/* grep.c - show lines matching regular expressions * * Copyright 2013 CE Strake * * See http://pubs.opengroup.org/onlinepubs/9699919799/utilities/grep.html * * Posix doesn't even specify -r: too many deviations to document. * TODO: -i is only ascii case insensitive, not unicode. USE_GREP(NEWTOY(grep, "(line-buffered)(color):;(exclude-dir)*S(exclude)*M(include)*ZzEFHIab(byte-offset)h(no-filename)ino(only-matching)rRsvwc(count)L(files-without-match)l(files-with-matches)q(quiet)(silent)e*f*C#B#A#m#x[!wx][!EF]", TOYFLAG_BIN|TOYFLAG_ARGFAIL(2)|TOYFLAG_LINEBUF)) USE_EGREP(OLDTOY(egrep, grep, TOYFLAG_BIN|TOYFLAG_ARGFAIL(2)|TOYFLAG_LINEBUF)) USE_FGREP(OLDTOY(fgrep, grep, TOYFLAG_BIN|TOYFLAG_ARGFAIL(2)|TOYFLAG_LINEBUF)) config GREP bool "grep" default y help usage: grep [-EFrivwcloqsHbhn] [-ABC NUM] [-m MAX] [-e REGEX]... [-MS PATTERN]... [-f REGFILE]... [FILE]... Show lines matching regular expressions. If no -e, first argument is regular expression to match. With no files (or "-" filename) read stdin. Returns 0 if matched, 1 if no match found, 2 for command errors. -e Regex(es) to match. -f File(s) of regexes to match (1 per line). file search: -r Recurse into subdirs -R Recurse following symlinks -M Match files (--include) -S Skip files (--exclude) -I Ignore binary files --exclude-dir=PATTERN Skip directories match type: -A Show NUM lines after -B Show NUM lines before match -C NUM lines context (A+B) -E extended regex syntax -F fixed (literal match) -a always text (not binary) -i case insensitive -m match MAX many lines -v invert match -w whole word (implies -E) -x whole line -z input NUL terminated display modes: (default: matched line) -L filenames with no match -Z output is NUL terminated -c count of matching lines -l filenames with a match -o only matching part -q quiet (errors only) -s silent (no error msg) output prefix (default: filename if checking more than 1 file) -H force filename -b byte offset of match -h hide filename -n line number of match config EGREP bool default y depends on GREP config FGREP bool default y depends on GREP */ #define FOR_grep #include "toys.h" GLOBALS( long m, A, B, C; struct arg_list *f, *e, *M, *S, *exclude_dir; char *color; char *purple, *cyan, *red, *green, *grey; struct double_list *reg; int found, tried, delim; struct arg_list **fixed; ) struct reg { struct reg *next, *prev; int rc; regex_t r; regmatch_t m; }; static void numdash(long num, char dash) { printf("%s%ld%s%c", TT.green, num, TT.cyan, dash); } // Emit line with various potential prefixes and delimiter static void outline(char *line, char dash, char *name, long lcount, long bcount, unsigned trim) { if (!trim && FLAG(o)) return; if (name && FLAG(H)) printf("%s%s%s%c", TT.purple, name, TT.cyan, dash); if (FLAG(c)) xprintf("%s%ld%c", TT.grey, lcount, TT.delim); else if (lcount && FLAG(n)) numdash(lcount, dash); if (bcount && FLAG(b)) numdash(bcount-1, dash); if (line) { if (FLAG(color)) xputsn(FLAG(o) ? TT.red : TT.grey); // support embedded NUL bytes in output xputsl(line, trim); xputc(TT.delim); } } // Show matches in one file static void