +#define iF (G1.iF ) +#define vhash (G1.vhash ) +#define ahash (G1.ahash ) +#define fdhash (G1.fdhash ) +#define fnhash (G1.fnhash ) +#define g_progname (G1.g_progname ) +#define g_lineno (G1.g_lineno ) +#define nfields (G1.nfields ) +#define maxfields (G1.maxfields ) +#define Fields (G1.Fields ) +#define g_cb (G1.g_cb ) +#define g_pos (G1.g_pos ) +#define g_buf (G1.g_buf ) +#define icase (G1.icase ) +#define exiting (G1.exiting ) +#define nextrec (G1.nextrec ) +#define nextfile (G1.nextfile ) +#define is_f0_split (G1.is_f0_split ) +#define t_info (G.t_info ) +#define t_tclass (G.t_tclass ) +#define t_string (G.t_string ) +#define t_lineno (G.t_lineno ) +#define t_rollback (G.t_rollback ) +#define intvar (G.intvar ) +#define fsplitter (G.fsplitter ) +#define rsplitter (G.rsplitter ) +#define INIT_G() do { \ + SET_PTR_TO_GLOBALS(xzalloc(sizeof(G1) + sizeof(G)) + sizeof(G1)); \ + G.next_token__ltclass = TC_OPTERM; \ + G.evaluate__seed = 1; \ +} while (0) + + +/* function prototypes */ +static void handle_special(var *); +static node *parse_expr(uint32_t); +static void chain_group(void); +static var *evaluate(node *, var *); +static rstream *next_input_file(void); +static int fmt_num(char *, int, const char *, double, int); +static int awk_exit(int) NORETURN; + +/* ---- error handling ---- */ + +static const char EMSG_INTERNAL_ERROR[] ALIGN1 = "Internal error"; +static const char EMSG_UNEXP_EOS[] ALIGN1 = "Unexpected end of string"; +static const char EMSG_UNEXP_TOKEN[] ALIGN1 = "Unexpected token"; +static const char EMSG_DIV_BY_ZERO[] ALIGN1 = "Division by zero"; +static const char EMSG_INV_FMT[] ALIGN1 = "Invalid format specifier"; +static const char EMSG_TOO_FEW_ARGS[] ALIGN1 = "Too few arguments for builtin"; +static const char EMSG_NOT_ARRAY[] ALIGN1 = "Not an array"; +static const char EMSG_POSSIBLE_ERROR[] ALIGN1 = "Possible syntax error"; +static const char EMSG_UNDEF_FUNC[] ALIGN1 = "Call to undefined function"; +#if !ENABLE_FEATURE_AWK_LIBM +static const char EMSG_NO_MATH[] ALIGN1 = "Math support is not compiled in"; +#endif + +static void zero_out_var(var * vp) +{ + memset(vp, 0, sizeof(*vp)); +} + +static void syntax_error(const char *const message) NORETURN; +static void syntax_error(const char *const message) +{ + bb_error_msg_and_die("%s:%i: %s", g_progname, g_lineno, message); +} + +/* ---- hash stuff ---- */ + +static unsigned hashidx(const char *name) +{ + unsigned idx = 0; + + while (*name) idx = *name++ + (idx << 6) - idx; + return idx; +} + +/* create new hash */ +static xhash *hash_init(void) +{ + xhash *newhash; + + newhash = xzalloc(sizeof(xhash)); + newhash->csize = FIRST_PRIME; + newhash->items = xzalloc(newhash->csize * sizeof(hash_item *)); + + return newhash; +} + +/* find item in hash, return ptr to data, NULL if not found */ +static void *hash_search(xhash *hash, const char *name) +{ + hash_item *hi; + + hi = hash->items [ hashidx(name) % hash->csize ]; + while (hi) { + if (strcmp(hi->name, name) == 0) + return &(hi->data); + hi = hi->next; + } + return NULL; +} + +/* grow hash if it becomes too big */ +static void hash_rebuild(xhash *hash) +{ + unsigned newsize, i, idx; + hash_item **newitems, *hi, *thi; + + if (hash->nprime == ARRAY_SIZE(PRIMES)) + return; + + newsize = PRIMES[hash->nprime++]; + newitems = xzalloc(newsize * sizeof(hash_item *)); + + for (i = 0; i < hash->csize; i++) { + hi = hash->items[i]; + while (hi) { + thi = hi; + hi = thi->next; + idx = hashidx(thi->name) % newsize; + thi->next = newitems[idx]; + newitems[idx] = thi; + } + } + + free(hash->items); + hash->csize = newsize; + hash->items = newitems; +} + +/* find item in hash, add it if necessary. Return ptr to data */ +static void *hash_find(xhash *hash, const char *name) +{ + hash_item *hi; + unsigned idx; + int l; + + hi = hash_search(hash, name); + if (!hi) { + if (++hash->nel / hash->csize > 10) + hash_rebuild(hash); + + l = strlen(name) + 1; + hi = xzalloc(sizeof(hash_item) + l); + memcpy(hi->name, name, l); + + idx = hashidx(name) % hash->csize; + hi->next = hash->items[idx]; + hash->items[idx] = hi; + hash->glen += l; + } + return &(hi->data); +} + +#define findvar(hash, name) ((var*) hash_find((hash), (name))) +#define newvar(name) ((var*) hash_find(vhash, (name))) +#define newfile(name) ((rstream*)hash_find(fdhash, (name))) +#define newfunc(name) ((func*) hash_find(fnhash, (name))) + +static void hash_remove(xhash *hash, const char *name) +{ + hash_item *hi, **phi; + + phi = &(hash->items[hashidx(name) % hash->csize]); + while (*phi) { + hi = *phi; + if (strcmp(hi->name, name) == 0) { + hash->glen -= (strlen(name) + 1); + hash->nel--; + *phi = hi->next; + free(hi); + break; + } + phi = &(hi->next); + } +} + +/* ------ some useful functions ------ */ + +static void skip_spaces(char **s) +{ + char *p = *s; + + while (1) { + if (*p == '\\' && p[1] == '\n') { + p++; + t_lineno++; + } else if (*p != ' ' && *p != '\t') { + break; + } + p++; + } + *s = p; +} + +static char *nextword(char **s) +{ + char *p = *s; + + while (*(*s)++) /* */; + + return p; +} + +static char nextchar(char **s) +{ + char c, *pps; + + c = *((*s)++); + pps = *s; + if (c == '\\') c = bb_process_escape_sequence((const char**)s); + if (c == '\\' && *s == pps) c = *((*s)++); + return c; +} + +static ALWAYS_INLINE int isalnum_(int c) +{ + return (isalnum(c) || c == '_'); +} + +static double my_strtod(char **pp) +{ +#if ENABLE_DESKTOP + if ((*pp)[0] == '0' + && ((((*pp)[1] | 0x20) == 'x') || isdigit((*pp)[1])) + ) { + return strtoull(*pp, pp, 0); + } +#endif + return strtod(*pp, pp); +} + +/* -------- working with variables (set/get/copy/etc) -------- */ + +static xhash *iamarray(var *v) +{ + var *a = v; + + while (a->type & VF_CHILD) + a = a->x.parent; + + if (!(a->type & VF_ARRAY)) { + a->type |= VF_ARRAY; + a->x.array = hash_init(); + } + return a->x.array; +} + +static void clear_array(xhash *array) +{ + unsigned i; + hash_item *hi, *thi; + + for (i = 0; i < array->csize; i++) { + hi = array->items[i]; + while (hi) { + thi = hi; + hi = hi->next; + free(thi->data.v.string); + free(thi); + } + array->items[i] = NULL; + } + array->glen = array->nel = 0; +} + +/* clear a variable */ +static var *clrvar(var *v) +{ + if (!(v->type & VF_FSTR)) + free(v->string); + + v->type &= VF_DONTTOUCH; + v->type |= VF_DIRTY; + v->string = NULL; + return v; +} + +/* assign string value to variable */ +static var *setvar_p(var *v, char *value) +{ + clrvar(v); + v->string = value; + handle_special(v); + return v; +} + +/* same as setvar_p but make a copy of string */ +static var *setvar_s(var *v, const char *value) +{ + return setvar_p(v, (value && *value) ? xstrdup(value) : NULL); +} + +/* same as setvar_s but set USER flag */ +static var *setvar_u(var *v, const char *value) +{ + setvar_s(v, value); + v->type |= VF_USER; + return v; +} + +/* set array element to user string */ +static void setari_u(var *a, int idx, const char *s) +{ + char sidx[sizeof(int)*3 + 1]; + var *v; + + sprintf(sidx, "%d", idx); + v = findvar(iamarray(a), sidx); + setvar_u(v, s); +} + +/* assign numeric value to variable */ +static var *setvar_i(var *v, double value) +{ + clrvar(v); + v->type |= VF_NUMBER; + v->number = value; + handle_special(v); + return v; +} + +static const char *getvar_s(var *v) +{ + /* if v is numeric and has no cached string, convert it to string */ + if ((v->type & (VF_NUMBER | VF_CACHED)) == VF_NUMBER) { + fmt_num(g_buf, MAXVARFMT, getvar_s(intvar[CONVFMT]), v->number, TRUE); + v->string = xstrdup(g_buf); + v->type |= VF_CACHED; + } + return (v->string == NULL) ? "" : v->string; +} + +static double getvar_i(var *v) +{ + char *s; + + if ((v->type & (VF_NUMBER | VF_CACHED)) == 0) { + v->number = 0; + s = v->string; + if (s && *s) { + v->number = my_strtod(&s); + if (v->type & VF_USER) { + skip_spaces(&s); + if (*s != '\0') + v->type &= ~VF_USER; + } + } else { + v->type &= ~VF_USER; + } + v->type |= VF_CACHED; + } + return v->number; +} + +/* Used for operands of bitwise ops */ +static unsigned long getvar_i_int(var *v) +{ + double d = getvar_i(v); + + /* Casting doubles to longs is undefined for values outside + * of target type range. Try to widen it as much as possible */ + if (d >= 0) + return (unsigned long)d; + /* Why? Think about d == -4294967295.0 (assuming 32bit longs) */ + return - (long) (unsigned long) (-d); +} + +static var *copyvar(var *dest, const var *src) +{ + if (dest != src) { + clrvar(dest); + dest->type |= (src->type & ~(VF_DONTTOUCH | VF_FSTR)); + dest->number = src->number; + if (src->string) + dest->string = xstrdup(src->string); + } + handle_special(dest); + return dest; +} + +static var *incvar(var *v) +{ + return setvar_i(v, getvar_i(v) + 1.); +} + +/* return true if v is number or numeric string */ +static int is_numeric(var *v) +{ + getvar_i(v); + return ((v->type ^ VF_DIRTY) & (VF_NUMBER | VF_USER | VF_DIRTY)); +} + +/* return 1 when value of v corresponds to true, 0 otherwise */ +static int istrue(var *v) +{ + if (is_numeric(v)) + return (v->number == 0) ? 0 : 1; + return (v->string && *(v->string)) ? 1 : 0; +} + +/* temporary variables allocator. Last allocated should be first freed */ +static var *nvalloc(int n) +{ + nvblock *pb = NULL; + var *v, *r; + int size; + + while (g_cb) { + pb = g_cb; + if ((g_cb->pos - g_cb->nv) + n <= g_cb->size) break; + g_cb = g_cb->next; + } + + if (!g_cb) { + size = (n <= MINNVBLOCK) ? MINNVBLOCK : n; + g_cb = xzalloc(sizeof(nvblock) + size * sizeof(var)); + g_cb->size = size; + g_cb->pos = g_cb->nv; + g_cb->prev = pb; + /*g_cb->next = NULL; - xzalloc did it */ + if (pb) pb->next = g_cb; + } + + v = r = g_cb->pos; + g_cb->pos += n; + + while (v < g_cb->pos) { + v->type = 0; + v->string = NULL; + v++; + } + + return r; +} + +static void nvfree(var *v) +{ + var *p; + + if (v < g_cb->nv || v >= g_cb->pos) + syntax_error(EMSG_INTERNAL_ERROR); + + for (p = v; p < g_cb->pos; p++) { + if ((p->type & (VF_ARRAY | VF_CHILD)) == VF_ARRAY) { + clear_array(iamarray(p)); + free(p->x.array->items); + free(p->x.array); + } + if (p->type & VF_WALK) + free(p->x.walker); + + clrvar(p); + } + + g_cb->pos = v; + while (g_cb->prev && g_cb->pos == g_cb->nv) { + g_cb = g_cb->prev; + } +} + +/* ------- awk program text parsing ------- */ + +/* Parse next token pointed by global pos, place results into global ttt. + * If token isn't expected, give away. Return token class + */ +static uint32_t next_token(uint32_t expected) +{ +#define concat_inserted (G.next_token__concat_inserted) +#define save_tclass (G.next_token__save_tclass) +#define save_info (G.next_token__save_info) +/* Initialized to TC_OPTERM: */ +#define ltclass (G.next_token__ltclass) + + char *p, *pp, *s; + const char *tl; + uint32_t tc; + const uint32_t *ti; + int l; + + if (t_rollback) { + t_rollback = FALSE; + + } else if (concat_inserted) { + concat_inserted = FALSE; + t_tclass = save_tclass; + t_info = save_info; + + } else { + p = g_pos; + readnext: + skip_spaces(&p); + g_lineno = t_lineno; + if (*p == '#') + while (*p != '\n' && *p != '\0') + p++; + + if (*p == '\n') + t_lineno++; + + if (*p == '\0') { + tc = TC_EOF; + + } else if (*p == '\"') { + /* it's a string */ + t_string = s = ++p; + while (*p != '\"') { + if (*p == '\0' || *p == '\n') + syntax_error(EMSG_UNEXP_EOS); + *(s++) = nextchar(&p); + } + p++; + *s = '\0'; + tc = TC_STRING; + + } else if ((expected & TC_REGEXP) && *p == '/') { + /* it's regexp */ + t_string = s = ++p; + while (*p != '/') { + if (*p == '\0' || *p == '\n') + syntax_error(EMSG_UNEXP_EOS); + *s = *p++; + if (*s++ == '\\') { + pp = p; + *(s-1) = bb_process_escape_sequence((const char **)&p); + if (*pp == '\\') + *s++ = '\\'; + if (p == pp) + *s++ = *p++; + } + } + p++; + *s = '\0'; + tc = TC_REGEXP; + + } else if (*p == '.' || isdigit(*p)) { + /* it's a number */ + t_double = my_strtod(&p); + if (*p == '.') + syntax_error(EMSG_UNEXP_TOKEN); + tc = TC_NUMBER; + + } else { + /* search for something known */ + tl = tokenlist; + tc = 0x00000001; + ti = tokeninfo; + while (*tl) { + l = *(tl++); + if (l == NTCC) { + tc <<= 1; + continue; + } + /* if token class is expected, token + * matches and it's not a longer word, + * then this is what we are looking for + */ + if ((tc & (expected | TC_WORD | TC_NEWLINE)) + && *tl == *p && strncmp(p, tl, l) == 0 + && !((tc & TC_WORD) && isalnum_(p[l])) + ) { + t_info = *ti; + p += l; + break; + } + ti++; + tl += l; + } + + if (!*tl) { + /* it's a name (var/array/function), + * otherwise it's something wrong + */ + if (!isalnum_(*p)) + syntax_error(EMSG_UNEXP_TOKEN); + + t_string = --p; + while (isalnum_(*(++p))) { + *(p-1) = *p; + } + *(p-1) = '\0'; + tc = TC_VARIABLE; + /* also consume whitespace between functionname and bracket */ + if (!(expected & TC_VARIABLE)) + skip_spaces(&p); + if (*p == '(') { + tc = TC_FUNCTION; + } else { + if (*p == '[') { + p++; + tc = TC_ARRAY; + } + } + } + } + g_pos = p; + + /* skipping newlines in some cases */ + if ((ltclass & TC_NOTERM) && (tc & TC_NEWLINE)) + goto readnext; + + /* insert concatenation operator when needed */ + if ((ltclass & TC_CONCAT1) && (tc & TC_CONCAT2) && (expected & TC_BINOP)) { + concat_inserted = TRUE; + save_tclass = tc; + save_info = t_info; + tc = TC_BINOP; + t_info = OC_CONCAT | SS | P(35); + } + + t_tclass = tc; + } + ltclass = t_tclass; + + /* Are we ready for this? */ + if (!(ltclass & expected)) + syntax_error((ltclass & (TC_NEWLINE | TC_EOF)) ? + EMSG_UNEXP_EOS : EMSG_UNEXP_TOKEN); + + return ltclass; +#undef concat_inserted +#undef save_tclass +#undef save_info +#undef ltclass +} + +static void rollback_token(void) +{ + t_rollback = TRUE; +} + +static node *new_node(uint32_t info) +{ + node *n; + + n = xzalloc(sizeof(node)); + n->info = info; + n->lineno = g_lineno; + return n; +} + +static node *mk_re_node(const char *s, node *n, regex_t *re) +{ + n->info = OC_REGEXP; + n->l.re = re; + n->r.ire = re + 1; + xregcomp(re, s, REG_EXTENDED); + xregcomp(re + 1, s, REG_EXTENDED | REG_ICASE); + + return n; +} + +static node *condition(void) +{ + next_token(TC_SEQSTART); + return parse_expr(TC_SEQTERM); +} + +/* parse expression terminated by given argument, return ptr + * to built subtree. Terminator is eaten by parse_expr */ +static node *parse_expr(uint32_t iexp) +{ + node sn; + node *cn = &sn; + node *vn, *glptr; + uint32_t tc, xtc; + var *v; + + sn.info = PRIMASK; + sn.r.n = glptr = NULL; + xtc = TC_OPERAND | TC_UOPPRE | TC_REGEXP | iexp; + + while (!((tc = next_token(xtc)) & iexp)) { + if (glptr && (t_info == (OC_COMPARE | VV | P(39) | 2))) { + /* input redirection (<) attached to glptr node */ + cn = glptr->l.n = new_node(OC_CONCAT | SS | P(37)); + cn->a.n = glptr; + xtc = TC_OPERAND | TC_UOPPRE; + glptr = NULL; + + } else if (tc & (TC_BINOP | TC_UOPPOST)) { + /* for binary and postfix-unary operators, jump back over + * previous operators with higher priority */ + vn = cn; + while ( ((t_info & PRIMASK) > (vn->a.n->info & PRIMASK2)) + || ((t_info == vn->info) && ((t_info & OPCLSMASK) == OC_COLON)) ) + vn = vn->a.n; + if ((t_info & OPCLSMASK) == OC_TERNARY) + t_info += P(6); + cn = vn->a.n->r.n = new_node(t_info); + cn->a.n = vn->a.n; + if (tc & TC_BINOP) { + cn->l.n = vn; + xtc = TC_OPERAND | TC_UOPPRE | TC_REGEXP; + if ((t_info & OPCLSMASK) == OC_PGETLINE) { + /* it's a pipe */ + next_token(TC_GETLINE); + /* give maximum priority to this pipe */ + cn->info &= ~PRIMASK; + xtc = TC_OPERAND | TC_UOPPRE | TC_BINOP | iexp; + } + } else { + cn->r.n = vn; + xtc = TC_OPERAND | TC_UOPPRE | TC_BINOP | iexp; + } + vn->a.n = cn; + + } else { + /* for operands and prefix-unary operators, attach them + * to last node */ + vn = cn; + cn = vn->r.n = new_node(t_info); + cn->a.n = vn; + xtc = TC_OPERAND | TC_UOPPRE | TC_REGEXP; + if (tc & (TC_OPERAND | TC_REGEXP)) { + xtc = TC_UOPPRE | TC_UOPPOST | TC_BINOP | TC_OPERAND | iexp; + /* one should be very careful with switch on tclass - + * only simple tclasses should be used! */ + switch (tc) { + case TC_VARIABLE: + case TC_ARRAY: + cn->info = OC_VAR; + v = hash_search(ahash, t_string); + if (v != NULL) { + cn->info = OC_FNARG; + cn->l.i = v->x.aidx; + } else { + cn->l.v = newvar(t_string); + } + if (tc & TC_ARRAY) { + cn->info |= xS; + cn->r.n = parse_expr(TC_ARRTERM); + } + break; + + case TC_NUMBER: + case TC_STRING: + cn->info = OC_VAR; + v = cn->l.v = xzalloc(sizeof(var)); + if (tc & TC_NUMBER) + setvar_i(v, t_double); + else + setvar_s(v, t_string); + break; + + case TC_REGEXP: + mk_re_node(t_string, cn, xzalloc(sizeof(regex_t)*2)); + break; + + case TC_FUNCTION: + cn->info = OC_FUNC; + cn->r.f = newfunc(t_string); + cn->l.n = condition(); + break; + + case TC_SEQSTART: + cn = vn->r.n = parse_expr(TC_SEQTERM); + cn->a.n = vn; + break; + + case TC_GETLINE: + glptr = cn; + xtc = TC_OPERAND | TC_UOPPRE | TC_BINOP | iexp; + break; + + case TC_BUILTIN: + cn->l.n = condition(); + break; + } + } + } + } + return sn.r.n; +} + +/* add node to chain. Return ptr to alloc'd node */ +static node *chain_node(uint32_t info) +{ + node *n; + + if (!seq->first) + seq->first = seq->last = new_node(0); + + if (seq->programname != g_progname) { + seq->programname = g_progname; + n = chain_node(OC_NEWSOURCE); + n->l.s = xstrdup(g_progname); + } + + n = seq->last; + n->info = info; + seq->last = n->a.n = new_node(OC_DONE); + + return n; +} + +static void chain_expr(uint32_t info) +{ + node *n; + + n = chain_node(info); + n->l.n = parse_expr(TC_OPTERM | TC_GRPTERM); + if (t_tclass & TC_GRPTERM) + rollback_token(); +} + +static node *chain_loop(node *nn) +{ + node *n, *n2, *save_brk, *save_cont; + + save_brk = break_ptr; + save_cont = continue_ptr; + + n = chain_node(OC_BR | Vx); + continue_ptr = new_node(OC_EXEC); + break_ptr = new_node(OC_EXEC); + chain_group(); + n2 = chain_node(OC_EXEC | Vx); + n2->l.n = nn; + n2->a.n = n; + continue_ptr->a.n = n2; + break_ptr->a.n = n->r.n = seq->last; + + continue_ptr = save_cont; + break_ptr = save_brk; + + return n; +} + +/* parse group and attach it to chain */ +static void chain_group(void) +{ + uint32_t c; + node *n, *n2, *n3; + + do { + c = next_token(TC_GRPSEQ); + } while (c & TC_NEWLINE); + + if (c & TC_GRPSTART) { + while (next_token(TC_GRPSEQ | TC_GRPTERM) != TC_GRPTERM) { + if (t_tclass & TC_NEWLINE) continue; + rollback_token(); + chain_group(); + } + } else if (c & (TC_OPSEQ | TC_OPTERM)) { + rollback_token(); + chain_expr(OC_EXEC | Vx); + } else { /* TC_STATEMNT */ + switch (t_info & OPCLSMASK) { + case ST_IF: + n = chain_node(OC_BR | Vx); + n->l.n = condition(); + chain_group(); + n2 = chain_node(OC_EXEC); + n->r.n = seq->last; + if (next_token(TC_GRPSEQ | TC_GRPTERM | TC_ELSE) == TC_ELSE) { + chain_group(); + n2->a.n = seq->last; + } else { + rollback_token(); + } + break; + + case ST_WHILE: + n2 = condition(); + n = chain_loop(NULL); + n->l.n = n2; + break; + + case ST_DO: + n2 = chain_node(OC_EXEC); + n = chain_loop(NULL); + n2->a.n = n->a.n; + next_token(TC_WHILE); + n->l.n = condition(); + break; + + case ST_FOR: + next_token(TC_SEQSTART); + n2 = parse_expr(TC_SEMICOL | TC_SEQTERM); + if (t_tclass & TC_SEQTERM) { /* for-in */ + if ((n2->info & OPCLSMASK) != OC_IN) + syntax_error(EMSG_UNEXP_TOKEN); + n = chain_node(OC_WALKINIT | VV); + n->l.n = n2->l.n; + n->r.n = n2->r.n; + n = chain_loop(NULL); + n->info = OC_WALKNEXT | Vx; + n->l.n = n2->l.n; + } else { /* for (;;) */ + n = chain_node(OC_EXEC | Vx); + n->l.n = n2; + n2 = parse_expr(TC_SEMICOL); + n3 = parse_expr(TC_SEQTERM); + n = chain_loop(n3); + n->l.n = n2; + if (!n2) + n->info = OC_EXEC; + } + break; + + case OC_PRINT: + case OC_PRINTF: + n = chain_node(t_info); + n->l.n = parse_expr(TC_OPTERM | TC_OUTRDR | TC_GRPTERM); + if (t_tclass & TC_OUTRDR) { + n->info |= t_info; + n->r.n = parse_expr(TC_OPTERM | TC_GRPTERM); + } + if (t_tclass & TC_GRPTERM) + rollback_token(); + break; + + case OC_BREAK: + n = chain_node(OC_EXEC); + n->a.n = break_ptr; + break; + + case OC_CONTINUE: + n = chain_node(OC_EXEC); + n->a.n = continue_ptr; + break; + + /* delete, next, nextfile, return, exit */ + default: + chain_expr(t_info); + } + } +} + +static void parse_program(char *p) +{ + uint32_t tclass; + node *cn; + func *f; + var *v; + + g_pos = p; + t_lineno = 1; + while ((tclass = next_token(TC_EOF | TC_OPSEQ | TC_GRPSTART | + TC_OPTERM | TC_BEGIN | TC_END | TC_FUNCDECL)) != TC_EOF) { + + if (tclass & TC_OPTERM) + continue; + + seq = &mainseq; + if (tclass & TC_BEGIN) { + seq = &beginseq; + chain_group(); + + } else if (tclass & TC_END) { + seq = &endseq; + chain_group(); + + } else if (tclass & TC_FUNCDECL) { + next_token(TC_FUNCTION); + g_pos++; + f = newfunc(t_string); + f->body.first = NULL; + f->nargs = 0; + while (next_token(TC_VARIABLE | TC_SEQTERM) & TC_VARIABLE) { + v = findvar(ahash, t_string); + v->x.aidx = (f->nargs)++; + + if (next_token(TC_COMMA | TC_SEQTERM) & TC_SEQTERM) + break; + } + seq = &(f->body); + chain_group(); + clear_array(ahash); + + } else if (tclass & TC_OPSEQ) { + rollback_token(); + cn = chain_node(OC_TEST); + cn->l.n = parse_expr(TC_OPTERM | TC_EOF | TC_GRPSTART); + if (t_tclass & TC_GRPSTART) { + rollback_token(); + chain_group(); + } else { + chain_node(OC_PRINT); + } + cn->r.n = mainseq.last; + + } else /* if (tclass & TC_GRPSTART) */ { + rollback_token(); + chain_group(); + } + } +} + + +/* -------- program execution part -------- */ + +static node *mk_splitter(const char *s, tsplitter *spl) +{ + regex_t *re, *ire; + node *n; + + re = &spl->re[0]; + ire = &spl->re[1]; + n = &spl->n; + if ((n->info & OPCLSMASK) == OC_REGEXP) { + regfree(re); + regfree(ire); // TODO: nuke ire, use re+1? + } + if (strlen(s) > 1) { + mk_re_node(s, n, re); + } else { + n->info = (uint32_t) *s; + } + + return n; +} + +/* use node as a regular expression. Supplied with node ptr and regex_t + * storage space. Return ptr to regex (if result points to preg, it should + * be later regfree'd manually + */ +static regex_t *as_regex(node *op, regex_t *preg) +{ + var *v; + const char *s; + + if ((op->info & OPCLSMASK) == OC_REGEXP) { + return icase ? op->r.ire : op->l.re; + } + v = nvalloc(1); + s = getvar_s(evaluate(op, v)); + xregcomp(preg, s, icase ? REG_EXTENDED | REG_ICASE : REG_EXTENDED); + nvfree(v); + return preg; +} + +/* gradually increasing buffer */ +static void qrealloc(char **b, int n, int *size) +{ + if (!*b || n >= *size) { + *size = n + (n>>1) + 80; + *b = xrealloc(*b, *size); + } +} + +/* resize field storage space */ +static void fsrealloc(int size) +{ + int i; + + if (size >= maxfields) { + i = maxfields; + maxfields = size + 16; + Fields = xrealloc(Fields, maxfields * sizeof(var)); + for (; i < maxfields; i++) { + Fields[i].type = VF_SPECIAL; + Fields[i].string = NULL; + } + } + + if (size < nfields) { + for (i = size; i < nfields; i++) { + clrvar(Fields + i); + } + } + nfields = size; +} + +static int awk_split(const char *s, node *spl, char **slist) +{ + int l, n = 0; + char c[4]; + char *s1; + regmatch_t pmatch[2]; // TODO: why [2]? [1] is enough... + + /* in worst case, each char would be a separate field */ + *slist = s1 = xzalloc(strlen(s) * 2 + 3); + strcpy(s1, s); + + c[0] = c[1] = (char)spl->info; + c[2] = c[3] = '\0'; + if (*getvar_s(intvar[RS]) == '\0') + c[2] = '\n'; + + if ((spl->info & OPCLSMASK) == OC_REGEXP) { /* regex split */ + if (!*s) + return n; /* "": zero fields */ + n++; /* at least one field will be there */ + do { + l = strcspn(s, c+2); /* len till next NUL or \n */ + if (regexec(icase ? spl->r.ire : spl->l.re, s, 1, pmatch, 0) == 0 + && pmatch[0].rm_so <= l + ) { + l = pmatch[0].rm_so; + if (pmatch[0].rm_eo == 0) { + l++; + pmatch[0].rm_eo++; + } + n++; /* we saw yet another delimiter */ + } else { + pmatch[0].rm_eo = l; + if (s[l]) pmatch[0].rm_eo++; + } + memcpy(s1, s, l); + s1[l] = '\0'; + nextword(&s1); + s += pmatch[0].rm_eo; + } while (*s); + return n; + } + if (c[0] == '\0') { /* null split */ + while (*s) { + *s1++ = *s++; + *s1++ = '\0'; + n++; + } + return n; + } + if (c[0] != ' ') { /* single-character split */ + if (icase) { + c[0] = toupper(c[0]); + c[1] = tolower(c[1]); + } + if (*s1) n++; + while ((s1 = strpbrk(s1, c))) { + *s1++ = '\0'; + n++; + } + return n; + } + /* space split */ + while (*s) { + s = skip_whitespace(s); + if (!*s) break; + n++; + while (*s && !isspace(*s)) + *s1++ = *s++; + *s1++ = '\0'; + } + return n; +} + +static void split_f0(void) +{ +/* static char *fstrings; */ +#define fstrings (G.split_f0__fstrings) + + int i, n; + char *s; + + if (is_f0_split) + return; + + is_f0_split = TRUE; + free(fstrings); + fsrealloc(0); + n = awk_split(getvar_s(intvar[F0]), &fsplitter.n, &fstrings); + fsrealloc(n); + s = fstrings; + for (i = 0; i < n; i++) { + Fields[i].string = nextword(&s); + Fields[i].type |= (VF_FSTR | VF_USER | VF_DIRTY); + } + + /* set NF manually to avoid side effects */ + clrvar(intvar[NF]); + intvar[NF]->type = VF_NUMBER | VF_SPECIAL; + intvar[NF]->number = nfields; +#undef fstrings +} + +/* perform additional actions when some internal variables changed */ +static void handle_special(var *v) +{ + int n; + char *b; + const char *sep, *s; + int sl, l, len, i, bsize; + + if (!(v->type & VF_SPECIAL)) + return; + + if (v == intvar[NF]) { + n = (int)getvar_i(v); + fsrealloc(n); + + /* recalculate $0 */ + sep = getvar_s(intvar[OFS]); + sl = strlen(sep); + b = NULL; + len = 0; + for (i = 0; i < n; i++) { + s = getvar_s(&Fields[i]); + l = strlen(s); + if (b) { + memcpy(b+len, sep, sl); + len += sl; + } + qrealloc(&b, len+l+sl, &bsize); + memcpy(b+len, s, l); + len += l; + } + if (b) + b[len] = '\0'; + setvar_p(intvar[F0], b); + is_f0_split = TRUE; + + } else if (v == intvar[F0]) { + is_f0_split = FALSE; + + } else if (v == intvar[FS]) { + mk_splitter(getvar_s(v), &fsplitter); + + } else if (v == intvar[RS]) { + mk_splitter(getvar_s(v), &rsplitter); + + } else if (v == intvar[IGNORECASE]) { + icase = istrue(v); + + } else { /* $n */ + n = getvar_i(intvar[NF]); + setvar_i(intvar[NF], n > v-Fields ? n : v-Fields+1); + /* right here v is invalid. Just to note... */ + } +} + +/* step through func/builtin/etc arguments */ +static node *nextarg(node **pn) +{ + node *n; + + n = *pn; + if (n && (n->info & OPCLSMASK) == OC_COMMA) { + *pn = n->r.n; + n = n->l.n; + } else { + *pn = NULL; + } + return n; +} + +static void hashwalk_init(var *v, xhash *array) +{ + char **w; + hash_item *hi; + unsigned i; + + if (v->type & VF_WALK) + free(v->x.walker); + + v->type |= VF_WALK; + w = v->x.walker = xzalloc(2 + 2*sizeof(char *) + array->glen); + w[0] = w[1] = (char *)(w + 2); + for (i = 0; i < array->csize; i++) { + hi = array->items[i]; + while (hi) { + strcpy(*w, hi->name); + nextword(w); + hi = hi->next; + } + } +} + +static int hashwalk_next(var *v) +{ + char **w; + + w = v->x.walker; + if (w[1] == w[0]) + return FALSE; + + setvar_s(v, nextword(w+1)); + return TRUE; +} + +/* evaluate node, return 1 when result is true, 0 otherwise */ +static int ptest(node *pattern) +{ + /* ptest__v is "static": to save stack space? */ + return istrue(evaluate(pattern, &G.ptest__v)); +} + +/* read next record from stream rsm into a variable v */ +static int awk_getline(rstream *rsm, var *v) +{ + char *b; + regmatch_t pmatch[2]; + int a, p, pp=0, size; + int fd, so, eo, r, rp; + char c, *m, *s; + + /* we're using our own buffer since we need access to accumulating + * characters + */ + fd = fileno(rsm->F); + m = rsm->buffer; + a = rsm->adv; + p = rsm->pos; + size = rsm->size; + c = (char) rsplitter.n.info; + rp = 0; + + if (!m) qrealloc(&m, 256, &size); + do { + b = m + a; + so = eo = p; + r = 1; + if (p > 0) { + if ((rsplitter.n.info & OPCLSMASK) == OC_REGEXP) { + if (regexec(icase ? rsplitter.n.r.ire : rsplitter.n.l.re, + b, 1, pmatch, 0) == 0) { + so = pmatch[0].rm_so; + eo = pmatch[0].rm_eo; + if (b[eo] != '\0') + break; + } + } else if (c != '\0') { + s = strchr(b+pp, c); + if (!s) s = memchr(b+pp, '\0', p - pp); + if (s) { + so = eo = s-b; + eo++; + break; + } + } else { + while (b[rp] == '\n') + rp++; + s = strstr(b+rp, "\n\n"); + if (s) { + so = eo = s-b; + while (b[eo] == '\n') eo++; + if (b[eo] != '\0') + break; + } + } + } + + if (a > 0) { + memmove(m, (const void *)(m+a), p+1); + b = m; + a = 0; + } + + qrealloc(&m, a+p+128, &size); + b = m + a; + pp = p; + p += safe_read(fd, b+p, size-p-1); + if (p < pp) { + p = 0; + r = 0; + setvar_i(intvar[ERRNO], errno); + } + b[p] = '\0'; + + } while (p > pp); + + if (p == 0) { + r--; + } else { + c = b[so]; b[so] = '\0'; + setvar_s(v, b+rp); + v->type |= VF_USER; + b[so] = c; + c = b[eo]; b[eo] = '\0'; + setvar_s(intvar[RT], b+so); + b[eo] = c; + } + + rsm->buffer = m; + rsm->adv = a + eo; + rsm->pos = p - eo; + rsm->size = size; + + return r; +} + +static int fmt_num(char *b, int size, const char *format, double n, int int_as_int) +{ + int r = 0; + char c; + const char *s = format; + + if (int_as_int && n == (int)n) { + r = snprintf(b, size, "%d", (int)n); + } else { + do { c = *s; } while (c && *++s); + if (strchr("diouxX", c)) { + r = snprintf(b, size, format, (int)n); + } else if (strchr("eEfgG", c)) { + r = snprintf(b, size, format, n); + } else { + syntax_error(EMSG_INV_FMT); + } + } + return r; +} + + +/* formatted output into an allocated buffer, return ptr to buffer */ +static char *awk_printf(node *n) +{ + char *b = NULL; + char *fmt, *s, *f; + const char *s1; + int i, j, incr, bsize; + char c, c1; + var *v, *arg; + + v = nvalloc(1); + fmt = f = xstrdup(getvar_s(evaluate(nextarg(&n), v))); + + i = 0; + while (*f) { + s = f; + while (*f && (*f != '%' || *(++f) == '%')) + f++; + while (*f && !isalpha(*f)) { + if (*f == '*') + syntax_error("%*x formats are not supported"); + f++; + } + + incr = (f - s) + MAXVARFMT; + qrealloc(&b, incr + i, &bsize); + c = *f; + if (c != '\0') f++; + c1 = *f; + *f = '\0'; + arg = evaluate(nextarg(&n), v); + + j = i; + if (c == 'c' || !c) { + i += sprintf(b+i, s, is_numeric(arg) ? + (char)getvar_i(arg) : *getvar_s(arg)); + } else if (c == 's') { + s1 = getvar_s(arg); + qrealloc(&b, incr+i+strlen(s1), &bsize); + i += sprintf(b+i, s, s1); + } else { + i += fmt_num(b+i, incr, s, getvar_i(arg), FALSE); + } + *f = c1; + + /* if there was an error while sprintf, return value is negative */ + if (i < j) i = j; + } + + b = xrealloc(b, i + 1); + free(fmt); + nvfree(v); + b[i] = '\0'; + return b; +} + +/* common substitution routine + * replace (nm) substring of (src) that match (n) with (repl), store + * result into (dest), return number of substitutions. If nm=0, replace + * all matches. If src or dst is NULL, use $0. If ex=TRUE, enable + * subexpression matching (\1-\9) + */ +static int awk_sub(node *rn, const char *repl, int nm, var *src, var *dest, int ex) +{ + char *ds = NULL; + const char *s; + const char *sp; + int c, i, j, di, rl, so, eo, nbs, n, dssize; + regmatch_t pmatch[10]; + regex_t sreg, *re; + + re = as_regex(rn, &sreg); + if (!src) src = intvar[F0]; + if (!dest) dest = intvar[F0]; + + i = di = 0; + sp = getvar_s(src); + rl = strlen(repl); + while (regexec(re, sp, 10, pmatch, sp==getvar_s(src) ? 