vi: allow regular expressions in ':s' commands

BusyBox vi has never supported the use of regular expressions in search/replace (':s') commands. Implement this using GNU regex when VI_REGEX_SEARCH is enabled. The implementation: - uses basic regular expressions, to match those used in the search command; - only supports substitution of back references ('\0' - '\9') in the replacement string. Any other character following a backslash is treated as that literal character. VI_REGEX_SEARCH isn't enabled in the default build. In that case: function old new delta colon 4036 4033 -3 ------------------------------------------------------------------------------ (add/remove: 0/0 grow/shrink: 0/1 up/down: 0/-3) Total: -3 bytes When VI_REGEX_SEARCH is enabled: function old new delta colon 4036 4378 +342 .rodata 108207 108229 +22 ------------------------------------------------------------------------------ (add/remove: 0/0 grow/shrink: 2/0 up/down: 364/0) Total: 364 bytes v2: Rebase. Code shrink. Ensure empty replacement string is null terminated. Signed-off-by: Andrey Dobrovolsky <andrey.dobrovolsky.odessa@gmail.com> Signed-off-by: Ron Yorston <rmy@pobox.com> Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
author: Denys Vlasenko <vda.linux@googlemail.com> 2021-07-13 14:38:20 +0200
committer: Denys Vlasenko <vda.linux@googlemail.com> 2021-07-13 14:38:20 +0200
commit: 95ac4a48f17c2fdd2a10524c0b399e3be72d8f42 (patch)
tree: 7877669d7835524e1152dfa890dc189faad61dc1
parent: c76c78740a19ed3b1f9c5910313460221096536a (diff)
download: busybox-w32-95ac4a48f17c2fdd2a10524c0b399e3be72d8f42.tar.gz
busybox-w32-95ac4a48f17c2fdd2a10524c0b399e3be72d8f42.tar.bz2
busybox-w32-95ac4a48f17c2fdd2a10524c0b399e3be72d8f42.zip
1 files changed, 116 insertions, 15 deletions
diff --git a/editors/vi.c b/editors/vi.c
index 2941b8ae4..070e0f55a 100644
--- a/editors/vi.c
+++ b/editors/vi.c
@@ -2677,6 +2677,59 @@ static char *expand_args(char *args)
 # endif
 #endif /* FEATURE_VI_COLON */
+#if ENABLE_FEATURE_VI_REGEX_SEARCH
+# define MAX_SUBPATTERN 10      // subpatterns \0 .. \9
+// If the return value is not NULL the caller should free R
+static char *regex_search(char *q, regex_t *preg, const char *Rorig,
+                                size_t *len_F, size_t *len_R, char **R)
+{
+        regmatch_t regmatch[MAX_SUBPATTERN], *cur_match;
+        char *found = NULL;
+        const char *t;
+        char *r;
+        regmatch[0].rm_so = 0;
+        regmatch[0].rm_eo = end_line(q) - q;
+        if (regexec(preg, q, MAX_SUBPATTERN, regmatch, REG_STARTEND) != 0)
+                return found;
+        found = q + regmatch[0].rm_so;
+        *len_F = regmatch[0].rm_eo - regmatch[0].rm_so;
+        *R = NULL;
+ fill_result:
+        // first pass calculates len_R, second fills R
+        *len_R = 0;
+        for (t = Rorig, r = *R; *t; t++) {
+                size_t len = 1; // default is to copy one char from replace pattern
+                const char *from = t;
+                if (*t == '\\') {
+                        from = ++t;     // skip backslash
+                        if (*t >= '0' && *t < '0' + MAX_SUBPATTERN) {
+                                cur_match = regmatch + (*t - '0');
+                                if (cur_match->rm_so >= 0) {
+                                        len = cur_match->rm_eo - cur_match->rm_so;
+                                        from = q + cur_match->rm_so;
+                                }
+                        }
+                }
+                *len_R += len;
+                if (*R) {
+                        memcpy(r, from, len);
+                        r += len;
+                        /* *r = '\0'; - xzalloc did it */
+                }
+        }
+        if (*R == NULL) {
+                *R = xzalloc(*len_R + 1);
+                goto fill_result;
+        }
+        return found;
+}
+#endif /* ENABLE_FEATURE_VI_REGEX_SEARCH */
 // buf must be no longer than MAX_INPUT_LEN!