do_grep(int fd, char *name) { long lcount = 0, mcount = 0, offset = 0, after = 0, before = 0, new = 1; struct double_list *dlb = 0; char *bars = 0; FILE *file; int bin = 0; if (!FLAG(r)) TT.tried++; if (!fd) name = "(standard input)"; // Only run binary file check on lseekable files. if (!FLAG(a) && !lseek(fd, 0, SEEK_CUR)) { char buf[256]; int len, i = 0; unsigned wc; // If the first 256 bytes don't parse as utf8, call it binary. if (0<(len = read(fd, buf, 256))) { lseek(fd, -len, SEEK_CUR); while (inext) shoe->rc = 0; // Loop to handle multiple matches in same line if (new) do { regmatch_t *mm = (void *)toybuf; struct arg_list *seek; mm->rm_so = mm->rm_eo = 0; rc = 1; // Handle "fixed" (literal) matches (if any) if (TT.e && *start) for (ss = start; ss-linenext) { if (*(pp = seek->arg)=='^' && !FLAG(F)) { if (ss!=start) continue; pp++; } for (ii = 0; pp[ii] && ss[ii]; ii++) { if (!FLAG(F)) { if (pp[ii]=='.') continue; if (pp[ii]=='\\' && pp[ii+1]) pp++; else if (pp[ii]=='$' && !pp[ii+1]) break; } if (FLAG(i)) { if (toupper(pp[ii])!=toupper(ss[ii])) break; } else if (pp[ii]!=ss[ii]) break; } if (pp[ii] && (pp[ii]!='$' || pp[ii+1] || ss[ii])) continue; mm->rm_eo = (mm->rm_so = ss-start)+ii; rc = 0; goto got; } if (FLAG(x)) break; } // Empty pattern always matches if (rc && *TT.fixed && !FLAG(o)) rc = 0; got: // Handle regex matches (if any) for (shoe = (void *)TT.reg; shoe; shoe = shoe->next) { // Do we need to re-check this regex? if (!shoe->rc) { shoe->m.rm_so -= move; shoe->m.rm_eo -= move; if (!matched || shoe->m.rm_so<0) shoe->rc = regexec0(&shoe->r, start, ulen-(start-line), 1, &shoe->m, start==line ? 0 : REG_NOTBOL); } // If we got a match, is it a _better_ match? if (!shoe->rc && (rc || shoe->m.rm_so < mm->rm_so || (shoe->m.rm_so == mm->rm_so && shoe->m.rm_eo >= mm->rm_eo))) { mm = &shoe->m; rc = 0; } } if (!rc && FLAG(o) && !mm->rm_eo && ulen>start-line) { move = 1; continue; } if (!rc && FLAG(x) && (mm->rm_so || ulen-(start-line)!=mm->rm_eo)) rc = 1; if (!rc && FLAG(w)) { char c = 0; if ((start+mm->rm_so)!=line) { c = start[mm->rm_so-1]; if (!isalnum(c) && c != '_') c = 0; } if (!c) { c = start[mm->rm_eo]; if (!isalnum(c) && c != '_') c = 0; } if (c) { move = mm->rm_so+1; continue; } } if (FLAG(v)) { if (FLAG(o)) { if (rc) mm->rm_eo = ulen-(start-line); else if (!mm->rm_so) { move = mm->rm_eo; continue; } else mm->rm_eo = mm->rm_so; } else { if (!rc) break; mm->rm_eo = ulen-(start-line); } mm->rm_so = 0; } else if (rc) break; // At least one line we didn't print since match while -ABC active if (bars) { xputs(bars); bars = 0; } matched++; TT.found = 1; // Are we NOT showing the matching text? if (FLAG(q)) { toys.exitval = 0; xexit(); } if (FLAG(L) || FLAG(l)) { if (FLAG(l)) xprintf("%s%c", name, '\n'*!FLAG(Z)); free(line); fclose(file); return; } if (!FLAG(c)) { long bcount = 1 + offset + (start-line) + (FLAG(o) ? mm->rm_so : 0); if (bin) printf("Binary file %s matches\n", name); else if (FLAG(o)) outline(start+mm->rm_so, ':', name, lcount, bcount, mm->rm_eo-mm->rm_so); else { while (dlb) { struct double_list *dl = dlist_pop(&dlb); unsigned *uu = (void *)(dl->data+(strlen(dl->data)|3)+1); outline(dl->data, '-', name, lcount-before, uu[0]+1, uu[1]); free(dl->data); free(dl); before--; } if (matched==1) outline(FLAG(color) ? 