0 : REG_NOTBOL) == 0) { + so = pmatch[0].rm_so; + eo = pmatch[0].rm_eo; + + qrealloc(&ds, di + eo + rl, &dssize); + memcpy(ds + di, sp, eo); + di += eo; + if (++i >= nm) { + /* replace */ + di -= (eo - so); + nbs = 0; + for (s = repl; *s; s++) { + ds[di++] = c = *s; + if (c == '\\') { + nbs++; + continue; + } + if (c == '&' || (ex && c >= '0' && c <= '9')) { + di -= ((nbs + 3) >> 1); + j = 0; + if (c != '&') { + j = c - '0'; + nbs++; + } + if (nbs % 2) { + ds[di++] = c; + } else { + n = pmatch[j].rm_eo - pmatch[j].rm_so; + qrealloc(&ds, di + rl + n, &dssize); + memcpy(ds + di, sp + pmatch[j].rm_so, n); + di += n; + } + } + nbs = 0; + } + } + + sp += eo; + if (i == nm) break; + if (eo == so) { + ds[di] = *sp++; + if (!ds[di++]) break; + } + } + + qrealloc(&ds, di + strlen(sp), &dssize); + strcpy(ds + di, sp); + setvar_p(dest, ds); + if (re == &sreg) regfree(re); + return i; +} + +static var *exec_builtin(node *op, var *res) +{ +#define tspl (G.exec_builtin__tspl) + + int (*to_xxx)(int); + var *tv; + node *an[4]; + var *av[4]; + const char *as[4]; + regmatch_t pmatch[2]; + regex_t sreg, *re; + node *spl; + uint32_t isr, info; + int nargs; + time_t tt; + char *s, *s1; + int i, l, ll, n; + + tv = nvalloc(4); + isr = info = op->info; + op = op->l.n; + + av[2] = av[3] = NULL; + for (i = 0; i < 4 && op; i++) { + an[i] = nextarg(&op); + if (isr & 0x09000000) av[i] = evaluate(an[i], &tv[i]); + if (isr & 0x08000000) as[i] = getvar_s(av[i]); + isr >>= 1; + } + + nargs = i; + if ((uint32_t)nargs < (info >> 30)) + syntax_error(EMSG_TOO_FEW_ARGS); + + switch (info & OPNMASK) { + + case B_a2: +#if ENABLE_FEATURE_AWK_LIBM + setvar_i(res, atan2(getvar_i(av[0]), getvar_i(av[1]))); +#else + syntax_error(EMSG_NO_MATH); +#endif + break; + + case B_sp: + if (nargs > 2) { + spl = (an[2]->info & OPCLSMASK) == OC_REGEXP ? + an[2] : mk_splitter(getvar_s(evaluate(an[2], &tv[2])), &tspl); + } else { + spl = &fsplitter.n; + } + + n = awk_split(as[0], spl, &s); + s1 = s; + clear_array(iamarray(av[1])); + for (i = 1; i <= n; i++) + setari_u(av[1], i, nextword(&s1)); + free(s); + setvar_i(res, n); + break; + + case B_ss: + l = strlen(as[0]); + i = getvar_i(av[1]) - 1; + if (i > l) i = l; + if (i < 0) i = 0; + n = (nargs > 2) ? getvar_i(av[2]) : l-i; + if (n < 0) n = 0; + s = xstrndup(as[0]+i, n); + setvar_p(res, s); + break; + + /* Bitwise ops must assume that operands are unsigned. GNU Awk 3.1.5: + * awk '{ print or(-1,1) }' gives "4.29497e+09", not "-2.xxxe+09" */ + case B_an: + setvar_i(res, getvar_i_int(av[0]) & getvar_i_int(av[1])); + break; + + case B_co: + setvar_i(res, ~getvar_i_int(av[0])); + break; + + case B_ls: + setvar_i(res, getvar_i_int(av[0]) << getvar_i_int(av[1])); + break; + + case B_or: + setvar_i(res, getvar_i_int(av[0]) | getvar_i_int(av[1])); + break; + + case B_rs: + setvar_i(res, getvar_i_int(av[0]) >> getvar_i_int(av[1])); + break; + + case B_xo: + setvar_i(res, getvar_i_int(av[0]) ^ getvar_i_int(av[1])); + break; + + case B_lo: + to_xxx = tolower; + goto lo_cont; + + case B_up: + to_xxx = toupper; + lo_cont: + s1 = s = xstrdup(as[0]); + while (*s1) { + *s1 = (*to_xxx)(*s1); + s1++; + } + setvar_p(res, s); + break; + + case B_ix: + n = 0; + ll = strlen(as[1]); + l = strlen(as[0]) - ll; + if (ll > 0 && l >= 0) { + if (!icase) { + s = strstr(as[0], as[1]); + if (s) n = (s - as[0]) + 1; + } else { + /* this piece of code is terribly slow and + * really should be rewritten + */ + for (i=0; i<=l; i++) { + if (strncasecmp(as[0]+i, as[1], ll) == 0) { + n = i+1; + break; + } + } + } + } + setvar_i(res, n); + break; + + case B_ti: + if (nargs > 1) + tt = getvar_i(av[1]); + else + time(&tt); + //s = (nargs > 0) ? as[0] : "%a %b %d %H:%M:%S %Z %Y"; + i = strftime(g_buf, MAXVARFMT, + ((nargs > 0) ? as[0] : "%a %b %d %H:%M:%S %Z %Y"), + localtime(&tt)); + g_buf[i] = '\0'; + setvar_s(res, g_buf); + break; + + case B_ma: + re = as_regex(an[1], &sreg); + n = regexec(re, as[0], 1, pmatch, 0); + if (n == 0) { + pmatch[0].rm_so++; + pmatch[0].rm_eo++; + } else { + pmatch[0].rm_so = 0; + pmatch[0].rm_eo = -1; + } + setvar_i(newvar("RSTART"), pmatch[0].rm_so); + setvar_i(newvar("RLENGTH"), pmatch[0].rm_eo - pmatch[0].rm_so); + setvar_i(res, pmatch[0].rm_so); + if (re == &sreg) regfree(re); + break; + + case B_ge: + awk_sub(an[0], as[1], getvar_i(av[2]), av[3], res, TRUE); + break; + + case B_gs: + setvar_i(res, awk_sub(an[0], as[1], 0, av[2], av[2], FALSE)); + break; + + case B_su: + setvar_i(res, awk_sub(an[0], as[1], 1, av[2], av[2], FALSE)); + break; + } + + nvfree(tv); + return res; +#undef tspl +} + +/* + * Evaluate node - the heart of the program. Supplied with subtree + * and place where to store result. returns ptr to result. + */ +#define XC(n) ((n) >> 8) + +static var *evaluate(node *op, var *res) +{ +/* This procedure is recursive so we should count every byte */ +#define fnargs (G.evaluate__fnargs) +/* seed is initialized to 1 */ +#define seed (G.evaluate__seed) +#define sreg (G.evaluate__sreg) + + node *op1; + var *v1; + union { + var *v; + const char *s; + double d; + int i; + } L, R; + uint32_t opinfo; + int opn; + union { + char *s; + rstream *rsm; + FILE *F; + var *v; + regex_t *re; + uint32_t info; + } X; + + if (!op) + return setvar_s(res, NULL); + + v1 = nvalloc(2); + + while (op) { + opinfo = op->info; + opn = (opinfo & OPNMASK); + g_lineno = op->lineno; + + /* execute inevitable things */ + op1 = op->l.n; + if (opinfo & OF_RES1) X.v = L.v = evaluate(op1, v1); + if (opinfo & OF_RES2) R.v = evaluate(op->r.n, v1+1); + if (opinfo & OF_STR1) L.s = getvar_s(L.v); + if (opinfo & OF_STR2) R.s = getvar_s(R.v); + if (opinfo & OF_NUM1) L.d = getvar_i(L.v); + + switch (XC(opinfo & OPCLSMASK)) { + + /* -- iterative node type -- */ + + /* test pattern */ + case XC( OC_TEST ): + if ((op1->info & OPCLSMASK) == OC_COMMA) { + /* it's range pattern */ + if ((opinfo & OF_CHECKED) || ptest(op1->l.n)) { + op->info |= OF_CHECKED; + if (ptest(op1->r.n)) + op->info &= ~OF_CHECKED; + + op = op->a.n; + } else { + op = op->r.n; + } + } else { + op = (ptest(op1)) ? op->a.n : op->r.n; + } + break; + + /* just evaluate an expression, also used as unconditional jump */ + case XC( OC_EXEC ): + break; + + /* branch, used in if-else and various loops */ + case XC( OC_BR ): + op = istrue(L.v) ? op->a.n : op->r.n; + break; + + /* initialize for-in loop */ + case XC( OC_WALKINIT ): + hashwalk_init(L.v, iamarray(R.v)); + break; + + /* get next array item */ + case XC( OC_WALKNEXT ): + op = hashwalk_next(L.v) ? op->a.n : op->r.n; + break; + + case XC( OC_PRINT ): + case XC( OC_PRINTF ): + X.F = stdout; + if (op->r.n) { + X.rsm = newfile(R.s); + if (!X.rsm->F) { + if (opn == '|') { + X.rsm->F = popen(R.s, "w"); + if (X.rsm->F == NULL) + bb_perror_msg_and_die("popen"); + X.rsm->is_pipe = 1; + } else { + X.rsm->F = xfopen(R.s, opn=='w' ? "w" : "a"); + } + } + X.F = X.rsm->F; + } + + if ((opinfo & OPCLSMASK) == OC_PRINT) { + if (!op1) { + fputs(getvar_s(intvar[F0]), X.F); + } else { + while (op1) { + L.v = evaluate(nextarg(&op1), v1); + if (L.v->type & VF_NUMBER) { + fmt_num(g_buf, MAXVARFMT, getvar_s(intvar[OFMT]), + getvar_i(L.v), TRUE); + fputs(g_buf, X.F); + } else { + fputs(getvar_s(L.v), X.F); + } + + if (op1) fputs(getvar_s(intvar[OFS]), X.F); + } + } + fputs(getvar_s(intvar[ORS]), X.F); + + } else { /* OC_PRINTF */ + L.s = awk_printf(op1); + fputs(L.s, X.F); + free((char*)L.s); + } + fflush(X.F); + break; + + case XC( OC_DELETE ): + X.info = op1->info & OPCLSMASK; + if (X.info == OC_VAR) { + R.v = op1->l.v; + } else if (X.info == OC_FNARG) { + R.v = &fnargs[op1->l.i]; + } else { + syntax_error(EMSG_NOT_ARRAY); + } + + if (op1->r.n) { + clrvar(L.v); + L.s = getvar_s(evaluate(op1->r.n, v1)); + hash_remove(iamarray(R.v), L.s); + } else { + clear_array(iamarray(R.v)); + } + break; + + case XC( OC_NEWSOURCE ): + g_progname = op->l.s; + break; + + case XC( OC_RETURN ): + copyvar(res, L.v); + break; + + case XC( OC_NEXTFILE ): + nextfile = TRUE; + case XC( OC_NEXT ): + nextrec = TRUE; + case XC( OC_DONE ): + clrvar(res); + break; + + case XC( OC_EXIT ): + awk_exit(L.d); + + /* -- recursive node type -- */ + + case XC( OC_VAR ): + L.v = op->l.v; + if (L.v == intvar[NF]) + split_f0(); + goto v_cont; + + case XC( OC_FNARG ): + L.v = &fnargs[op->l.i]; + v_cont: + res = op->r.n ? findvar(iamarray(L.v), R.s) : L.v; + break; + + case XC( OC_IN ): + setvar_i(res, hash_search(iamarray(R.v), L.s) ? 1 : 0); + break; + + case XC( OC_REGEXP ): + op1 = op; + L.s = getvar_s(intvar[F0]); + goto re_cont; + + case XC( OC_MATCH ): + op1 = op->r.n; + re_cont: + X.re = as_regex(op1, &sreg); + R.i = regexec(X.re, L.s, 0, NULL, 0); + if (X.re == &sreg) regfree(X.re); + setvar_i(res, (R.i == 0 ? 1 : 0) ^ (opn == '!' ? 1 : 0)); + break; + + case XC( OC_MOVE ): + /* if source is a temporary string, jusk relink it to dest */ + if (R.v == v1+1 && R.v->string) { + res = setvar_p(L.v, R.v->string); + R.v->string = NULL; + } else { + res = copyvar(L.v, R.v); + } + break; + + case XC( OC_TERNARY ): + if ((op->r.n->info & OPCLSMASK) != OC_COLON) + syntax_error(EMSG_POSSIBLE_ERROR); + res = evaluate(istrue(L.v) ? op->r.n->l.n : op->r.n->r.n, res); + break; + + case XC( OC_FUNC ): + if (!op->r.f->body.first) + syntax_error(EMSG_UNDEF_FUNC); + + X.v = R.v = nvalloc(op->r.f->nargs+1); + while (op1) { + L.v = evaluate(nextarg(&op1), v1); + copyvar(R.v, L.v); + R.v->type |= VF_CHILD; + R.v->x.parent = L.v; + if (++R.v - X.v >= op->r.f->nargs) + break; + } + + R.v = fnargs; + fnargs = X.v; + + L.s = g_progname; + res = evaluate(op->r.f->body.first, res); + g_progname = L.s; + + nvfree(fnargs); + fnargs = R.v; + break; + + case XC( OC_GETLINE ): + case XC( OC_PGETLINE ): + if (op1) { + X.rsm = newfile(L.s); + if (!X.rsm->F) { + if ((opinfo & OPCLSMASK) == OC_PGETLINE) { + X.rsm->F = popen(L.s, "r"); + X.rsm->is_pipe = TRUE; + } else { + X.rsm->F = fopen_for_read(L.s); /* not xfopen! */ + } + } + } else { + if (!iF) iF = next_input_file(); + X.rsm = iF; + } + + if (!X.rsm->F) { + setvar_i(intvar[ERRNO], errno); + setvar_i(res, -1); + break; + } + + if (!op->r.n) + R.v = intvar[F0]; + + L.i = awk_getline(X.rsm, R.v); + if (L.i > 0) { + if (!op1) { + incvar(intvar[FNR]); + incvar(intvar[NR]); + } + } + setvar_i(res, L.i); + break; + + /* simple builtins */ + case XC( OC_FBLTIN ): + switch (opn) { + + case F_in: + R.d = (int)L.d; + break; + + case F_rn: + R.d = (double)rand() / (double)RAND_MAX; + break; +#if ENABLE_FEATURE_AWK_LIBM + case F_co: + R.d = cos(L.d); + break; + + case F_ex: + R.d = exp(L.d); + break; + + case F_lg: + R.d = log(L.d); + break; + + case F_si: + R.d = sin(L.d); + break; + + case F_sq: + R.d = sqrt(L.d); + break; +#else + case F_co: + case F_ex: + case F_lg: + case F_si: + case F_sq: + syntax_error(EMSG_NO_MATH); + break; +#endif + case F_sr: + R.d = (double)seed; + seed = op1 ? (unsigned)L.d : (unsigned)time(NULL); + srand(seed); + break; + + case F_ti: + R.d = time(NULL); + break; + + case F_le: + if (!op1) + L.s = getvar_s(intvar[F0]); + R.d = strlen(L.s); + break; + + case F_sy: + fflush(NULL); + R.d = (ENABLE_FEATURE_ALLOW_EXEC && L.s && *L.s) + ? (system(L.s) >> 8) : 0; + break; + + case F_ff: + if (!op1) + fflush(stdout); + else { + if (L.s && *L.s) { + X.rsm = newfile(L.s); + fflush(X.rsm->F); + } else { + fflush(NULL); + } + } + break; + + case F_cl: + X.rsm = (rstream *)hash_search(fdhash, L.s); + if (X.rsm) { + R.i = X.rsm->is_pipe ? pclose(X.rsm->F) : fclose(X.rsm->F); + free(X.rsm->buffer); + hash_remove(fdhash, L.s); + } + if (R.i != 0) + setvar_i(intvar[ERRNO], errno); + R.d = (double)R.i; + break; + } + setvar_i(res, R.d); + break; + + case XC( OC_BUILTIN ): + res = exec_builtin(op, res); + break; + + case XC( OC_SPRINTF ): + setvar_p(res, awk_printf(op1)); + break; + + case XC( OC_UNARY ): + X.v = R.v; + L.d = R.d = getvar_i(R.v); + switch (opn) { + case 'P': + L.d = ++R.d; + goto r_op_change; + case 'p': + R.d++; + goto r_op_change; + case 'M': + L.d = --R.d; + goto r_op_change; + case 'm': + R.d--; + goto r_op_change; + case '!': + L.d = istrue(X.v) ? 0 : 1; + break; + case '-': + L.d = -R.d; + break; + r_op_change: + setvar_i(X.v, R.d); + } + setvar_i(res, L.d); + break; + + case XC( OC_FIELD ): + R.i = (int)getvar_i(R.v); + if (R.i == 0) { + res = intvar[F0]; + } else { + split_f0(); + if (R.i > nfields) + fsrealloc(R.i); + res = &Fields[R.i - 1]; + } + break; + + /* concatenation (" ") and index joining (",") */ + case XC( OC_CONCAT ): + case XC( OC_COMMA ): + opn = strlen(L.s) + strlen(R.s) + 2; + X.s = xmalloc(opn); + strcpy(X.s, L.s); + if ((opinfo & OPCLSMASK) == OC_COMMA) { + L.s = getvar_s(intvar[SUBSEP]); + X.s = xrealloc(X.s, opn + strlen(L.s)); + strcat(X.s, L.s); + } + strcat(X.s, R.s); + setvar_p(res, X.s); + break; + + case XC( OC_LAND ): + setvar_i(res, istrue(L.v) ? ptest(op->r.n) : 0); + break; + + case XC( OC_LOR ): + setvar_i(res, istrue(L.v) ? 1 : ptest(op->r.n)); + break; + + case XC( OC_BINARY ): + case XC( OC_REPLACE ): + R.d = getvar_i(R.v); + switch (opn) { + case '+': + L.d += R.d; + break; + case '-': + L.d -= R.d; + break; + case '*': + L.d *= R.d; + break; + case '/': + if (R.d == 0) syntax_error(EMSG_DIV_BY_ZERO); + L.d /= R.d; + break; + case '&': +#if ENABLE_FEATURE_AWK_LIBM + L.d = pow(L.d, R.d); +#else + syntax_error(EMSG_NO_MATH); +#endif + break; + case '%': + if (R.d == 0) syntax_error(EMSG_DIV_BY_ZERO); + L.d -= (int)(L.d / R.d) * R.d; + break; + } + res = setvar_i(((opinfo & OPCLSMASK) == OC_BINARY) ? res : X.v, L.d); + break; + + case XC( OC_COMPARE ): + if (is_numeric(L.v) && is_numeric(R.v)) { + L.d = getvar_i(L.v) - getvar_i(R.v); + } else { + L.s = getvar_s(L.v); + R.s = getvar_s(R.v); + L.d = icase ? strcasecmp(L.s, R.s) : strcmp(L.s, R.s); + } + switch (opn & 0xfe) { + case 0: + R.i = (L.d > 0); + break; + case 2: + R.i = (L.d >= 0); + break; + case 4: + R.i = (L.d == 0); + break; + } + setvar_i(res, (opn & 0x1 ? R.i : !R.i) ? 1 : 0); + break; + + default: + syntax_error(EMSG_POSSIBLE_ERROR); + } + if ((opinfo & OPCLSMASK) <= SHIFT_TIL_THIS) + op = op->a.n; + if ((opinfo & OPCLSMASK) >= RECUR_FROM_THIS) + break; + if (nextrec) + break; + } + nvfree(v1); + return res; +#undef fnargs +#undef seed +#undef sreg +} + + +/* -------- main & co. -------- */ + +static int awk_exit(int r) +{ + var tv; + unsigned i; + hash_item *hi; + + zero_out_var(&tv); + + if (!exiting) { + exiting = TRUE; + nextrec = FALSE; + evaluate(endseq.first, &tv); + } + + /* waiting for children */ + for (i = 0; i < fdhash->csize; i++) { + hi = fdhash->items[i]; + while (hi) { + if (hi->data.rs.F && hi->data.rs.is_pipe) + pclose(hi->data.rs.F); + hi = hi->next; + } + } + + exit(r); +} + +/* if expr looks like "var=value", perform assignment and return 1, + * otherwise return 0 */ +static int is_assignment(const char *expr) +{ + char *exprc, *s, *s0, *s1; + + exprc = xstrdup(expr); + if (!isalnum_(*exprc) || (s = strchr(exprc, '=')) == NULL) { + free(exprc); + return FALSE; + } + + *(s++) = '\0'; + s0 = s1 = s; + while (*s) + *(s1++) = nextchar(&s); + + *s1 = '\0'; + setvar_u(newvar(exprc), s0); + free(exprc); + return TRUE; +} + +/* switch to next input file */ +static rstream *next_input_file(void) +{ +#define rsm (G.next_input_file__rsm) +#define files_happen (G.next_input_file__files_happen) + + FILE *F = NULL; + const char *fname, *ind; + + if (rsm.F) fclose(rsm.F); + rsm.F = NULL; + rsm.pos = rsm.adv = 0; + + do { + if (getvar_i(intvar[ARGIND])+1 >= getvar_i(intvar[ARGC])) { + if (files_happen) + return NULL; + fname = "-"; + F = stdin; + } else { + ind = getvar_s(incvar(intvar[ARGIND])); + fname = getvar_s(findvar(iamarray(intvar[ARGV]), ind)); + if (fname && *fname && !is_assignment(fname)) + F = xfopen_stdin(fname); + } + } while (!F); + + files_happen = TRUE; + setvar_s(intvar[FILENAME], fname); + rsm.F = F; + return &rsm; +#undef rsm +#undef files_happen +} + +int awk_main(int argc, char **argv) MAIN_EXTERNALLY_VISIBLE; +int awk_main(int argc, char **argv) +{ + unsigned opt; + char *opt_F, *opt_W; + llist_t *list_v = NULL; + llist_t *list_f = NULL; + int i, j; + var *v; + var tv; + char **envp; + char *vnames = (char *)vNames; /* cheat */ + char *vvalues = (char *)vValues; + + INIT_G(); + + /* Undo busybox.c, or else strtod may eat ','! This breaks parsing: + * $1,$2 == '$1,' '$2', NOT '$1' ',' '$2' */ + if (ENABLE_LOCALE_SUPPORT) + setlocale(LC_NUMERIC, "C"); + + zero_out_var(&tv); + + /* allocate global buffer */ + g_buf = xmalloc(MAXVARFMT + 1); + + vhash = hash_init(); + ahash = hash_init(); + fdhash = hash_init(); + fnhash = hash_init(); + + /* initialize variables */ + for (i = 0; *vnames; i++) { + intvar[i] = v = newvar(nextword(&vnames)); + if (*vvalues != '\377') + setvar_s(v, nextword(&vvalues)); + else + setvar_i(v, 0); + + if (*vnames == '*') { + v->type |= VF_SPECIAL; + vnames++; + } + } + + handle_special(intvar[FS]); + handle_special(intvar[RS]); + + newfile("/dev/stdin")->F = stdin; + newfile("/dev/stdout")->F = stdout; + newfile("/dev/stderr")->F = stderr; + + /* Huh, people report that sometimes environ is NULL. Oh well. */ + if (environ) for (envp = environ; *envp; envp++) { + /* environ is writable, thus we don't strdup it needlessly */ + char *s = *envp; + char *s1 = strchr(s, '='); + if (s1) { + *s1 = '\0'; + /* Both findvar and setvar_u take const char* + * as 2nd arg -> environment is not trashed */ + setvar_u(findvar(iamarray(intvar[ENVIRON]), s), s1 + 1); + *s1 = '='; + } + } + opt_complementary = "v::f::"; /* -v and -f can occur multiple times */ + opt = getopt32(argv, "F:v:f:W:", &opt_F, &list_v, &list_f, &opt_W); + argv += optind; + argc -= optind; + if (opt & 0x1) + setvar_s(intvar[FS], opt_F); // -F + while (list_v) { /* -v */ + if (!is_assignment(llist_pop(&list_v))) + bb_show_usage(); + } + if (list_f) { /* -f */ + do { + char *s = NULL; + FILE *from_file; + + g_progname = llist_pop(&list_f); + from_file = xfopen_stdin(g_progname); + /* one byte is reserved for some trick in next_token */ + for (i = j = 1; j > 0; i += j) { + s = xrealloc(s, i + 4096); + j = fread(s + i, 1, 4094, from_file); + } + s[i] = '\0'; + fclose(from_file); + parse_program(s + 1); + free(s); + } while (list_f); + } else { // no -f: take program from 1st parameter + if (!argc) + bb_show_usage(); + g_progname = "cmd. line"; + parse_program(*argv++); + argc--; + } + if (opt & 0x8) // -W + bb_error_msg("warning: unrecognized option '-W %s' ignored", opt_W); + + /* fill in ARGV array */ + setvar_i(intvar[ARGC], argc + 1); + setari_u(intvar[ARGV], 0, "awk"); + i = 0; + while (*argv) + setari_u(intvar[ARGV], ++i, *argv++); + + evaluate(beginseq.first, &tv); + if (!mainseq.first && !endseq.first) + awk_exit(EXIT_SUCCESS); + + /* input file could already be opened in BEGIN block */ + if (!iF) iF = next_input_file(); + + /* passing through input files */ + while (iF) { + nextfile = FALSE; + setvar_i(intvar[FNR], 0); + + while ((i = awk_getline(iF, intvar[F0])) > 0) { + nextrec = FALSE; + incvar(intvar[NR]); + incvar(intvar[FNR]); + evaluate(mainseq.first, &tv); + + if (nextfile) + break; + } + + if (i < 0) + syntax_error(strerror(errno)); + + iF = next_input_file(); + } + + awk_exit(EXIT_SUCCESS); + /*return 0;*/ +} diff --git a/editors/cmp.c b/editors/cmp.c new file mode 100644 index 0000000..2e98e6e --- /dev/null +++ b/editors/cmp.c @@ -0,0 +1,135 @@ +/* vi: set sw=4 ts=4: */ +/* + * Mini cmp implementation for busybox + * + * Copyright (C) 2000,2001 by Matt Kraai <kraai@alumni.carnegiemellon.edu> + * + * Licensed under GPLv2 or later, see file LICENSE in this tarball for details. + */ + +/* BB_AUDIT SUSv3 (virtually) compliant -- uses nicer GNU format for -l. */ +/* http://www.opengroup.org/onlinepubs/007904975/utilities/cmp.html */ + +/* Mar 16, 2003 Manuel Novoa III (mjn3@codepoet.org) + * + * Original version majorly reworked for SUSv3 compliance, bug fixes, and + * size optimizations. Changes include: + * 1) Now correctly distinguishes between errors and actual file differences. + * 2) Proper handling of '-' args. + * 3) Actual error checking of i/o. + * 4) Accept SUSv3 -l option. Note that we use the slightly nicer gnu format + * in the '-l' case. + */ + +#include "libbb.h" + +static const char fmt_eof[] ALIGN1 = "cmp: EOF on %s\n"; +static const char fmt_differ[] ALIGN1 = "%s %s differ: char %"OFF_FMT"d, line %d\n"; +// This fmt_l_opt uses gnu-isms. SUSv3 would be "%.0s%.0s%"OFF_FMT"d %o %o\n" +static const char fmt_l_opt[] ALIGN1 = "%.0s%.0s%"OFF_FMT"d %3o %3o\n"; + +static const char opt_chars[] ALIGN1 = "sl"; +#define CMP_OPT_s (1<<0) +#define CMP_OPT_l (1<<1) + +int cmp_main(int argc, char **argv) MAIN_EXTERNALLY_VISIBLE; +int cmp_main(int argc UNUSED_PARAM, char **argv) +{ + FILE *fp1, *fp2, *outfile = stdout; + const char *filename1, *filename2 = "-"; + USE_DESKTOP(off_t skip1 = 0, skip2 = 0;) + off_t char_pos = 0; + int line_pos = 1; /* Hopefully won't overflow... */ + const char *fmt; + int c1, c2; + unsigned opt; + int retval = 0; + + xfunc_error_retval = 2; /* 1 is returned if files are different. */ + + opt_complementary = "-1" + USE_DESKTOP(":?4") + SKIP_DESKTOP(":?2") + ":l--s:s--l"; + opt = getopt32(argv, opt_chars); + argv += optind; + + filename1 = *argv; + fp1 = xfopen_stdin(filename1); + + if (*++argv) { + filename2 = *argv; +#if ENABLE_DESKTOP + if (*++argv) { + skip1 = XATOOFF(*argv); + if (*++argv) { + skip2 = XATOOFF(*argv); + } + } +#endif + } + + fp2 = xfopen_stdin(filename2); + if (fp1 == fp2) { /* Paranoia check... stdin == stdin? */ + /* Note that we don't bother reading stdin. Neither does gnu wc. + * But perhaps we should, so that other apps down the chain don't + * get the input. Consider 'echo hello | (cmp - - && cat -)'. + */ + return 0; + } + + if (opt & CMP_OPT_l) + fmt = fmt_l_opt; + else + fmt = fmt_differ; + +#if ENABLE_DESKTOP + while (skip1) { getc(fp1); skip1--; } + while (skip2) { getc(fp2); skip2--; } +#endif + do { + c1 = getc(fp1); + c2 = getc(fp2); + ++char_pos; + if (c1 != c2) { /* Remember: a read error may have occurred. */ + retval = 1; /* But assume the files are different for now. */ + if (c2 == EOF) { + /* We know that fp1 isn't at EOF or in an error state. But to + * save space below, things are setup to expect an EOF in fp1 + * if an EOF occurred. So, swap things around. + */ + fp1 = fp2; + filename1 = filename2; + c1 = c2; + } + if (c1 == EOF) { + die_if_ferror(fp1, filename1); + fmt = fmt_eof; /* Well, no error, so it must really be EOF. */ + outfile = stderr; + /* There may have been output to stdout (option -l), so + * make sure we fflush before writing to stderr. */ + xfflush_stdout(); + } + if (!(opt & CMP_OPT_s)) { + if (opt & CMP_OPT_l) { + line_pos = c1; /* line_pos is unused in the -l case. */ + } + fprintf(outfile, fmt, filename1, filename2, char_pos, line_pos, c2); + if (opt) { /* This must be -l since not -s. */ + /* If we encountered an EOF, + * the while check will catch it. */ + continue; + } + } + break; + } + if (c1 == '\n') { + ++line_pos; + } + } while (c1 != EOF); + + die_if_ferror(fp1, filename1); + die_if_ferror(fp2, filename2); + + fflush_stdout_and_exit(retval); +} diff --git a/editors/diff.c b/editors/diff.c new file mode 100644 index 0000000..0e96c84 --- /dev/null +++ b/editors/diff.c @@ -0,0 +1,1344 @@ +/* vi: set sw=4 ts=4: */ +/* + * Mini diff implementation for busybox, adapted from OpenBSD diff. + * + * Copyright (C) 2006 by Robert Sullivan <cogito.ergo.cogito@hotmail.com> + * Copyright (c) 2003 Todd C. Miller <Todd.Miller@courtesan.com> + * + * Sponsored in part by the Defense Advanced Research Projects + * Agency (DARPA) and Air Force Research Laboratory, Air Force + * Materiel Command, USAF, under agreement number F39502-99-1-0512. + * + * Licensed under GPLv2 or later, see file LICENSE in this tarball for details. + */ + +#include "libbb.h" + +// #define FSIZE_MAX 32768 + +/* NOINLINEs added to prevent gcc from merging too much into diffreg() + * (it bites more than it can (efficiently) chew). */ + +/* + * Output flags + */ +enum { + /* Print a header/footer between files */ + /* D_HEADER = 1, - unused */ + /* Treat file as empty (/dev/null) */ + D_EMPTY1 = 2 * ENABLE_FEATURE_DIFF_DIR, + D_EMPTY2 = 4 * ENABLE_FEATURE_DIFF_DIR, +}; + +/* + * Status values for print_status() and diffreg() return values + * Guide: + * D_SAME - files are the same + * D_DIFFER - files differ + * D_BINARY - binary files differ + * D_COMMON - subdirectory common to both dirs + * D_ONLY - file only exists in one dir + * D_ISDIR1 - path1 a dir, path2 a file + * D_ISDIR2 - path1 a file, path2 a dir + * D_ERROR - error occurred + * D_SKIPPED1 - skipped path1 as it is a special file + * D_SKIPPED2 - skipped path2 as it is a special file + */ +#define D_SAME 0 +#define D_DIFFER (1 << 0) +#define D_BINARY (1 << 1) +#define D_COMMON (1 << 2) +/*#define D_ONLY (1 << 3) - unused */ +#define D_ISDIR1 (1 << 4) +#define D_ISDIR2 (1 << 5) +#define D_ERROR (1 << 6) +#define D_SKIPPED1 (1 << 7) +#define D_SKIPPED2 (1 << 8) + +/* Command line options */ +#define FLAG_a (1 << 0) +#define FLAG_b (1 << 1) +#define FLAG_d (1 << 2) +#define FLAG_i (1 << 3) +#define FLAG_L (1 << 4) +#define FLAG_N (1 << 5) +#define FLAG_q (1 << 6) +#define FLAG_r (1 << 7) +#define FLAG_s (1 << 8) +#define FLAG_S (1 << 9) +#define FLAG_t (1 << 10) +#define FLAG_T (1 << 11) +#define FLAG_U (1 << 12) +#define FLAG_w (1 << 13) + + +struct cand { + int x; + int y; + int pred; +}; + +struct line { + int serial; + int value; +}; + +/* + * The following struct is used to record change information + * doing a "context" or "unified" diff. (see routine "change" to + * understand the highly mnemonic field names) + */ +struct context_vec { + int a; /* start line in old file */ + int b; /* end line in old file */ + int c; /* start line in new file */ + int d; /* end line in new file */ +}; + + +#define g_read_buf bb_common_bufsiz1 + +struct globals { + bool anychange; + smallint exit_status; + int opt_U_context; + size_t max_context; /* size of context_vec_start */ + USE_FEATURE_DIFF_DIR(int dl_count;) + USE_FEATURE_DIFF_DIR(char **dl;) + char *opt_S_start; + const char *label1; + const char *label2; + int *J; /* will be overlaid on class */ + int clen; + int pref, suff; /* length of prefix and suffix */ + int nlen[2]; + int slen[2]; + int clistlen; /* the length of clist */ + struct cand *clist; /* merely a free storage pot for candidates */ + long *ixnew; /* will be overlaid on nfile[1] */ + long *ixold; /* will be overlaid on klist */ + struct line *nfile[2]; + struct line *sfile[2]; /* shortened by pruning common prefix/suffix */ + struct context_vec *context_vec_start; + struct context_vec *context_vec_end; + struct context_vec *context_vec_ptr; + char *tempname1, *tempname2; + struct stat stb1, stb2; +}; +#define G (*ptr_to_globals) +#define anychange (G.anychange ) +#define exit_status (G.exit_status ) +#define opt_U_context (G.opt_U_context ) +#define max_context (G.max_context ) +#define dl_count (G.dl_count ) +#define dl (G.dl ) +#define opt_S_start (G.opt_S_start ) +#define label1 (G.label1 ) +#define label2 (G.label2 ) +#define J (G.J ) +#define clen (G.clen ) +#define pref (G.pref ) +#define suff (G.suff ) +#define nlen (G.nlen ) +#define slen (G.slen ) +#define clistlen (G.clistlen ) +#define clist (G.clist ) +#define ixnew (G.ixnew ) +#define ixold (G.ixold ) +#define nfile (G.nfile ) +#define sfile (G.sfile ) +#define context_vec_start (G.context_vec_start ) +#define context_vec_end (G.context_vec_end ) +#define context_vec_ptr (G.context_vec_ptr ) +#define stb1 (G.stb1 ) +#define stb2 (G.stb2 ) +#define tempname1 (G.tempname1 ) +#define tempname2 (G.tempname2 ) +#define INIT_G() do { \ + SET_PTR_TO_GLOBALS(xzalloc(sizeof(G))); \ + opt_U_context = 3; \ + max_context = 64; \ +} while (0) + + +#if ENABLE_FEATURE_DIFF_DIR +static void print_only(const char *path, const char *entry) +{ + printf("Only in %s: %s\n", path, entry); +} +#endif + + +static void print_status(int val, char *_path1, char *_path2) +{ + /*const char *const _entry = entry ? entry : "";*/ + /*char *const _path1 = entry ? concat_path_file(path1, _entry) : path1;*/ + /*char *const _path2 = entry ? concat_path_file(path2, _entry) : path2;*/ + + switch (val) { +/* case D_ONLY: + print_only(path1, entry); + break; +*/ + case D_COMMON: + printf("Common subdirectories: %s and %s\n", _path1, _path2); + break; + case D_BINARY: + printf("Binary files %s and %s differ\n", _path1, _path2); + break; + case D_DIFFER: + if (option_mask32 & FLAG_q) + printf("Files %s and %s differ\n", _path1, _path2); + break; + case D_SAME: + if (option_mask32 & FLAG_s) + printf("Files %s and %s are identical\n", _path1, _path2); + break; + case D_ISDIR1: + printf("File %s is a %s while file %s is a %s\n", + _path1, "directory", _path2, "regular file"); + break; + case D_ISDIR2: + printf("File %s is a %s while file %s is a %s\n", + _path1, "regular file", _path2, "directory"); + break; + case D_SKIPPED1: + printf("File %s is not a regular file or directory and was skipped\n", + _path1); + break; + case D_SKIPPED2: + printf("File %s is not a regular file or directory and was skipped\n", + _path2); + break; + } +/* + if (entry) { + free(_path1); + free(_path2); + } +*/ +} + + +/* Read line, return its nonzero hash. Return 0 if EOF. + * + * Hash function taken from Robert Sedgewick, Algorithms in C, 3d ed., p 578. + */ +static ALWAYS_INLINE int fiddle_sum(int sum, int t) +{ + return sum * 127 + t; +} +static int readhash(FILE *fp) +{ + int i, t, space; + int sum; + + sum = 1; + space = 0; + i = 0; + if (!(option_mask32 & (FLAG_b | FLAG_w))) { + while ((t = getc(fp)) != '\n') { + if (t == EOF) { + if (i == 0) + return 0; + break; + } + sum = fiddle_sum(sum, t); + i = 1; + } + } else { + while (1) { + switch (t = getc(fp)) { + case '\t': + case '\r': + case '\v': + case '\f': + case ' ': + space = 1; + continue; + default: + if (space && !(option_mask32 & FLAG_w)) { + i = 1; + space = 0; + } + sum = fiddle_sum(sum, t); + i = 1; + continue; + case EOF: + if (i == 0) + return 0; + /* FALLTHROUGH */ + case '\n': + break; + } + break; + } + } + /* + * There is a remote possibility that we end up with a zero sum. + * Zero is used as an EOF marker, so return 1 instead. + */ + return (sum == 0 ? 1 : sum); +} + + +/* Our diff implementation is using seek. + * When we meet non-seekable file, we must make a temp copy. + */ +static char *make_temp(FILE *f, struct stat *sb) +{ + char *name; + int fd; + + if (S_ISREG(sb->st_mode) || S_ISBLK(sb->st_mode)) + return NULL; + name = xstrdup("/tmp/difXXXXXX"); + fd = mkstemp(name); + if (fd < 0) + bb_perror_msg_and_die("mkstemp"); + if (bb_copyfd_eof(fileno(f), fd) < 0) { + clean_up: + unlink(name); + xfunc_die(); /* error message is printed by bb_copyfd_eof */ + } + fstat(fd, sb); + close(fd); + if (freopen(name, "r+", f) == NULL) { + bb_perror_msg("freopen"); + goto clean_up; + } + return name; +} + + +/* + * Check to see if the given files differ. + * Returns 0 if they are the same, 1 if different, and -1 on error. + */ +static NOINLINE int files_differ(FILE *f1, FILE *f2) +{ + size_t i, j; + + /* Prevent making copies for "/dev/null" (too common) */ + /* Deal with input from pipes etc */ + tempname1 = make_temp(f1, &stb1); + tempname2 = make_temp(f2, &stb2); + if (stb1.st_size != stb2.st_size) { + return 1; + } + while (1) { + i = fread(g_read_buf, 1, COMMON_BUFSIZE/2, f1); + j = fread(g_read_buf + COMMON_BUFSIZE/2, 1, COMMON_BUFSIZE/2, f2); + if (i != j) + return 1; + if (i == 0) + return (ferror(f1) || ferror(f2)) ? -1 : 0; + if (memcmp(g_read_buf, + g_read_buf + COMMON_BUFSIZE/2, i) != 0) + return 1; + } +} + + +static void prepare(int i, FILE *fp /*, off_t filesize*/) +{ + struct line *p; + int h; + size_t j, sz; + + rewind(fp); + + /*sz = (filesize <= FSIZE_MAX ? filesize : FSIZE_MAX) / 25;*/ + /*if (sz < 100)*/ + sz = 100; + + p = xmalloc((sz + 3) * sizeof(p[0])); + j = 0; + while ((h = readhash(fp)) != 0) { /* while not EOF */ + if (j == sz) { + sz = sz * 3 / 2; + p = xrealloc(p, (sz + 3) * sizeof(p[0])); + } + p[++j].value = h; + } + nlen[i] = j; + nfile[i] = p; +} + + +static void prune(void) +{ + int i, j; + + for (pref = 0; pref < nlen[0] && pref < nlen[1] && + nfile[0][pref + 1].value == nfile[1][pref + 1].value; pref++) + continue; + for (suff = 0; suff < nlen[0] - pref && suff < nlen[1] - pref && + nfile[0][nlen[0] - suff].value == nfile[1][nlen[1] - suff].value; + suff++) + continue; + for (j = 0; j < 2; j++) { + sfile[j] = nfile[j] + pref; + slen[j] = nlen[j] - pref - suff; + for (i = 0; i <= slen[j]; i++) + sfile[j][i].serial = i; + } +} + + +static void equiv(struct line *a, int n, struct line *b, int m, int *c) +{ + int i, j; + + i = j = 1; + while (i <= n && j <= m) { + if (a[i].value < b[j].value) + a[i++].value = 0; + else if (a[i].value == b[j].value) + a[i++].value = j; + else + j++; + } + while (i <= n) + a[i++].value = 0; + b[m + 1].value = 0; + j = 0; + while (++j <= m) { + c[j] = -b[j].serial; + while (b[j + 1].value == b[j].value) { + j++; + c[j] = b[j].serial; + } + } + c[j] = -1; +} + + +static int isqrt(int n) +{ + int y, x; + + if (n == 0) + return 0; + x = 1; + do { + y = x; + x = n / x; + x += y; + x /= 2; + } while ((x - y) > 1 || (x - y) < -1); + + return x; +} + + +static int newcand(int x, int y, int pred) +{ + struct cand *q; + + if (clen == clistlen) { + clistlen = clistlen * 11 / 10; + clist = xrealloc(clist, clistlen * sizeof(struct cand)); + } + q = clist + clen; + q->x = x; + q->y = y; + q->pred = pred; + return clen++; +} + + +static int search(int *c, int k, int y) +{ + int i, j, l, t; + + if (clist[c[k]].y < y) /* quick look for typical case */ + return k + 1; + i = 0; + j = k + 1; + while (1) { + l = i + j; + if ((l >>= 1) <= i) + break; + t = clist[c[l]].y; + if (t > y) + j = l; + else if (t < y) + i = l; + else + return l; + } + return l + 1; +} + + +static int stone(int *a, int n, int *b, int *c) +{ + int i, k, y, j, l; + int oldc, tc, oldl; + unsigned int numtries; +#if ENABLE_FEATURE_DIFF_MINIMAL + const unsigned int bound = + (option_mask32 & FLAG_d) ? UINT_MAX : MAX(256, isqrt(n)); +#else + const unsigned int bound = MAX(256, isqrt(n)); +#endif + + k = 0; + c[0] = newcand(0, 0, 0); + for (i = 1; i <= n; i++) { + j = a[i]; + if (j == 0) + continue; + y = -b[j]; + oldl = 0; + oldc = c[0]; + numtries = 0; + do { + if (y <= clist[oldc].y) + continue; + l = search(c, k, y); + if (l != oldl + 1) + oldc = c[l - 1]; + if (l <= k) { + if (clist[c[l]].y <= y) + continue; + tc = c[l]; + c[l] = newcand(i, y, oldc); + oldc = tc; + oldl = l; + numtries++; + } else { + c[l] = newcand(i, y, oldc); + k++; + break; + } + } while ((y = b[++j]) > 0 && numtries < bound); + } + return k; +} + + +static void unravel(int p) +{ + struct cand *q; + int i; + + for (i = 0; i <= nlen[0]; i++) + J[i] = i <= pref ? i : i > nlen[0] - suff ? i + nlen[1] - nlen[0] : 0; + for (q = clist + p; q->y != 0; q = clist + q->pred) + J[q->x + pref] = q->y + pref; +} + + +static void unsort(struct line *f, int l, int *b) +{ + int *a, i; + + a = xmalloc((l + 1) * sizeof(int)); + for (i = 1; i <= l; i++) + a[f[i].serial] = f[i].value; + for (i = 1; i <= l; i++) + b[i] = a[i]; + free(a); +} + + +static int skipline(FILE *f) +{ + int i, c; + + for (i = 1; (c = getc(f)) != '\n' && c != EOF; i++) + continue; + return i; +} + + +/* + * Check does double duty: + * 1. ferret out any fortuitous correspondences due + * to confounding by hashing (which result in "jackpot") + * 2. collect random access indexes to the two files + */ +static NOINLINE void check(FILE *f1, FILE *f2) +{ + int i, j, jackpot, c, d; + long ctold, ctnew; + + rewind(f1); + rewind(f2); + j = 1; + ixold[0] = ixnew[0] = 0; + jackpot = 0; + ctold = ctnew = 0; + for (i = 1; i <= nlen[0]; i++) { + if (J[i] == 0) { + ixold[i] = ctold += skipline(f1); + continue; + } + while (j < J[i]) { + ixnew[j] = ctnew += skipline(f2); + j++; + } + if (option_mask32 & (FLAG_b | FLAG_w | FLAG_i)) { + while (1) { + c = getc(f1); + d = getc(f2); + /* + * GNU diff ignores a missing newline + * in one file if bflag || wflag. + */ + if ((option_mask32 & (FLAG_b | FLAG_w)) + && ((c == EOF && d == '\n') || (c == '\n' && d == EOF)) + ) { + break; + } + ctold++; + ctnew++; + if ((option_mask32 & FLAG_b) && isspace(c) && isspace(d)) { + do { + if (c == '\n') + break; + ctold++; + c = getc(f1); + } while (isspace(c)); + do { + if (d == '\n') + break; + ctnew++; + d = getc(f2); + } while (isspace(d)); + } else if (option_mask32 & FLAG_w) { + while (isspace(c) && c != '\n') { + c = getc(f1); + ctold++; + } + while (isspace(d) && d != '\n') { + d = getc(f2); + ctnew++; + } + } + if (c != d) { + jackpot++; + J[i] = 0; + if (c != '\n' && c != EOF) + ctold += skipline(f1); + if (d != '\n' && c != EOF) + ctnew += skipline(f2); + break; + } + if (c == '\n' || c == EOF) + break; + } + } else { + while (1) { + ctold++; + ctnew++; + c = getc(f1); + d = getc(f2); + if (c != d) { + J[i] = 0; + if (c != '\n' && c != EOF) + ctold += skipline(f1); +/* was buggy? "if (d != '\n' && c != EOF)" */ + if (d != '\n' && d != EOF) + ctnew += skipline(f2); + break; + } + if (c == '\n' || c == EOF) + break; + } + } + ixold[i] = ctold; + ixnew[j] = ctnew; + j++; + } + for (; j <= nlen[1]; j++) + ixnew[j] = ctnew += skipline(f2); +} + + +/* shellsort CACM #201 */ +static void sort(struct line *a, int n) +{ + struct line *ai, *aim, w; + int j, m = 0, k; + + if (n == 0) + return; + for (j = 1; j <= n; j *= 2) + m = 2 * j - 1; + for (m /= 2; m != 0; m /= 2) { + k = n - m; + for (j = 1; j <= k; j++) { + for (ai = &a[j]; ai > a; ai -= m) { + aim = &ai[m]; + if (aim < ai) + break; /* wraparound */ + if (aim->value > ai[0].value + || (aim->value == ai[0].value && aim->serial > ai[0].serial) + ) { + break; + } + w.value = ai[0].value; + ai[0].value = aim->value; + aim->value = w.value; + w.serial = ai[0].serial; + ai[0].serial = aim->serial; + aim->serial = w.serial; + } + } + } +} + + +static void uni_range(int a, int b) +{ + if (a < b) + printf("%d,%d", a, b - a + 1); + else if (a == b) + printf("%d", b); + else + printf("%d,0", b); +} + + +static void fetch(long *f, int a, int b, FILE *lb, int ch) +{ + int i, j, c, lastc, col, nc; + + if (a > b) + return; + for (i = a; i <= b; i++) { + fseek(lb, f[i - 1], SEEK_SET); + nc = f[i] - f[i - 1]; + if (ch != '\0') { + putchar(ch); + if (option_mask32 & FLAG_T) + putchar('\t'); + } + col = 0; + for (j = 0, lastc = '\0'; j < nc; j++, lastc = c) { + c = getc(lb); + if (c == EOF) { + printf("\n\\ No newline at end of file\n"); + return; + } + if (c == '\t' && (option_mask32 & FLAG_t)) { + do { + putchar(' '); + } while (++col & 7); + } else { + putchar(c); + col++; + } + } + } +} + + +#if ENABLE_FEATURE_DIFF_BINARY +static int asciifile(FILE *f) +{ + int i, cnt; + + if (option_mask32 & FLAG_a) + return 1; + rewind(f); + cnt = fread(g_read_buf, 1, COMMON_BUFSIZE, f); + for (i = 0; i < cnt; i++) { + if (!isprint(g_read_buf[i]) + && !isspace(g_read_buf[i]) + ) { + return 0; + } + } + return 1; +} +#else +#define asciifile(f) 1 +#endif + + +/* dump accumulated "unified" diff changes */ +static void dump_unified_vec(FILE *f1, FILE *f2) +{ + struct context_vec *cvp = context_vec_start; + int lowa, upb, lowc, upd; + int a, b, c, d; + char ch; + + if (context_vec_start > context_vec_ptr) + return; + + b = d = 0; /* gcc */ + lowa = MAX(1, cvp->a - opt_U_context); + upb = MIN(nlen[0], context_vec_ptr->b + opt_U_context); + lowc = MAX(1, cvp->c - opt_U_context); + upd = MIN(nlen[1], context_vec_ptr->d + opt_U_context); + + printf("@@ -"); + uni_range(lowa, upb); + printf(" +"); + uni_range(lowc, upd); + printf(" @@\n"); + + /* + * Output changes in "unified" diff format--the old and new lines + * are printed together. + */ + for (; cvp <= context_vec_ptr; cvp++) { + a = cvp->a; + b = cvp->b; + c = cvp->c; + d = cvp->d; + + /* + * c: both new and old changes + * d: only changes in the old file + * a: only changes in the new file + */ + if (a <= b && c <= d) + ch = 'c'; + else + ch = (a <= b) ? 