0 : line, ':', name, lcount, bcount, ulen); if (FLAG(color)) { xputsn(TT.grey); if (mm->rm_so) xputsl(start, mm->rm_so); xputsn(TT.red); xputsl(start+mm->rm_so, mm->rm_eo-mm->rm_so); } if (TT.A) after = TT.A+1; } } if (mm->rm_so == (move = mm->rm_eo)) break; } while (*(start += move)); offset += len; if (matched) { // Finish off pending line color fragment. if (FLAG(color) && !FLAG(o)) { xputsn(TT.grey); if (ulen > start-line) xputsl(start, ulen-(start-line)); xputc(TT.delim); } mcount++; } else { int discard = (after || TT.B); if (after && --after) { outline(line, '-', name, lcount, 0, ulen); discard = 0; } if (discard && TT.B) { unsigned *uu, ul = (ulen|3)+1; line = xrealloc(line, ul+8); uu = (void *)(line+ul); uu[0] = offset-len; uu[1] = ulen; dlist_add(&dlb, line); line = 0; if (++before>TT.B) { struct double_list *dl; dl = dlist_pop(&dlb); free(dl->data); free(dl); before--; } else discard = 0; } // If we discarded a line while displaying context, show bars before next // line (but don't show them now in case that was last match in file) if (discard && mcount) bars = "--"; } free(line); if (FLAG(m) && mcount >= TT.m) { if (!after) break; new = 0; } } if (FLAG(L)) xprintf("%s%c", name, TT.delim); else if (FLAG(c)) outline(0, ':', name, mcount, 0, 1); // loopfiles will also close the fd, but this frees an (opaque) struct. fclose(file); llist_traverse(dlb, llist_free_double); } static int lensort(struct arg_list **a, struct arg_list **b) { long la = strlen((*a)->arg), lb = strlen((*b)->arg); if (lalb) return 1; return 0; } static void parse_regex(void) { struct arg_list *al, *new, *list = NULL, **last; char *s, *ss, *special = "\\.^$[()|*+?{"; int len, ii, key; // Add all -f lines to -e list. (Yes, this is leaking allocation context for // exit to free. Not supporting nofork for this command any time soon.) al = TT.f ? TT.f : TT.e; while (al) { if (TT.f) { if (!*(s = xreadfile(al->arg, 0, 0))) { free(s); s = 0; } else if (*(ss = s+strlen(s)-1)=='\n') *ss = 0; } else s = al->arg; // Advance, when we run out of -f switch to -e. al = al->next; if (!al && TT.f) { TT.f = 0; al = TT.e; } if (!s) continue; // NOTE: even with -z, -f is still \n delimited. Blank line = match all // Split lines at \n, add individual lines to new list. do { if ((ss = strchr(s, '\n'))) *(ss++) = 0; new = xmalloc(sizeof(struct arg_list)); new->next = list; new->arg = s; list = new; s = ss; } while (s); } TT.e = list; // Convert to regex where appropriate for (last = &TT.e; *last;) { // Can we use the fast path? s = (*last)->arg; if ('.'!=*s && !FLAG(F) && strcmp(s, "^$")) for (; *s; s++) { if (*s=='\\') { if (!s[1] || !strchr(special, *++s)) break; if (!FLAG(E) && *s=='(') break; } else if (*s>127 || strchr(special+4, *s)) break; } // Add entry to fast path (literal-ish match) or slow path (regexec) if (!