'd' : 'a'; +#if 0 + switch (ch) { + case 'c': +// fetch() seeks! + fetch(ixold, lowa, a - 1, f1, ' '); + fetch(ixold, a, b, f1, '-'); + fetch(ixnew, c, d, f2, '+'); + break; + case 'd': + fetch(ixold, lowa, a - 1, f1, ' '); + fetch(ixold, a, b, f1, '-'); + break; + case 'a': + fetch(ixnew, lowc, c - 1, f2, ' '); + fetch(ixnew, c, d, f2, '+'); + break; + } +#else + if (ch == 'c' || ch == 'd') { + fetch(ixold, lowa, a - 1, f1, ' '); + fetch(ixold, a, b, f1, '-'); + } + if (ch == 'a') + fetch(ixnew, lowc, c - 1, f2, ' '); + if (ch == 'c' || ch == 'a') + fetch(ixnew, c, d, f2, '+'); +#endif + lowa = b + 1; + lowc = d + 1; + } + fetch(ixnew, d + 1, upd, f2, ' '); + + context_vec_ptr = context_vec_start - 1; +} + + +static void print_header(const char *file1, const char *file2) +{ + if (label1) + printf("--- %s\n", label1); + else + printf("--- %s\t%s", file1, ctime(&stb1.st_mtime)); + if (label2) + printf("+++ %s\n", label2); + else + printf("+++ %s\t%s", file2, ctime(&stb2.st_mtime)); +} + + +/* + * Indicate that there is a difference between lines a and b of the from file + * to get to lines c to d of the to file. If a is greater than b then there + * are no lines in the from file involved and this means that there were + * lines appended (beginning at b). If c is greater than d then there are + * lines missing from the to file. + */ +static void change(char *file1, FILE *f1, char *file2, FILE *f2, + int a, int b, int c, int d) +{ + if ((a > b && c > d) || (option_mask32 & FLAG_q)) { + anychange = 1; + return; + } + + /* + * Allocate change records as needed. + */ + if (context_vec_ptr == context_vec_end - 1) { + ptrdiff_t offset = context_vec_ptr - context_vec_start; + + max_context <<= 1; + context_vec_start = xrealloc(context_vec_start, + max_context * sizeof(struct context_vec)); + context_vec_end = context_vec_start + max_context; + context_vec_ptr = context_vec_start + offset; + } + if (anychange == 0) { + /* + * Print the context/unidiff header first time through. + */ + print_header(file1, file2); + } else if (a > context_vec_ptr->b + (2 * opt_U_context) + 1 + && c > context_vec_ptr->d + (2 * opt_U_context) + 1 + ) { + /* + * If this change is more than 'context' lines from the + * previous change, dump the record and reset it. + */ +// dump_unified_vec() seeks! + dump_unified_vec(f1, f2); + } + context_vec_ptr++; + context_vec_ptr->a = a; + context_vec_ptr->b = b; + context_vec_ptr->c = c; + context_vec_ptr->d = d; + anychange = 1; +} + + +static void output(char *file1, FILE *f1, char *file2, FILE *f2) +{ + /* Note that j0 and j1 can't be used as they are defined in math.h. + * This also allows the rather amusing variable 'j00'... */ + int m, i0, i1, j00, j01; + + rewind(f1); + rewind(f2); + m = nlen[0]; + J[0] = 0; + J[m + 1] = nlen[1] + 1; + for (i0 = 1; i0 <= m; i0 = i1 + 1) { + while (i0 <= m && J[i0] == J[i0 - 1] + 1) + i0++; + j00 = J[i0 - 1] + 1; + i1 = i0 - 1; + while (i1 < m && J[i1 + 1] == 0) + i1++; + j01 = J[i1 + 1] - 1; + J[i1] = j01; +// change() seeks! + change(file1, f1, file2, f2, i0, i1, j00, j01); + } + if (m == 0) { +// change() seeks! + change(file1, f1, file2, f2, 1, 0, 1, nlen[1]); + } + if (anychange != 0 && !(option_mask32 & FLAG_q)) { +// dump_unified_vec() seeks! + dump_unified_vec(f1, f2); + } +} + +/* + * The following code uses an algorithm due to Harold Stone, + * which finds a pair of longest identical subsequences in + * the two files. + * + * The major goal is to generate the match vector J. + * J[i] is the index of the line in file1 corresponding + * to line i in file0. J[i] = 0 if there is no + * such line in file1. + * + * Lines are hashed so as to work in core. All potential + * matches are located by sorting the lines of each file + * on the hash (called "value"). In particular, this + * collects the equivalence classes in file1 together. + * Subroutine equiv replaces the value of each line in + * file0 by the index of the first element of its + * matching equivalence in (the reordered) file1. + * To save space equiv squeezes file1 into a single + * array member in which the equivalence classes + * are simply concatenated, except that their first + * members are flagged by changing sign. + * + * Next the indices that point into member are unsorted into + * array class according to the original order of file0. + * + * The cleverness lies in routine stone. This marches + * through the lines of file0, developing a vector klist + * of "k-candidates". At step i a k-candidate is a matched + * pair of lines x,y (x in file0, y in file1) such that + * there is a common subsequence of length k + * between the first i lines of file0 and the first y + * lines of file1, but there is no such subsequence for + * any smaller y. x is the earliest possible mate to y + * that occurs in such a subsequence. + * + * Whenever any of the members of the equivalence class of + * lines in file1 matable to a line in file0 has serial number + * less than the y of some k-candidate, that k-candidate + * with the smallest such y is replaced. The new + * k-candidate is chained (via pred) to the current + * k-1 candidate so that the actual subsequence can + * be recovered. When a member has serial number greater + * that the y of all k-candidates, the klist is extended. + * At the end, the longest subsequence is pulled out + * and placed in the array J by unravel + * + * With J in hand, the matches there recorded are + * checked against reality to assure that no spurious + * matches have crept in due to hashing. If they have, + * they are broken, and "jackpot" is recorded--a harmless + * matter except that a true match for a spuriously + * mated line may now be unnecessarily reported as a change. + * + * Much of the complexity of the program comes simply + * from trying to minimize core utilization and + * maximize the range of doable problems by dynamically + * allocating what is needed and reusing what is not. + * The core requirements for problems larger than somewhat + * are (in words) 2*length(file0) + length(file1) + + * 3*(number of k-candidates installed), typically about + * 6n words for files of length n. + */ +/* NB: files can be not REGular. The only sure thing that they + * are not both DIRectories. */ +static unsigned diffreg(char *file1, char *file2, int flags) +{ + int *member; /* will be overlaid on nfile[1] */ + int *class; /* will be overlaid on nfile[0] */ + int *klist; /* will be overlaid on nfile[0] after class */ + FILE *f1; + FILE *f2; + unsigned rval; + int i; + + anychange = 0; + context_vec_ptr = context_vec_start - 1; + tempname1 = tempname2 = NULL; + + /* Is any of them a directory? Then it's simple */ + if (S_ISDIR(stb1.st_mode) != S_ISDIR(stb2.st_mode)) + return (S_ISDIR(stb1.st_mode) ? D_ISDIR1 : D_ISDIR2); + + /* None of them are directories */ + rval = D_SAME; + + if (flags & D_EMPTY1) + /* can't be stdin, but xfopen_stdin() is smaller code */ + f1 = xfopen_stdin(bb_dev_null); + else + f1 = xfopen_stdin(file1); + if (flags & D_EMPTY2) + f2 = xfopen_stdin(bb_dev_null); + else + f2 = xfopen_stdin(file2); + + /* NB: if D_EMPTY1/2 is set, other file is always a regular file, + * not pipe/fifo/chardev/etc - D_EMPTY is used by "diff -r" only, + * and it never diffs non-ordinary files in subdirs. */ + if (!(flags & (D_EMPTY1 | D_EMPTY2))) { + /* Quick check whether they are different */ + /* NB: copies non-REG files to tempfiles and fills tempname1/2 */ + i = files_differ(f1, f2); + if (i != 1) { /* not different? */ + if (i != 0) /* error? */ + exit_status |= 2; + goto closem; + } + } + + if (!asciifile(f1) || !asciifile(f2)) { + rval = D_BINARY; + exit_status |= 1; + goto closem; + } + +// Rewind inside! + prepare(0, f1 /*, stb1.st_size*/); + prepare(1, f2 /*, stb2.st_size*/); + prune(); + sort(sfile[0], slen[0]); + sort(sfile[1], slen[1]); + + member = (int *) nfile[1]; + equiv(sfile[0], slen[0], sfile[1], slen[1], member); +//TODO: xrealloc_vector? + member = xrealloc(member, (slen[1] + 2) * sizeof(int)); + + class = (int *) nfile[0]; + unsort(sfile[0], slen[0], class); + class = xrealloc(class, (slen[0] + 2) * sizeof(int)); + + klist = xmalloc((slen[0] + 2) * sizeof(int)); + clen = 0; + clistlen = 100; + clist = xmalloc(clistlen * sizeof(struct cand)); + i = stone(class, slen[0], member, klist); + free(member); + free(class); + + J = xrealloc(J, (nlen[0] + 2) * sizeof(int)); + unravel(klist[i]); + free(clist); + free(klist); + + ixold = xrealloc(ixold, (nlen[0] + 2) * sizeof(long)); + ixnew = xrealloc(ixnew, (nlen[1] + 2) * sizeof(long)); +// Rewind inside! + check(f1, f2); +// Rewind inside! + output(file1, f1, file2, f2); + + closem: + if (anychange) { + exit_status |= 1; + if (rval == D_SAME) + rval = D_DIFFER; + } + fclose_if_not_stdin(f1); + fclose_if_not_stdin(f2); + if (tempname1) { + unlink(tempname1); + free(tempname1); + } + if (tempname2) { + unlink(tempname2); + free(tempname2); + } + return rval; +} + + +#if ENABLE_FEATURE_DIFF_DIR +static void do_diff(char *dir1, char *path1, char *dir2, char *path2) +{ + int flags = 0; /*D_HEADER;*/ + int val; + char *fullpath1 = NULL; /* if -N */ + char *fullpath2 = NULL; + + if (path1) + fullpath1 = concat_path_file(dir1, path1); + if (path2) + fullpath2 = concat_path_file(dir2, path2); + + if (!fullpath1 || stat(fullpath1, &stb1) != 0) { + flags |= D_EMPTY1; + memset(&stb1, 0, sizeof(stb1)); + if (path2) { + free(fullpath1); + fullpath1 = concat_path_file(dir1, path2); + } + } + if (!fullpath2 || stat(fullpath2, &stb2) != 0) { + flags |= D_EMPTY2; + memset(&stb2, 0, sizeof(stb2)); + stb2.st_mode = stb1.st_mode; + if (path1) { + free(fullpath2); + fullpath2 = concat_path_file(dir2, path1); + } + } + + if (stb1.st_mode == 0) + stb1.st_mode = stb2.st_mode; + + if (S_ISDIR(stb1.st_mode) && S_ISDIR(stb2.st_mode)) { + printf("Common subdirectories: %s and %s\n", fullpath1, fullpath2); + goto ret; + } + + if (!S_ISREG(stb1.st_mode) && !S_ISDIR(stb1.st_mode)) + val = D_SKIPPED1; + else if (!S_ISREG(stb2.st_mode) && !S_ISDIR(stb2.st_mode)) + val = D_SKIPPED2; + else { + /* Both files are either REGular or DIRectories */ + val = diffreg(fullpath1, fullpath2, flags); + } + + print_status(val, fullpath1, fullpath2 /*, NULL*/); + ret: + free(fullpath1); + free(fullpath2); +} +#endif + + +#if ENABLE_FEATURE_DIFF_DIR +/* This function adds a filename to dl, the directory listing. */ +static int FAST_FUNC add_to_dirlist(const char *filename, + struct stat *sb UNUSED_PARAM, + void *userdata, + int depth UNUSED_PARAM) +{ + dl = xrealloc_vector(dl, 5, dl_count); + dl[dl_count] = xstrdup(filename + (int)(ptrdiff_t)userdata); + dl_count++; + return TRUE; +} + + +/* This returns a sorted directory listing. */ +static char **get_recursive_dirlist(char *path) +{ + dl_count = 0; + dl = xzalloc(sizeof(dl[0])); + + /* We need to trim root directory prefix. + * Using void *userdata to specify its length, + * add_to_dirlist will remove it. */ + if (option_mask32 & FLAG_r) { + recursive_action(path, ACTION_RECURSE|ACTION_FOLLOWLINKS, + add_to_dirlist, /* file_action */ + NULL, /* dir_action */ + (void*)(ptrdiff_t)(strlen(path) + 1), + 0); + } else { + DIR *dp; + struct dirent *ep; + + dp = warn_opendir(path); + while ((ep = readdir(dp))) { + if (!strcmp(ep->d_name, "..") || LONE_CHAR(ep->d_name, '.')) + continue; + add_to_dirlist(ep->d_name, NULL, (void*)(int)0, 0); + } + closedir(dp); + } + + /* Sort dl alphabetically. */ + qsort_string_vector(dl, dl_count); + + dl[dl_count] = NULL; + return dl; +} + + +static void diffdir(char *p1, char *p2) +{ + char **dirlist1, **dirlist2; + char *dp1, *dp2; + int pos; + + /* Check for trailing slashes. */ + dp1 = last_char_is(p1, '/'); + if (dp1 != NULL) + *dp1 = '\0'; + dp2 = last_char_is(p2, '/'); + if (dp2 != NULL) + *dp2 = '\0'; + + /* Get directory listings for p1 and p2. */ + dirlist1 = get_recursive_dirlist(p1); + dirlist2 = get_recursive_dirlist(p2); + + /* If -S was set, find the starting point. */ + if (opt_S_start) { + while (*dirlist1 != NULL && strcmp(*dirlist1, opt_S_start) < 0) + dirlist1++; + while (*dirlist2 != NULL && strcmp(*dirlist2, opt_S_start) < 0) + dirlist2++; + if ((*dirlist1 == NULL) || (*dirlist2 == NULL)) + bb_error_msg(bb_msg_invalid_arg, "NULL", "-S"); + } + + /* Now that both dirlist1 and dirlist2 contain sorted directory + * listings, we can start to go through dirlist1. If both listings + * contain the same file, then do a normal diff. Otherwise, behaviour + * is determined by whether the -N flag is set. */ + while (*dirlist1 != NULL || *dirlist2 != NULL) { + dp1 = *dirlist1; + dp2 = *dirlist2; + pos = dp1 == NULL ? 1 : (dp2 == NULL ? -1 : strcmp(dp1, dp2)); + if (pos == 0) { + do_diff(p1, dp1, p2, dp2); + dirlist1++; + dirlist2++; + } else if (pos < 0) { + if (option_mask32 & FLAG_N) + do_diff(p1, dp1, p2, NULL); + else + print_only(p1, dp1); + dirlist1++; + } else { + if (option_mask32 & FLAG_N) + do_diff(p1, NULL, p2, dp2); + else + print_only(p2, dp2); + dirlist2++; + } + } +} +#endif + + +int diff_main(int argc, char **argv) MAIN_EXTERNALLY_VISIBLE; +int diff_main(int argc UNUSED_PARAM, char **argv) +{ + int gotstdin = 0; + char *f1, *f2; + llist_t *L_arg = NULL; + + INIT_G(); + + /* exactly 2 params; collect multiple -L <label>; -U N */ + opt_complementary = "=2:L::U+"; + getopt32(argv, "abdiL:NqrsS:tTU:wu" + "p" /* ignored (for compatibility) */, + &L_arg, &opt_S_start, &opt_U_context); + /*argc -= optind;*/ + argv += optind; + while (L_arg) { + if (label1 && label2) + bb_show_usage(); + if (label1) /* then label2 is NULL */ + label2 = label1; + label1 = llist_pop(&L_arg); + } + + /* + * Do sanity checks, fill in stb1 and stb2 and call the appropriate + * driver routine. Both drivers use the contents of stb1 and stb2. + */ + f1 = argv[0]; + f2 = argv[1]; + if (LONE_DASH(f1)) { + fstat(STDIN_FILENO, &stb1); + gotstdin++; + } else + xstat(f1, &stb1); + if (LONE_DASH(f2)) { + fstat(STDIN_FILENO, &stb2); + gotstdin++; + } else + xstat(f2, &stb2); + + if (gotstdin && (S_ISDIR(stb1.st_mode) || S_ISDIR(stb2.st_mode))) + bb_error_msg_and_die("can't compare stdin to a directory"); + + if (S_ISDIR(stb1.st_mode) && S_ISDIR(stb2.st_mode)) { +#if ENABLE_FEATURE_DIFF_DIR + diffdir(f1, f2); + return exit_status; +#else + bb_error_msg_and_die("no support for directory comparison"); +#endif + } + + if (S_ISDIR(stb1.st_mode)) { /* "diff dir file" */ + /* NB: "diff dir dir2/dir3/file" must become + * "diff dir/file dir2/dir3/file" */ + char *slash = strrchr(f2, '/'); + f1 = concat_path_file(f1, slash ? slash + 1 : f2); + xstat(f1, &stb1); + } + if (S_ISDIR(stb2.st_mode)) { + char *slash = strrchr(f1, '/'); + f2 = concat_path_file(f2, slash ? slash + 1 : f1); + xstat(f2, &stb2); + } + + /* diffreg can get non-regular files here, + * they are not both DIRestories */ + print_status((gotstdin > 1 ? D_SAME : diffreg(f1, f2, 0)), + f1, f2 /*, NULL*/); + return exit_status; +} diff --git a/editors/ed.c b/editors/ed.c new file mode 100644 index 0000000..9084a17 --- /dev/null +++ b/editors/ed.c @@ -0,0 +1,1049 @@ +/* vi: set sw=4 ts=4: */ +/* + * Copyright (c) 2002 by David I. Bell + * Permission is granted to use, distribute, or modify this source, + * provided that this copyright notice remains intact. + * + * The "ed" built-in command (much simplified) + */ + +#include "libbb.h" + +typedef struct LINE { + struct LINE *next; + struct LINE *prev; + int len; + char data[1]; +} LINE; + + +#define searchString bb_common_bufsiz1 + +enum { + USERSIZE = sizeof(searchString) > 1024 ? 1024 + : sizeof(searchString) - 1, /* max line length typed in by user */ + INITBUF_SIZE = 1024, /* initial buffer size */ +}; + +struct globals { + int curNum; + int lastNum; + int bufUsed; + int bufSize; + LINE *curLine; + char *bufBase; + char *bufPtr; + char *fileName; + LINE lines; + smallint dirty; + int marks[26]; +}; +#define G (*ptr_to_globals) +#define curLine (G.curLine ) +#define bufBase (G.bufBase ) +#define bufPtr (G.bufPtr ) +#define fileName (G.fileName ) +#define curNum (G.curNum ) +#define lastNum (G.lastNum ) +#define bufUsed (G.bufUsed ) +#define bufSize (G.bufSize ) +#define dirty (G.dirty ) +#define lines (G.lines ) +#define marks (G.marks ) +#define INIT_G() do { \ + SET_PTR_TO_GLOBALS(xzalloc(sizeof(G))); \ +} while (0) + + +static void doCommands(void); +static void subCommand(const char *cmd, int num1, int num2); +static int getNum(const char **retcp, smallint *retHaveNum, int *retNum); +static int setCurNum(int num); +static void addLines(int num); +static int insertLine(int num, const char *data, int len); +static void deleteLines(int num1, int num2); +static int printLines(int num1, int num2, int expandFlag); +static int writeLines(const char *file, int num1, int num2); +static int readLines(const char *file, int num); +static int searchLines(const char *str, int num1, int num2); +static LINE *findLine(int num); +static int findString(const LINE *lp, const char * str, int len, int offset); + + +static int bad_nums(int num1, int num2, const char *for_what) +{ + if ((num1 < 1) || (num2 > lastNum) || (num1 > num2)) { + bb_error_msg("bad line range for %s", for_what); + return 1; + } + return 0; +} + + +static char *skip_blank(const char *cp) +{ + while (isblank(*cp)) + cp++; + return (char *)cp; +} + + +int ed_main(int argc, char **argv) MAIN_EXTERNALLY_VISIBLE; +int ed_main(int argc UNUSED_PARAM, char **argv) +{ + INIT_G(); + + bufSize = INITBUF_SIZE; + bufBase = xmalloc(bufSize); + bufPtr = bufBase; + lines.next = &lines; + lines.prev = &lines; + + if (argv[1]) { + fileName = xstrdup(argv[1]); + if (!readLines(fileName, 1)) { + return EXIT_SUCCESS; + } + if (lastNum) + setCurNum(1); + dirty = FALSE; + } + + doCommands(); + return EXIT_SUCCESS; +} + +/* + * Read commands until we are told to stop. + */ +static void doCommands(void) +{ + const char *cp; + char *endbuf, buf[USERSIZE]; + int len, num1, num2; + smallint have1, have2; + + while (TRUE) { + /* Returns: + * -1 on read errors or EOF, or on bare Ctrl-D. + * 0 on ctrl-C, + * >0 length of input string, including terminating '\n' + */ + len = read_line_input(": ", buf, sizeof(buf), NULL); + if (len <= 0) + return; + endbuf = &buf[len - 1]; + while ((endbuf > buf) && isblank(endbuf[-1])) + endbuf--; + *endbuf = '\0'; + + cp = skip_blank(buf); + have1 = FALSE; + have2 = FALSE; + + if ((curNum == 0) && (lastNum > 0)) { + curNum = 1; + curLine = lines.next; + } + + if (!getNum(&cp, &have1, &num1)) + continue; + + cp = skip_blank(cp); + + if (*cp == ',') { + cp++; + if (!getNum(&cp, &have2, &num2)) + continue; + if (!have1) + num1 = 1; + if (!have2) + num2 = lastNum; + have1 = TRUE; + have2 = TRUE; + } + if (!have1) + num1 = curNum; + if (!have2) + num2 = num1; + + switch (*cp++) { + case 'a': + addLines(num1 + 1); + break; + + case 'c': + deleteLines(num1, num2); + addLines(num1); + break; + + case 'd': + deleteLines(num1, num2); + break; + + case 'f': + if (*cp && !isblank(*cp)) { + bb_error_msg("bad file command"); + break; + } + cp = skip_blank(cp); + if (*cp == '\0') { + if (fileName) + printf("\"%s\"\n", fileName); + else + printf("No file name\n"); + break; + } + free(fileName); + fileName = xstrdup(cp); + break; + + case 'i': + addLines(num1); + break; + + case 'k': + cp = skip_blank(cp); + if ((*cp < 'a') || (*cp > 'z') || cp[1]) { + bb_error_msg("bad mark name"); + break; + } + marks[*cp - 'a'] = num2; + break; + + case 'l': + printLines(num1, num2, TRUE); + break; + + case 'p': + printLines(num1, num2, FALSE); + break; + + case 'q': + cp = skip_blank(cp); + if (have1 || *cp) { + bb_error_msg("bad quit command"); + break; + } + if (!dirty) + return; + len = read_line_input("Really quit? ", buf, 16, NULL); + /* read error/EOF - no way to continue */ + if (len < 0) + return; + cp = skip_blank(buf); + if ((*cp | 0x20) == 'y') /* Y or y */ + return; + break; + + case 'r': + if (*cp && !isblank(*cp)) { + bb_error_msg("bad read command"); + break; + } + cp = skip_blank(cp); + if (*cp == '\0') { + bb_error_msg("no file name"); + break; + } + if (!have1) + num1 = lastNum; + if (readLines(cp, num1 + 1)) + break; + if (fileName == NULL) + fileName = xstrdup(cp); + break; + + case 's': + subCommand(cp, num1, num2); + break; + + case 'w': + if (*cp && !isblank(*cp)) { + bb_error_msg("bad write command"); + break; + } + cp = skip_blank(cp); + if (!have1) { + num1 = 1; + num2 = lastNum; + } + if (*cp == '\0') + cp = fileName; + if (cp == NULL) { + bb_error_msg("no file name specified"); + break; + } + writeLines(cp, num1, num2); + break; + + case 'z': + switch (*cp) { + case '-': + printLines(curNum - 21, curNum, FALSE); + break; + case '.': + printLines(curNum - 11, curNum + 10, FALSE); + break; + default: + printLines(curNum, curNum + 21, FALSE); + break; + } + break; + + case '.': + if (have1) { + bb_error_msg("no arguments allowed"); + break; + } + printLines(curNum, curNum, FALSE); + break; + + case '-': + if (setCurNum(curNum - 1)) + printLines(curNum, curNum, FALSE); + break; + + case '=': + printf("%d\n", num1); + break; + case '\0': + if (have1) { + printLines(num2, num2, FALSE); + break; + } + if (setCurNum(curNum + 1)) + printLines(curNum, curNum, FALSE); + break; + + default: + bb_error_msg("unimplemented command"); + break; + } + } +} + + +/* + * Do the substitute command. + * The current line is set to the last substitution done. + */ +static void subCommand(const char *cmd, int num1, int num2) +{ + char *cp, *oldStr, *newStr, buf[USERSIZE]; + int delim, oldLen, newLen, deltaLen, offset; + LINE *lp, *nlp; + int globalFlag, printFlag, didSub, needPrint; + + if (bad_nums(num1, num2, "substitute")) + return; + + globalFlag = FALSE; + printFlag = FALSE; + didSub = FALSE; + needPrint = FALSE; + + /* + * Copy the command so we can modify it. + */ + strcpy(buf, cmd); + cp = buf; + + if (isblank(*cp) || (*cp == '\0')) { + bb_error_msg("bad delimiter for substitute"); + return; + } + + delim = *cp++; + oldStr = cp; + + cp = strchr(cp, delim); + if (cp == NULL) { + bb_error_msg("missing 2nd delimiter for substitute"); + return; + } + + *cp++ = '\0'; + + newStr = cp; + cp = strchr(cp, delim); + + if (cp) + *cp++ = '\0'; + else + cp = (char*)""; + + while (*cp) switch (*cp++) { + case 'g': + globalFlag = TRUE; + break; + case 'p': + printFlag = TRUE; + break; + default: + bb_error_msg("unknown option for substitute"); + return; + } + + if (*oldStr == '\0') { + if (searchString[0] == '\0') { + bb_error_msg("no previous search string"); + return; + } + oldStr = searchString; + } + + if (oldStr != searchString) + strcpy(searchString, oldStr); + + lp = findLine(num1); + if (lp == NULL) + return; + + oldLen = strlen(oldStr); + newLen = strlen(newStr); + deltaLen = newLen - oldLen; + offset = 0; + nlp = NULL; + + while (num1 <= num2) { + offset = findString(lp, oldStr, oldLen, offset); + + if (offset < 0) { + if (needPrint) { + printLines(num1, num1, FALSE); + needPrint = FALSE; + } + offset = 0; + lp = lp->next; + num1++; + continue; + } + + needPrint = printFlag; + didSub = TRUE; + dirty = TRUE; + + /* + * If the replacement string is the same size or shorter + * than the old string, then the substitution is easy. + */ + if (deltaLen <= 0) { + memcpy(&lp->data[offset], newStr, newLen); + if (deltaLen) { + memcpy(&lp->data[offset + newLen], + &lp->data[offset + oldLen], + lp->len - offset - oldLen); + + lp->len += deltaLen; + } + offset += newLen; + if (globalFlag) + continue; + if (needPrint) { + printLines(num1, num1, FALSE); + needPrint = FALSE; + } + lp = lp->next; + num1++; + continue; + } + + /* + * The new string is larger, so allocate a new line + * structure and use that. Link it in in place of + * the old line structure. + */ + nlp = malloc(sizeof(LINE) + lp->len + deltaLen); + if (nlp == NULL) { + bb_error_msg("cannot get memory for line"); + return; + } + + nlp->len = lp->len + deltaLen; + + memcpy(nlp->data, lp->data, offset); + memcpy(&nlp->data[offset], newStr, newLen); + memcpy(&nlp->data[offset + newLen], + &lp->data[offset + oldLen], + lp->len - offset - oldLen); + + nlp->next = lp->next; + nlp->prev = lp->prev; + nlp->prev->next = nlp; + nlp->next->prev = nlp; + + if (curLine == lp) + curLine = nlp; + + free(lp); + lp = nlp; + + offset += newLen; + + if (globalFlag) + continue; + + if (needPrint) { + printLines(num1, num1, FALSE); + needPrint = FALSE; + } + + lp = lp->next; + num1++; + } + + if (!didSub) + bb_error_msg("no substitutions found for \"%s\"", oldStr); +} + + +/* + * Search a line for the specified string starting at the specified + * offset in the line. Returns the offset of the found string, or -1. + */ +static int findString(const LINE *lp, const char *str, int len, int offset) +{ + int left; + const char *cp, *ncp; + + cp = &lp->data[offset]; + left = lp->len - offset; + + while (left >= len) { + ncp = memchr(cp, *str, left); + if (ncp == NULL) + return -1; + left -= (ncp - cp); + if (left < len) + return -1; + cp = ncp; + if (memcmp(cp, str, len) == 0) + return (cp - lp->data); + cp++; + left--; + } + + return -1; +} + + +/* + * Add lines which are typed in by the user. + * The lines are inserted just before the specified line number. + * The lines are terminated by a line containing a single dot (ugly!), + * or by an end of file. + */ +static void addLines(int num) +{ + int len; + char buf[USERSIZE + 1]; + + while (1) { + /* Returns: + * -1 on read errors or EOF, or on bare Ctrl-D. + * 0 on ctrl-C, + * >0 length of input string, including terminating '\n' + */ + len = read_line_input("", buf, sizeof(buf), NULL); + if (len <= 0) { + /* Previously, ctrl-C was exiting to shell. + * Now we exit to ed prompt. Is in important? */ + return; + } + if ((buf[0] == '.') && (buf[1] == '\n') && (buf[2] == '\0')) + return; + if (!insertLine(num++, buf, len)) + return; + } +} + + +/* + * Parse a line number argument if it is present. This is a sum + * or difference of numbers, '.', '$', 'x, or a search string. + * Returns TRUE if successful (whether or not there was a number). + * Returns FALSE if there was a parsing error, with a message output. + * Whether there was a number is returned indirectly, as is the number. + * The character pointer which stopped the scan is also returned. + */ +static int getNum(const char **retcp, smallint *retHaveNum, int *retNum) +{ + const char *cp; + char *endStr, str[USERSIZE]; + int value, num; + smallint haveNum, minus; + + cp = *retcp; + value = 0; + haveNum = FALSE; + minus = 0; + + while (TRUE) { + cp = skip_blank(cp); + + switch (*cp) { + case '.': + haveNum = TRUE; + num = curNum; + cp++; + break; + + case '$': + haveNum = TRUE; + num = lastNum; + cp++; + break; + + case '\'': + cp++; + if ((*cp < 'a') || (*cp > 'z')) { + bb_error_msg("bad mark name"); + return FALSE; + } + haveNum = TRUE; + num = marks[*cp++ - 'a']; + break; + + case '/': + strcpy(str, ++cp); + endStr = strchr(str, '/'); + if (endStr) { + *endStr++ = '\0'; + cp += (endStr - str); + } else + cp = ""; + num = searchLines(str, curNum, lastNum); + if (num == 0) + return FALSE; + haveNum = TRUE; + break; + + default: + if (!isdigit(*cp)) { + *retcp = cp; + *retHaveNum = haveNum; + *retNum = value; + return TRUE; + } + num = 0; + while (isdigit(*cp)) + num = num * 10 + *cp++ - '0'; + haveNum = TRUE; + break; + } + + value += (minus ? -num : num); + + cp = skip_blank(cp); + + switch (*cp) { + case '-': + minus = 1; + cp++; + break; + + case '+': + minus = 0; + cp++; + break; + + default: + *retcp = cp; + *retHaveNum = haveNum; + *retNum = value; + return TRUE; + } + } +} + + +/* + * Read lines from a file at the specified line number. + * Returns TRUE if the file was successfully read. + */ +static int readLines(const char *file, int num) +{ + int fd, cc; + int len, lineCount, charCount; + char *cp; + + if ((num < 1) || (num > lastNum + 1)) { + bb_error_msg("bad line for read"); + return FALSE; + } + + fd = open(file, 0); + if (fd < 0) { + perror(file); + return FALSE; + } + + bufPtr = bufBase; + bufUsed = 0; + lineCount = 0; + charCount = 0; + cc = 0; + + printf("\"%s\", ", file); + fflush(stdout); + + do { + cp = memchr(bufPtr, '\n', bufUsed); + + if (cp) { + len = (cp - bufPtr) + 1; + if (!insertLine(num, bufPtr, len)) { + close(fd); + return FALSE; + } + bufPtr += len; + bufUsed -= len; + charCount += len; + lineCount++; + num++; + continue; + } + + if (bufPtr != bufBase) { + memcpy(bufBase, bufPtr, bufUsed); + bufPtr = bufBase + bufUsed; + } + + if (bufUsed >= bufSize) { + len = (bufSize * 3) / 2; + cp = realloc(bufBase, len); + if (cp == NULL) { + bb_error_msg("no memory for buffer"); + close(fd); + return FALSE; + } + bufBase = cp; + bufPtr = bufBase + bufUsed; + bufSize = len; + } + + cc = safe_read(fd, bufPtr, bufSize - bufUsed); + bufUsed += cc; + bufPtr = bufBase; + + } while (cc > 0); + + if (cc < 0) { + perror(file); + close(fd); + return FALSE; + } + + if (bufUsed) { + if (!insertLine(num, bufPtr, bufUsed)) { + close(fd); + return -1; + } + lineCount++; + charCount += bufUsed; + } + + close(fd); + + printf("%d lines%s, %d chars\n", lineCount, + (bufUsed ? " (incomplete)" : ""), charCount); + + return TRUE; +} + + +/* + * Write the specified lines out to the specified file. + * Returns TRUE if successful, or FALSE on an error with a message output. + */ +static int writeLines(const char *file, int num1, int num2) +{ + LINE *lp; + int fd, lineCount, charCount; + + if (bad_nums(num1, num2, "write")) + return FALSE; + + lineCount = 0; + charCount = 0; + + fd = creat(file, 0666); + if (fd < 0) { + perror(file); + return FALSE; + } + + printf("\"%s\", ", file); + fflush(stdout); + + lp = findLine(num1); + if (lp == NULL) { + close(fd); + return FALSE; + } + + while (num1++ <= num2) { + if (full_write(fd, lp->data, lp->len) != lp->len) { + perror(file); + close(fd); + return FALSE; + } + charCount += lp->len; + lineCount++; + lp = lp->next; + } + + if (close(fd) < 0) { + perror(file); + return FALSE; + } + + printf("%d lines, %d chars\n", lineCount, charCount); + return TRUE; +} + + +/* + * Print lines in a specified range. + * The last line printed becomes the current line. + * If expandFlag is TRUE, then the line is printed specially to + * show magic characters. + */ +static int printLines(int num1, int num2, int expandFlag) +{ + const LINE *lp; + const char *cp; + int ch, count; + + if (bad_nums(num1, num2, "print")) + return FALSE; + + lp = findLine(num1); + if (lp == NULL) + return FALSE; + + while (num1 <= num2) { + if (!expandFlag) { + write(STDOUT_FILENO, lp->data, lp->len); + setCurNum(num1++); + lp = lp->next; + continue; + } + + /* + * Show control characters and characters with the + * high bit set specially. + */ + cp = lp->data; + count = lp->len; + + if ((count > 0) && (cp[count - 1] == '\n')) + count--; + + while (count-- > 0) { + ch = (unsigned char) *cp++; + fputc_printable(ch | PRINTABLE_META, stdout); + } + + fputs("$\n", stdout); + + setCurNum(num1++); + lp = lp->next; + } + + return TRUE; +} + + +/* + * Insert a new line with the specified text. + * The line is inserted so as to become the specified line, + * thus pushing any existing and further lines down one. + * The inserted line is also set to become the current line. + * Returns TRUE if successful. + */ +static int insertLine(int num, const char *data, int len) +{ + LINE *newLp, *lp; + + if ((num < 1) || (num > lastNum + 1)) { + bb_error_msg("inserting at bad line number"); + return FALSE; + } + + newLp = malloc(sizeof(LINE) + len - 1); + if (newLp == NULL) { + bb_error_msg("failed to allocate memory for line"); + return FALSE; + } + + memcpy(newLp->data, data, len); + newLp->len = len; + + if (num > lastNum) + lp = &lines; + else { + lp = findLine(num); + if (lp == NULL) { + free((char *) newLp); + return FALSE; + } + } + + newLp->next = lp; + newLp->prev = lp->prev; + lp->prev->next = newLp; + lp->prev = newLp; + + lastNum++; + dirty = TRUE; + return setCurNum(num); +} + + +/* + * Delete lines from the given range. + */ +static void deleteLines(int num1, int num2) +{ + LINE *lp, *nlp, *plp; + int count; + + if (bad_nums(num1, num2, "delete")) + return; + + lp = findLine(num1); + if (lp == NULL) + return; + + if ((curNum >= num1) && (curNum <= num2)) { + if (num2 < lastNum) + setCurNum(num2 + 1); + else if (num1 > 1) + setCurNum(num1 - 1); + else + curNum = 0; + } + + count = num2 - num1 + 1; + if (curNum > num2) + curNum -= count; + lastNum -= count; + + while (count-- > 0) { + nlp = lp->next; + plp = lp->prev; + plp->next = nlp; + nlp->prev = plp; + free(lp); + lp = nlp; + } + + dirty = TRUE; +} + + +/* + * Search for a line which contains the specified string. + * If the string is "", then the previously searched for string + * is used. The currently searched for string is saved for future use. + * Returns the line number which matches, or 0 if there was no match + * with an error printed. + */ +static int searchLines(const char *str, int num1, int num2) +{ + const LINE *lp; + int len; + + if (bad_nums(num1, num2, "search")) + return 0; + + if (*str == '\0') { + if (searchString[0] == '\0') { + bb_error_msg("no previous search string"); + return 0; + } + str = searchString; + } + + if (str != searchString) + strcpy(searchString, str); + + len = strlen(str); + + lp = findLine(num1); + if (lp == NULL) + return 0; + + while (num1 <= num2) { + if (findString(lp, str, len, 0) >= 0) + return num1; + num1++; + lp = lp->next; + } + + bb_error_msg("cannot find string \"%s\"", str); + return 0; +} + + +/* + * Return a pointer to the specified line number. + */ +static LINE *findLine(int num) +{ + LINE *lp; + int lnum; + + if ((num < 1) || (num > lastNum)) { + bb_error_msg("line number %d does not exist", num); + return NULL; + } + + if (curNum <= 0) { + curNum = 1; + curLine = lines.next; + } + + if (num == curNum) + return curLine; + + lp = curLine; + lnum = curNum; + if (num < (curNum / 2)) { + lp = lines.next; + lnum = 1; + } else if (num > ((curNum + lastNum) / 2)) { + lp = lines.prev; + lnum = lastNum; + } + + while (lnum < num) { + lp = lp->next; + lnum++; + } + + while (lnum > num) { + lp = lp->prev; + lnum--; + } + return lp; +} + + +/* + * Set the current line number. + * Returns TRUE if successful. + */ +static int setCurNum(int num) +{ + LINE *lp; + + lp = findLine(num); + if (lp == NULL) + return FALSE; + curNum = num; + curLine = lp; + return TRUE; +} diff --git a/editors/patch.c b/editors/patch.c new file mode 100644 index 0000000..e8482a7 --- /dev/null +++ b/editors/patch.c @@ -0,0 +1,254 @@ +/* vi: set sw=4 ts=4: */ +/* + * busybox patch applet to handle the unified diff format. + * Copyright (C) 2003 Glenn McGrath + * + * Licensed under the GPL v2 or later, see the file LICENSE in this tarball. + * + * This applet is written to work with patches generated by GNU diff, + * where there is equivalent functionality busybox patch shall behave + * as per GNU patch. + * + * There is a SUSv3 specification for patch, however it looks to be + * incomplete, it doesnt even mention unified diff format. + * http://www.opengroup.org/onlinepubs/007904975/utilities/patch.html + * + * Issues + * - Non-interactive + * - Patches must apply cleanly or patch (not just one hunk) will fail. + * - Reject file isnt saved + */ + +#include "libbb.h" + +static unsigned copy_lines(FILE *src_stream, FILE *dst_stream, unsigned lines_count) +{ + while (src_stream && lines_count) { + char *line; + line = xmalloc_fgets(src_stream); + if (line == NULL) { + break; + } + if (fputs(line, dst_stream) == EOF) { + bb_perror_msg_and_die("error writing to new file"); + } + free(line); + lines_count--; + } + return lines_count; +} + +/* If patch_level is -1 it will remove all directory names + * char *line must be greater than 4 chars + * returns NULL if the file doesnt exist or error + * returns malloc'ed filename + * NB: frees 1st argument! + */ +static char *extract_filename(char *line, int patch_level, const char *pat) +{ + char *temp = NULL, *filename_start_ptr = line + 4; + + if (strncmp(line, pat, 4) == 0) { + /* Terminate string at end of source filename */ + line[strcspn(line, "\t\n\r")] = '\0'; + + /* Skip over (patch_level) number of leading directories */ + while (patch_level--) { + temp = strchr(filename_start_ptr, '/'); + if (!temp) + break; + filename_start_ptr = temp + 1; + } + temp = xstrdup(filename_start_ptr); + } + free(line); + return temp; +} + +int patch_main(int argc, char **argv) MAIN_EXTERNALLY_VISIBLE; +int patch_main(int argc UNUSED_PARAM, char **argv) +{ + struct stat saved_stat; + char *patch_line; + FILE *patch_file; + int patch_level; + int ret = 0; + char plus = '+'; + + xfunc_error_retval = 2; + { + const char *p = "-1"; + const char *i = "-"; /* compat */ + if (getopt32(argv, "p:i:R", &p, &i) & 4) + plus = '-'; + patch_level = xatoi(p); /* can be negative! */ + patch_file = xfopen_stdin(i); + } + + patch_line = xmalloc_fgetline(patch_file); + while (patch_line) { + FILE *src_stream; + FILE *dst_stream; + //char *old_filename; + char *new_filename; + char *backup_filename; + unsigned src_cur_line = 1; + unsigned dst_cur_line = 0; + unsigned dst_beg_line; + unsigned bad_hunk_count = 0; + unsigned hunk_count = 0; + smallint copy_trailing_lines_flag = 0; + + /* Skip everything upto the "---" marker + * No need to parse the lines "Only in <dir>", and "diff <args>" + */ + do { + /* Extract the filename used before the patch was generated */ + new_filename = extract_filename(patch_line, patch_level, "--- "); + // was old_filename above + patch_line = xmalloc_fgetline(patch_file); + if (!patch_line) goto quit; + } while (!new_filename); + free(new_filename); // "source" filename is irrelevant + + new_filename = extract_filename(patch_line, patch_level, "+++ "); + if (!new_filename) { + bb_error_msg_and_die("invalid patch"); + } + + /* Get access rights from the file to be patched */ + if (stat(new_filename, &saved_stat) != 0) { + char *slash = strrchr(new_filename, '/'); + if (slash) { + /* Create leading directories */ + *slash = '\0'; + bb_make_directory(new_filename, -1, FILEUTILS_RECUR); + *slash = '/'; + } + backup_filename = NULL; + src_stream = NULL; + saved_stat.st_mode = 0644; + } else { + backup_filename = xasprintf("%s.orig", new_filename); + xrename(new_filename, backup_filename); + src_stream = xfopen_for_read(backup_filename); + } + dst_stream = xfopen_for_write(new_filename); + fchmod(fileno(dst_stream), saved_stat.st_mode); + + printf("patching file %s\n", new_filename); + + /* Handle all hunks for this file */ + patch_line = xmalloc_fgets(patch_file); + while (patch_line) { + unsigned count; + unsigned src_beg_line; + unsigned hunk_offset_start; + unsigned src_last_line = 1; + unsigned dst_last_line = 1; + + if ((sscanf(patch_line, "@@ -%d,%d +%d,%d", &src_beg_line, &src_last_line, &dst_beg_line, &dst_last_line) < 3) + && (sscanf(patch_line, "@@ -%d +%d,%d", &src_beg_line, &dst_beg_line, &dst_last_line) < 2) + ) { + /* No more hunks for this file */ + break; + } + if (plus != '+') { + /* reverse patch */ + unsigned tmp = src_last_line; + src_last_line = dst_last_line; + dst_last_line = tmp; + tmp = src_beg_line; + src_beg_line = dst_beg_line; + dst_beg_line = tmp; + } + hunk_count++; + + if (src_beg_line && dst_beg_line) { + /* Copy unmodified lines upto start of hunk */ + /* src_beg_line will be 0 if it's a new file */ + count = src_beg_line - src_cur_line; + if (copy_lines(src_stream, dst_stream, count)) { + bb_error_msg_and_die("bad src file"); + } + src_cur_line += count; + dst_cur_line += count; + copy_trailing_lines_flag = 1; + } + src_last_line += hunk_offset_start = src_cur_line; + dst_last_line += dst_cur_line; + + while (1) { + free(patch_line); + patch_line = xmalloc_fgets(patch_file); + if (patch_line == NULL) + break; /* EOF */ + if ((*patch_line != '-') && (*patch_line != '+') + && (*patch_line != ' ') + ) { + break; /* End of hunk */ + } + if (*patch_line != plus) { /* '-' or ' ' */ + char *src_line = NULL; + if (src_cur_line == src_last_line) + break; + if (src_stream) { + src_line = xmalloc_fgets(src_stream); + if (src_line) { + int diff = strcmp(src_line, patch_line + 1); + src_cur_line++; + free(src_line); + if (diff) + src_line = NULL; + } + } + if (!src_line) { + bb_error_msg("hunk #%u FAILED at %u", hunk_count, hunk_offset_start); + bad_hunk_count++; + break; + } + if (*patch_line != ' ') { /* '-' */ + continue; + } + } + if (dst_cur_line == dst_last_line) + break; + fputs(patch_line + 1, dst_stream); + dst_cur_line++; + } /* end of while loop handling one hunk */ + } /* end of while loop handling one file */ + + /* Cleanup last patched file */ + if (copy_trailing_lines_flag) { + copy_lines(src_stream, dst_stream, (unsigned)(-1)); + } + if (src_stream) { + fclose(src_stream); + } + fclose(dst_stream); + if (bad_hunk_count) { + ret = 1; + bb_error_msg("%u out of %u hunk FAILED", bad_hunk_count, hunk_count); + } else { + /* It worked, we can remove the backup */ + if (backup_filename) { + unlink(backup_filename); + } + if ((dst_cur_line == 0) || (dst_beg_line == 0)) { + /* The new patched file is empty, remove it */ + xunlink(new_filename); + // /* old_filename and new_filename may be the same file */ + // unlink(old_filename); + } + } + free(backup_filename); + //free(old_filename); + free(new_filename); + } /* end of "while there are patch lines" */ + quit: + /* 0 = SUCCESS + * 1 = Some hunks failed + * 2 = More serious problems (exited earlier) + */ + return ret; +} diff --git a/editors/sed.c b/editors/sed.c new file mode 100644 index 0000000..eb31f7d --- /dev/null +++ b/editors/sed.c @@ -0,0 +1,1349 @@ +/* vi: set sw=4 ts=4: */ +/* + * sed.c - very minimalist version of sed + * + * Copyright (C) 1999,2000,2001 by Lineo, inc. and Mark Whitley + * Copyright (C) 1999,2000,2001 by Mark Whitley <markw@codepoet.org> + * Copyright (C) 2002 Matt Kraai + * Copyright (C) 2003 by Glenn McGrath + * Copyright (C) 2003,2004 by Rob Landley <rob@landley.net> + * + * MAINTAINER: Rob Landley <rob@landley.net> + * + * Licensed under GPL version 2, see file LICENSE in this tarball for details. + */ + +/* Code overview. + + Files are laid out to avoid unnecessary function declarations. So for + example, every function add_cmd calls occurs before add_cmd in this file. + + add_cmd() is called on each line of sed command text (from a file or from + the command line). It calls get_address() and parse_cmd_args(). The + resulting sed_cmd_t structures are appended to a linked list + (G.sed_cmd_head/G.sed_cmd_tail). + + add_input_file() adds a FILE* to the list of input files. We need to + know all input sources ahead of time to find the last line for the $ match. + + process_files() does actual sedding, reading data lines from each input FILE * + (which could be stdin) and applying the sed command list (sed_cmd_head) to + each of the resulting lines. + + sed_main() is where external code calls into this, with a command line. +*/ + + +/* + Supported features and commands in this version of sed: + + - comments ('#') + - address matching: num|/matchstr/[,num|/matchstr/|$]command + - commands: (p)rint, (d)elete, (s)ubstitue (with g & I flags) + - edit commands: (a)ppend, (i)nsert, (c)hange + - file commands: (r)ead + - backreferences in substitution expressions (\0, \1, \2...\9) + - grouped commands: {cmd1;cmd2} + - transliteration (y/source-chars/dest-chars/) + - pattern space hold space storing / swapping (g, h, x) + - labels / branching (: label, b, t, T) + + (Note: Specifying an address (range) to match is *optional*; commands + default to the whole pattern space if no specific address match was + requested.) + + Todo: + - Create a wrapper around regex to make libc's regex conform with sed + + Reference http://www.opengroup.org/onlinepubs/007904975/utilities/sed.html +*/ + +#include "libbb.h" +#include "xregex.h" + +/* Each sed command turns into one of these structures. */ +typedef struct sed_cmd_s { + /* Ordered by alignment requirements: currently 36 bytes on x86 */ + struct sed_cmd_s *next; /* Next command (linked list, NULL terminated) */ + + /* address storage */ + regex_t *beg_match; /* sed -e '/match/cmd' */ + regex_t *end_match; /* sed -e '/match/,/end_match/cmd' */ + regex_t *sub_match; /* For 's/sub_match/string/' */ + int beg_line; /* 'sed 1p' 0 == apply commands to all lines */ + int end_line; /* 'sed 1,3p' 0 == one line only. -1 = last line ($) */ + + FILE *sw_file; /* File (sw) command writes to, -1 for none. */ + char *string; /* Data string for (saicytb) commands. */ + + unsigned which_match; /* (s) Which match to replace (0 for all) */ + + /* Bitfields (gcc won't group them if we don't) */ + unsigned invert:1; /* the '!' after the address */ + unsigned in_match:1; /* Next line also included in match? */ + unsigned sub_p:1; /* (s) print option */ + + char sw_last_char; /* Last line written by (sw) had no '\n' */ + + /* GENERAL FIELDS */ + char cmd; /* The command char: abcdDgGhHilnNpPqrstwxy:={} */ +} sed_cmd_t; + +static const char semicolon_whitespace[] ALIGN1 = "; \n\r\t\v"; + +struct globals { + /* options */ + int be_quiet, regex_type; + FILE *nonstdout; + char *outname, *hold_space; + + /* List of input files */ + int input_file_count, current_input_file; + FILE **input_file_list; + + regmatch_t regmatch[10]; + regex_t *previous_regex_ptr; + + /* linked list of sed commands */ + sed_cmd_t sed_cmd_head, *sed_cmd_tail; + + /* Linked list of append lines */ + llist_t *append_head; + + char *add_cmd_line; + + struct pipeline { + char *buf; /* Space to hold string */ + int idx; /* Space used */ + int len; /* Space allocated */ + } pipeline; +}; +#define G (*(struct globals*)&bb_common_bufsiz1) +void BUG_sed_globals_too_big(void); +#define INIT_G() do { \ + if (sizeof(struct globals) > COMMON_BUFSIZE) \ + BUG_sed_globals_too_big(); \ + G.sed_cmd_tail = &G.sed_cmd_head; \ +} while (0) + + +#if ENABLE_FEATURE_CLEAN_UP +static void sed_free_and_close_stuff(void) +{ + sed_cmd_t *sed_cmd = G.sed_cmd_head.next; + + llist_free(G.append_head, free); + + while (sed_cmd) { + sed_cmd_t *sed_cmd_next = sed_cmd->next; + + if (sed_cmd->sw_file) + xprint_and_close_file(sed_cmd->sw_file); + + if (sed_cmd->beg_match) { + regfree(sed_cmd->beg_match); + free(sed_cmd->beg_match); + } + if (sed_cmd->end_match) { + regfree(sed_cmd->end_match); + free(sed_cmd->end_match); + } + if (sed_cmd->sub_match) { + regfree(sed_cmd->sub_match); + free(sed_cmd->sub_match); + } + free(sed_cmd->string); + free(sed_cmd); + sed_cmd = sed_cmd_next; + } + + free(G.hold_space); + + while (G.current_input_file < G.input_file_count) + fclose(G.input_file_list[G.current_input_file++]); +} +#else +void sed_free_and_close_stuff(void); +#endif + +/* If something bad happens during -i operation, delete temp file */ + +static void cleanup_outname(void) +{ + if (G.outname) unlink(G.outname); +} + +/* strcpy, replacing "\from" with 'to'. If to is NUL, replacing "\any" with 'any' */ + +static void parse_escapes(char *dest, const char *string, int len, char from, char to) +{ + int i = 0; + + while (i < len) { + if (string[i] == '\\') { + if (!to || string[i+1] == from) { + *dest++ = to ? to : string[i+1]; + i += 2; + continue; + } + *dest++ = string[i++]; + } + /* TODO: is it safe wrt a string with trailing '\\' ? */ + *dest++ = string[i++]; + } + *dest = '\0'; +} + +static char *copy_parsing_escapes(const char *string, int len) +{ + char *dest = xmalloc(len + 1); + + parse_escapes(dest, string, len, 'n', '\n'); + /* GNU sed also recognizes \t */ + parse_escapes(dest, dest, strlen(dest), 't', '\t'); + return dest; +} + + +/* + * index_of_next_unescaped_regexp_delim - walks left to right through a string + * beginning at a specified index and returns the index of the next regular + * expression delimiter (typically a forward slash ('/')) not preceded by + * a backslash ('\'). A negative delimiter disables square bracket checking. + */ +static int index_of_next_unescaped_regexp_delim(int delimiter, const char *str) +{ + int bracket = -1; + int escaped = 0; + int idx = 0; + char ch; + + if (delimiter < 0) { + bracket--; + delimiter = -delimiter; + } + + for (; (ch = str[idx]); idx++) { + if (bracket >= 0) { + if (ch == ']' && !(bracket == idx - 1 || (bracket == idx - 2 + && str[idx - 1] == '^'))) + bracket = -1; + } else if (escaped) + escaped = 0; + else if (ch == '\\') + escaped = 1; + else if (bracket == -1 && ch == '[') + bracket = idx; + else if (ch == delimiter) + return idx; + } + + /* if we make it to here, we've hit the end of the string */ + bb_error_msg_and_die("unmatched '%c'", delimiter); +} + +/* + * Returns the index of the third delimiter + */ +static int parse_regex_delim(const char *cmdstr, char **match, char **replace) +{ + const char *cmdstr_ptr = cmdstr; + char delimiter; + int idx = 0; + + /* verify that the 's' or 'y' is followed by something. That something + * (typically a 'slash') is now our regexp delimiter... */ + if (*cmdstr == '\0') + bb_error_msg_and_die("bad format in substitution expression"); + delimiter = *cmdstr_ptr++; + + /* save the match string */ + idx = index_of_next_unescaped_regexp_delim(delimiter, cmdstr_ptr); + *match = copy_parsing_escapes(cmdstr_ptr, idx); + + /* save the replacement string */ + cmdstr_ptr += idx + 1; + idx = index_of_next_unescaped_regexp_delim(-delimiter, cmdstr_ptr); + *replace = copy_parsing_escapes(cmdstr_ptr, idx); + + return ((cmdstr_ptr - cmdstr) + idx); +} + +/* + * returns the index in the string just past where the address ends. + */ +static int get_address(const char *my_str, int *linenum, regex_t ** regex) +{ + const char *pos = my_str; + + if (isdigit(*my_str)) { + *linenum = strtol(my_str, (char**)&pos, 10); + /* endstr shouldnt ever equal NULL */ + } else if (*my_str == '$') { + *linenum = -1; + pos++; + } else if (*my_str == '/' || *my_str == '\\') { + int next; + char delimiter; + char *temp; + + delimiter = '/'; + if (*my_str == '\\') delimiter = *++pos; + next = index_of_next_unescaped_regexp_delim(delimiter, ++pos); + temp = copy_parsing_escapes(pos, next); + *regex = xmalloc(sizeof(regex_t)); + xregcomp(*regex, temp, G.regex_type|REG_NEWLINE); + free(temp); + /* Move position to next character after last delimiter */ + pos += (next+1); + } + return pos - my_str; +} + +/* Grab a filename. Whitespace at start is skipped, then goes to EOL. */ +static int parse_file_cmd(/*sed_cmd_t *sed_cmd,*/ const char *filecmdstr, char **retval) +{ + int start = 0, idx, hack = 0; + + /* Skip whitespace, then grab filename to end of line */ + while (isspace(filecmdstr[start])) + start++; + idx = start; + while (filecmdstr[idx] && filecmdstr[idx] != '\n') + idx++; + + /* If lines glued together, put backslash back. */ + if (filecmdstr[idx] == '\n') + hack = 1; + if (idx == start) + bb_error_msg_and_die("empty filename"); + *retval = xstrndup(filecmdstr+start, idx-start+hack+1); + if (hack) + (*retval)[idx] = '\\'; + + return idx; +} + +static int parse_subst_cmd(sed_cmd_t *sed_cmd, const char *substr) +{ + int cflags = G.regex_type; + char *match; + int idx; + + /* + * A substitution command should look something like this: + * s/match/replace/ #gIpw + * || | ||| + * mandatory optional + */ + idx = parse_regex_delim(substr, &match, &sed_cmd->string); + + /* determine the number of back references in the match string */ + /* Note: we compute this here rather than in the do_subst_command() + * function to save processor time, at the expense of a little more memory + * (4 bits) per sed_cmd */ + + /* process the flags */ + + sed_cmd->which_match = 1; + while (substr[++idx]) { + /* Parse match number */ + if (isdigit(substr[idx])) { + if (match[0] != '^') { + /* Match 0 treated as all, multiple matches we take the last one. */ + const char *pos = substr + idx; +/* FIXME: error check? */ + sed_cmd->which_match = (unsigned)strtol(substr+idx, (char**) &pos, 10); + idx = pos - substr; + } + continue; + } + /* Skip spaces */ + if (isspace(substr[idx])) continue; + + switch (substr[idx]) { + /* Replace all occurrences */ + case 'g': + if (match[0] != '^') + sed_cmd->which_match = 0; + break; + /* Print pattern space */ + case 'p': + sed_cmd->sub_p = 1; + break; + /* Write to file */ + case 'w': + { + char *temp; + idx += parse_file_cmd(/*sed_cmd,*/ substr+idx, &temp); + break; + } + /* Ignore case (gnu exension) */ + case 'I': + cflags |= REG_ICASE; + break; + /* Comment */ + case '#': + while (substr[++idx]) /*skip all*/; + /* Fall through */ + /* End of command */ + case ';': + case '}': + goto out; + default: + bb_error_msg_and_die("bad option in substitution expression"); + } + } +out: + /* compile the match string into a regex */ + if (*match != '\0') { + /* If match is empty, we use last regex used at runtime */ + sed_cmd->sub_match = xmalloc(sizeof(regex_t)); + xregcomp(sed_cmd->sub_match, match, cflags); + } + free(match); + + return idx; +} + +/* + * Process the commands arguments + */ +static const char *parse_cmd_args(sed_cmd_t *sed_cmd, const char *cmdstr) +{ + /* handle (s)ubstitution command */ + if (sed_cmd->cmd == 's') + cmdstr += parse_subst_cmd(sed_cmd, cmdstr); + /* handle edit cmds: (a)ppend, (i)nsert, and (c)hange */ + else if (strchr("aic", sed_cmd->cmd)) { + if ((sed_cmd->end_line || sed_cmd->end_match) && sed_cmd->cmd != 'c') + bb_error_msg_and_die + ("only a beginning address can be specified for edit commands"); + for (;;) { + if (*cmdstr == '\n' || *cmdstr == '\\') { + cmdstr++; + break; + } else if (isspace(*cmdstr)) + cmdstr++; + else + break; + } + sed_cmd->string = xstrdup(cmdstr); + /* "\anychar" -> "anychar" */ + parse_escapes(sed_cmd->string, sed_cmd->string, strlen(cmdstr), '\0', '\0'); + cmdstr += strlen(cmdstr); + /* handle file cmds: (r)ead */ + } else if (strchr("rw", sed_cmd->cmd)) { + if (sed_cmd->end_line || sed_cmd->end_match) + bb_error_msg_and_die("command only uses one address"); + cmdstr += parse_file_cmd(/*sed_cmd,*/ cmdstr, &sed_cmd->string); + if (sed_cmd->cmd == 'w') { + sed_cmd->sw_file = xfopen_for_write(sed_cmd->string); + sed_cmd->sw_last_char = '\n'; + } + /* handle branch commands */ + } else if (strchr(":btT", sed_cmd->cmd)) { + int length; + + cmdstr = skip_whitespace(cmdstr); + length = strcspn(cmdstr, semicolon_whitespace); + if (length) { + sed_cmd->string = xstrndup(cmdstr, length); + cmdstr += length; + } + } + /* translation command */ + else if (sed_cmd->cmd == 'y') { + char *match, *replace; + int i = cmdstr[0]; + + cmdstr += parse_regex_delim(cmdstr, &match, &replace)+1; + /* \n already parsed, but \delimiter needs unescaping. */ + parse_escapes(match, match, strlen(match), i, i); + parse_escapes(replace, replace, strlen(replace), i, i); + + sed_cmd->string = xzalloc((strlen(match) + 1) * 2); + for (i = 0; match[i] && replace[i]; i++) { + sed_cmd->string[i*2] = match[i]; + sed_cmd->string[i*2+1] = replace[i]; + } + free(match); + free(replace); + } + /* if it wasnt a single-letter command that takes no arguments + * then it must be an invalid command. + */ + else if (strchr("dDgGhHlnNpPqx={}", sed_cmd->cmd) == 0) { + bb_error_msg_and_die("unsupported command %c", sed_cmd->cmd); + } + + /* give back whatever's left over */ + return cmdstr; +} + + +/* Parse address+command sets, skipping comment lines. */ + +static void add_cmd(const char *cmdstr) +{ + sed_cmd_t *sed_cmd; + int temp; + + /* Append this line to any unfinished line from last time. */ + if (G.add_cmd_line) { + char *tp = xasprintf("%s\n%s", G.add_cmd_line, cmdstr); + free(G.add_cmd_line); + cmdstr = G.add_cmd_line = tp; + } + + /* If this line ends with backslash, request next line. */ + temp = strlen(cmdstr); + if (temp && cmdstr[--temp] == '\\') { + if (!G.add_cmd_line) + G.add_cmd_line = xstrdup(cmdstr); + G.add_cmd_line[temp] = '\0'; + return; + } + + /* Loop parsing all commands in this line. */ + while (*cmdstr) { + /* Skip leading whitespace and semicolons */ + cmdstr += strspn(cmdstr, semicolon_whitespace); + + /* If no more commands, exit. */ + if (!*cmdstr) break; + + /* if this is a comment, jump past it and keep going */ + if (*cmdstr == '#') { + /* "#n" is the same as using -n on the command line */ + if (cmdstr[1] == 'n') + G.be_quiet++; + cmdstr = strpbrk(cmdstr, "\n\r"); + if (!cmdstr) break; + continue; + } + + /* parse the command + * format is: [addr][,addr][!]cmd + * |----||-----||-| + * part1 part2 part3 + */ + + sed_cmd = xzalloc(sizeof(sed_cmd_t)); + + /* first part (if present) is an address: either a '$', a number or a /regex/ */ + cmdstr += get_address(cmdstr, &sed_cmd->beg_line, &sed_cmd->beg_match); + + /* second part (if present) will begin with a comma */ + if (*cmdstr == ',') { + int idx; + + cmdstr++; + idx = get_address(cmdstr, &sed_cmd->end_line, &sed_cmd->end_match); + if (!idx) + bb_error_msg_and_die("no address after comma"); + cmdstr += idx; + } + + /* skip whitespace before the command */ + cmdstr = skip_whitespace(cmdstr); + + /* Check for inversion flag */ + if (*cmdstr == '!') { + sed_cmd->invert = 1; + cmdstr++; + + /* skip whitespace before the command */ + cmdstr = skip_whitespace(cmdstr); + } + + /* last part (mandatory) will be a command */ + if (!*cmdstr) + bb_error_msg_and_die("missing command"); + sed_cmd->cmd = *(cmdstr++); + cmdstr = parse_cmd_args(sed_cmd, cmdstr); + + /* Add the command to the command array */ + G.sed_cmd_tail->next = sed_cmd; + G.sed_cmd_tail = G.sed_cmd_tail->next; + } + + /* If we glued multiple lines together, free the memory. */ + free(G.add_cmd_line); + G.add_cmd_line = NULL; +} + +/* Append to a string, reallocating memory as necessary. */ + +#define PIPE_GROW 64 + +static void pipe_putc(char c) +{ + if (G.pipeline.idx == G.pipeline.len) { + G.pipeline.buf = xrealloc(G.pipeline.buf, + G.pipeline.len + PIPE_GROW); + G.pipeline.len += PIPE_GROW; + } + G.pipeline.buf[G.pipeline.idx++] = c; +} + +static void do_subst_w_backrefs(char *line, char *replace) +{ + int i,j; + + /* go through the replacement string */ + for (i = 0; replace[i]; i++) { + /* if we find a backreference (\1, \2, etc.) print the backref'ed * text */ + if (replace[i] == '\\') { + unsigned backref = replace[++i] - '0'; + if (backref <= 9) { + /* print out the text held in G.regmatch[backref] */ + if (G.regmatch[backref].rm_so != -1) { + j = G.regmatch[backref].rm_so; + while (j < G.regmatch[backref].rm_eo) + pipe_putc(line[j++]); + } + continue; + } + /* I _think_ it is impossible to get '\' to be + * the last char in replace string. Thus we dont check + * for replace[i] == NUL. (counterexample anyone?) */ + /* if we find a backslash escaped character, print the character */ + pipe_putc(replace[i]); + continue; + } + /* if we find an unescaped '&' print out the whole matched text. */ + if (replace[i] == '&') { + j = G.regmatch[0].rm_so; + while (j < G.regmatch[0].rm_eo) + pipe_putc(line[j++]); + continue; + } + /* Otherwise just output the character. */ + pipe_putc(replace[i]); + } +} + +static int do_subst_command(sed_cmd_t *sed_cmd, char **line) +{ + char *oldline = *line; + int altered = 0; + unsigned match_count = 0; + regex_t *current_regex; + + /* Handle empty regex. */ + if (sed_cmd->sub_match == NULL) { + current_regex = G.previous_regex_ptr; + if (!current_regex) + bb_error_msg_and_die("no previous regexp"); + } else + G.previous_regex_ptr = current_regex = sed_cmd->sub_match; + + /* Find the first match */ + if (REG_NOMATCH == regexec(current_regex, oldline, 10, G.regmatch, 0)) + return 0; + + /* Initialize temporary output buffer. */ + G.pipeline.buf = xmalloc(PIPE_GROW); + G.pipeline.len = PIPE_GROW; + G.pipeline.idx = 0; + + /* Now loop through, substituting for matches */ + do { + int i; + + /* Work around bug in glibc regexec, demonstrated by: + echo " a.b" | busybox sed 's [^ .]* x g' + The match_count check is so not to break + echo "hi" | busybox sed 's/^/!/g' */ + if (!G.regmatch[0].rm_so && !G.regmatch[0].rm_eo && match_count) { + pipe_putc(*oldline++); + continue; + } + + match_count++; + + /* If we aren't interested in this match, output old line to + end of match and continue */ + if (sed_cmd->which_match + && (sed_cmd->which_match != match_count) + ) { + for (i = 0; i < G.regmatch[0].rm_eo; i++) + pipe_putc(*oldline++); + continue; + } + + /* print everything before the match */ + for (i = 0; i < G.regmatch[0].rm_so; i++) + pipe_putc(oldline[i]); + + /* then print the substitution string */ + do_subst_w_backrefs(oldline, sed_cmd->string); + + /* advance past the match */ + oldline += G.regmatch[0].rm_eo; + /* flag that something has changed */ + altered++; + + /* if we're not doing this globally, get out now */ + if (sed_cmd->which_match) + break; + } while (*oldline && (regexec(current_regex, oldline, 10, G.regmatch, 0) != REG_NOMATCH)); + + /* Copy rest of string into output pipeline */ + + while (*oldline) + pipe_putc(*oldline++); + pipe_putc(0); + + free(*line); + *line = G.pipeline.buf; + return altered; +} + +/* Set command pointer to point to this label. (Does not handle null label.) */ +static sed_cmd_t *branch_to(char *label) +{ + sed_cmd_t *sed_cmd; + + for (sed_cmd = G.sed_cmd_head.next; sed_cmd; sed_cmd = sed_cmd->next) { + if (sed_cmd->cmd == ':' && sed_cmd->string && !strcmp(sed_cmd->string, label)) { + return sed_cmd; + } + } + bb_error_msg_and_die("can't find label for jump to '%s'", label); +} + +static void append(char *s) +{ + llist_add_to_end(&G.append_head, xstrdup(s)); +} + +static void flush_append(void) +{ + char *data; + + /* Output appended lines. */ + while ((data = (char *)llist_pop(&G.append_head))) { + fprintf(G.nonstdout, "%s\n", data); + free(data); + } +} + +static void add_input_file(FILE *file) +{ + G.input_file_list = xrealloc_vector(G.input_file_list, 2, G.input_file_count); + G.input_file_list[G.input_file_count++] = file; +} + +/* Get next line of input from G.input_file_list, flushing append buffer and + * noting if we ran out of files without a newline on the last line we read. + */ +enum { + NO_EOL_CHAR = 1, + LAST_IS_NUL = 2, +}; +static char *get_next_line(char *gets_char) +{ + char *temp = NULL; + int len; + char gc; + + flush_append(); + + /* will be returned if last line in the file + * doesn't end with either '\n' or '\0' */ + gc = NO_EOL_CHAR; + while (G.current_input_file < G.input_file_count) { + FILE *fp = G.input_file_list[G.current_input_file]; + /* Read line up to a newline or NUL byte, inclusive, + * return malloc'ed char[]. length of the chunk read + * is stored in len. NULL if EOF/error */ + temp = bb_get_chunk_from_file(fp, &len); + if (temp) { + /* len > 0 here, it's ok to do temp[len-1] */ + char c = temp[len-1]; + if (c == '\n' || c == '\0') { + temp[len-1] = '\0'; + gc = c; + if (c == '\0') { + int ch = fgetc(fp); + if (ch != EOF) + ungetc(ch, fp); + else + gc = LAST_IS_NUL; + } + } + /* else we put NO_EOL_CHAR into *gets_char */ + break; + + /* NB: I had the idea of peeking next file(s) and returning + * NO_EOL_CHAR only if it is the *last* non-empty + * input file. But there is a case where this won't work: + * file1: "a woo\nb woo" + * file2: "c no\nd no" + * sed -ne 's/woo/bang/p' input1 input2 => "a bang\nb bang" + * (note: *no* newline after "b bang"!) */ + } + /* Close this file and advance to next one */ + fclose(fp); + G.current_input_file++; + } + *gets_char = gc; + return temp; +} + +/* Output line of text. */ +/* Note: + * The tricks with NO_EOL_CHAR and last_puts_char are there to emulate gnu sed. + * Without them, we had this: + * echo -n thingy >z1 + * echo -n again >z2 + * >znull + * sed "s/i/z/" z1 z2 znull | hexdump -vC + * output: + * gnu sed 4.1.5: + * 00000000 74 68 7a 6e 67 79 0a 61 67 61 7a 6e |thzngy.agazn| + * bbox: + * 00000000 74 68 7a 6e 67 79 61 67 61 7a 6e |thzngyagazn| + */ +static void puts_maybe_newline(char *s, FILE *file, char *last_puts_char, char last_gets_char) +{ + char lpc = *last_puts_char; + + /* Need to insert a '\n' between two files because first file's + * last line wasn't terminated? */ + if (lpc != '\n' && lpc != '\0') { + fputc('\n', file); + lpc = '\n'; + } + fputs(s, file); + + /* 'x' - just something which is not '\n', '\0' or NO_EOL_CHAR */ + if (s[0]) + lpc = 'x'; + + /* had trailing '\0' and it was last char of file? */ + if (last_gets_char == LAST_IS_NUL) { + fputc('\0', file); + lpc = 'x'; /* */ + } else + /* had trailing '\n' or '\0'? */ + if (last_gets_char != NO_EOL_CHAR) { + fputc(last_gets_char, file); + lpc = last_gets_char; + } + + if (ferror(file)) { + xfunc_error_retval = 4; /* It's what gnu sed exits with... */ + bb_error_msg_and_die(bb_msg_write_error); + } + *last_puts_char = lpc; +} + +#define sed_puts(s, n) (puts_maybe_newline(s, G.nonstdout, &last_puts_char, n)) + +static int beg_match(sed_cmd_t *sed_cmd, const char *pattern_space) +{ + int retval = sed_cmd->beg_match && !regexec(sed_cmd->beg_match, pattern_space, 0, NULL, 0); + if (retval) + G.previous_regex_ptr = sed_cmd->beg_match; + return retval; +} + +/* Process all the lines in all the files */ + +static void process_files(void) +{ + char *pattern_space, *next_line; + int linenum = 0; + char last_puts_char = '\n'; + char last_gets_char, next_gets_char; + sed_cmd_t *sed_cmd; + int substituted; + + /* Prime the pump */ + next_line = get_next_line(&next_gets_char); + + /* go through every line in each file */ + again: + substituted = 0; + + /* Advance to next line. Stop if out of lines. */ + pattern_space = next_line; + if (!pattern_space) return; + last_gets_char = next_gets_char; + + /* Read one line in advance so we can act on the last line, + * the '$' address */ + next_line = get_next_line(&next_gets_char); + linenum++; + restart: + /* for every line, go through all the commands */ + for (sed_cmd = G.sed_cmd_head.next; sed_cmd; sed_cmd = sed_cmd->next) { + int old_matched, matched; + + old_matched = sed_cmd->in_match; + + /* Determine if this command matches this line: */ + + /* Are we continuing a previous multi-line match? */ + sed_cmd->in_match = sed_cmd->in_match + /* Or is no range necessary? */ + || (!sed_cmd->beg_line && !sed_cmd->end_line + && !sed_cmd->beg_match && !sed_cmd->end_match) + /* Or did we match the start of a numerical range? */ + || (sed_cmd->beg_line > 0 && (sed_cmd->beg_line == linenum)) + /* Or does this line match our begin address regex? */ + || (beg_match(sed_cmd, pattern_space)) + /* Or did we match last line of input? */ + || (sed_cmd->beg_line == -1 && next_line == NULL); + + /* Snapshot the value */ + + matched = sed_cmd->in_match; + + /* Is this line the end of the current match? */ + + if (matched) { + sed_cmd->in_match = !