*s || FLAG(F)) last = &((*last)->next); else { struct reg *shoe; dlist_add_nomalloc(&TT.reg, (void *)(shoe = xmalloc(sizeof(struct reg)))); xregcomp(&shoe->r, (*last)->arg, REG_EXTENDED*FLAG(E)|REG_ICASE*FLAG(i)); al = *last; *last = (*last)->next; free(al); } } dlist_terminate(TT.reg); // Sort fast path patterns into buckets by first character for (al = TT.e; al; al = new) { new = al->next; if (FLAG(F)) key = 0; else { key = '^'==*al->arg; if ('\\'==al->arg[key]) key++; else if ('$'==al->arg[key] && !al->arg[key+1]) key++; } key = al->arg[key]; if (FLAG(i)) key = toupper(key); al->next = TT.fixed[key]; TT.fixed[key] = al; } // Sort each fast path pattern set by length so first hit is longest match if (TT.e) for (key = 0; key<256; key++) { if (!TT.fixed[key]) continue; for (len = 0, al = TT.fixed[key]; al; al = al->next) len++; last = xmalloc(len*sizeof(void *)); for (len = 0, al = TT.fixed[key]; al; al = al->next) last[len++] = al; qsort(last, len, sizeof(void *), (void *)lensort); for (ii = 0; iinext = ii ? last[ii-1] : 0; TT.fixed[key] = last[len-1]; free(last); } } static int do_grep_r(struct dirtree *new) { struct arg_list *al; char *name; if (!new->parent) TT.tried++; if (!dirtree_notdotdot(new)) return 0; if (S_ISDIR(new->st.st_mode)) { for (al = TT.exclude_dir; al; al = al->next) if (!fnmatch(al->arg, new->name, 0)) return 0; return DIRTREE_RECURSE|DIRTREE_SYMFOLLOW*FLAG(R); } if (TT.S || TT.M) { for (al = TT.S; al; al = al->next) if (!fnmatch(al->arg, new->name, 0)) return 0; if (TT.M) { for (al = TT.M; al; al = al->next) if (!fnmatch(al->arg, new->name, 0)) break; if (!al) return 0; } } // "grep -r onefile" doesn't show filenames, but "grep -r onedir" should. if (new->parent && !FLAG(h)) toys.optflags |= FLAG_H; name = dirtree_path(new, 0); do_grep(openat(dirtree_parentfd(new), new->name, 0), name); free(name); return 0; } void grep_main(void) { char **ss = toys.optargs; TT.fixed = xzalloc(256*sizeof(*TT.fixed)); if (FLAG(color) && (!TT.color || !strcmp(TT.color, "auto")) && !isatty(1)) toys.optflags &= ~FLAG_color; if (FLAG(color)) { TT.purple = "\e[35m"; TT.cyan = "\e[36m"; TT.red = "\e[1;31m"; TT.green = "\e[32m"; TT.grey = "\e[m"; } else TT.purple = TT.cyan = TT.red = TT.green = TT.grey = ""; if (FLAG(R)) toys.optflags |= FLAG_r; // Grep exits with 2 for errors toys.exitval = 2; if (!TT.A) TT.A = TT.C; if (!TT.B) TT.B = TT.C; TT.delim = '\n' * !FLAG(z); // Handle egrep and fgrep if (*toys.which->name == 'e') toys.optflags |= FLAG_E; if (*toys.which->name == 'f') toys.optflags |= FLAG_F; if (!TT.e && !TT.f) { if (!*ss) error_exit("no REGEX"); TT.e = xzalloc(sizeof(struct arg_list)); TT.e->arg = *(ss++); toys.optc--; } parse_regex(); if (!FLAG(h) && toys.optc>1) toys.optflags |= FLAG_H; if (FLAG(s)) { close(2); xopen_stdio("/dev/null", O_RDWR); } if (FLAG(r)) { // Iterate through -r arguments. Use "." as default if none provided. for (ss = *ss ? ss : (char *[]){".", 0}; *ss; ss++) { if (!strcmp(*ss, "-")) do_grep(0, *ss); else dirtree_read(*ss, do_grep_r); } } else loopfiles_rw(ss, O_RDONLY|WARN_ONLY, 0, do_grep); if (TT.tried >= toys.optc || (FLAG(q)&&TT.found)) toys.exitval = !TT.found; }