( + /* has the ending line come, or is this a single address command? */ + (sed_cmd->end_line ? + sed_cmd->end_line == -1 ? + !next_line + : (sed_cmd->end_line <= linenum) + : !sed_cmd->end_match + ) + /* or does this line matches our last address regex */ + || (sed_cmd->end_match && old_matched + && (regexec(sed_cmd->end_match, + pattern_space, 0, NULL, 0) == 0)) + ); + } + + /* Skip blocks of commands we didn't match. */ + if (sed_cmd->cmd == '{') { + if (sed_cmd->invert ? matched : !matched) { + while (sed_cmd->cmd != '}') { + sed_cmd = sed_cmd->next; + if (!sed_cmd) + bb_error_msg_and_die("unterminated {"); + } + } + continue; + } + + /* Okay, so did this line match? */ + if (sed_cmd->invert ? !matched : matched) { + /* Update last used regex in case a blank substitute BRE is found */ + if (sed_cmd->beg_match) { + G.previous_regex_ptr = sed_cmd->beg_match; + } + + /* actual sedding */ + switch (sed_cmd->cmd) { + + /* Print line number */ + case '=': + fprintf(G.nonstdout, "%d\n", linenum); + break; + + /* Write the current pattern space up to the first newline */ + case 'P': + { + char *tmp = strchr(pattern_space, '\n'); + + if (tmp) { + *tmp = '\0'; + /* TODO: explain why '\n' below */ + sed_puts(pattern_space, '\n'); + *tmp = '\n'; + break; + } + /* Fall Through */ + } + + /* Write the current pattern space to output */ + case 'p': + /* NB: we print this _before_ the last line + * (of current file) is printed. Even if + * that line is nonterminated, we print + * '\n' here (gnu sed does the same) */ + sed_puts(pattern_space, '\n'); + break; + /* Delete up through first newline */ + case 'D': + { + char *tmp = strchr(pattern_space, '\n'); + + if (tmp) { + tmp = xstrdup(tmp+1); + free(pattern_space); + pattern_space = tmp; + goto restart; + } + } + /* discard this line. */ + case 'd': + goto discard_line; + + /* Substitute with regex */ + case 's': + if (!do_subst_command(sed_cmd, &pattern_space)) + break; + substituted |= 1; + + /* handle p option */ + if (sed_cmd->sub_p) + sed_puts(pattern_space, last_gets_char); + /* handle w option */ + if (sed_cmd->sw_file) + puts_maybe_newline( + pattern_space, sed_cmd->sw_file, + &sed_cmd->sw_last_char, last_gets_char); + break; + + /* Append line to linked list to be printed later */ + case 'a': + append(sed_cmd->string); + break; + + /* Insert text before this line */ + case 'i': + sed_puts(sed_cmd->string, '\n'); + break; + + /* Cut and paste text (replace) */ + case 'c': + /* Only triggers on last line of a matching range. */ + if (!sed_cmd->in_match) + sed_puts(sed_cmd->string, NO_EOL_CHAR); + goto discard_line; + + /* Read file, append contents to output */ + case 'r': + { + FILE *rfile; + + rfile = fopen_for_read(sed_cmd->string); + if (rfile) { + char *line; + + while ((line = xmalloc_fgetline(rfile)) + != NULL) + append(line); + xprint_and_close_file(rfile); + } + + break; + } + + /* Write pattern space to file. */ + case 'w': + puts_maybe_newline( + pattern_space, sed_cmd->sw_file, + &sed_cmd->sw_last_char, last_gets_char); + break; + + /* Read next line from input */ + case 'n': + if (!G.be_quiet) + sed_puts(pattern_space, last_gets_char); + if (next_line) { + free(pattern_space); + pattern_space = next_line; + last_gets_char = next_gets_char; + next_line = get_next_line(&next_gets_char); + substituted = 0; + linenum++; + break; + } + /* fall through */ + + /* Quit. End of script, end of input. */ + case 'q': + /* Exit the outer while loop */ + free(next_line); + next_line = NULL; + goto discard_commands; + + /* Append the next line to the current line */ + case 'N': + { + int len; + /* If no next line, jump to end of script and exit. */ + if (next_line == NULL) { + /* Jump to end of script and exit */ + free(next_line); + next_line = NULL; + goto discard_line; + /* append next_line, read new next_line. */ + } + len = strlen(pattern_space); + pattern_space = realloc(pattern_space, len + strlen(next_line) + 2); + pattern_space[len] = '\n'; + strcpy(pattern_space + len+1, next_line); + last_gets_char = next_gets_char; + next_line = get_next_line(&next_gets_char); + linenum++; + break; + } + + /* Test/branch if substitution occurred */ + case 't': + if (!substituted) break; + substituted = 0; + /* Fall through */ + /* Test/branch if substitution didn't occur */ + case 'T': + if (substituted) break; + /* Fall through */ + /* Branch to label */ + case 'b': + if (!sed_cmd->string) goto discard_commands; + else sed_cmd = branch_to(sed_cmd->string); + break; + /* Transliterate characters */ + case 'y': + { + int i, j; + + for (i = 0; pattern_space[i]; i++) { + for (j = 0; sed_cmd->string[j]; j += 2) { + if (pattern_space[i] == sed_cmd->string[j]) { + pattern_space[i] = sed_cmd->string[j + 1]; + break; + } + } + } + + break; + } + case 'g': /* Replace pattern space with hold space */ + free(pattern_space); + pattern_space = xstrdup(G.hold_space ? G.hold_space : ""); + break; + case 'G': /* Append newline and hold space to pattern space */ + { + int pattern_space_size = 2; + int hold_space_size = 0; + + if (pattern_space) + pattern_space_size += strlen(pattern_space); + if (G.hold_space) + hold_space_size = strlen(G.hold_space); + pattern_space = xrealloc(pattern_space, + pattern_space_size + hold_space_size); + if (pattern_space_size == 2) + pattern_space[0] = 0; + strcat(pattern_space, "\n"); + if (G.hold_space) + strcat(pattern_space, G.hold_space); + last_gets_char = '\n'; + + break; + } + case 'h': /* Replace hold space with pattern space */ + free(G.hold_space); + G.hold_space = xstrdup(pattern_space); + break; + case 'H': /* Append newline and pattern space to hold space */ + { + int hold_space_size = 2; + int pattern_space_size = 0; + + if (G.hold_space) + hold_space_size += strlen(G.hold_space); + if (pattern_space) + pattern_space_size = strlen(pattern_space); + G.hold_space = xrealloc(G.hold_space, + hold_space_size + pattern_space_size); + + if (hold_space_size == 2) + *G.hold_space = 0; + strcat(G.hold_space, "\n"); + if (pattern_space) + strcat(G.hold_space, pattern_space); + + break; + } + case 'x': /* Exchange hold and pattern space */ + { + char *tmp = pattern_space; + pattern_space = G.hold_space ? : xzalloc(1); + last_gets_char = '\n'; + G.hold_space = tmp; + break; + } + } + } + } + + /* + * exit point from sedding... + */ + discard_commands: + /* we will print the line unless we were told to be quiet ('-n') + or if the line was suppressed (ala 'd'elete) */ + if (!G.be_quiet) + sed_puts(pattern_space, last_gets_char); + + /* Delete and such jump here. */ + discard_line: + flush_append(); + free(pattern_space); + + goto again; +} + +/* It is possible to have a command line argument with embedded + * newlines. This counts as multiple command lines. + * However, newline can be escaped: 's/e/z\<newline>z/' + * We check for this. + */ + +static void add_cmd_block(char *cmdstr) +{ + char *sv, *eol; + + cmdstr = sv = xstrdup(cmdstr); + do { + eol = strchr(cmdstr, '\n'); + next: + if (eol) { + /* Count preceding slashes */ + int slashes = 0; + char *sl = eol; + + while (sl != cmdstr && *--sl == '\\') + slashes++; + /* Odd number of preceding slashes - newline is escaped */ + if (slashes & 1) { + overlapping_strcpy(eol - 1, eol); + eol = strchr(eol, '\n'); + goto next; + } + *eol = '\0'; + } + add_cmd(cmdstr); + cmdstr = eol + 1; + } while (eol); + free(sv); +} + +int sed_main(int argc, char **argv) MAIN_EXTERNALLY_VISIBLE; +int sed_main(int argc UNUSED_PARAM, char **argv) +{ + enum { + OPT_in_place = 1 << 0, + }; + unsigned opt; + llist_t *opt_e, *opt_f; + int status = EXIT_SUCCESS; + + INIT_G(); + + /* destroy command strings on exit */ + if (ENABLE_FEATURE_CLEAN_UP) atexit(sed_free_and_close_stuff); + + /* Lie to autoconf when it starts asking stupid questions. */ + if (argv[1] && !strcmp(argv[1], "--version")) { + puts("This is not GNU sed version 4.0"); + return 0; + } + + /* do normal option parsing */ + opt_e = opt_f = NULL; + opt_complementary = "e::f::" /* can occur multiple times */ + "nn"; /* count -n */ + opt = getopt32(argv, "irne:f:", &opt_e, &opt_f, + &G.be_quiet); /* counter for -n */ + //argc -= optind; + argv += optind; + if (opt & OPT_in_place) { // -i + atexit(cleanup_outname); + } + if (opt & 0x2) G.regex_type |= REG_EXTENDED; // -r + //if (opt & 0x4) G.be_quiet++; // -n + while (opt_e) { // -e + add_cmd_block(llist_pop(&opt_e)); + } + while (opt_f) { // -f + char *line; + FILE *cmdfile; + cmdfile = xfopen_for_read(llist_pop(&opt_f)); + while ((line = xmalloc_fgetline(cmdfile)) != NULL) { + add_cmd(line); + free(line); + } + fclose(cmdfile); + } + /* if we didn't get a pattern from -e or -f, use argv[0] */ + if (!(opt & 0x18)) { + if (!*argv) + bb_show_usage(); + add_cmd_block(*argv++); + } + /* Flush any unfinished commands. */ + add_cmd(""); + + /* By default, we write to stdout */ + G.nonstdout = stdout; + + /* argv[0..(argc-1)] should be names of file to process. If no + * files were specified or '-' was specified, take input from stdin. + * Otherwise, we process all the files specified. */ + if (argv[0] == NULL) { + if (opt & OPT_in_place) + bb_error_msg_and_die(bb_msg_requires_arg, "-i"); + add_input_file(stdin); + process_files(); + } else { + int i; + FILE *file; + + for (i = 0; argv[i]; i++) { + struct stat statbuf; + int nonstdoutfd; + + if (LONE_DASH(argv[i]) && !(opt & OPT_in_place)) { + add_input_file(stdin); + process_files(); + continue; + } + file = fopen_or_warn(argv[i], "r"); + if (!file) { + status = EXIT_FAILURE; + continue; + } + if (!(opt & OPT_in_place)) { + add_input_file(file); + continue; + } + + G.outname = xasprintf("%sXXXXXX", argv[i]); + nonstdoutfd = mkstemp(G.outname); + if (-1 == nonstdoutfd) + bb_perror_msg_and_die("cannot create temp file %s", G.outname); + G.nonstdout = fdopen(nonstdoutfd, "w"); + + /* Set permissions of output file */ + + fstat(fileno(file), &statbuf); + fchmod(nonstdoutfd, statbuf.st_mode); + add_input_file(file); + process_files(); + fclose(G.nonstdout); + + G.nonstdout = stdout; + /* unlink(argv[i]); */ + xrename(G.outname, argv[i]); + free(G.outname); + G.outname = NULL; + } + if (G.input_file_count > G.current_input_file) + process_files(); + } + + return status; +} diff --git a/editors/sed1line.txt b/editors/sed1line.txt new file mode 100644 index 0000000..11a2e36 --- /dev/null +++ b/editors/sed1line.txt @@ -0,0 +1,425 @@ +http://www.student.northpark.edu/pemente/sed/sed1line.txt +------------------------------------------------------------------------- +HANDY ONE-LINERS FOR SED (Unix stream editor) Apr. 26, 2004 +compiled by Eric Pement - pemente[at]northpark[dot]edu version 5.4 +Latest version of this file is usually at: + http://sed.sourceforge.net/sed1line.txt + http://www.student.northpark.edu/pemente/sed/sed1line.txt +This file is also available in Portuguese at: + http://www.lrv.ufsc.br/wmaker/sed_ptBR.html + +FILE SPACING: + + # double space a file + sed G + + # double space a file which already has blank lines in it. Output file + # should contain no more than one blank line between lines of text. + sed '/^$/d;G' + + # triple space a file + sed 'G;G' + + # undo double-spacing (assumes even-numbered lines are always blank) + sed 'n;d' + + # insert a blank line above every line which matches "regex" + sed '/regex/{x;p;x;}' + + # insert a blank line below every line which matches "regex" + sed '/regex/G' + + # insert a blank line above and below every line which matches "regex" + sed '/regex/{x;p;x;G;}' + +NUMBERING: + + # number each line of a file (simple left alignment). Using a tab (see + # note on '\t' at end of file) instead of space will preserve margins. + sed = filename | sed 'N;s/\n/\t/' + + # number each line of a file (number on left, right-aligned) + sed = filename | sed 'N; s/^/ /; s/ *\(.\{6,\}\)\n/\1 /' + + # number each line of file, but only print numbers if line is not blank + sed '/./=' filename | sed '/./N; s/\n/ /' + + # count lines (emulates "wc -l") + sed -n '$=' + +TEXT CONVERSION AND SUBSTITUTION: + + # IN UNIX ENVIRONMENT: convert DOS newlines (CR/LF) to Unix format + sed 's/.$//' # assumes that all lines end with CR/LF + sed 's/^M$//' # in bash/tcsh, press Ctrl-V then Ctrl-M + sed 's/\x0D$//' # gsed 3.02.80, but top script is easier + + # IN UNIX ENVIRONMENT: convert Unix newlines (LF) to DOS format + sed "s/$/`echo -e \\\r`/" # command line under ksh + sed 's/$'"/`echo \\\r`/" # command line under bash + sed "s/$/`echo \\\r`/" # command line under zsh + sed 's/$/\r/' # gsed 3.02.80 + + # IN DOS ENVIRONMENT: convert Unix newlines (LF) to DOS format + sed "s/$//" # method 1 + sed -n p # method 2 + + # IN DOS ENVIRONMENT: convert DOS newlines (CR/LF) to Unix format + # Can only be done with UnxUtils sed, version 4.0.7 or higher. + # Cannot be done with other DOS versions of sed. Use "tr" instead. + sed "s/\r//" infile >outfile # UnxUtils sed v4.0.7 or higher + tr -d \r <infile >outfile # GNU tr version 1.22 or higher + + # delete leading whitespace (spaces, tabs) from front of each line + # aligns all text flush left + sed 's/^[ \t]*//' # see note on '\t' at end of file + + # delete trailing whitespace (spaces, tabs) from end of each line + sed 's/[ \t]*$//' # see note on '\t' at end of file + + # delete BOTH leading and trailing whitespace from each line + sed 's/^[ \t]*//;s/[ \t]*$//' + + # insert 5 blank spaces at beginning of each line (make page offset) + sed 's/^/ /' + + # align all text flush right on a 79-column width + sed -e :a -e 's/^.\{1,78\}$/ &/;ta' # set at 78 plus 1 space + + # center all text in the middle of 79-column width. In method 1, + # spaces at the beginning of the line are significant, and trailing + # spaces are appended at the end of the line. In method 2, spaces at + # the beginning of the line are discarded in centering the line, and + # no trailing spaces appear at the end of lines. + sed -e :a -e 's/^.\{1,77\}$/ & /;ta' # method 1 + sed -e :a -e 's/^.\{1,77\}$/ &/;ta' -e 's/\( *\)\1/\1/' # method 2 + + # substitute (find and replace) "foo" with "bar" on each line + sed 's/foo/bar/' # replaces only 1st instance in a line + sed 's/foo/bar/4' # replaces only 4th instance in a line + sed 's/foo/bar/g' # replaces ALL instances in a line + sed 's/\(.*\)foo\(.*foo\)/\1bar\2/' # replace the next-to-last case + sed 's/\(.*\)foo/\1bar/' # replace only the last case + + # substitute "foo" with "bar" ONLY for lines which contain "baz" + sed '/baz/s/foo/bar/g' + + # substitute "foo" with "bar" EXCEPT for lines which contain "baz" + sed '/baz/!s/foo/bar/g' + + # change "scarlet" or "ruby" or "puce" to "red" + sed 's/scarlet/red/g;s/ruby/red/g;s/puce/red/g' # most seds + gsed 's/scarlet\|ruby\|puce/red/g' # GNU sed only + + # reverse order of lines (emulates "tac") + # bug/feature in HHsed v1.5 causes blank lines to be deleted + sed '1!G;h;$!d' # method 1 + sed -n '1!G;h;$p' # method 2 + + # reverse each character on the line (emulates "rev") + sed '/\n/!G;s/\(.\)\(.*\n\)/&\2\1/;//D;s/.//' + + # join pairs of lines side-by-side (like "paste") + sed '$!N;s/\n/ /' + + # if a line ends with a backslash, append the next line to it + sed -e :a -e '/\\$/N; s/\\\n//; ta' + + # if a line begins with an equal sign, append it to the previous line + # and replace the "=" with a single space + sed -e :a -e '$!N;s/\n=/ /;ta' -e 'P;D' + + # add commas to numeric strings, changing "1234567" to "1,234,567" + gsed ':a;s/\B[0-9]\{3\}\>/,&/;ta' # GNU sed + sed -e :a -e 's/\(.*[0-9]\)\([0-9]\{3\}\)/\1,\2/;ta' # other seds + + # add commas to numbers with decimal points and minus signs (GNU sed) + gsed ':a;s/\(^\|[^0-9.]\)\([0-9]\+\)\([0-9]\{3\}\)/\1\2,\3/g;ta' + + # add a blank line every 5 lines (after lines 5, 10, 15, 20, etc.) + gsed '0~5G' # GNU sed only + sed 'n;n;n;n;G;' # other seds + +SELECTIVE PRINTING OF CERTAIN LINES: + + # print first 10 lines of file (emulates behavior of "head") + sed 10q + + # print first line of file (emulates "head -1") + sed q + + # print the last 10 lines of a file (emulates "tail") + sed -e :a -e '$q;N;11,$D;ba' + + # print the last 2 lines of a file (emulates "tail -2") + sed '$!N;$!D' + + # print the last line of a file (emulates "tail -1") + sed '$!d' # method 1 + sed -n '$p' # method 2 + + # print only lines which match regular expression (emulates "grep") + sed -n '/regexp/p' # method 1 + sed '/regexp/!d' # method 2 + + # print only lines which do NOT match regexp (emulates "grep -v") + sed -n '/regexp/!p' # method 1, corresponds to above + sed '/regexp/d' # method 2, simpler syntax + + # print the line immediately before a regexp, but not the line + # containing the regexp + sed -n '/regexp/{g;1!p;};h' + + # print the line immediately after a regexp, but not the line + # containing the regexp + sed -n '/regexp/{n;p;}' + + # print 1 line of context before and after regexp, with line number + # indicating where the regexp occurred (similar to "grep -A1 -B1") + sed -n -e '/regexp/{=;x;1!p;g;$!N;p;D;}' -e h + + # grep for AAA and BBB and CCC (in any order) + sed '/AAA/!d; /BBB/!d; /CCC/!d' + + # grep for AAA and BBB and CCC (in that order) + sed '/AAA.*BBB.*CCC/!d' + + # grep for AAA or BBB or CCC (emulates "egrep") + sed -e '/AAA/b' -e '/BBB/b' -e '/CCC/b' -e d # most seds + gsed '/AAA\|BBB\|CCC/!d' # GNU sed only + + # print paragraph if it contains AAA (blank lines separate paragraphs) + # HHsed v1.5 must insert a 'G;' after 'x;' in the next 3 scripts below + sed -e '/./{H;$!d;}' -e 'x;/AAA/!d;' + + # print paragraph if it contains AAA and BBB and CCC (in any order) + sed -e '/./{H;$!d;}' -e 'x;/AAA/!d;/BBB/!d;/CCC/!d' + + # print paragraph if it contains AAA or BBB or CCC + sed -e '/./{H;$!d;}' -e 'x;/AAA/b' -e '/BBB/b' -e '/CCC/b' -e d + gsed '/./{H;$!d;};x;/AAA\|BBB\|CCC/b;d' # GNU sed only + + # print only lines of 65 characters or longer + sed -n '/^.\{65\}/p' + + # print only lines of less than 65 characters + sed -n '/^.\{65\}/!p' # method 1, corresponds to above + sed '/^.\{65\}/d' # method 2, simpler syntax + + # print section of file from regular expression to end of file + sed -n '/regexp/,$p' + + # print section of file based on line numbers (lines 8-12, inclusive) + sed -n '8,12p' # method 1 + sed '8,12!d' # method 2 + + # print line number 52 + sed -n '52p' # method 1 + sed '52!d' # method 2 + sed '52q;d' # method 3, efficient on large files + + # beginning at line 3, print every 7th line + gsed -n '3~7p' # GNU sed only + sed -n '3,${p;n;n;n;n;n;n;}' # other seds + + # print section of file between two regular expressions (inclusive) + sed -n '/Iowa/,/Montana/p' # case sensitive + +SELECTIVE DELETION OF CERTAIN LINES: + + # print all of file EXCEPT section between 2 regular expressions + sed '/Iowa/,/Montana/d' + + # delete duplicate, consecutive lines from a file (emulates "uniq"). + # First line in a set of duplicate lines is kept, rest are deleted. + sed '$!N; /^\(.*\)\n\1$/!P; D' + + # delete duplicate, nonconsecutive lines from a file. Beware not to + # overflow the buffer size of the hold space, or else use GNU sed. + sed -n 'G; s/\n/&&/; /^\([ -~]*\n\).*\n\1/d; s/\n//; h; P' + + # delete all lines except duplicate lines (emulates "uniq -d"). + sed '$!N; s/^\(.*\)\n\1$/\1/; t; D' + + # delete the first 10 lines of a file + sed '1,10d' + + # delete the last line of a file + sed '$d' + + # delete the last 2 lines of a file + sed 'N;$!P;$!D;$d' + + # delete the last 10 lines of a file + sed -e :a -e '$d;N;2,10ba' -e 'P;D' # method 1 + sed -n -e :a -e '1,10!{P;N;D;};N;ba' # method 2 + + # delete every 8th line + gsed '0~8d' # GNU sed only + sed 'n;n;n;n;n;n;n;d;' # other seds + + # delete ALL blank lines from a file (same as "grep '.' ") + sed '/^$/d' # method 1 + sed '/./!d' # method 2 + + # delete all CONSECUTIVE blank lines from file except the first; also + # deletes all blank lines from top and end of file (emulates "cat -s") + sed '/./,/^$/!d' # method 1, allows 0 blanks at top, 1 at EOF + sed '/^$/N;/\n$/D' # method 2, allows 1 blank at top, 0 at EOF + + # delete all CONSECUTIVE blank lines from file except the first 2: + sed '/^$/N;/\n$/N;//D' + + # delete all leading blank lines at top of file + sed '/./,$!d' + + # delete all trailing blank lines at end of file + sed -e :a -e '/^\n*$/{$d;N;ba' -e '}' # works on all seds + sed -e :a -e '/^\n*$/N;/\n$/ba' # ditto, except for gsed 3.02* + + # delete the last line of each paragraph + sed -n '/^$/{p;h;};/./{x;/./p;}' + +SPECIAL APPLICATIONS: + + # remove nroff overstrikes (char, backspace) from man pages. The 'echo' + # command may need an -e switch if you use Unix System V or bash shell. + sed "s/.`echo \\\b`//g" # double quotes required for Unix environment + sed 's/.^H//g' # in bash/tcsh, press Ctrl-V and then Ctrl-H + sed 's/.\x08//g' # hex expression for sed v1.5 + + # get Usenet/e-mail message header + sed '/^$/q' # deletes everything after first blank line + + # get Usenet/e-mail message body + sed '1,/^$/d' # deletes everything up to first blank line + + # get Subject header, but remove initial "Subject: " portion + sed '/^Subject: */!d; s///;q' + + # get return address header + sed '/^Reply-To:/q; /^From:/h; /./d;g;q' + + # parse out the address proper. Pulls out the e-mail address by itself + # from the 1-line return address header (see preceding script) + sed 's/ *(.*)//; s/>.*//; s/.*[:<] *//' + + # add a leading angle bracket and space to each line (quote a message) + sed 's/^/> /' + + # delete leading angle bracket & space from each line (unquote a message) + sed 's/^> //' + + # remove most HTML tags (accommodates multiple-line tags) + sed -e :a -e 's/<[^>]*>//g;/</N;//ba' + + # extract multi-part uuencoded binaries, removing extraneous header + # info, so that only the uuencoded portion remains. Files passed to + # sed must be passed in the proper order. Version 1 can be entered + # from the command line; version 2 can be made into an executable + # Unix shell script. (Modified from a script by Rahul Dhesi.) + sed '/^end/,/^begin/d' file1 file2 ... fileX | uudecode # vers. 1 + sed '/^end/,/^begin/d' "$@" | uudecode # vers. 2 + + # zip up each .TXT file individually, deleting the source file and + # setting the name of each .ZIP file to the basename of the .TXT file + # (under DOS: the "dir /b" switch returns bare filenames in all caps). + echo @echo off >zipup.bat + dir /b *.txt | sed "s/^\(.*\)\.TXT/pkzip -mo \1 \1.TXT/" >>zipup.bat + +TYPICAL USE: Sed takes one or more editing commands and applies all of +them, in sequence, to each line of input. After all the commands have +been applied to the first input line, that line is output and a second +input line is taken for processing, and the cycle repeats. The +preceding examples assume that input comes from the standard input +device (i.e, the console, normally this will be piped input). One or +more filenames can be appended to the command line if the input does +not come from stdin. Output is sent to stdout (the screen). Thus: + + cat filename | sed '10q' # uses piped input + sed '10q' filename # same effect, avoids a useless "cat" + sed '10q' filename > newfile # redirects output to disk + +For additional syntax instructions, including the way to apply editing +commands from a disk file instead of the command line, consult "sed & +awk, 2nd Edition," by Dale Dougherty and Arnold Robbins (O'Reilly, +1997; http://www.ora.com), "UNIX Text Processing," by Dale Dougherty +and Tim O'Reilly (Hayden Books, 1987) or the tutorials by Mike Arst +distributed in U-SEDIT2.ZIP (many sites). To fully exploit the power +of sed, one must understand "regular expressions." For this, see +"Mastering Regular Expressions" by Jeffrey Friedl (O'Reilly, 1997). +The manual ("man") pages on Unix systems may be helpful (try "man +sed", "man regexp", or the subsection on regular expressions in "man +ed"), but man pages are notoriously difficult. They are not written to +teach sed use or regexps to first-time users, but as a reference text +for those already acquainted with these tools. + +QUOTING SYNTAX: The preceding examples use single quotes ('...') +instead of double quotes ("...") to enclose editing commands, since +sed is typically used on a Unix platform. Single quotes prevent the +Unix shell from intrepreting the dollar sign ($) and backquotes +(`...`), which are expanded by the shell if they are enclosed in +double quotes. Users of the "csh" shell and derivatives will also need +to quote the exclamation mark (!) with the backslash (i.e., \!) to +properly run the examples listed above, even within single quotes. +Versions of sed written for DOS invariably require double quotes +("...") instead of single quotes to enclose editing commands. + +USE OF '\t' IN SED SCRIPTS: For clarity in documentation, we have used +the expression '\t' to indicate a tab character (0x09) in the scripts. +However, most versions of sed do not recognize the '\t' abbreviation, +so when typing these scripts from the command line, you should press +the TAB key instead. '\t' is supported as a regular expression +metacharacter in awk, perl, and HHsed, sedmod, and GNU sed v3.02.80. + +VERSIONS OF SED: Versions of sed do differ, and some slight syntax +variation is to be expected. In particular, most do not support the +use of labels (:name) or branch instructions (b,t) within editing +commands, except at the end of those commands. We have used the syntax +which will be portable to most users of sed, even though the popular +GNU versions of sed allow a more succinct syntax. When the reader sees +a fairly long command such as this: + + sed -e '/AAA/b' -e '/BBB/b' -e '/CCC/b' -e d + +it is heartening to know that GNU sed will let you reduce it to: + + sed '/AAA/b;/BBB/b;/CCC/b;d' # or even + sed '/AAA\|BBB\|CCC/b;d' + +In addition, remember that while many versions of sed accept a command +like "/one/ s/RE1/RE2/", some do NOT allow "/one/! s/RE1/RE2/", which +contains space before the 's'. Omit the space when typing the command. + +OPTIMIZING FOR SPEED: If execution speed needs to be increased (due to +large input files or slow processors or hard disks), substitution will +be executed more quickly if the "find" expression is specified before +giving the "s/.../.../" instruction. Thus: + + sed 's/foo/bar/g' filename # standard replace command + sed '/foo/ s/foo/bar/g' filename # executes more quickly + sed '/foo/ s//bar/g' filename # shorthand sed syntax + +On line selection or deletion in which you only need to output lines +from the first part of the file, a "quit" command (q) in the script +will drastically reduce processing time for large files. Thus: + + sed -n '45,50p' filename # print line nos. 45-50 of a file + sed -n '51q;45,50p' filename # same, but executes much faster + +If you have any additional scripts to contribute or if you find errors +in this document, please send e-mail to the compiler. Indicate the +version of sed you used, the operating system it was compiled for, and +the nature of the problem. Various scripts in this file were written +or contributed by: + + Al Aab <af137@freenet.toronto.on.ca> # "seders" list moderator + Edgar Allen <era@sky.net> # various + Yiorgos Adamopoulos <adamo@softlab.ece.ntua.gr> + Dale Dougherty <dale@songline.com> # author of "sed & awk" + Carlos Duarte <cdua@algos.inesc.pt> # author of "do it with sed" + Eric Pement <pemente@northpark.edu> # author of this document + Ken Pizzini <ken@halcyon.com> # author of GNU sed v3.02 + S.G. Ravenhall <stew.ravenhall@totalise.co.uk> # great de-html script + Greg Ubben <gsu@romulus.ncsc.mil> # many contributions & much help +------------------------------------------------------------------------- diff --git a/editors/sed_summary.htm b/editors/sed_summary.htm new file mode 100644 index 0000000..34e72b0 --- /dev/null +++ b/editors/sed_summary.htm @@ -0,0 +1,223 @@ +<html> + +<head><title>Command Summary for sed (sed & awk, Second Edition)</title> +</head> + +<body> + +<h2>Command Summary for sed</h2> + +<dl> + +<dt><b>: </b> <b> :</b><em>label</em></dt> +<dd>Label a line in the script for the transfer of control by +<b>b</b> or <b>t</b>. +<em>label</em> may contain up to seven characters. +(The POSIX standard says that an implementation can allow longer +labels if it wishes to. GNU sed allows labels to be of any length.) +</p></dd> + + +<dt><b>=</b> [<em>address</em>]<b>=</b></dt> +<dd>Write to standard output the line number of addressed line.</p></dd> + + +<dt><b>a</b> [<em>address</em>]<b>a\</b></dt> +<dd><em>text</em></p> + +<p>Append <em>text</em> +following each line matched by <em>address</em>. If +<em>text</em> goes over more than one line, newlines +must be "hidden" by preceding them with a backslash. The +<em>text</em> will be terminated by the first +newline that is not hidden in this way. The +<em>text</em> is not available in the pattern space +and subsequent commands cannot be applied to it. The results of this +command are sent to standard output when the list of editing commands +is finished, regardless of what happens to the current line in the +pattern space.</p></dd> + + +<dt><b>b</b> [<em>address1</em>[,<em>address2</em>]]<b>b</b>[<em>label</em>]</dt> +<dd>Transfer control unconditionally (branch) to +<b>:</b><em>label</em> elsewhere in +script. That is, the command following the +<em>label</em> is the next command applied to the +current line. If no <em>label</em> is specified, +control falls through to the end of the script, so no more commands +are applied to the current line.</p></dd> + + +<dt><b>c</b> [<em>address1</em>[,<em>address2</em>]]<b>c\</b></dt> +<dd><em>text</em></p> + +<p>Replace (change) the lines selected by the address with +<em>text</em>. When a range of lines is specified, +all lines as a group are replaced by a single copy of +<em>text</em>. The newline following each line of +<em>text</em> must be escaped by a backslash, except +the last line. The contents of the pattern space are, in effect, +deleted and no subsequent editing commands can be applied to it (or to +<em>text</em>).</p></dd> + + +<dt><b>d</b> [<em>address1</em>[,<em>address2</em>]]<b>d</b></dt> +<dd>Delete line(s) from pattern space. Thus, the line is not passed to standard +output. A new line of input is read and editing resumes with first +command in script.</p></dd> + + +<dt><b>D</b> [<em>address1</em>[,<em>address2</em>]]<b>D</b></dt> +<dd>Delete first part (up to embedded newline) of multiline pattern space created +by <b>N</b> command and resume editing with first command in +script. If this command empties the pattern space, then a new line +of input is read, as if the <b>d</b> command had been executed.</p></dd> + + +<dt><b>g</b> [<em>address1</em>[,<em>address2</em>]]<b>g</b></dt> +<dd>Copy (get) contents of hold space (see <b>h</b> or +<b>H</b> command) into the pattern space, wiping out +previous contents.</p></dd> + + +<dt><b>G</b> [<em>address1</em>[,<em>address2</em>]]<b>G</b></dt> +<dd>Append newline followed by contents of hold space (see +<b>h</b> or <b>H</b> command) to contents of +the pattern space. If hold space is empty, a newline is still +appended to the pattern space.</p></dd> + + +<dt><b>h</b> [<em>address1</em>[,<em>address2</em>]]<b>h</b></dt> +<dd>Copy pattern space into hold space, a special temporary buffer. +Previous contents of hold space are wiped out.</p></dd> + + +<dt><b>H</b> [<em>address1</em>[,<em>address2</em>]]<b>H</b></dt> +<dd>Append newline and contents of pattern space to contents of the hold +space. Even if hold space is empty, this command still appends the +newline first.</p></dd> + + +<dt><b>i</b> [<em>address1</em>]<b>i\</b></dt> +<dd><em>text</em></p> + +<p>Insert <em>text</em> before each line matched by +<em>address</em>. (See <b>a</b> for +details on <em>text</em>.)</p></dd> + + +<dt><b>l</b> [<em>address1</em>[,<em>address2</em>]]<b>l</b></dt> +<dd>List the contents of the pattern space, showing nonprinting characters +as ASCII codes. Long lines are wrapped.</p></dd> + + +<dt><b>n</b> [<em>address1</em>[,<em>address2</em>]]<b>n</b></dt> +<dd>Read next line of input into pattern space. Current line is sent to +standard output. New line becomes current line and increments line +counter. Control passes to command following <b>n</b> +instead of resuming at the top of the script.</p></dd> + + +<dt><b>N</b> [<em>address1</em>[,<em>address2</em>]]<b>N</b></dt> +<dd>Append next input line to contents of pattern space; the new line is +separated from the previous contents of the pattern space by a newline. +(This command is designed to allow pattern matches across two +lines. Using \n to match the embedded newline, you can match +patterns across multiple lines.)</p></dd> + + +<dt><b>p</b> [<em>address1</em>[,<em>address2</em>]]<b>p</b></dt> +<dd>Print the addressed line(s). Note that this can result in duplicate +output unless default output is suppressed by using "#n" or +the <span class="option">-n</span> + +command-line option. Typically used before commands that change flow +control (<b>d</b>, <b>n</b>, +<b>b</b>) and might prevent the current line from being +output.</p></dd> + + +<dt><b>P</b> [<em>address1</em>[,<em>address2</em>]]<b>P</b></dt> +<dd>Print first part (up to embedded newline) of multiline pattern space +created by <b>N</b> command. Same as <b>p</b> +if <b>N</b> has not been applied to a line.</p></dd> + + +<dt><b>q</b> [<em>address</em>]<b>q</b></dt> +<dd>Quit when <em>address</em> is encountered. The +addressed line is first written to output (if default output is not +suppressed), along with any text appended to it by previous +<b>a</b> or <b>r</b> commands.</p></dd> + + +<dt><b>r</b> [<em>address</em>]<b>r</b> <em>file</em></dt> +<dd>Read contents of <em>file</em> and append after the +contents of the pattern space. Exactly one space must be put between +<b>r</b> and the filename.</p></dd> + + +<dt><b>s</b> [<em>address1</em>[,<em>address2</em>]]<b>s</b>/<em>pattern</em>/<em>replacement</em>/[<em>flags</em>]</dt> +<dd>Substitute <em>replacement</em> for +<em>pattern</em> on each addressed line. If pattern +addresses are used, the pattern <b>//</b> represents the +last pattern address specified. The following flags can be specified:</p> + + <dl> + + <dt><b>n</b></dt> + <dd>Replace <em>n</em>th instance of + /<em>pattern</em>/ on each addressed line. + <em>n</em> is any number in the range 1 to 512, and + the default is 1.</p></dd> + + <dt><b>g</b></dt> + <dd>Replace all instances of /<em>pattern</em>/ on each + addressed line, not just the first instance.</p></dd> + + <dt><b>I</b></dt> + <dd>Matching is case-insensitive.<p></p></dd> + + <dt><b>p</b></dt> + <dd>Print the line if a successful substitution is done. If several + successful substitutions are done, multiple copies of the line will be + printed.</p></dd> + + <dt><b>w</b> <em>file</em></dt> + <dd>Write the line to <em>file</em> if a replacement + was done. A maximum of 10 different <em>files</em> can be opened.</p></dd> + + </dl> + +</dd> + + +<dt><b>t</b> [<em>address1</em>[,<em>address2</em>]]<b>t </b>[<em>label</em>]</dt> +<dd>Test if successful substitutions have been made on addressed lines, +and if so, branch to line marked by :<em>label</em>. +(See <b>b</b> and <b>:</b>.) If label is not +specified, control falls through to bottom of script.</p></dd> + + +<dt><b>w</b> [<em>address1</em>[,<em>address2</em>]]<b>w</b> <em>file</em></dt> +<dd>Append contents of pattern space to <em>file</em>. +This action occurs when the command is encountered rather than when +the pattern space is output. Exactly one space must separate the +<b>w</b> and the filename. A maximum of 10 different +files can be opened in a script. This command will create the file if +it does not exist; if the file exists, its contents will be +overwritten each time the script is executed. Multiple write commands +that direct output to the same file append to the end of the file.</p></dd> + + +<dt><b>x</b> [<em>address1</em>[,<em>address2</em>]]<b>x</b></dt> +<dd>Exchange contents of the pattern space with the contents of the hold +space.</p></dd> + + +<dt><b>y</b> [<em>address1</em>[,<em>address2</em>]]<b>y</b>/<em>abc</em>/<em>xyz</em>/</dt> +<dd>Transform each character by position in string +<em>abc</em> to its equivalent in string +<em>xyz</em>.</p></dd> + + +</dl> diff --git a/editors/vi.c b/editors/vi.c new file mode 100644 index 0000000..92c069d --- /dev/null +++ b/editors/vi.c @@ -0,0 +1,3954 @@ +/* vi: set sw=4 ts=4: */ +/* + * tiny vi.c: A small 'vi' clone + * Copyright (C) 2000, 2001 Sterling Huxley <sterling@europa.com> + * + * Licensed under the GPL v2 or later, see the file LICENSE in this tarball. + */ + +/* + * Things To Do: + * EXINIT + * $HOME/.exrc and ./.exrc + * add magic to search /foo.*bar + * add :help command + * :map macros + * if mark[] values were line numbers rather than pointers + * it would be easier to change the mark when add/delete lines + * More intelligence in refresh() + * ":r !cmd" and "!cmd" to filter text through an external command + * A true "undo" facility + * An "ex" line oriented mode- maybe using "cmdedit" + */ + +#include "libbb.h" + +/* the CRASHME code is unmaintained, and doesn't currently build */ +#define ENABLE_FEATURE_VI_CRASHME 0 + + +#if ENABLE_LOCALE_SUPPORT + +#if ENABLE_FEATURE_VI_8BIT +#define Isprint(c) isprint(c) +#else +#define Isprint(c) (isprint(c) && (unsigned char)(c) < 0x7f) +#endif + +#else + +/* 0x9b is Meta-ESC */ +#if ENABLE_FEATURE_VI_8BIT +#define Isprint(c) ((unsigned char)(c) >= ' ' && (c) != 0x7f && (unsigned char)(c) != 0x9b) +#else +#define Isprint(c) ((unsigned char)(c) >= ' ' && (unsigned char)(c) < 0x7f) +#endif + +#endif + + +enum { + MAX_TABSTOP = 32, // sanity limit + // User input len. Need not be extra big. + // Lines in file being edited *can* be bigger than this. + MAX_INPUT_LEN = 128, + // Sanity limits. We have only one buffer of this size. + MAX_SCR_COLS = CONFIG_FEATURE_VI_MAX_LEN, + MAX_SCR_ROWS = CONFIG_FEATURE_VI_MAX_LEN, +}; + +/* vt102 typical ESC sequence */ +/* terminal standout start/normal ESC sequence */ +static const char SOs[] ALIGN1 = "\033[7m"; +static const char SOn[] ALIGN1 = "\033[0m"; +/* terminal bell sequence */ +static const char bell[] ALIGN1 = "\007"; +/* Clear-end-of-line and Clear-end-of-screen ESC sequence */ +static const char Ceol[] ALIGN1 = "\033[0K"; +static const char Ceos[] ALIGN1 = "\033[0J"; +/* Cursor motion arbitrary destination ESC sequence */ +static const char CMrc[] ALIGN1 = "\033[%d;%dH"; +/* Cursor motion up and down ESC sequence */ +static const char CMup[] ALIGN1 = "\033[A"; +static const char CMdown[] ALIGN1 = "\n"; + +#if ENABLE_FEATURE_VI_DOT_CMD || ENABLE_FEATURE_VI_YANKMARK +// cmds modifying text[] +// vda: removed "aAiIs" as they switch us into insert mode +// and remembering input for replay after them makes no sense +static const char modifying_cmds[] = "cCdDJoOpPrRxX<>~"; +#endif + +enum { + YANKONLY = FALSE, + YANKDEL = TRUE, + FORWARD = 1, // code depends on "1" for array index + BACK = -1, // code depends on "-1" for array index + LIMITED = 0, // how much of text[] in char_search + FULL = 1, // how much of text[] in char_search + + S_BEFORE_WS = 1, // used in skip_thing() for moving "dot" + S_TO_WS = 2, // used in skip_thing() for moving "dot" + S_OVER_WS = 3, // used in skip_thing() for moving "dot" + S_END_PUNCT = 4, // used in skip_thing() for moving "dot" + S_END_ALNUM = 5, // used in skip_thing() for moving "dot" +}; + + +/* vi.c expects chars to be unsigned. */ +/* busybox build system provides that, but it's better */ +/* to audit and fix the source */ + +struct globals { + /* many references - keep near the top of globals */ + char *text, *end; // pointers to the user data in memory + char *dot; // where all the action takes place + int text_size; // size of the allocated buffer + + /* the rest */ + smallint vi_setops; +#define VI_AUTOINDENT 1 +#define VI_SHOWMATCH 2 +#define VI_IGNORECASE 4 +#define VI_ERR_METHOD 8 +#define autoindent (vi_setops & VI_AUTOINDENT) +#define showmatch (vi_setops & VI_SHOWMATCH ) +#define ignorecase (vi_setops & VI_IGNORECASE) +/* indicate error with beep or flash */ +#define err_method (vi_setops & VI_ERR_METHOD) + +#if ENABLE_FEATURE_VI_READONLY + smallint readonly_mode; +#define SET_READONLY_FILE(flags) ((flags) |= 0x01) +#define SET_READONLY_MODE(flags) ((flags) |= 0x02) +#define UNSET_READONLY_FILE(flags) ((flags) &= 0xfe) +#else +#define SET_READONLY_FILE(flags) ((void)0) +#define SET_READONLY_MODE(flags) ((void)0) +#define UNSET_READONLY_FILE(flags) ((void)0) +#endif + + smallint editing; // >0 while we are editing a file + // [code audit says "can be 0, 1 or 2 only"] + smallint cmd_mode; // 0=command 1=insert 2=replace + int file_modified; // buffer contents changed (counter, not flag!) + int last_file_modified; // = -1; + int fn_start; // index of first cmd line file name + int save_argc; // how many file names on cmd line + int cmdcnt; // repetition count + unsigned rows, columns; // the terminal screen is this size + int crow, ccol; // cursor is on Crow x Ccol + int offset; // chars scrolled off the screen to the left + int have_status_msg; // is default edit status needed? + // [don't make smallint!] + int last_status_cksum; // hash of current status line + char *current_filename; + char *screenbegin; // index into text[], of top line on the screen + char *screen; // pointer to the virtual screen buffer + int screensize; // and its size + int tabstop; + int last_forward_char; // last char searched for with 'f' (int because of Unicode) + char erase_char; // the users erase character + char last_input_char; // last char read from user + + smalluint chars_to_parse; +#if ENABLE_FEATURE_VI_DOT_CMD + smallint adding2q; // are we currently adding user input to q + int lmc_len; // length of last_modifying_cmd + char *ioq, *ioq_start; // pointer to string for get_one_char to "read" +#endif +#if ENABLE_FEATURE_VI_OPTIMIZE_CURSOR + int last_row; // where the cursor was last moved to +#endif +#if ENABLE_FEATURE_VI_USE_SIGNALS || ENABLE_FEATURE_VI_CRASHME + int my_pid; +#endif +#if ENABLE_FEATURE_VI_SEARCH + char *last_search_pattern; // last pattern from a '/' or '?' search +#endif + + /* former statics */ +#if ENABLE_FEATURE_VI_YANKMARK + char *edit_file__cur_line; +#endif + int refresh__old_offset; + int format_edit_status__tot; + + /* a few references only */ +#if ENABLE_FEATURE_VI_YANKMARK + int YDreg, Ureg; // default delete register and orig line for "U" + char *reg[28]; // named register a-z, "D", and "U" 0-25,26,27 + char *mark[28]; // user marks points somewhere in text[]- a-z and previous context '' + char *context_start, *context_end; +#endif +#if ENABLE_FEATURE_VI_USE_SIGNALS + sigjmp_buf restart; // catch_sig() +#endif + struct termios term_orig, term_vi; // remember what the cooked mode was +#if ENABLE_FEATURE_VI_COLON + char *initial_cmds[3]; // currently 2 entries, NULL terminated +#endif + // Should be just enough to hold a key sequence, + // but CRASHME mode uses it as generated command buffer too +#if ENABLE_FEATURE_VI_CRASHME + char readbuffer[128]; +#else + char readbuffer[KEYCODE_BUFFER_SIZE]; +#endif +#define STATUS_BUFFER_LEN 200 + char status_buffer[STATUS_BUFFER_LEN]; // messages to the user +#if ENABLE_FEATURE_VI_DOT_CMD + char last_modifying_cmd[MAX_INPUT_LEN]; // last modifying cmd for "." +#endif + char get_input_line__buf[MAX_INPUT_LEN]; /* former static */ + + char scr_out_buf[MAX_SCR_COLS + MAX_TABSTOP * 2]; +}; +#define G (*ptr_to_globals) +#define text (G.text ) +#define text_size (G.text_size ) +#define end (G.end ) +#define dot (G.dot ) +#define reg (G.reg ) + +#define vi_setops (G.vi_setops ) +#define editing (G.editing ) +#define cmd_mode (G.cmd_mode ) +#define file_modified (G.file_modified ) +#define last_file_modified (G.last_file_modified ) +#define fn_start (G.fn_start ) +#define save_argc (G.save_argc ) +#define cmdcnt (G.cmdcnt ) +#define rows (G.rows ) +#define columns (G.columns ) +#define crow (G.crow ) +#define ccol (G.ccol ) +#define offset (G.offset ) +#define status_buffer (G.status_buffer ) +#define have_status_msg (G.have_status_msg ) +#define last_status_cksum (G.last_status_cksum ) +#define current_filename (G.current_filename ) +#define screen (G.screen ) +#define screensize (G.screensize ) +#define screenbegin (G.screenbegin ) +#define tabstop (G.tabstop ) +#define last_forward_char (G.last_forward_char ) +#define erase_char (G.erase_char ) +#define last_input_char (G.last_input_char ) +#define chars_to_parse (G.chars_to_parse ) +#if ENABLE_FEATURE_VI_READONLY +#define readonly_mode (G.readonly_mode ) +#else +#define readonly_mode 0 +#endif +#define adding2q (G.adding2q ) +#define lmc_len (G.lmc_len ) +#define ioq (G.ioq ) +#define ioq_start (G.ioq_start ) +#define last_row (G.last_row ) +#define my_pid (G.my_pid ) +#define last_search_pattern (G.last_search_pattern) + +#define edit_file__cur_line (G.edit_file__cur_line) +#define refresh__old_offset (G.refresh__old_offset) +#define format_edit_status__tot (G.format_edit_status__tot) + +#define YDreg (G.YDreg ) +#define Ureg (G.Ureg ) +#define mark (G.mark ) +#define context_start (G.context_start ) +#define context_end (G.context_end ) +#define restart (G.restart ) +#define term_orig (G.term_orig ) +#define term_vi (G.term_vi ) +#define initial_cmds (G.initial_cmds ) +#define readbuffer (G.readbuffer ) +#define scr_out_buf (G.scr_out_buf ) +#define last_modifying_cmd (G.last_modifying_cmd ) +#define get_input_line__buf (G.get_input_line__buf) + +#define INIT_G() do { \ + SET_PTR_TO_GLOBALS(xzalloc(sizeof(G))); \ + last_file_modified = -1; \ + /* "" but has space for 2 chars: */ \ + USE_FEATURE_VI_SEARCH(last_search_pattern = xzalloc(2);) \ +} while (0) + + +static int init_text_buffer(char *); // init from file or create new +static void edit_file(char *); // edit one file +static void do_cmd(int); // execute a command +static int next_tabstop(int); +static void sync_cursor(char *, int *, int *); // synchronize the screen cursor to dot +static char *begin_line(char *); // return pointer to cur line B-o-l +static char *end_line(char *); // return pointer to cur line E-o-l +static char *prev_line(char *); // return pointer to prev line B-o-l +static char *next_line(char *); // return pointer to next line B-o-l +static char *end_screen(void); // get pointer to last char on screen +static int count_lines(char *, char *); // count line from start to stop +static char *find_line(int); // find begining of line #li +static char *move_to_col(char *, int); // move "p" to column l +static void dot_left(void); // move dot left- dont leave line +static void dot_right(void); // move dot right- dont leave line +static void dot_begin(void); // move dot to B-o-l +static void dot_end(void); // move dot to E-o-l +static void dot_next(void); // move dot to next line B-o-l +static void dot_prev(void); // move dot to prev line B-o-l +static void dot_scroll(int, int); // move the screen up or down +static void dot_skip_over_ws(void); // move dot pat WS +static void dot_delete(void); // delete the char at 'dot' +static char *bound_dot(char *); // make sure text[0] <= P < "end" +static char *new_screen(int, int); // malloc virtual screen memory +static char *char_insert(char *, char); // insert the char c at 'p' +static char *stupid_insert(char *, char); // stupidly insert the char c at 'p' +static int find_range(char **, char **, char); // return pointers for an object +static int st_test(char *, int, int, char *); // helper for skip_thing() +static char *skip_thing(char *, int, int, int); // skip some object +static char *find_pair(char *, char); // find matching pair () [] {} +static char *text_hole_delete(char *, char *); // at "p", delete a 'size' byte hole +static char *text_hole_make(char *, int); // at "p", make a 'size' byte hole +static char *yank_delete(char *, char *, int, int); // yank text[] into register then delete +static void show_help(void); // display some help info +static void rawmode(void); // set "raw" mode on tty +static void cookmode(void); // return to "cooked" mode on tty +// sleep for 'h' 1/100 seconds, return 1/0 if stdin is (ready for read)/(not ready) +static int mysleep(int); +static int readit(void); // read (maybe cursor) key from stdin +static int get_one_char(void); // read 1 char from stdin +static int file_size(const char *); // what is the byte size of "fn" +#if ENABLE_FEATURE_VI_READONLY +static int file_insert(const char *, char *, int); +#else +static int file_insert(const char *, char *); +#endif +static int file_write(char *, char *, char *); +#if !ENABLE_FEATURE_VI_OPTIMIZE_CURSOR +#define place_cursor(a, b, optimize) place_cursor(a, b) +#endif +static void place_cursor(int, int, int); +static void screen_erase(void); +static void clear_to_eol(void); +static void clear_to_eos(void); +static void go_bottom_and_clear_to_eol(void); +static void standout_start(void); // send "start reverse video" sequence +static void standout_end(void); // send "end reverse video" sequence +static void flash(int); // flash the terminal screen +static void show_status_line(void); // put a message on the bottom line +static void status_line(const char *, ...); // print to status buf +static void status_line_bold(const char *, ...); +static void not_implemented(const char *); // display "Not implemented" message +static int format_edit_status(void); // format file status on status line +static void redraw(int); // force a full screen refresh +static char* format_line(char* /*, int*/); +static void refresh(int); // update the terminal from screen[] + +static void Indicate_Error(void); // use flash or beep to indicate error +#define indicate_error(c) Indicate_Error() +static void Hit_Return(void); + +#if ENABLE_FEATURE_VI_SEARCH +static char *char_search(char *, const char *, int, int); // search for pattern starting at p +static int mycmp(const char *, const char *, int); // string cmp based in "ignorecase" +#endif +#if ENABLE_FEATURE_VI_COLON +static char *get_one_address(char *, int *); // get colon addr, if present +static char *get_address(char *, int *, int *); // get two colon addrs, if present +static void colon(char *); // execute the "colon" mode cmds +#endif +#if ENABLE_FEATURE_VI_USE_SIGNALS +static void winch_sig(int); // catch window size changes +static void suspend_sig(int); // catch ctrl-Z +static void catch_sig(int); // catch ctrl-C and alarm time-outs +#endif +#if ENABLE_FEATURE_VI_DOT_CMD +static void start_new_cmd_q(char); // new queue for command +static void end_cmd_q(void); // stop saving input chars +#else +#define end_cmd_q() ((void)0) +#endif +#if ENABLE_FEATURE_VI_SETOPTS +static void showmatching(char *); // show the matching pair () [] {} +#endif +#if ENABLE_FEATURE_VI_YANKMARK || (ENABLE_FEATURE_VI_COLON && ENABLE_FEATURE_VI_SEARCH) || ENABLE_FEATURE_VI_CRASHME +static char *string_insert(char *, char *); // insert the string at 'p' +#endif +#if ENABLE_FEATURE_VI_YANKMARK +static char *text_yank(char *, char *, int); // save copy of "p" into a register +static char what_reg(void); // what is letter of current YDreg +static void check_context(char); // remember context for '' command +#endif +#if ENABLE_FEATURE_VI_CRASHME +static void crash_dummy(); +static void crash_test(); +static int crashme = 0; +#endif + + +static void write1(const char *out) +{ + fputs(out, stdout); +} + +int vi_main(int argc, char **argv) MAIN_EXTERNALLY_VISIBLE; +int vi_main(int argc, char **argv) +{ + int c; + + INIT_G(); + +#if ENABLE_FEATURE_VI_USE_SIGNALS || ENABLE_FEATURE_VI_CRASHME + my_pid = getpid(); +#endif +#if ENABLE_FEATURE_VI_CRASHME + srand((long) my_pid); +#endif +#ifdef NO_SUCH_APPLET_YET + /* If we aren't "vi", we are "view" */ + if (ENABLE_FEATURE_VI_READONLY && applet_name[2]) { + SET_READONLY_MODE(readonly_mode); + } +#endif + + vi_setops = VI_AUTOINDENT | VI_SHOWMATCH | VI_IGNORECASE; + // 1- process $HOME/.exrc file (not inplemented yet) + // 2- process EXINIT variable from environment + // 3- process command line args +#if ENABLE_FEATURE_VI_COLON + { + char *p = getenv("EXINIT"); + if (p && *p) + initial_cmds[0] = xstrndup(p, MAX_INPUT_LEN); + } +#endif + while ((c = getopt(argc, argv, "hCRH" USE_FEATURE_VI_COLON("c:"))) != -1) { + switch (c) { +#if ENABLE_FEATURE_VI_CRASHME + case 'C': + crashme = 1; + break; +#endif +#if ENABLE_FEATURE_VI_READONLY + case 'R': // Read-only flag + SET_READONLY_MODE(readonly_mode); + break; +#endif +#if ENABLE_FEATURE_VI_COLON + case 'c': // cmd line vi command + if (*optarg) + initial_cmds[initial_cmds[0] != 0] = xstrndup(optarg, MAX_INPUT_LEN); + break; +#endif + case 'H': + show_help(); + /* fall through */ + default: + bb_show_usage(); + return 1; + } + } + + // The argv array can be used by the ":next" and ":rewind" commands + // save optind. + fn_start = optind; // remember first file name for :next and :rew + save_argc = argc; + + //----- This is the main file handling loop -------------- + if (optind >= argc) { + edit_file(0); + } else { + for (; optind < argc; optind++) { + edit_file(argv[optind]); + } + } + //----------------------------------------------------------- + + return 0; +} + +/* read text from file or create an empty buf */ +/* will also update current_filename */ +static int init_text_buffer(char *fn) +{ + int rc; + int size = file_size(fn); // file size. -1 means does not exist. + + /* allocate/reallocate text buffer */ + free(text); + text_size = size + 10240; + screenbegin = dot = end = text = xzalloc(text_size); + + if (fn != current_filename) { + free(current_filename); + current_filename = xstrdup(fn); + } + if (size < 0) { + // file dont exist. Start empty buf with dummy line + char_insert(text, '\n'); + rc = 0; + } else { + rc = file_insert(fn, text + USE_FEATURE_VI_READONLY(, 1)); + } + file_modified = 0; + last_file_modified = -1; +#if ENABLE_FEATURE_VI_YANKMARK + /* init the marks. */ + memset(mark, 0, sizeof(mark)); +#endif + return rc; +} + +static void edit_file(char *fn) +{ +#if ENABLE_FEATURE_VI_YANKMARK +#define cur_line edit_file__cur_line +#endif + int c; + int size; +#if ENABLE_FEATURE_VI_USE_SIGNALS + int sig; +#endif + + editing = 1; // 0 = exit, 1 = one file, 2 = multiple files + rawmode(); + rows = 24; + columns = 80; + size = 0; + if (ENABLE_FEATURE_VI_WIN_RESIZE) { + get_terminal_width_height(0, &columns, &rows); + if (rows > MAX_SCR_ROWS) rows = MAX_SCR_ROWS; + if (columns > MAX_SCR_COLS) columns = MAX_SCR_COLS; + } + new_screen(rows, columns); // get memory for virtual screen + init_text_buffer(fn); + +#if ENABLE_FEATURE_VI_YANKMARK + YDreg = 26; // default Yank/Delete reg + Ureg = 27; // hold orig line for "U" cmd + mark[26] = mark[27] = text; // init "previous context" +#endif + + last_forward_char = last_input_char = '\0'; + crow = 0; + ccol = 0; + +#if ENABLE_FEATURE_VI_USE_SIGNALS + catch_sig(0); + signal(SIGWINCH, winch_sig); + signal(SIGTSTP, suspend_sig); + sig = sigsetjmp(restart, 1); + if (sig != 0) { + screenbegin = dot = text; + } +#endif + + cmd_mode = 0; // 0=command 1=insert 2='R'eplace + cmdcnt = 0; + tabstop = 8; + offset = 0; // no horizontal offset + c = '\0'; +#if ENABLE_FEATURE_VI_DOT_CMD + free(ioq_start); + ioq = ioq_start = NULL; + lmc_len = 0; + adding2q = 0; +#endif + +#if ENABLE_FEATURE_VI_COLON + { + char *p, *q; + int n = 0; + + while ((p = initial_cmds[n])) { + do { + q = p; + p = strchr(q, '\n'); + if (p) + while (*p == '\n') + *p++ = '\0'; + if (*q) + colon(q); + } while (p); + free(initial_cmds[n]); + initial_cmds[n] = NULL; + n++; + } + } +#endif + redraw(FALSE); // dont force every col re-draw + //------This is the main Vi cmd handling loop ----------------------- + while (editing > 0) { +#if ENABLE_FEATURE_VI_CRASHME + if (crashme > 0) { + if ((end - text) > 1) { + crash_dummy(); // generate a random command + } else { + crashme = 0; + dot = string_insert(text, "\n\n##### Ran out of text to work on. #####\n\n"); // insert the string + refresh(FALSE); + } + } +#endif + last_input_char = c = get_one_char(); // get a cmd from user +#if ENABLE_FEATURE_VI_YANKMARK + // save a copy of the current line- for the 'U" command + if (begin_line(dot) != cur_line) { + cur_line = begin_line(dot); + text_yank(begin_line(dot), end_line(dot), Ureg); + } +#endif +#if ENABLE_FEATURE_VI_DOT_CMD + // These are commands that change text[]. + // Remember the input for the "." command + if (!adding2q && ioq_start == NULL + && cmd_mode == 0 // command mode + && c > '\0' // exclude NUL and non-ASCII chars + && c < 0x7f // (Unicode and such) + && strchr(modifying_cmds, c) + ) { + start_new_cmd_q(c); + } +#endif + do_cmd(c); // execute the user command + + // poll to see if there is input already waiting. if we are + // not able to display output fast enough to keep up, skip + // the display update until we catch up with input. + if (!chars_to_parse && mysleep(0) == 0) { + // no input pending - so update output + refresh(FALSE); + show_status_line(); + } +#if ENABLE_FEATURE_VI_CRASHME + if (crashme > 0) + crash_test(); // test editor variables +#endif + } + //------------------------------------------------------------------- + + go_bottom_and_clear_to_eol(); + cookmode(); +#undef cur_line +} + +//----- The Colon commands ------------------------------------- +#if ENABLE_FEATURE_VI_COLON +static char *get_one_address(char *p, int *addr) // get colon addr, if present +{ + int st; + char *q; + USE_FEATURE_VI_YANKMARK(char c;) + USE_FEATURE_VI_SEARCH(char *pat;) + + *addr = -1; // assume no addr + if (*p == '.') { // the current line + p++; + q = begin_line(dot); + *addr = count_lines(text, q); + } +#if ENABLE_FEATURE_VI_YANKMARK + else if (*p == '\'') { // is this a mark addr + p++; + c = tolower(*p); + p++; + if (c >= 'a' && c <= 'z') { + // we have a mark + c = c - 'a'; + q = mark[(unsigned char) c]; + if (q != NULL) { // is mark valid + *addr = count_lines(text, q); // count lines + } + } + } +#endif +#if ENABLE_FEATURE_VI_SEARCH + else if (*p == '/') { // a search pattern + q = strchrnul(++p, '/'); + pat = xstrndup(p, q - p); // save copy of pattern + p = q; + if (*p == '/') + p++; + q = char_search(dot, pat, FORWARD, FULL); + if (q != NULL) { + *addr = count_lines(text, q); + } + free(pat); + } +#endif + else if (*p == '$') { // the last line in file + p++; + q = begin_line(end - 1); + *addr = count_lines(text, q); + } else if (isdigit(*p)) { // specific line number + sscanf(p, "%d%n", addr, &st); + p += st; + } else { + // unrecognised address - assume -1 + *addr = -1; + } + return p; +} + +static char *get_address(char *p, int *b, int *e) // get two colon addrs, if present +{ + //----- get the address' i.e., 1,3 'a,'b ----- + // get FIRST addr, if present + while (isblank(*p)) + p++; // skip over leading spaces + if (*p == '%') { // alias for 1,$ + p++; + *b = 1; + *e = count_lines(text, end-1); + goto ga0; + } + p = get_one_address(p, b); + while (isblank(*p)) + p++; + if (*p == ',') { // is there a address separator + p++; + while (isblank(*p)) + p++; + // get SECOND addr, if present + p = get_one_address(p, e); + } + ga0: + while (isblank(*p)) + p++; // skip over trailing spaces + return p; +} + +#if ENABLE_FEATURE_VI_SET && ENABLE_FEATURE_VI_SETOPTS +static void setops(const char *args, const char *opname, int flg_no, + const char *short_opname, int opt) +{ + const char *a = args + flg_no; + int l = strlen(opname) - 1; /* opname have + ' ' */ + + if (strncasecmp(a, opname, l) == 0 + || strncasecmp(a, short_opname, 2) == 0 + ) { + if (flg_no) + vi_setops &= ~opt; + else + vi_setops |= opt; + } +} +#endif + +// buf must be no longer than MAX_INPUT_LEN! +static void colon(char *buf) +{ + char c, *orig_buf, *buf1, *q, *r; + char *fn, cmd[MAX_INPUT_LEN], args[MAX_INPUT_LEN]; + int i, l, li, ch, b, e; + int useforce, forced = FALSE; + + // :3154 // if (-e line 3154) goto it else stay put + // :4,33w! foo // write a portion of buffer to file "foo" + // :w // write all of buffer to current file + // :q // quit + // :q! // quit- dont care about modified file + // :'a,'z!sort -u // filter block through sort + // :'f // goto mark "f" + // :'fl // list literal the mark "f" line + // :.r bar // read file "bar" into buffer before dot + // :/123/,/abc/d // delete lines from "123" line to "abc" line + // :/xyz/ // goto the "xyz" line + // :s/find/replace/ // substitute pattern "find" with "replace" + // :!<cmd> // run <cmd> then return + // + + if (!buf[0]) + goto vc1; + if (*buf == ':') + buf++; // move past the ':' + + li = ch = i = 0; + b = e = -1; + q = text; // assume 1,$ for the range + r = end - 1; + li = count_lines(text, end - 1); + fn = current_filename; + + // look for optional address(es) :. :1 :1,9 :'q,'a :% + buf = get_address(buf, &b, &e); + + // remember orig command line + orig_buf = buf; + + // get the COMMAND into cmd[] + buf1 = cmd; + while (*buf != '\0') { + if (isspace(*buf)) + break; + *buf1++ = *buf++; + } + *buf1 = '\0'; + // get any ARGuments + while (isblank(*buf)) + buf++; + strcpy(args, buf); + useforce = FALSE; + buf1 = last_char_is(cmd, '!'); + if (buf1) { + useforce = TRUE; + *buf1 = '\0'; // get rid of ! + } + if (b >= 0) { + // if there is only one addr, then the addr + // is the line number of the single line the + // user wants. So, reset the end + // pointer to point at end of the "b" line + q = find_line(b); // what line is #b + r = end_line(q); + li = 1; + } + if (e >= 0) { + // we were given two addrs. change the + // end pointer to the addr given by user. + r = find_line(e); // what line is #e + r = end_line(r); + li = e - b + 1; + } + // ------------ now look for the command ------------ + i = strlen(cmd); + if (i == 0) { // :123CR goto line #123 + if (b >= 0) { + dot = find_line(b); // what line is #b + dot_skip_over_ws(); + } + } +#if ENABLE_FEATURE_ALLOW_EXEC + else if (strncmp(cmd, "!", 1) == 0) { // run a cmd + int retcode; + // :!ls run the <cmd> + go_bottom_and_clear_to_eol(); + cookmode(); + retcode = system(orig_buf + 1); // run the cmd + if (retcode) + printf("\nshell returned %i\n\n", retcode); + rawmode(); + Hit_Return(); // let user see results + } +#endif + else if (strncmp(cmd, "=", i) == 0) { // where is the address + if (b < 0) { // no addr given- use defaults + b = e = count_lines(text, dot); + } + status_line("%d", b); + } else if (strncasecmp(cmd, "delete", i) == 0) { // delete lines + if (b < 0) { // no addr given- use defaults + q = begin_line(dot); // assume .,. for the range + r = end_line(dot); + } + dot = yank_delete(q, r, 1, YANKDEL); // save, then delete lines + dot_skip_over_ws(); + } else if (strncasecmp(cmd, "edit", i) == 0) { // Edit a file + // don't edit, if the current file has been modified + if (file_modified && !useforce) { + status_line_bold("No write since last change (:edit! overrides)"); + goto vc1; + } + if (args[0]) { + // the user supplied a file name + fn = args; + } else if (current_filename && current_filename[0]) { + // no user supplied name- use the current filename + // fn = current_filename; was set by default + } else { + // no user file name, no current name- punt + status_line_bold("No current filename"); + goto vc1; + } + + if (init_text_buffer(fn) < 0) + goto vc1; + +#if ENABLE_FEATURE_VI_YANKMARK + if (Ureg >= 0 && Ureg < 28 && reg[Ureg] != 0) { + free(reg[Ureg]); // free orig line reg- for 'U' + reg[Ureg]= 0; + } + if (YDreg >= 0 && YDreg < 28 && reg[YDreg] != 0) { + free(reg[YDreg]); // free default yank/delete register + reg[YDreg]= 0; + } +#endif + // how many lines in text[]? + li = count_lines(text, end - 1); + status_line("\"%s\"%s" + USE_FEATURE_VI_READONLY("%s") + " %dL, %dC", current_filename, + (file_size(fn) < 0 ? " [New file]" : ""), + USE_FEATURE_VI_READONLY( + ((readonly_mode) ? " [Readonly]" : ""), + ) + li, ch); + } else if (strncasecmp(cmd, "file", i) == 0) { // what File is this + if (b != -1 || e != -1) { + not_implemented("No address allowed on this command"); + goto vc1; + } + if (args[0]) { + // user wants a new filename + free(current_filename); + current_filename = xstrdup(args); + } else { + // user wants file status info + last_status_cksum = 0; // force status update + } + } else if (strncasecmp(cmd, "features", i) == 0) { // what features are available + // print out values of all features + go_bottom_and_clear_to_eol(); + cookmode(); + show_help(); + rawmode(); + Hit_Return(); + } else if (strncasecmp(cmd, "list", i) == 0) { // literal print line + if (b < 0) { // no addr given- use defaults + q = begin_line(dot); // assume .,. for the range + r = end_line(dot); + } + go_bottom_and_clear_to_eol(); + puts("\r"); + for (; q <= r; q++) { + int c_is_no_print; + + c = *q; + c_is_no_print = (c & 0x80) && !Isprint(c); + if (c_is_no_print) { + c = '.'; + standout_start(); + } + if (c == '\n') { + write1("$\r"); + } else if (c < ' ' || c == 127) { + bb_putchar('^'); + if (c == 127) + c = '?'; + else + c += '@'; + } + bb_putchar(c); + if (c_is_no_print) + standout_end(); + } +#if ENABLE_FEATURE_VI_SET + vc2: +#endif + Hit_Return(); + } else if (strncasecmp(cmd, "quit", i) == 0 // Quit + || strncasecmp(cmd, "next", i) == 0 // edit next file + ) { + if (useforce) { + // force end of argv list + if (*cmd == 'q') { + optind = save_argc; + } + editing = 0; + goto vc1; + } + // don't exit if the file been modified + if (file_modified) { + status_line_bold("No write since last change (:%s! overrides)", + (*cmd == 'q' ? "quit" : "next")); + goto vc1; + } + // are there other file to edit + if (*cmd == 'q' && optind < save_argc - 1) { + status_line_bold("%d more file to edit", (save_argc - optind - 1)); + goto vc1; + } + if (*cmd == 'n' && optind >= save_argc - 1) { + status_line_bold("No more files to edit"); + goto vc1; + } + editing = 0; + } else if (strncasecmp(cmd, "read", i) == 0) { // read file into text[] + fn = args; + if (!fn[0]) { + status_line_bold("No filename given"); + goto vc1; + } + if (b < 0) { // no addr given- use defaults + q = begin_line(dot); // assume "dot" + } + // read after current line- unless user said ":0r foo" + if (b != 0) + q = next_line(q); + ch = file_insert(fn, q USE_FEATURE_VI_READONLY(, 0)); + if (ch < 0) + goto vc1; // nothing was inserted + // how many lines in text[]? + li = count_lines(q, q + ch - 1); + status_line("\"%s\"" + USE_FEATURE_VI_READONLY("%s") + " %dL, %dC", fn, + USE_FEATURE_VI_READONLY((readonly_mode ? " [Readonly]" : ""),) + li, ch); + if (ch > 0) { + // if the insert is before "dot" then we need to update + if (q <= dot) + dot += ch; + file_modified++; + } + } else if (strncasecmp(cmd, "rewind", i) == 0) { // rewind cmd line args + if (file_modified && !useforce) { + status_line_bold("No write since last change (:rewind! overrides)"); + } else { + // reset the filenames to edit + optind = fn_start - 1; + editing = 0; + } +#if ENABLE_FEATURE_VI_SET + } else if (strncasecmp(cmd, "set", i) == 0) { // set or clear features +#if ENABLE_FEATURE_VI_SETOPTS + char *argp; +#endif + i = 0; // offset into args + // only blank is regarded as args delmiter. What about tab '\t' ? + if (!args[0] || strcasecmp(args, "all") == 0) { + // print out values of all options + go_bottom_and_clear_to_eol(); + printf("----------------------------------------\r\n"); +#if ENABLE_FEATURE_VI_SETOPTS + if (!autoindent) + printf("no"); + printf("autoindent "); + if (!err_method) + printf("no"); + printf("flash "); + if (!ignorecase) + printf("no"); + printf("ignorecase "); + if (!showmatch) + printf("no"); + printf("showmatch "); + printf("tabstop=%d ", tabstop); +#endif + printf("\r\n"); + goto vc2; + } +#if ENABLE_FEATURE_VI_SETOPTS + argp = args; + while (*argp) { + if (strncasecmp(argp, "no", 2) == 0) + i = 2; // ":set noautoindent" + setops(argp, "autoindent ", i, "ai", VI_AUTOINDENT); + setops(argp, "flash ", i, "fl", VI_ERR_METHOD); + setops(argp, "ignorecase ", i, "ic", VI_IGNORECASE); + setops(argp, "showmatch ", i, "ic", VI_SHOWMATCH); + /* tabstopXXXX */ + if (strncasecmp(argp + i, "tabstop=%d ", 7) == 0) { + sscanf(strchr(argp + i, '='), "tabstop=%d" + 7, &ch); + if (ch > 0 && ch <= MAX_TABSTOP) + tabstop = ch; + } + while (*argp && *argp != ' ') + argp++; // skip to arg delimiter (i.e. blank) + while (*argp && *argp == ' ') + argp++; // skip all delimiting blanks + } +#endif /* FEATURE_VI_SETOPTS */ +#endif /* FEATURE_VI_SET */ +#if ENABLE_FEATURE_VI_SEARCH + } else if (strncasecmp(cmd, "s", 1) == 0) { // substitute a pattern with a replacement pattern + char *ls, *F, *R; + int gflag; + + // F points to the "find" pattern + // R points to the "replace" pattern + // replace the cmd line delimiters "/" with NULLs + gflag = 0; // global replace flag + c = orig_buf[1]; // what is the delimiter + F = orig_buf + 2; // start of "find" + R = strchr(F, c); // middle delimiter + if (!R) goto colon_s_fail; + *R++ = '\0'; // terminate "find" + buf1 = strchr(R, c); + if (!buf1) goto colon_s_fail; + *buf1++ = '\0'; // terminate "replace" + if (*buf1 == 'g') { // :s/foo/bar/g + buf1++; + gflag++; // turn on gflag + } + q = begin_line(q); + if (b < 0) { // maybe :s/foo/bar/ + q = begin_line(dot); // start with cur line + b = count_lines(text, q); // cur line number + } + if (e < 0) + e = b; // maybe :.s/foo/bar/ + for (i = b; i <= e; i++) { // so, :20,23 s \0 find \0 replace \0 + ls = q; // orig line start + vc4: + buf1 = char_search(q, F, FORWARD, LIMITED); // search cur line only for "find" + if (buf1) { + // we found the "find" pattern - delete it + text_hole_delete(buf1, buf1 + strlen(F) - 1); + // inset the "replace" patern + string_insert(buf1, R); // insert the string + // check for "global" :s/foo/bar/g + if (gflag == 1) { + if ((buf1 + strlen(R)) < end_line(ls)) { + q = buf1 + strlen(R); + goto vc4; // don't let q move past cur line + } + } + } + q = next_line(ls); + } +#endif /* FEATURE_VI_SEARCH */ + } else if (strncasecmp(cmd, "version", i) == 0) { // show software version + status_line(BB_VER " " BB_BT); + } else if (strncasecmp(cmd, "write", i) == 0 // write text to file + || strncasecmp(cmd, "wq", i) == 0 + || strncasecmp(cmd, "wn", i) == 0 + || strncasecmp(cmd, "x", i) == 0 + ) { + // is there a file name to write to? + if (args[0]) { + fn = args; + } +#if ENABLE_FEATURE_VI_READONLY + if (readonly_mode && !useforce) { + status_line_bold("\"%s\" File is read only", fn); + goto vc3; + } +#endif + // how many lines in text[]? + li = count_lines(q, r); + ch = r - q + 1; + // see if file exists- if not, its just a new file request + if (useforce) { + // if "fn" is not write-able, chmod u+w + // sprintf(syscmd, "chmod u+w %s", fn); + // system(syscmd); + forced = TRUE; + } + l = file_write(fn, q, r); + if (useforce && forced) { + // chmod u-w + // sprintf(syscmd, "chmod u-w %s", fn); + // system(syscmd); + forced = FALSE; + } + if (l < 0) { + if (l == -1) + status_line_bold("\"%s\" %s", fn, strerror(errno)); + } else { + status_line("\"%s\" %dL, %dC", fn, li, l); + if (q == text && r == end - 1 && l == ch) { + file_modified = 0; + last_file_modified = -1; + } + if ((cmd[0] == 'x' || cmd[1] == 'q' || cmd[1] == 'n' || + cmd[0] == 'X' || cmd[1] == 'Q' || cmd[1] == 'N') + && l == ch) { + editing = 0; + } + } +#if ENABLE_FEATURE_VI_READONLY + vc3:; +#endif +#if ENABLE_FEATURE_VI_YANKMARK + } else if (strncasecmp(cmd, "yank", i) == 0) { // yank lines + if (b < 0) { // no addr given- use defaults + q = begin_line(dot); // assume .,. for the range + r = end_line(dot); + } + text_yank(q, r, YDreg); + li = count_lines(q, r); + status_line("Yank %d lines (%d chars) into [%c]", + li, strlen(reg[YDreg]), what_reg()); +#endif + } else { + // cmd unknown + not_implemented(cmd); + } + vc1: + dot = bound_dot(dot); // make sure "dot" is valid + return; +#if ENABLE_FEATURE_VI_SEARCH + colon_s_fail: + status_line(":s expression missing delimiters"); +#endif +} + +#endif /* FEATURE_VI_COLON */ + +static void Hit_Return(void) +{ + int c; + + standout_start(); + write1("[Hit return to continue]"); + standout_end(); + while ((c = get_one_char()) != '\n' && c != '\r') + continue; + redraw(TRUE); // force redraw all +} + +static int next_tabstop(int col) +{ + return col + ((tabstop - 1) - (col % tabstop)); +} + +//----- Synchronize the cursor to Dot -------------------------- +static void sync_cursor(char *d, int *row, int *col) +{ + char *beg_cur; // begin and end of "d" line + char *tp; + int cnt, ro, co; + + beg_cur = begin_line(d); // first char of cur line + + if (beg_cur < screenbegin) { + // "d" is before top line on screen + // how many lines do we have to move + cnt = count_lines(beg_cur, screenbegin); + sc1: + screenbegin = beg_cur; + if (cnt > (rows - 1) / 2) { + // we moved too many lines. put "dot" in middle of screen + for (cnt = 0; cnt < (rows - 1) / 2; cnt++) { + screenbegin = prev_line(screenbegin); + } + } + } else { + char *end_scr; // begin and end of screen + end_scr = end_screen(); // last char of screen + if (beg_cur > end_scr) { + // "d" is after bottom line on screen + // how many lines do we have to move + cnt = count_lines(end_scr, beg_cur); + if (cnt > (rows - 1) / 2) + goto sc1; // too many lines + for (ro = 0; ro < cnt - 1; ro++) { + // move screen begin the same amount + screenbegin = next_line(screenbegin); + // now, move the end of screen + end_scr = next_line(end_scr); + end_scr = end_line(end_scr); + } + } + } + // "d" is on screen- find out which row + tp = screenbegin; + for (ro = 0; ro < rows - 1; ro++) { // drive "ro" to correct row + if (tp == beg_cur) + break; + tp = next_line(tp); + } + + // find out what col "d" is on + co = 0; + while (tp < d) { // drive "co" to correct column + if (*tp == '\n') //vda || *tp == '\0') + break; + if (*tp == '\t') { + // handle tabs like real vi + if (d == tp && cmd_mode) { + break; + } + co = next_tabstop(co); + } else if ((unsigned char)*tp < ' ' || *tp == 0x7f) { + co++; // display as ^X, use 2 columns + } + co++; + tp++; + } + + // "co" is the column where "dot" is. + // The screen has "columns" columns. + // The currently displayed columns are 0+offset -- columns+ofset + // |-------------------------------------------------------------| + // ^ ^ ^ + // offset | |------- columns ----------------| + // + // If "co" is already in this range then we do not have to adjust offset + // but, we do have to subtract the "offset" bias from "co". + // If "co" is outside this range then we have to change "offset". + // If the first char of a line is a tab the cursor will try to stay + // in column 7, but we have to set offset to 0. + + if (co < 0 + offset) { + offset = co; + } + if (co >= columns + offset) { + offset = co - columns + 1; + } + // if the first char of the line is a tab, and "dot" is sitting on it + // force offset to 0. + if (d == beg_cur && *d == '\t') { + offset = 0; + } + co -= offset; + + *row = ro; + *col = co; +} + +//----- Text Movement Routines --------------------------------- +static char *begin_line(char *p) // return pointer to first char cur line +{ + if (p > text) { + p = memrchr(text, '\n', p - text); + if (!p) + return text; + return p + 1; + } + return p; +} + +static char *end_line(char *p) // return pointer to NL of cur line +{ + if (p < end - 1) { + p = memchr(p, '\n', end - p - 1); + if (!p) + return end - 1; + } + return p; +} + +static char *dollar_line(char *p) // return pointer to just before NL line +{ + p = end_line(p); + // Try to stay off of the Newline + if (*p == '\n' && (p - begin_line(p)) > 0) + p--; + return p; +} + +static char *prev_line(char *p) // return pointer first char prev line +{ + p = begin_line(p); // goto begining of cur line + if (p > text && p[-1] == '\n') + p--; // step to prev line + p = begin_line(p); // goto begining of prev line + return p; +} + +static char *next_line(char *p) // return pointer first char next line +{ + p = end_line(p); + if (p < end - 1 && *p == '\n') + p++; // step to next line + return p; +} + +//----- Text Information Routines ------------------------------ +static char *end_screen(void) +{ + char *q; + int cnt; + + // find new bottom line + q = screenbegin; + for (cnt = 0; cnt < rows - 2; cnt++) + q = next_line(q); + q = end_line(q); + return q; +} + +// count line from start to stop +static int count_lines(char *start, char *stop) +{ + char *q; + int cnt; + + if (stop < start) { // start and stop are backwards- reverse them + q = start; + start = stop; + stop = q; + } + cnt = 0; + stop = end_line(stop); + while (start <= stop && start <= end - 1) { + start = end_line(start); + if (*start == '\n') + cnt++; + start++; + } + return cnt; +} + +static char *find_line(int li) // find begining of line #li +{ + char *q; + + for (q = text; li > 1; li--) { + q = next_line(q); + } + return q; +} + +//----- Dot Movement Routines ---------------------------------- +static void dot_left(void) +{ + if (dot > text && dot[-1] != '\n') + dot--; +} + +static void dot_right(void) +{ + if (dot < end - 1 && *dot != '\n') + dot++; +} + +static void dot_begin(void) +{ + dot = begin_line(dot); // return pointer to first char cur line +} + +static void dot_end(void) +{ + dot = end_line(dot); // return pointer to last char cur line +} + +static char *move_to_col(char *p, int l) +{ + int co; + + p = begin_line(p); + co = 0; + while (co < l && p < end) { + if (*p == '\n') //vda || *p == '\0') + break; + if (*p == '\t') { + co = next_tabstop(co); + } else if (*p < ' ' || *p == 127) { + co++; // display as ^X, use 2 columns + } + co++; + p++; + } + return p; +} + +static void dot_next(void) +{ + dot = next_line(dot); +} + +static void dot_prev(void) +{ + dot = prev_line(dot); +} + +static void dot_scroll(int cnt, int dir) +{ + char *q; + + for (; cnt > 0; cnt--) { + if (dir < 0) { + // scroll Backwards + // ctrl-Y scroll up one line + screenbegin = prev_line(screenbegin); + } else { + // scroll Forwards + // ctrl-E scroll down one line + screenbegin = next_line(screenbegin); + } + } + // make sure "dot" stays on the screen so we dont scroll off + if (dot < screenbegin) + dot = screenbegin; + q = end_screen(); // find new bottom line + if (dot > q) + dot = begin_line(q); // is dot is below bottom line? + dot_skip_over_ws(); +} + +static void dot_skip_over_ws(void) +{ + // skip WS + while (isspace(*dot) && *dot != '\n' && dot < end - 1) + dot++; +} + +static void dot_delete(void) // delete the char at 'dot' +{ + text_hole_delete(dot, dot); +} + +static char *bound_dot(char *p) // make sure text[0] <= P < "end" +{ + if (p >= end && end > text) { + p = end - 1; + indicate_error('1'); + } + if (p < text) { + p = text; + indicate_error('2'); + } + return p; +} + +//----- Helper Utility Routines -------------------------------- + +//---------------------------------------------------------------- +//----- Char Routines -------------------------------------------- +/* Chars that are part of a word- + * 0123456789_ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz + * Chars that are Not part of a word (stoppers) + * !"#$%&'()*+,-./:;<=>?@[\]^`{|}~ + * Chars that are WhiteSpace + * TAB NEWLINE VT FF RETURN SPACE + * DO NOT COUNT NEWLINE AS WHITESPACE + */ + +static char *new_screen(int ro, int co) +{ + int li; + + free(screen); + screensize = ro * co + 8; + screen = xmalloc(screensize); + // initialize the new screen. assume this will be a empty file. + screen_erase(); + // non-existent text[] lines start with a tilde (~). + for (li = 1; li < ro - 1; li++) { + screen[(li * co) + 0] = '~'; + } + return screen; +} + +#if ENABLE_FEATURE_VI_SEARCH +static int mycmp(const char *s1, const char *s2, int len) +{ + int i; + + i = strncmp(s1, s2, len); + if (ENABLE_FEATURE_VI_SETOPTS && ignorecase) { + i = strncasecmp(s1, s2, len); + } + return i; +} + +// search for pattern starting at p +static char *char_search(char *p, const char *pat, int dir, int range) +{ +#ifndef REGEX_SEARCH + char *start, *stop; + int len; + + len = strlen(pat); + if (dir == FORWARD) { + stop = end - 1; // assume range is p - end-1 + if (range == LIMITED) + stop = next_line(p); // range is to next line + for (start = p; start < stop; start++) { + if (mycmp(start, pat, len) == 0) { + return start; + } + } + } else if (dir == BACK) { + stop = text; // assume range is text - p + if (range == LIMITED) + stop = prev_line(p); // range is to prev line + for (start = p - len; start >= stop; start--) { + if (mycmp(start, pat, len) == 0) { + return start; + } + } + } + // pattern not found + return NULL; +#else /* REGEX_SEARCH */ + char *q; + struct re_pattern_buffer preg; + int i; + int size, range; + + re_syntax_options = RE_SYNTAX_POSIX_EXTENDED; + preg.translate = 0; + preg.fastmap = 0; + preg.buffer = 0; + preg.allocated = 0; + + // assume a LIMITED forward search + q = next_line(p); + q = end_line(q); + q = end - 1; + if (dir == BACK) { + q = prev_line(p); + q = text; + } + // count the number of chars to search over, forward or backward + size = q - p; + if (size < 0) + size = p - q; + // RANGE could be negative if we are searching backwards + range = q - p; + + q = re_compile_pattern(pat, strlen(pat), &preg); + if (q != 0) { + // The pattern was not compiled + status_line_bold("bad search pattern: \"%s\": %s", pat, q); + i = 0; // return p if pattern not compiled + goto cs1; + } + + q = p; + if (range < 0) { + q = p - size; + if (q < text) + q = text; + } + // search for the compiled pattern, preg, in p[] + // range < 0- search backward + // range > 0- search forward + // 0 < start < size + // re_search() < 0 not found or error + // re_search() > 0 index of found pattern + // struct pattern char int int int struct reg + // re_search (*pattern_buffer, *string, size, start, range, *regs) + i = re_search(&preg, q, size, 0, range, 0); + if (i == -1) { + p = 0; + i = 0; // return NULL if pattern not found + } + cs1: + if (dir == FORWARD) { + p = p + i; + } else { + p = p - i; + } + return p; +#endif /* REGEX_SEARCH */ +} +#endif /* FEATURE_VI_SEARCH */ + +static char *char_insert(char *p, char c) // insert the char c at 'p' +{ + if (c == 22) { // Is this an ctrl-V? + p = stupid_insert(p, '^'); // use ^ to indicate literal next + p--; // backup onto ^ + refresh(FALSE); // show the ^ + c = get_one_char(); + *p = c; + p++; + file_modified++; + } else if (c == 27) { // Is this an ESC? + cmd_mode = 0; + cmdcnt = 0; + end_cmd_q(); // stop adding to q + last_status_cksum = 0; // force status update + if ((p[-1] != '\n') && (dot > text)) { + p--; + } + } else if (c == erase_char || c == 8 || c == 127) { // Is this a BS + // 123456789 + if ((p[-1] != '\n') && (dot>text)) { + p--; + p = text_hole_delete(p, p); // shrink buffer 1 char + } + } else { + // insert a char into text[] + char *sp; // "save p" + + if (c == 13) + c = '\n'; // translate \r to \n + sp = p; // remember addr of insert + p = stupid_insert(p, c); // insert the char +#if ENABLE_FEATURE_VI_SETOPTS + if (showmatch && strchr(")]}", *sp) != NULL) { + showmatching(sp); + } + if (autoindent && c == '\n') { // auto indent the new line + char *q; + + q = prev_line(p); // use prev line as templet + for (; isblank(*q); q++) { + p = stupid_insert(p, *q); // insert the char + } + } +#endif + } + return p; +} + +static char *stupid_insert(char *p, char c) // stupidly insert the char c at 'p' +{ + p = text_hole_make(p, 1); + *p = c; + //file_modified++; - done by text_hole_make() + return p + 1; +} + +static int find_range(char **start, char **stop, char c) +{ + char *save_dot, *p, *q, *t; + int cnt, multiline = 0; + + save_dot = dot; + p = q = dot; + + if (strchr("cdy><", c)) { + // these cmds operate on whole lines + p = q = begin_line(p); + for (cnt = 1; cnt < cmdcnt; cnt++) { + q = next_line(q); + } + q = end_line(q); + } else if (strchr("^%$0bBeEfth\b\177", c)) { + // These cmds operate on char positions + do_cmd(c); // execute movement cmd + q = dot; + } else if (strchr("wW", c)) { + do_cmd(c); // execute movement cmd + // if we are at the next word's first char + // step back one char + // but check the possibilities when it is true + if (dot > text && ((isspace(dot[-1]) && !isspace(dot[0])) + || (ispunct(dot[-1]) && !ispunct(dot[0])) + || (isalnum(dot[-1]) && !isalnum(dot[0])))) + dot--; // move back off of next word + if (dot > text && *dot == '\n') + dot--; // stay off NL + q = dot; + } else if (strchr("H-k{", c)) { + // these operate on multi-lines backwards + q = end_line(dot); // find NL + do_cmd(c); // execute movement cmd + dot_begin(); + p = dot; + } else if (strchr("L+j}\r\n", c)) { + // these operate on multi-lines forwards + p = begin_line(dot); + do_cmd(c); // execute movement cmd + dot_end(); // find NL + q = dot; + } else { + // nothing -- this causes any other values of c to + // represent the one-character range under the + // cursor. this is correct for ' ' and 'l', but + // perhaps no others. + // + } + if (q < p) { + t = q; + q = p; + p = t; + } + + // backward char movements don't include start position + if (q > p && strchr("^0bBh\b\177", c)) q--; + + multiline = 0; + for (t = p; t <= q; t++) { + if (*t == '\n') { + multiline = 1; + break; + } + } + + *start = p; + *stop = q; + dot = save_dot; + return multiline; +} + +static int st_test(char *p, int type, int dir, char *tested) +{ + char c, c0, ci; + int test, inc; + + inc = dir; + c = c0 = p[0]; + ci = p[inc]; + test = 0; + + if (type == S_BEFORE_WS) { + c = ci; + test = ((!isspace(c)) || c == '\n'); + } + if (type == S_TO_WS) { + c = c0; + test = ((!isspace(c)) || c == '\n'); + } + if (type == S_OVER_WS) { + c = c0; + test = ((isspace(c))); + } + if (type == S_END_PUNCT) { + c = ci; + test = ((ispunct(c))); + } + if (type == S_END_ALNUM) { + c = ci; + test = ((isalnum(c)) || c == '_'); + } + *tested = c; + return test; +} + +static char *skip_thing(char *p, int linecnt, int dir, int type) +{ + char c; + + while (st_test(p, type, dir, &c)) { + // make sure we limit search to correct number of lines + if (c == '\n' && --linecnt < 1) + break; + if (dir >= 0 && p >= end - 1) + break; + if (dir < 0 && p <= text) + break; + p += dir; // move to next char + } + return p; +} + +// find matching char of pair () [] {} +static char *find_pair(char *p, const char c) +{ + char match, *q; + int dir, level; + + match = ')'; + level = 1; + dir = 1; // assume forward + switch (c) { + case '(': match = ')'; break; + case '[': match = ']'; break; + case '{': match = '}'; break; + case ')': match = '('; dir = -1; break; + case ']': match = '['; dir = -1; break; + case '}': match = '{'; dir = -1; break; + } + for (q = p + dir; text <= q && q < end; q += dir) { + // look for match, count levels of pairs (( )) + if (*q == c) + level++; // increase pair levels + if (*q == match) + level--; // reduce pair level + if (level == 0) + break; // found matching pair + } + if (level != 0) + q = NULL; // indicate no match + return q; +} + +#if ENABLE_FEATURE_VI_SETOPTS +// show the matching char of a pair, () [] {} +static void showmatching(char *p) +{ + char *q, *save_dot; + + // we found half of a pair + q = find_pair(p, *p); // get loc of matching char + if (q == NULL) { + indicate_error('3'); // no matching char + } else { + // "q" now points to matching pair + save_dot = dot; // remember where we are + dot = q; // go to new loc + refresh(FALSE); // let the user see it + mysleep(40); // give user some time + dot = save_dot; // go back to old loc + refresh(FALSE); + } +} +#endif /* FEATURE_VI_SETOPTS */ + +// open a hole in text[] +static char *text_hole_make(char *p, int size) // at "p", make a 'size' byte hole +{ + if (size <= 0) + return p; + end += size; // adjust the new END + if (end >= (text + text_size)) { + char *new_text; + text_size += end - (text + text_size) + 10240; + new_text = xrealloc(text, text_size); + screenbegin = new_text + (screenbegin - text); + dot = new_text + (dot - text); + end = new_text + (end - text); + p = new_text + (p - text); + text = new_text; + } + memmove(p + size, p, end - size - p); + memset(p, ' ', size); // clear new hole + file_modified++; + return p; +} + +// close a hole in text[] +static char *text_hole_delete(char *p, char *q) // delete "p" through "q", inclusive +{ + char *src, *dest; + int cnt, hole_size; + + // move forwards, from beginning + // assume p <= q + src = q + 1; + dest = p; + if (q < p) { // they are backward- swap them + src = p + 1; + dest = q; + } + hole_size = q - p + 1; + cnt = end - src; + if (src < text || src > end) + goto thd0; + if (dest < text || dest >= end) + goto thd0; + if (src >= end) + goto thd_atend; // just delete the end of the buffer + memmove(dest, src, cnt); + thd_atend: + end = end - hole_size; // adjust the new END + if (dest >= end) + dest = end - 1; // make sure dest in below end-1 + if (end <= text) + dest = end = text; // keep pointers valid + file_modified++; + thd0: + return dest; +} + +// copy text into register, then delete text. +// if dist <= 0, do not include, or go past, a NewLine +// +static char *yank_delete(char *start, char *stop, int dist, int yf) +{ + char *p; + + // make sure start <= stop + if (start > stop) { + // they are backwards, reverse them + p = start; + start = stop; + stop = p; + } + if (dist <= 0) { + // we cannot cross NL boundaries + p = start; + if (*p == '\n') + return p; + // dont go past a NewLine + for (; p + 1 <= stop; p++) { + if (p[1] == '\n') { + stop = p; // "stop" just before NewLine + break; + } + } + } + p = start; +#if ENABLE_FEATURE_VI_YANKMARK + text_yank(start, stop, YDreg); +#endif + if (yf == YANKDEL) { + p = text_hole_delete(start, stop); + } // delete lines + return p; +} + +static void show_help(void) +{ + puts("These features are available:" +#if ENABLE_FEATURE_VI_SEARCH + "\n\tPattern searches with / and ?" +#endif +#if ENABLE_FEATURE_VI_DOT_CMD + "\n\tLast command repeat with \'.\'" +#endif +#if ENABLE_FEATURE_VI_YANKMARK + "\n\tLine marking with 'x" + "\n\tNamed buffers with \"x" +#endif +#if ENABLE_FEATURE_VI_READONLY + "\n\tReadonly if vi is called as \"view\"" + "\n\tReadonly with -R command line arg" +#endif +#if ENABLE_FEATURE_VI_SET + "\n\tSome colon mode commands with \':\'" +#endif +#if ENABLE_FEATURE_VI_SETOPTS + "\n\tSettable options with \":set\"" +#endif +#if ENABLE_FEATURE_VI_USE_SIGNALS + "\n\tSignal catching- ^C" + "\n\tJob suspend and resume with ^Z" +#endif +#if ENABLE_FEATURE_VI_WIN_RESIZE + "\n\tAdapt to window re-sizes" +#endif + ); +} + +#if ENABLE_FEATURE_VI_DOT_CMD +static void start_new_cmd_q(char c) +{ + // get buffer for new cmd + // if there is a current cmd count put it in the buffer first + if (cmdcnt > 0) { + lmc_len = sprintf(last_modifying_cmd, "%d%c", cmdcnt, c); + } else { // just save char c onto queue + last_modifying_cmd[0] = c; + lmc_len = 1; + } + adding2q = 1; +} + +static void end_cmd_q(void) +{ +#if ENABLE_FEATURE_VI_YANKMARK + YDreg = 26; // go back to default Yank/Delete reg +#endif + adding2q = 0; +} +#endif /* FEATURE_VI_DOT_CMD */ + +#if ENABLE_FEATURE_VI_YANKMARK \ + || (ENABLE_FEATURE_VI_COLON && ENABLE_FEATURE_VI_SEARCH) \ + || ENABLE_FEATURE_VI_CRASHME +static char *string_insert(char *p, char *s) // insert the string at 'p' +{ + int cnt, i; + + i = strlen(s); + text_hole_make(p, i); + strncpy(p, s, i); + for (cnt = 0; *s != '\0'; s++) { + if (*s == '\n') + cnt++; + } +#if ENABLE_FEATURE_VI_YANKMARK + status_line("Put %d lines (%d chars) from [%c]", cnt, i, what_reg()); +#endif + return p; +} +#endif + +#if ENABLE_FEATURE_VI_YANKMARK +static char *text_yank(char *p, char *q, int dest) // copy text into a register +{ + char *t; + int cnt; + + if (q < p) { // they are backwards- reverse them + t = q; + q = p; + p = t; + } + cnt = q - p + 1; + t = reg[dest]; + free(t); // if already a yank register, free it + t = xmalloc(cnt + 1); // get a new register + memset(t, '\0', cnt + 1); // clear new text[] + strncpy(t, p, cnt); // copy text[] into bufer + reg[dest] = t; + return p; +} + +static char what_reg(void) +{ + char c; + + c = 'D'; // default to D-reg + if (0 <= YDreg && YDreg <= 25) + c = 'a' + (char) YDreg; + if (YDreg == 26) + c = 'D'; + if (YDreg == 27) + c = 'U'; + return c; +} + +static void check_context(char cmd) +{ + // A context is defined to be "modifying text" + // Any modifying command establishes a new context. + + if (dot < context_start || dot > context_end) { + if (strchr(modifying_cmds, cmd) != NULL) { + // we are trying to modify text[]- make this the current context + mark[27] = mark[26]; // move cur to prev + mark[26] = dot; // move local to cur + context_start = prev_line(prev_line(dot)); + context_end = next_line(next_line(dot)); + //loiter= start_loiter= now; + } + } +} + +static char *swap_context(char *p) // goto new context for '' command make this the current context +{ + char *tmp; + + // the current context is in mark[26] + // the previous context is in mark[27] + // only swap context if other context is valid + if (text <= mark[27] && mark[27] <= end - 1) { + tmp = mark[27]; + mark[27] = mark[26]; + mark[26] = tmp; + p = mark[26]; // where we are going- previous context + context_start = prev_line(prev_line(prev_line(p))); + context_end = next_line(next_line(next_line(p))); + } + return p; +} +#endif /* FEATURE_VI_YANKMARK */ + +//----- Set terminal attributes -------------------------------- +static void rawmode(void) +{ + tcgetattr(0, &term_orig); + term_vi = term_orig; + term_vi.c_lflag &= (~ICANON & ~ECHO); // leave ISIG ON- allow intr's + term_vi.c_iflag &= (~IXON & ~ICRNL); + term_vi.c_oflag &= (~ONLCR); + term_vi.c_cc[VMIN] = 1; + term_vi.c_cc[VTIME] = 0; + erase_char = term_vi.c_cc[VERASE]; + tcsetattr_stdin_TCSANOW(&term_vi); +} + +static void cookmode(void) +{ + fflush(stdout); + tcsetattr_stdin_TCSANOW(&term_orig); +} + +//----- Come here when we get a window resize signal --------- +#if ENABLE_FEATURE_VI_USE_SIGNALS +static void winch_sig(int sig UNUSED_PARAM) +{ + // FIXME: do it in main loop!!! + signal(SIGWINCH, winch_sig); + if (ENABLE_FEATURE_VI_WIN_RESIZE) { + get_terminal_width_height(0, &columns, &rows); + if (rows > MAX_SCR_ROWS) rows = MAX_SCR_ROWS; + if (columns > MAX_SCR_COLS) columns = MAX_SCR_COLS; + } + new_screen(rows, columns); // get memory for virtual screen + redraw(TRUE); // re-draw the screen +} + +//----- Come here when we get a continue signal ------------------- +static void cont_sig(int sig UNUSED_PARAM) +{ + rawmode(); // terminal to "raw" + last_status_cksum = 0; // force status update + redraw(TRUE); // re-draw the screen + + signal(SIGTSTP, suspend_sig); + signal(SIGCONT, SIG_DFL); + kill(my_pid, SIGCONT); // huh? why? we are already "continued"... +} + +//----- Come here when we get a Suspend signal ------------------- +static void suspend_sig(int sig UNUSED_PARAM) +{ + go_bottom_and_clear_to_eol(); + cookmode(); // terminal to "cooked" + + signal(SIGCONT, cont_sig); + signal(SIGTSTP, SIG_DFL); + kill(my_pid, SIGTSTP); +} + +//----- Come here when we get a signal --------------------------- +static void catch_sig(int sig) +{ + signal(SIGINT, catch_sig); + if (sig) + siglongjmp(restart, sig); +} +#endif /* FEATURE_VI_USE_SIGNALS */ + +static int mysleep(int hund) // sleep for 'h' 1/100 seconds +{ + struct pollfd pfd[1]; + + pfd[0].fd = 0; + pfd[0].events = POLLIN; + return safe_poll(pfd, 1, hund*10) > 0; +} + +//----- IO Routines -------------------------------------------- +static int readit(void) // read (maybe cursor) key from stdin +{ + int c; + + fflush(stdout); + c = read_key(STDIN_FILENO, &chars_to_parse, readbuffer); + if (c == -1) { // EOF/error + go_bottom_and_clear_to_eol(); + cookmode(); // terminal to "cooked" + bb_error_msg_and_die("can't read user input"); + } + return c; +} + +//----- IO Routines -------------------------------------------- +static int get_one_char(void) +{ + int c; + +#if ENABLE_FEATURE_VI_DOT_CMD + if (!adding2q) { + // we are not adding to the q. + // but, we may be reading from a q + if (ioq == 0) { + // there is no current q, read from STDIN + c = readit(); // get the users input + } else { + // there is a queue to get chars from first + // careful with correct sign expansion! + c = (unsigned char)*ioq++; + if (c == '\0') { + // the end of the q, read from STDIN + free(ioq_start); + ioq_start = ioq = 0; + c = readit(); // get the users input + } + } + } else { + // adding STDIN chars to q + c = readit(); // get the users input + if (lmc_len >= MAX_INPUT_LEN - 1) { + status_line_bold("last_modifying_cmd overrun"); + } else { + // add new char to q + last_modifying_cmd[lmc_len++] = c; + } + } +#else + c = readit(); // get the users input +#endif /* FEATURE_VI_DOT_CMD */ + return c; +} + +// Get input line (uses "status line" area) +static char *get_input_line(const char *prompt) +{ + // char [MAX_INPUT_LEN] +#define buf get_input_line__buf + + int c; + int i; + + strcpy(buf, prompt); + last_status_cksum = 0; // force status update + go_bottom_and_clear_to_eol(); + write1(prompt); // write out the :, /, or ? prompt + + i = strlen(buf); + while (i < MAX_INPUT_LEN) { + c = get_one_char(); + if (c == '\n' || c == '\r' || c == 27) + break; // this is end of input + if (c == erase_char || c == 8 || c == 127) { + // user wants to erase prev char + buf[--i] = '\0'; + write1("\b \b"); // erase char on screen + if (i <= 0) // user backs up before b-o-l, exit + break; + } else if (c > 0 && c < 256) { // exclude Unicode + // (TODO: need to handle Unicode) + buf[i] = c; + buf[++i] = '\0'; + bb_putchar(c); + } + } + refresh(FALSE); + return buf; +#undef buf +} + +static int file_size(const char *fn) // what is the byte size of "fn" +{ + struct stat st_buf; + int cnt; + + cnt = -1; + if (fn && fn[0] && stat(fn, &st_buf) == 0) // see if file exists + cnt = (int) st_buf.st_size; + return cnt; +} + +static int file_insert(const char *fn, char *p + USE_FEATURE_VI_READONLY(, int update_ro_status)) +{ + int cnt = -1; + int fd, size; + struct stat statbuf; + + /* Validate file */ + if (stat(fn, &statbuf) < 0) { + status_line_bold("\"%s\" %s", fn, strerror(errno)); + goto fi0; + } + if (!S_ISREG(statbuf.st_mode)) { + // This is not a regular file + status_line_bold("\"%s\" Not a regular file", fn); + goto fi0; + } + if (p < text || p > end) { + status_line_bold("Trying to insert file outside of memory"); + goto fi0; + } + + // read file to buffer + fd = open(fn, O_RDONLY); + if (fd < 0) { + status_line_bold("\"%s\" %s", fn, strerror(errno)); + goto fi0; + } + size = statbuf.st_size; + p = text_hole_make(p, size); + cnt = safe_read(fd, p, size); + if (cnt < 0) { + status_line_bold("\"%s\" %s", fn, strerror(errno)); + p = text_hole_delete(p, p + size - 1); // un-do buffer insert + } else if (cnt < size) { + // There was a partial read, shrink unused space text[] + p = text_hole_delete(p + cnt, p + (size - cnt) - 1); // un-do buffer insert + status_line_bold("cannot read all of file \"%s\"", fn); + } + if (cnt >= size) + file_modified++; + close(fd); + fi0: +#if ENABLE_FEATURE_VI_READONLY + if (update_ro_status + && ((access(fn, W_OK) < 0) || + /* root will always have access() + * so we check fileperms too */ + !(statbuf.st_mode & (S_IWUSR | S_IWGRP | S_IWOTH)) + ) + ) { + SET_READONLY_FILE(readonly_mode); + } +#endif + return cnt; +} + +static int file_write(char *fn, char *first, char *last) +{ + int fd, cnt, charcnt; + + if (fn == 0) { + status_line_bold("No current filename"); + return -2; + } + charcnt = 0; + /* By popular request we do not open file with O_TRUNC, + * but instead ftruncate() it _after_ successful write. + * Might reduce amount of data lost on power fail etc. + */ + fd = open(fn, (O_WRONLY | O_CREAT), 0666); + if (fd < 0) + return -1; + cnt = last - first + 1; + charcnt = full_write(fd, first, cnt); + ftruncate(fd, charcnt); + if (charcnt == cnt) { + // good write + //file_modified = FALSE; + } else { + charcnt = 0; + } + close(fd); + return charcnt; +} + +//----- Terminal Drawing --------------------------------------- +// The terminal is made up of 'rows' line of 'columns' columns. +// classically this would be 24 x 80. +// screen coordinates +// 0,0 ... 0,79 +// 1,0 ... 1,79 +// . ... . +// . ... . +// 22,0 ... 22,79 +// 23,0 ... 23,79 <- status line + +//----- Move the cursor to row x col (count from 0, not 1) ------- +static void place_cursor(int row, int col, int optimize) +{ + char cm1[sizeof(CMrc) + sizeof(int)*3 * 2]; +#if ENABLE_FEATURE_VI_OPTIMIZE_CURSOR + enum { + SZ_UP = sizeof(CMup), + SZ_DN = sizeof(CMdown), + SEQ_SIZE = SZ_UP > SZ_DN ? SZ_UP : SZ_DN, + }; + char cm2[SEQ_SIZE * 5 + 32]; // bigger than worst case size +#endif + char *cm; + + if (row < 0) row = 0; + if (row >= rows) row = rows - 1; + if (col < 0) col = 0; + if (col >= columns) col = columns - 1; + + //----- 1. Try the standard terminal ESC sequence + sprintf(cm1, CMrc, row + 1, col + 1); + cm = cm1; + +#if ENABLE_FEATURE_VI_OPTIMIZE_CURSOR + if (optimize && col < 16) { + char *screenp; + int Rrow = last_row; + int diff = Rrow - row; + + if (diff < -5 || diff > 5) + goto skip; + + //----- find the minimum # of chars to move cursor ------------- + //----- 2. Try moving with discreet chars (Newline, [back]space, ...) + cm2[0] = '\0'; + + // move to the correct row + while (row < Rrow) { + // the cursor has to move up + strcat(cm2, CMup); + Rrow--; + } + while (row > Rrow) { + // the cursor has to move down + strcat(cm2, CMdown); + Rrow++; + } + + // now move to the correct column + strcat(cm2, "\r"); // start at col 0 + // just send out orignal source char to get to correct place + screenp = &screen[row * columns]; // start of screen line + strncat(cm2, screenp, col); + + // pick the shortest cursor motion to send out + if (strlen(cm2) < strlen(cm)) { + cm = cm2; + } + skip: ; + } + last_row = row; +#endif /* FEATURE_VI_OPTIMIZE_CURSOR */ + write1(cm); +} + +//----- Erase from cursor to end of line ----------------------- +static void clear_to_eol(void) +{ + write1(Ceol); // Erase from cursor to end of line +} + +static void go_bottom_and_clear_to_eol(void) +{ + place_cursor(rows - 1, 0, FALSE); // go to bottom of screen + clear_to_eol(); // erase to end of line +} + +//----- Erase from cursor to end of screen ----------------------- +static void clear_to_eos(void) +{ + write1(Ceos); // Erase from cursor to end of screen +} + +//----- Start standout mode ------------------------------------ +static void standout_start(void) // send "start reverse video" sequence +{ + write1(SOs); // Start reverse video mode +} + +//----- End standout mode -------------------------------------- +static void standout_end(void) // send "end reverse video" sequence +{ + write1(SOn); // End reverse video mode +} + +//----- Flash the screen -------------------------------------- +static void flash(int h) +{ + standout_start(); // send "start reverse video" sequence + redraw(TRUE); + mysleep(h); + standout_end(); // send "end reverse video" sequence + redraw(TRUE); +} + +static void Indicate_Error(void) +{ +#if ENABLE_FEATURE_VI_CRASHME + if (crashme > 0) + return; // generate a random command +#endif + if (!err_method) { + write1(bell); // send out a bell character + } else { + flash(10); + } +} + +//----- Screen[] Routines -------------------------------------- +//----- Erase the Screen[] memory ------------------------------ +static void screen_erase(void) +{ + memset(screen, ' ', screensize); // clear new screen +} + +static int bufsum(char *buf, int count) +{ + int sum = 0; + char *e = buf + count; + + while (buf < e) + sum += (unsigned char) *buf++; + return sum; +} + +//----- Draw the status line at bottom of the screen ------------- +static void show_status_line(void) +{ + int cnt = 0, cksum = 0; + + // either we already have an error or status message, or we + // create one. + if (!have_status_msg) { + cnt = format_edit_status(); + cksum = bufsum(status_buffer, cnt); + } + if (have_status_msg || ((cnt > 0 && last_status_cksum != cksum))) { + last_status_cksum = cksum; // remember if we have seen this line + go_bottom_and_clear_to_eol(); + write1(status_buffer); + if (have_status_msg) { + if (((int)strlen(status_buffer) - (have_status_msg - 1)) > + (columns - 1) ) { + have_status_msg = 0; + Hit_Return(); + } + have_status_msg = 0; + } + place_cursor(crow, ccol, FALSE); // put cursor back in correct place + } + fflush(stdout); +} + +//----- format the status buffer, the bottom line of screen ------ +// format status buffer, with STANDOUT mode +static void status_line_bold(const char *format, ...) +{ + va_list args; + + va_start(args, format); + strcpy(status_buffer, SOs); // Terminal standout mode on + vsprintf(status_buffer + sizeof(SOs)-1, format, args); + strcat(status_buffer, SOn); // Terminal standout mode off + va_end(args); + + have_status_msg = 1 + sizeof(SOs) + sizeof(SOn) - 2; +} + +// format status buffer +static void status_line(const char *format, ...) +{ + va_list args; + + va_start(args, format); + vsprintf(status_buffer, format, args); + va_end(args); + + have_status_msg = 1; +} + +// copy s to buf, convert unprintable +static void print_literal(char *buf, const char *s) +{ + unsigned char c; + char b[2]; + + b[1] = '\0'; + buf[0] = '\0'; + if (!s[0]) + s = "(NULL)"; + for (; *s; s++) { + int c_is_no_print; + + c = *s; + c_is_no_print = (c & 0x80) && !Isprint(c); + if (c_is_no_print) { + strcat(buf, SOn); + c = '.'; + } + if (c < ' ' || c == 127) { + strcat(buf, "^"); + if (c == 127) + c = '?'; + else + c += '@'; + } + b[0] = c; + strcat(buf, b); + if (c_is_no_print) + strcat(buf, SOs); + if (*s == '\n') + strcat(buf, "$"); + if (strlen(buf) > MAX_INPUT_LEN - 10) // paranoia + break; + } +} + +static void not_implemented(const char *s) +{ + char buf[MAX_INPUT_LEN]; + + print_literal(buf, s); + status_line_bold("\'%s\' is not implemented", buf); +} + +// show file status on status line +static int format_edit_status(void) +{ + static const char cmd_mode_indicator[] ALIGN1 = "-IR-"; + +#define tot format_edit_status__tot + + int cur, percent, ret, trunc_at; + + // file_modified is now a counter rather than a flag. this + // helps reduce the amount of line counting we need to do. + // (this will cause a mis-reporting of modified status + // once every MAXINT editing operations.) + + // it would be nice to do a similar optimization here -- if + // we haven't done a motion that could have changed which line + // we're on, then we shouldn't have to do this count_lines() + cur = count_lines(text, dot); + + // reduce counting -- the total lines can't have + // changed if we haven't done any edits. + if (file_modified != last_file_modified) { + tot = cur + count_lines(dot, end - 1) - 1; + last_file_modified = file_modified; + } + + // current line percent + // ------------- ~~ ---------- + // total lines 100 + if (tot > 0) { + percent = (100 * cur) / tot; + } else { + cur = tot = 0; + percent = 100; + } + + trunc_at = columns < STATUS_BUFFER_LEN-1 ? + columns : STATUS_BUFFER_LEN-1; + + ret = snprintf(status_buffer, trunc_at+1, +#if ENABLE_FEATURE_VI_READONLY + "%c %s%s%s %d/%d %d%%", +#else + "%c %s%s %d/%d %d%%", +#endif + cmd_mode_indicator[cmd_mode & 3], + (current_filename != NULL ? current_filename : "No file"), +#if ENABLE_FEATURE_VI_READONLY + (readonly_mode ? " [Readonly]" : ""), +#endif + (file_modified ? " [Modified]" : ""), + cur, tot, percent); + + if (ret >= 0 && ret < trunc_at) + return ret; /* it all fit */ + + return trunc_at; /* had to truncate */ +#undef tot +} + +//----- Force refresh of all Lines ----------------------------- +static void redraw(int full_screen) +{ + place_cursor(0, 0, FALSE); // put cursor in correct place + clear_to_eos(); // tell terminal to erase display + screen_erase(); // erase the internal screen buffer + last_status_cksum = 0; // force status update + refresh(full_screen); // this will redraw the entire display + show_status_line(); +} + +//----- Format a text[] line into a buffer --------------------- +static char* format_line(char *src /*, int li*/) +{ + unsigned char c; + int co; + int ofs = offset; + char *dest = scr_out_buf; // [MAX_SCR_COLS + MAX_TABSTOP * 2] + + c = '~'; // char in col 0 in non-existent lines is '~' + co = 0; + while (co < columns + tabstop) { + // have we gone past the end? + if (src < end) { + c = *src++; + if (c == '\n') + break; + if ((c & 0x80) && !Isprint(c)) { + c = '.'; + } + if (c < ' ' || c == 0x7f) { + if (c == '\t') { + c = ' '; + // co % 8 != 7 + while ((co % tabstop) != (tabstop - 1)) { + dest[co++] = c; + } + } else { + dest[co++] = '^'; + if (c == 0x7f) + c = '?'; + else + c += '@'; // Ctrl-X -> 'X' + } + } + } + dest[co++] = c; + // discard scrolled-off-to-the-left portion, + // in tabstop-sized pieces + if (ofs >= tabstop && co >= tabstop) { + memmove(dest, dest + tabstop, co); + co -= tabstop; + ofs -= tabstop; + } + if (src >= end) + break; + } + // check "short line, gigantic offset" case + if (co < ofs) + ofs = co; + // discard last scrolled off part + co -= ofs; + dest += ofs; + // fill the rest with spaces + if (co < columns) + memset(&dest[co], ' ', columns - co); + return dest; +} + +//----- Refresh the changed screen lines ----------------------- +// Copy the source line from text[] into the buffer and note +// if the current screenline is different from the new buffer. +// If they differ then that line needs redrawing on the terminal. +// +static void refresh(int full_screen) +{ +#define old_offset refresh__old_offset + + int li, changed; + char *tp, *sp; // pointer into text[] and screen[] + + if (ENABLE_FEATURE_VI_WIN_RESIZE) { + unsigned c = columns, r = rows; + get_terminal_width_height(0, &columns, &rows); + if (rows > MAX_SCR_ROWS) rows = MAX_SCR_ROWS; + if (columns > MAX_SCR_COLS) columns = MAX_SCR_COLS; + full_screen |= (c - columns) | (r - rows); + } + sync_cursor(dot, &crow, &ccol); // where cursor will be (on "dot") + tp = screenbegin; // index into text[] of top line + + // compare text[] to screen[] and mark screen[] lines that need updating + for (li = 0; li < rows - 1; li++) { + int cs, ce; // column start & end + char *out_buf; + // format current text line + out_buf = format_line(tp /*, li*/); + + // skip to the end of the current text[] line + if (tp < end) { + char *t = memchr(tp, '\n', end - tp); + if (!t) t = end - 1; + tp = t + 1; + } + + // see if there are any changes between vitual screen and out_buf + changed = FALSE; // assume no change + cs = 0; + ce = columns - 1; + sp = &screen[li * columns]; // start of screen line + if (full_screen) { + // force re-draw of every single column from 0 - columns-1 + goto re0; + } + // compare newly formatted buffer with virtual screen + // look forward for first difference between buf and screen + for (; cs <= ce; cs++) { + if (out_buf[cs] != sp[cs]) { + changed = TRUE; // mark for redraw + break; + } + } + + // look backward for last difference between out_buf and screen + for (; ce >= cs; ce--) { + if (out_buf[ce] != sp[ce]) { + changed = TRUE; // mark for redraw + break; + } + } + // now, cs is index of first diff, and ce is index of last diff + + // if horz offset has changed, force a redraw + if (offset != old_offset) { + re0: + changed = TRUE; + } + + // make a sanity check of columns indexes + if (cs < 0) cs = 0; + if (ce > columns - 1) ce = columns - 1; + if (cs > ce) { cs = 0; ce = columns - 1; } + // is there a change between vitual screen and out_buf + if (changed) { + // copy changed part of buffer to virtual screen + memcpy(sp+cs, out_buf+cs, ce-cs+1); + + // move cursor to column of first change + //if (offset != old_offset) { + // // place_cursor is still too stupid + // // to handle offsets correctly + // place_cursor(li, cs, FALSE); + //} else { + place_cursor(li, cs, TRUE); + //} + + // write line out to terminal + fwrite(&sp[cs], ce - cs + 1, 1, stdout); + } + } + + place_cursor(crow, ccol, TRUE); + + old_offset = offset; +#undef old_offset +} + +//--------------------------------------------------------------------- +//----- the Ascii Chart ----------------------------------------------- +// +// 00 nul 01 soh 02 stx 03 etx 04 eot 05 enq 06 ack 07 bel +// 08 bs 09 ht 0a nl 0b vt 0c np 0d cr 0e so 0f si +// 10 dle 11 dc1 12 dc2 13 dc3 14 dc4 15 nak 16 syn 17 etb +// 18 can 19 em 1a sub 1b esc 1c fs 1d gs 1e rs 1f us +// 20 sp 21 ! 22 " 23 # 24 $ 25 % 26 & 27 ' +// 28 ( 29 ) 2a * 2b + 2c , 2d - 2e . 2f / +// 30 0 31 1 32 2 33 3 34 4 35 5 36 6 37 7 +// 38 8 39 9 3a : 3b ; 3c < 3d = 3e > 3f ? +// 40 @ 41 A 42 B 43 C 44 D 45 E 46 F 47 G +// 48 H 49 I 4a J 4b K 4c L 4d M 4e N 4f O +// 50 P 51 Q 52 R 53 S 54 T 55 U 56 V 57 W +// 58 X 59 Y 5a Z 5b [ 5c \ 5d ] 5e ^ 5f _ +// 60 ` 61 a 62 b 63 c 64 d 65 e 66 f 67 g +// 68 h 69 i 6a j 6b k 6c l 6d m 6e n 6f o +// 70 p 71 q 72 r 73 s 74 t 75 u 76 v 77 w +// 78 x 79 y 7a z 7b { 7c | 7d } 7e ~ 7f del +//--------------------------------------------------------------------- + +//----- Execute a Vi Command ----------------------------------- +static void do_cmd(int c) +{ + const char *msg = msg; // for compiler + char *p, *q, *save_dot; + char buf[12]; + int dir; + int cnt, i, j; + int c1; + +// c1 = c; // quiet the compiler +// cnt = yf = 0; // quiet the compiler +// msg = p = q = save_dot = buf; // quiet the compiler + memset(buf, '\0', 12); + + show_status_line(); + + /* if this is a cursor key, skip these checks */ + switch (c) { + case KEYCODE_UP: + case KEYCODE_DOWN: + case KEYCODE_LEFT: + case KEYCODE_RIGHT: + case KEYCODE_HOME: + case KEYCODE_END: + case KEYCODE_PAGEUP: + case KEYCODE_PAGEDOWN: + case KEYCODE_DELETE: + goto key_cmd_mode; + } + + if (cmd_mode == 2) { + // flip-flop Insert/Replace mode + if (c == KEYCODE_INSERT) + goto dc_i; + // we are 'R'eplacing the current *dot with new char + if (*dot == '\n') { + // don't Replace past E-o-l + cmd_mode = 1; // convert to insert + } else { + if (1 <= c || Isprint(c)) { + if (c != 27) + dot = yank_delete(dot, dot, 0, YANKDEL); // delete char + dot = char_insert(dot, c); // insert new char + } + goto dc1; + } + } + if (cmd_mode == 1) { + // hitting "Insert" twice means "R" replace mode + if (c == KEYCODE_INSERT) goto dc5; + // insert the char c at "dot" + if (1 <= c || Isprint(c)) { + dot = char_insert(dot, c); + } + goto dc1; + } + + key_cmd_mode: + switch (c) { + //case 0x01: // soh + //case 0x09: // ht + //case 0x0b: // vt + //case 0x0e: // so + //case 0x0f: // si + //case 0x10: // dle + //case 0x11: // dc1 + //case 0x13: // dc3 +#if ENABLE_FEATURE_VI_CRASHME + case 0x14: // dc4 ctrl-T + crashme = (crashme == 0) ? 1 : 0; + break; +#endif + //case 0x16: // syn + //case 0x17: // etb + //case 0x18: // can + //case 0x1c: // fs + //case 0x1d: // gs + //case 0x1e: // rs + //case 0x1f: // us + //case '!': // !- + //case '#': // #- + //case '&': // &- + //case '(': // (- + //case ')': // )- + //case '*': // *- + //case '=': // =- + //case '@': // @- + //case 'F': // F- + //case 'K': // K- + //case 'Q': // Q- + //case 'S': // S- + //case 'T': // T- + //case 'V': // V- + //case '[': // [- + //case '\\': // \- + //case ']': // ]- + //case '_': // _- + //case '`': // `- + //case 'u': // u- FIXME- there is no undo + //case 'v': // v- + default: // unrecognised command + buf[0] = c; + buf[1] = '\0'; + if (c < ' ') { + buf[0] = '^'; + buf[1] = c + '@'; + buf[2] = '\0'; + } + not_implemented(buf); + end_cmd_q(); // stop adding to q + case 0x00: // nul- ignore + break; + case 2: // ctrl-B scroll up full screen + case KEYCODE_PAGEUP: // Cursor Key Page Up + dot_scroll(rows - 2, -1); + break; + case 4: // ctrl-D scroll down half screen + dot_scroll((rows - 2) / 2, 1); + break; + case 5: // ctrl-E scroll down one line + dot_scroll(1, 1); + break; + case 6: // ctrl-F scroll down full screen + case KEYCODE_PAGEDOWN: // Cursor Key Page Down + dot_scroll(rows - 2, 1); + break; + case 7: // ctrl-G show current status + last_status_cksum = 0; // force status update + break; + case 'h': // h- move left + case KEYCODE_LEFT: // cursor key Left + case 8: // ctrl-H- move left (This may be ERASE char) + case 0x7f: // DEL- move left (This may be ERASE char) + if (cmdcnt-- > 1) { + do_cmd(c); + } // repeat cnt + dot_left(); + break; + case 10: // Newline ^J + case 'j': // j- goto next line, same col + case KEYCODE_DOWN: // cursor key Down + if (cmdcnt-- > 1) { + do_cmd(c); + } // repeat cnt + dot_next(); // go to next B-o-l + dot = move_to_col(dot, ccol + offset); // try stay in same col + break; + case 12: // ctrl-L force redraw whole screen + case 18: // ctrl-R force redraw + place_cursor(0, 0, FALSE); // put cursor in correct place + clear_to_eos(); // tel terminal to erase display + mysleep(10); + screen_erase(); // erase the internal screen buffer + last_status_cksum = 0; // force status update + refresh(TRUE); // this will redraw the entire display + break; + case 13: // Carriage Return ^M + case '+': // +- goto next line + if (cmdcnt-- > 1) { + do_cmd(c); + } // repeat cnt + dot_next(); + dot_skip_over_ws(); + break; + case 21: // ctrl-U scroll up half screen + dot_scroll((rows - 2) / 2, -1); + break; + case 25: // ctrl-Y scroll up one line + dot_scroll(1, -1); + break; + case 27: // esc + if (cmd_mode == 0) + indicate_error(c); + cmd_mode = 0; // stop insrting + end_cmd_q(); + last_status_cksum = 0; // force status update + break; + case ' ': // move right + case 'l': // move right + case KEYCODE_RIGHT: // Cursor Key Right + if (cmdcnt-- > 1) { + do_cmd(c); + } // repeat cnt + dot_right(); + break; +#if ENABLE_FEATURE_VI_YANKMARK + case '"': // "- name a register to use for Delete/Yank + c1 = (get_one_char() | 0x20) - 'a'; // | 0x20 is tolower() + if ((unsigned)c1 <= 25) { // a-z? + YDreg = c1; + } else { + indicate_error(c); + } + break; + case '\'': // '- goto a specific mark + c1 = (get_one_char() | 0x20) - 'a'; + if ((unsigned)c1 <= 25) { // a-z? + // get the b-o-l + q = mark[c1]; + if (text <= q && q < end) { + dot = q; + dot_begin(); // go to B-o-l + dot_skip_over_ws(); + } + } else if (c1 == '\'') { // goto previous context + dot = swap_context(dot); // swap current and previous context + dot_begin(); // go to B-o-l + dot_skip_over_ws(); + } else { + indicate_error(c); + } + break; + case 'm': // m- Mark a line + // this is really stupid. If there are any inserts or deletes + // between text[0] and dot then this mark will not point to the + // correct location! It could be off by many lines! + // Well..., at least its quick and dirty. + c1 = (get_one_char() | 0x20) - 'a'; + if ((unsigned)c1 <= 25) { // a-z? + // remember the line + mark[c1] = dot; + } else { + indicate_error(c); + } + break; + case 'P': // P- Put register before + case 'p': // p- put register after + p = reg[YDreg]; + if (p == 0) { + status_line_bold("Nothing in register %c", what_reg()); + break; + } + // are we putting whole lines or strings + if (strchr(p, '\n') != NULL) { + if (c == 'P') { + dot_begin(); // putting lines- Put above + } + if (c == 'p') { + // are we putting after very last line? + if (end_line(dot) == (end - 1)) { + dot = end; // force dot to end of text[] + } else { + dot_next(); // next line, then put before + } + } + } else { + if (c == 'p') + dot_right(); // move to right, can move to NL + } + dot = string_insert(dot, p); // insert the string + end_cmd_q(); // stop adding to q + break; + case 'U': // U- Undo; replace current line with original version + if (reg[Ureg] != 0) { + p = begin_line(dot); + q = end_line(dot); + p = text_hole_delete(p, q); // delete cur line + p = string_insert(p, reg[Ureg]); // insert orig line + dot = p; + dot_skip_over_ws(); + } + break; +#endif /* FEATURE_VI_YANKMARK */ + case '$': // $- goto end of line + case KEYCODE_END: // Cursor Key End + if (cmdcnt-- > 1) { + do_cmd(c); + } // repeat cnt + dot = end_line(dot); + break; + case '%': // %- find matching char of pair () [] {} + for (q = dot; q < end && *q != '\n'; q++) { + if (strchr("()[]{}", *q) != NULL) { + // we found half of a pair + p = find_pair(q, *q); + if (p == NULL) { + indicate_error(c); + } else { + dot = p; + } + break; + } + } + if (*q == '\n') + indicate_error(c); + break; + case 'f': // f- forward to a user specified char + last_forward_char = get_one_char(); // get the search char + // + // dont separate these two commands. 'f' depends on ';' + // + //**** fall through to ... ';' + case ';': // ;- look at rest of line for last forward char + if (cmdcnt-- > 1) { + do_cmd(';'); + } // repeat cnt + if (last_forward_char == 0) + break; + q = dot + 1; + while (q < end - 1 && *q != '\n' && *q != last_forward_char) { + q++; + } + if (*q == last_forward_char) + dot = q; + break; + case ',': // repeat latest 'f' in opposite direction + if (cmdcnt-- > 1) { + do_cmd(','); + } // repeat cnt + if (last_forward_char == 0) + break; + q = dot - 1; + while (q >= text && *q != '\n' && *q != last_forward_char) { + q--; + } + if (q >= text && *q == last_forward_char) + dot = q; + break; + + case '-': // -- goto prev line + if (cmdcnt-- > 1) { + do_cmd(c); + } // repeat cnt + dot_prev(); + dot_skip_over_ws(); + break; +#if ENABLE_FEATURE_VI_DOT_CMD + case '.': // .- repeat the last modifying command + // Stuff the last_modifying_cmd back into stdin + // and let it be re-executed. + if (lmc_len > 0) { + last_modifying_cmd[lmc_len] = 0; + ioq = ioq_start = xstrdup(last_modifying_cmd); + } + break; +#endif +#if ENABLE_FEATURE_VI_SEARCH + case '?': // /- search for a pattern + case '/': // /- search for a pattern + buf[0] = c; + buf[1] = '\0'; + q = get_input_line(buf); // get input line- use "status line" + if (q[0] && !q[1]) { + if (last_search_pattern[0]) + last_search_pattern[0] = c; + goto dc3; // if no pat re-use old pat + } + if (q[0]) { // strlen(q) > 1: new pat- save it and find + // there is a new pat + free(last_search_pattern); + last_search_pattern = xstrdup(q); + goto dc3; // now find the pattern + } + // user changed mind and erased the "/"- do nothing + break; + case 'N': // N- backward search for last pattern + if (cmdcnt-- > 1) { + do_cmd(c); + } // repeat cnt + dir = BACK; // assume BACKWARD search + p = dot - 1; + if (last_search_pattern[0] == '?') { + dir = FORWARD; + p = dot + 1; + } + goto dc4; // now search for pattern + break; + case 'n': // n- repeat search for last pattern + // search rest of text[] starting at next char + // if search fails return orignal "p" not the "p+1" address + if (cmdcnt-- > 1) { + do_cmd(c); + } // repeat cnt + dc3: + dir = FORWARD; // assume FORWARD search + p = dot + 1; + if (last_search_pattern[0] == '?') { + dir = BACK; + p = dot - 1; + } + dc4: + q = char_search(p, last_search_pattern + 1, dir, FULL); + if (q != NULL) { + dot = q; // good search, update "dot" + msg = ""; + goto dc2; + } + // no pattern found between "dot" and "end"- continue at top + p = text; + if (dir == BACK) { + p = end - 1; + } + q = char_search(p, last_search_pattern + 1, dir, FULL); + if (q != NULL) { // found something + dot = q; // found new pattern- goto it + msg = "search hit BOTTOM, continuing at TOP"; + if (dir == BACK) { + msg = "search hit TOP, continuing at BOTTOM"; + } + } else { + msg = "Pattern not found"; + } + dc2: + if (*msg) + status_line_bold("%s", msg); + break; + case '{': // {- move backward paragraph + q = char_search(dot, "\n\n", BACK, FULL); + if (q != NULL) { // found blank line + dot = next_line(q); // move to next blank line + } + break; + case '}': // }- move forward paragraph + q = char_search(dot, "\n\n", FORWARD, FULL); + if (q != NULL) { // found blank line + dot = next_line(q); // move to next blank line + } + break; +#endif /* FEATURE_VI_SEARCH */ + case '0': // 0- goto begining of line + case '1': // 1- + case '2': // 2- + case '3': // 3- + case '4': // 4- + case '5': // 5- + case '6': // 6- + case '7': // 7- + case '8': // 8- + case '9': // 9- + if (c == '0' && cmdcnt < 1) { + dot_begin(); // this was a standalone zero + } else { + cmdcnt = cmdcnt * 10 + (c - '0'); // this 0 is part of a number + } + break; + case ':': // :- the colon mode commands + p = get_input_line(":"); // get input line- use "status line" +#if ENABLE_FEATURE_VI_COLON + colon(p); // execute the command +#else + if (*p == ':') + p++; // move past the ':' + cnt = strlen(p); + if (cnt <= 0) + break; + if (strncasecmp(p, "quit", cnt) == 0 + || strncasecmp(p, "q!", cnt) == 0 // delete lines + ) { + if (file_modified && p[1] != '!') { + status_line_bold("No write since last change (:quit! overrides)"); + } else { + editing = 0; + } + } else if (strncasecmp(p, "write", cnt) == 0 + || strncasecmp(p, "wq", cnt) == 0 + || strncasecmp(p, "wn", cnt) == 0 + || strncasecmp(p, "x", cnt) == 0 + ) { + cnt = file_write(current_filename, text, end - 1); + if (cnt < 0) { + if (cnt == -1) + status_line_bold("Write error: %s", strerror(errno)); + } else { + file_modified = 0; + last_file_modified = -1; + status_line("\"%s\" %dL, %dC", current_filename, count_lines(text, end - 1), cnt); + if (p[0] == 'x' || p[1] == 'q' || p[1] == 'n' + || p[0] == 'X' || p[1] == 'Q' || p[1] == 'N' + ) { + editing = 0; + } + } + } else if (strncasecmp(p, "file", cnt) == 0) { + last_status_cksum = 0; // force status update + } else if (sscanf(p, "%d", &j) > 0) { + dot = find_line(j); // go to line # j + dot_skip_over_ws(); + } else { // unrecognised cmd + not_implemented(p); + } +#endif /* !FEATURE_VI_COLON */ + break; + case '<': // <- Left shift something + case '>': // >- Right shift something + cnt = count_lines(text, dot); // remember what line we are on + c1 = get_one_char(); // get the type of thing to delete + find_range(&p, &q, c1); + yank_delete(p, q, 1, YANKONLY); // save copy before change + p = begin_line(p); + q = end_line(q); + i = count_lines(p, q); // # of lines we are shifting + for ( ; i > 0; i--, p = next_line(p)) { + if (c == '<') { + // shift left- remove tab or 8 spaces + if (*p == '\t') { + // shrink buffer 1 char + text_hole_delete(p, p); + } else if (*p == ' ') { + // we should be calculating columns, not just SPACE + for (j = 0; *p == ' ' && j < tabstop; j++) { + text_hole_delete(p, p); + } + } + } else if (c == '>') { + // shift right -- add tab or 8 spaces + char_insert(p, '\t'); + } + } + dot = find_line(cnt); // what line were we on + dot_skip_over_ws(); + end_cmd_q(); // stop adding to q + break; + case 'A': // A- append at e-o-l + dot_end(); // go to e-o-l + //**** fall through to ... 'a' + case 'a': // a- append after current char + if (*dot != '\n') + dot++; + goto dc_i; + break; + case 'B': // B- back a blank-delimited Word + case 'E': // E- end of a blank-delimited word + case 'W': // W- forward a blank-delimited word + if (cmdcnt-- > 1) { + do_cmd(c); + } // repeat cnt + dir = FORWARD; + if (c == 'B') + dir = BACK; + if (c == 'W' || isspace(dot[dir])) { + dot = skip_thing(dot, 1, dir, S_TO_WS); + dot = skip_thing(dot, 2, dir, S_OVER_WS); + } + if (c != 'W') + dot = skip_thing(dot, 1, dir, S_BEFORE_WS); + break; + case 'C': // C- Change to e-o-l + case 'D': // D- delete to e-o-l + save_dot = dot; + dot = dollar_line(dot); // move to before NL + // copy text into a register and delete + dot = yank_delete(save_dot, dot, 0, YANKDEL); // delete to e-o-l + if (c == 'C') + goto dc_i; // start inserting +#if ENABLE_FEATURE_VI_DOT_CMD + if (c == 'D') + end_cmd_q(); // stop adding to q +#endif + break; + case 'g': // 'gg' goto a line number (vim) (default: very first line) + c1 = get_one_char(); + if (c1 != 'g') { + buf[0] = 'g'; + buf[1] = c1; // TODO: if Unicode? + buf[2] = '\0'; + not_implemented(buf); + break; + } + if (cmdcnt == 0) + cmdcnt = 1; + /* fall through */ + case 'G': // G- goto to a line number (default= E-O-F) + dot = end - 1; // assume E-O-F + if (cmdcnt > 0) { + dot = find_line(cmdcnt); // what line is #cmdcnt + } + dot_skip_over_ws(); + break; + case 'H': // H- goto top line on screen + dot = screenbegin; + if (cmdcnt > (rows - 1)) { + cmdcnt = (rows - 1); + } + if (cmdcnt-- > 1) { + do_cmd('+'); + } // repeat cnt + dot_skip_over_ws(); + break; + case 'I': // I- insert before first non-blank + dot_begin(); // 0 + dot_skip_over_ws(); + //**** fall through to ... 'i' + case 'i': // i- insert before current char + case KEYCODE_INSERT: // Cursor Key Insert + dc_i: + cmd_mode = 1; // start insrting + break; + case 'J': // J- join current and next lines together + if (cmdcnt-- > 2) { + do_cmd(c); + } // repeat cnt + dot_end(); // move to NL + if (dot < end - 1) { // make sure not last char in text[] + *dot++ = ' '; // replace NL with space + file_modified++; + while (isblank(*dot)) { // delete leading WS + dot_delete(); + } + } + end_cmd_q(); // stop adding to q + break; + case 'L': // L- goto bottom line on screen + dot = end_screen(); + if (cmdcnt > (rows - 1)) { + cmdcnt = (rows - 1); + } + if (cmdcnt-- > 1) { + do_cmd('-'); + } // repeat cnt + dot_begin(); + dot_skip_over_ws(); + break; + case 'M': // M- goto middle line on screen + dot = screenbegin; + for (cnt = 0; cnt < (rows-1) / 2; cnt++) + dot = next_line(dot); + break; + case 'O': // O- open a empty line above + // 0i\n ESC -i + p = begin_line(dot); + if (p[-1] == '\n') { + dot_prev(); + case 'o': // o- open a empty line below; Yes, I know it is in the middle of the "if (..." + dot_end(); + dot = char_insert(dot, '\n'); + } else { + dot_begin(); // 0 + dot = char_insert(dot, '\n'); // i\n ESC + dot_prev(); // - + } + goto dc_i; + break; + case 'R': // R- continuous Replace char + dc5: + cmd_mode = 2; + break; + case KEYCODE_DELETE: + c = 'x'; + // fall through + case 'X': // X- delete char before dot + case 'x': // x- delete the current char + case 's': // s- substitute the current char + if (cmdcnt-- > 1) { + do_cmd(c); + } // repeat cnt + dir = 0; + if (c == 'X') + dir = -1; + if (dot[dir] != '\n') { + if (c == 'X') + dot--; // delete prev char + dot = yank_delete(dot, dot, 0, YANKDEL); // delete char + } + if (c == 's') + goto dc_i; // start insrting + end_cmd_q(); // stop adding to q + break; + case 'Z': // Z- if modified, {write}; exit + // ZZ means to save file (if necessary), then exit + c1 = get_one_char(); + if (c1 != 'Z') { + indicate_error(c); + break; + } + if (file_modified) { + if (ENABLE_FEATURE_VI_READONLY && readonly_mode) { + status_line_bold("\"%s\" File is read only", current_filename); + break; + } + cnt = file_write(current_filename, text, end - 1); + if (cnt < 0) { + if (cnt == -1) + status_line_bold("Write error: %s", strerror(errno)); + } else if (cnt == (end - 1 - text + 1)) { + editing = 0; + } + } else { + editing = 0; + } + break; + case '^': // ^- move to first non-blank on line + dot_begin(); + dot_skip_over_ws(); + break; + case 'b': // b- back a word + case 'e': // e- end of word + if (cmdcnt-- > 1) { + do_cmd(c); + } // repeat cnt + dir = FORWARD; + if (c == 'b') + dir = BACK; + if ((dot + dir) < text || (dot + dir) > end - 1) + break; + dot += dir; + if (isspace(*dot)) { + dot = skip_thing(dot, (c == 'e') ? 2 : 1, dir, S_OVER_WS); + } + if (isalnum(*dot) || *dot == '_') { + dot = skip_thing(dot, 1, dir, S_END_ALNUM); + } else if (ispunct(*dot)) { + dot = skip_thing(dot, 1, dir, S_END_PUNCT); + } + break; + case 'c': // c- change something + case 'd': // d- delete something +#if ENABLE_FEATURE_VI_YANKMARK + case 'y': // y- yank something + case 'Y': // Y- Yank a line +#endif + { + int yf, ml, whole = 0; + yf = YANKDEL; // assume either "c" or "d" +#if ENABLE_FEATURE_VI_YANKMARK + if (c == 'y' || c == 'Y') + yf = YANKONLY; +#endif + c1 = 'y'; + if (c != 'Y') + c1 = get_one_char(); // get the type of thing to delete + // determine range, and whether it spans lines + ml = find_range(&p, &q, c1); + if (c1 == 27) { // ESC- user changed mind and wants out + c = c1 = 27; // Escape- do nothing + } else if (strchr("wW", c1)) { + if (c == 'c') { + // don't include trailing WS as part of word + while (isblank(*q)) { + if (q <= text || q[-1] == '\n') + break; + q--; + } + } + dot = yank_delete(p, q, ml, yf); // delete word + } else if (strchr("^0bBeEft%$ lh\b\177", c1)) { + // partial line copy text into a register and delete + dot = yank_delete(p, q, ml, yf); // delete word + } else if (strchr("cdykjHL+-{}\r\n", c1)) { + // whole line copy text into a register and delete + dot = yank_delete(p, q, ml, yf); // delete lines + whole = 1; + } else { + // could not recognize object + c = c1 = 27; // error- + ml = 0; + indicate_error(c); + } + if (ml && whole) { + if (c == 'c') { + dot = char_insert(dot, '\n'); + // on the last line of file don't move to prev line + if (whole && dot != (end-1)) { + dot_prev(); + } + } else if (c == 'd') { + dot_begin(); + dot_skip_over_ws(); + } + } + if (c1 != 27) { + // if CHANGING, not deleting, start inserting after the delete + if (c == 'c') { + strcpy(buf, "Change"); + goto dc_i; // start inserting + } + if (c == 'd') { + strcpy(buf, "Delete"); + } +#if ENABLE_FEATURE_VI_YANKMARK + if (c == 'y' || c == 'Y') { + strcpy(buf, "Yank"); + } + p = reg[YDreg]; + q = p + strlen(p); + for (cnt = 0; p <= q; p++) { + if (*p == '\n') + cnt++; + } + status_line("%s %d lines (%d chars) using [%c]", + buf, cnt, strlen(reg[YDreg]), what_reg()); +#endif + end_cmd_q(); // stop adding to q + } + break; + } + case 'k': // k- goto prev line, same col + case KEYCODE_UP: // cursor key Up + if (cmdcnt-- > 1) { + do_cmd(c); + } // repeat cnt + dot_prev(); + dot = move_to_col(dot, ccol + offset); // try stay in same col + break; + case 'r': // r- replace the current char with user input + c1 = get_one_char(); // get the replacement char + if (*dot != '\n') { + *dot = c1; + file_modified++; + } + end_cmd_q(); // stop adding to q + break; + case 't': // t- move to char prior to next x + last_forward_char = get_one_char(); + do_cmd(';'); + if (*dot == last_forward_char) + dot_left(); + last_forward_char = 0; + break; + case 'w': // w- forward a word + if (cmdcnt-- > 1) { + do_cmd(c); + } // repeat cnt + if (isalnum(*dot) || *dot == '_') { // we are on ALNUM + dot = skip_thing(dot, 1, FORWARD, S_END_ALNUM); + } else if (ispunct(*dot)) { // we are on PUNCT + dot = skip_thing(dot, 1, FORWARD, S_END_PUNCT); + } + if (dot < end - 1) + dot++; // move over word + if (isspace(*dot)) { + dot = skip_thing(dot, 2, FORWARD, S_OVER_WS); + } + break; + case 'z': // z- + c1 = get_one_char(); // get the replacement char + cnt = 0; + if (c1 == '.') + cnt = (rows - 2) / 2; // put dot at center + if (c1 == '-') + cnt = rows - 2; // put dot at bottom + screenbegin = begin_line(dot); // start dot at top + dot_scroll(cnt, -1); + break; + case '|': // |- move to column "cmdcnt" + dot = move_to_col(dot, cmdcnt - 1); // try to move to column + break; + case '~': // ~- flip the case of letters a-z -> A-Z + if (cmdcnt-- > 1) { + do_cmd(c); + } // repeat cnt + if (islower(*dot)) { + *dot = toupper(*dot); + file_modified++; + } else if (isupper(*dot)) { + *dot = tolower(*dot); + file_modified++; + } + dot_right(); + end_cmd_q(); // stop adding to q + break; + //----- The Cursor and Function Keys ----------------------------- + case KEYCODE_HOME: // Cursor Key Home + dot_begin(); + break; + // The Fn keys could point to do_macro which could translate them +#if 0 + case KEYCODE_FUN1: // Function Key F1 + case KEYCODE_FUN2: // Function Key F2 + case KEYCODE_FUN3: // Function Key F3 + case KEYCODE_FUN4: // Function Key F4 + case KEYCODE_FUN5: // Function Key F5 + case KEYCODE_FUN6: // Function Key F6 + case KEYCODE_FUN7: // Function Key F7 + case KEYCODE_FUN8: // Function Key F8 + case KEYCODE_FUN9: // Function Key F9 + case KEYCODE_FUN10: // Function Key F10 + case KEYCODE_FUN11: // Function Key F11 + case KEYCODE_FUN12: // Function Key F12 + break; +#endif + } + + dc1: + // if text[] just became empty, add back an empty line + if (end == text) { + char_insert(text, '\n'); // start empty buf with dummy line + dot = text; + } + // it is OK for dot to exactly equal to end, otherwise check dot validity + if (dot != end) { + dot = bound_dot(dot); // make sure "dot" is valid + } +#if ENABLE_FEATURE_VI_YANKMARK + check_context(c); // update the current context +#endif + + if (!isdigit(c)) + cmdcnt = 0; // cmd was not a number, reset cmdcnt + cnt = dot - begin_line(dot); + // Try to stay off of the Newline + if (*dot == '\n' && cnt > 0 && cmd_mode == 0) + dot--; +} + +/* NB! the CRASHME code is unmaintained, and doesn't currently build */ +#if ENABLE_FEATURE_VI_CRASHME +static int totalcmds = 0; +static int Mp = 85; // Movement command Probability +static int Np = 90; // Non-movement command Probability +static int Dp = 96; // Delete command Probability +static int Ip = 97; // Insert command Probability +static int Yp = 98; // Yank command Probability +static int Pp = 99; // Put command Probability +static int M = 0, N = 0, I = 0, D = 0, Y = 0, P = 0, U = 0; +static const char chars[20] = "\t012345 abcdABCD-=.$"; +static const char *const words[20] = { + "this", "is", "a", "test", + "broadcast", "the", "emergency", "of", + "system", "quick", "brown", "fox", + "jumped", "over", "lazy", "dogs", + "back", "January", "Febuary", "March" +}; +static const char *const lines[20] = { + "You should have received a copy of the GNU General Public License\n", + "char c, cm, *cmd, *cmd1;\n", + "generate a command by percentages\n", + "Numbers may be typed as a prefix to some commands.\n", + "Quit, discarding changes!\n", + "Forced write, if permission originally not valid.\n", + "In general, any ex or ed command (such as substitute or delete).\n", + "I have tickets available for the Blazers vs LA Clippers for Monday, Janurary 1 at 1:00pm.\n", + "Please get w/ me and I will go over it with you.\n", + "The following is a list of scheduled, committed changes.\n", + "1. Launch Norton Antivirus (Start, Programs, Norton Antivirus)\n", + "Reminder....Town Meeting in Central Perk cafe today at 3:00pm.\n", + "Any question about transactions please contact Sterling Huxley.\n", + "I will try to get back to you by Friday, December 31.\n", + "This Change will be implemented on Friday.\n", + "Let me know if you have problems accessing this;\n", + "Sterling Huxley recently added you to the access list.\n", + "Would you like to go to lunch?\n", + "The last command will be automatically run.\n", + "This is too much english for a computer geek.\n", +}; +static char *multilines[20] = { + "You should have received a copy of the GNU General Public License\n", + "char c, cm, *cmd, *cmd1;\n", + "generate a command by percentages\n", + "Numbers may be typed as a prefix to some commands.\n", + "Quit, discarding changes!\n", + "Forced write, if permission originally not valid.\n", + "In general, any ex or ed command (such as substitute or delete).\n", + "I have tickets available for the Blazers vs LA Clippers for Monday, Janurary 1 at 1:00pm.\n", + "Please get w/ me and I will go over it with you.\n", + "The following is a list of scheduled, committed changes.\n", + "1. Launch Norton Antivirus (Start, Programs, Norton Antivirus)\n", + "Reminder....Town Meeting in Central Perk cafe today at 3:00pm.\n", + "Any question about transactions please contact Sterling Huxley.\n", + "I will try to get back to you by Friday, December 31.\n", + "This Change will be implemented on Friday.\n", + "Let me know if you have problems accessing this;\n", + "Sterling Huxley recently added you to the access list.\n", + "Would you like to go to lunch?\n", + "The last command will be automatically run.\n", + "This is too much english for a computer geek.\n", +}; + +// create a random command to execute +static void crash_dummy() +{ + static int sleeptime; // how long to pause between commands + char c, cm, *cmd, *cmd1; + int i, cnt, thing, rbi, startrbi, percent; + + // "dot" movement commands + cmd1 = " \n\r\002\004\005\006\025\0310^$-+wWeEbBhjklHL"; + + // is there already a command running? + if (chars_to_parse > 0) + goto cd1; + cd0: + startrbi = rbi = 0; + sleeptime = 0; // how long to pause between commands + memset(readbuffer, '\0', sizeof(readbuffer)); + // generate a command by percentages + percent = (int) lrand48() % 100; // get a number from 0-99 + if (percent < Mp) { // Movement commands + // available commands + cmd = cmd1; + M++; + } else if (percent < Np) { // non-movement commands + cmd = "mz<>\'\""; // available commands + N++; + } else if (percent < Dp) { // Delete commands + cmd = "dx"; // available commands + D++; + } else if (percent < Ip) { // Inset commands + cmd = "iIaAsrJ"; // available commands + I++; + } else if (percent < Yp) { // Yank commands + cmd = "yY"; // available commands + Y++; + } else if (percent < Pp) { // Put commands + cmd = "pP"; // available commands + P++; + } else { + // We do not know how to handle this command, try again + U++; + goto cd0; + } + // randomly pick one of the available cmds from "cmd[]" + i = (int) lrand48() % strlen(cmd); + cm = cmd[i]; + if (strchr(":\024", cm)) + goto cd0; // dont allow colon or ctrl-T commands + readbuffer[rbi++] = cm; // put cmd into input buffer + + // now we have the command- + // there are 1, 2, and multi char commands + // find out which and generate the rest of command as necessary + if (strchr("dmryz<>\'\"", cm)) { // 2-char commands + cmd1 = " \n\r0$^-+wWeEbBhjklHL"; + if (cm == 'm' || cm == '\'' || cm == '\"') { // pick a reg[] + cmd1 = "abcdefghijklmnopqrstuvwxyz"; + } + thing = (int) lrand48() % strlen(cmd1); // pick a movement command + c = cmd1[thing]; + readbuffer[rbi++] = c; // add movement to input buffer + } + if (strchr("iIaAsc", cm)) { // multi-char commands + if (cm == 'c') { + // change some thing + thing = (int) lrand48() % strlen(cmd1); // pick a movement command + c = cmd1[thing]; + readbuffer[rbi++] = c; // add movement to input buffer + } + thing = (int) lrand48() % 4; // what thing to insert + cnt = (int) lrand48() % 10; // how many to insert + for (i = 0; i < cnt; i++) { + if (thing == 0) { // insert chars + readbuffer[rbi++] = chars[((int) lrand48() % strlen(chars))]; + } else if (thing == 1) { // insert words + strcat(readbuffer, words[(int) lrand48() % 20]); + strcat(readbuffer, " "); + sleeptime = 0; // how fast to type + } else if (thing == 2) { // insert lines + strcat(readbuffer, lines[(int) lrand48() % 20]); + sleeptime = 0; // how fast to type + } else { // insert multi-lines + strcat(readbuffer, multilines[(int) lrand48() % 20]); + sleeptime = 0; // how fast to type + } + } + strcat(readbuffer, "\033"); + } + chars_to_parse = strlen(readbuffer); + cd1: + totalcmds++; + if (sleeptime > 0) + mysleep(sleeptime); // sleep 1/100 sec +} + +// test to see if there are any errors +static void crash_test() +{ + static time_t oldtim; + + time_t tim; + char d[2], msg[80]; + + msg[0] = '\0'; + if (end < text) { + strcat(msg, "end<text "); + } + if (end > textend) { + strcat(msg, "end>textend "); + } + if (dot < text) { + strcat(msg, "dot<text "); + } + if (dot > end) { + strcat(msg, "dot>end "); + } + if (screenbegin < text) { + strcat(msg, "screenbegin<text "); + } + if (screenbegin > end - 1) { + strcat(msg, "screenbegin>end-1 "); + } + + if (msg[0]) { + printf("\n\n%d: \'%c\' %s\n\n\n%s[Hit return to continue]%s", + totalcmds, last_input_char, msg, SOs, SOn); + fflush(stdout); + while (safe_read(STDIN_FILENO, d, 1) > 0) { + if (d[0] == '\n' || d[0] == '\r') + break; + } + } + tim = time(NULL); + if (tim >= (oldtim + 3)) { + sprintf(status_buffer, + "Tot=%d: M=%d N=%d I=%d D=%d Y=%d P=%d U=%d size=%d", + totalcmds, M, N, I, D, Y, P, U, end - text + 1); + oldtim = tim; + } +} +#endif |