 static void colon(char *buf)
 {
@@ -3084,6 +3137,14 @@ static void colon(char *buf)
 #  if ENABLE_FEATURE_VI_VERBOSE_STATUS
                int last_line = 0, lines = 0;
 #  endif
+#  if ENABLE_FEATURE_VI_REGEX_SEARCH
+                regex_t preg;
+                int cflags;
+                char *Rorig;
+#   if ENABLE_FEATURE_VI_UNDO
+                int undo = 0;
+#   endif
+#  endif
                // F points to the "find" pattern
                // R points to the "replace" pattern
@@ -3100,7 +3161,6 @@ static void colon(char *buf)
                        *flags++ = '\0';        // terminate "replace"
                        gflag = *flags;
                }
-                len_R = strlen(R);
                if (len_F) {    // save "find" as last search pattern
                        free(last_search_pattern);
@@ -3122,31 +3182,68 @@ static void colon(char *buf)
                        b = e;
                }
+#  if ENABLE_FEATURE_VI_REGEX_SEARCH
+                Rorig = R;
+                cflags = 0;
+                if (ignorecase)
+                        cflags = REG_ICASE;
+                memset(&preg, 0, sizeof(preg));
+                if (regcomp(&preg, F, cflags) != 0) {
+                        status_line(":s bad search pattern");
+                        goto regex_search_end;
+                }
+#  else
+                len_R = strlen(R);
+#  endif
                for (i = b; i <= e; i++) {      // so, :20,23 s \0 find \0 replace \0
                        char *ls = q;           // orig line start
                        char *found;
 vc4:
+#  if ENABLE_FEATURE_VI_REGEX_SEARCH
+                        found = regex_search(q, &preg, Rorig, &len_F, &len_R, &R);
+#  else
                        found = char_search(q, F, (FORWARD << 1) | LIMITED);    // search cur line only for "find"
+#  endif
                        if (found) {
                                uintptr_t bias;
                                // we found the "find" pattern - delete it
                                // For undo support, the first item should not be chained
-                                text_hole_delete(found, found + len_F - 1,
+                                // This needs to be handled differently depending on
-                                                        subs ? ALLOW_UNDO_CHAIN: ALLOW_UNDO);
+                                // whether or not regex support is enabled.
-                                // can't do this above, no undo => no third argument
+#  if ENABLE_FEATURE_VI_REGEX_SEARCH
-                                subs++;
+#   define TEST_LEN_F len_F     // len_F may be zero
-#  if ENABLE_FEATURE_VI_VERBOSE_STATUS
+#   define TEST_UNDO1 undo++
-                                if (last_line != i) {
+#   define TEST_UNDO2 undo++
-                                        last_line = i;
+#  else
-                                        ++lines;
+#   define TEST_LEN_F 1         // len_F is never zero
+#   define TEST_UNDO1 subs
+#   define TEST_UNDO2 1
+#  endif
+                                if (TEST_LEN_F) // match can be empty, no delete needed
+                                        text_hole_delete(found, found + len_F - 1,
+                                                                TEST_UNDO1 ? ALLOW_UNDO_CHAIN: ALLOW_UNDO);
+                                if (len_R) {    // insert the "replace" pattern, if required
+                                        bias = string_insert(found, R,
+                                                                TEST_UNDO2 ? ALLOW_UNDO_CHAIN: ALLOW_UNDO);
+                                        found += bias;
+                                        ls += bias;
+                                        dot = ls;
+                                        //q += bias; - recalculated anyway
                                }
+#  if ENABLE_FEATURE_VI_REGEX_SEARCH
+                                free(R);
 #  endif
-                                // insert the "replace" patern
+                                if (TEST_LEN_F || len_R) {
-                                bias = string_insert(found, R, ALLOW_UNDO_CHAIN);
+                                        dot = ls;
-                                found += bias;
+                                        subs++;
-                                ls += bias;
+#  if ENABLE_FEATURE_VI_VERBOSE_STATUS
-                                dot = ls;
+                                        if (last_line != i) {
-                                //q += bias; - recalculated anyway
+                                                last_line = i;
+                                                ++lines;
+                                        }
+#  endif
+                                }
                                // check for "global"  :s/foo/bar/g
                                if (gflag == 'g') {
                                        if ((found + len_R) < end_line(ls)) {
@@ -3166,6 +3263,10 @@ static void colon(char *buf)
                                status_line("%d substitutions on %d lines", subs, lines);
 #  endif
                }
+#  if ENABLE_FEATURE_VI_REGEX_SEARCH
+ regex_search_end:
+                regfree(&preg);
+#  endif
 # endif /* FEATURE_VI_SEARCH */
        } else if (strncmp(cmd, "version", i) == 0) {  // show software version
                status_line(BB_VER);
author	Denys Vlasenko <vda.linux@googlemail.com>	2021-07-13 14:38:20 +0200
committer	Denys Vlasenko <vda.linux@googlemail.com>	2021-07-13 14:38:20 +0200
commit	95ac4a48f17c2fdd2a10524c0b399e3be72d8f42 (patch)
tree	7877669d7835524e1152dfa890dc189faad61dc1
parent	c76c78740a19ed3b1f9c5910313460221096536a (diff)
download	busybox-w32-95ac4a48f17c2fdd2a10524c0b399e3be72d8f42.tar.gz busybox-w32-95ac4a48f17c2fdd2a10524c0b399e3be72d8f42.tar.bz2 busybox-w32-95ac4a48f17c2fdd2a10524c0b399e3be72d8f42.zip

diff --git a/editors/vi.c b/editors/vi.c index 2941b8ae4..070e0f55a 100644 --- a/editors/vi.c +++ b/editors/vi.c
@@ -2677,6 +2677,59 @@ static char expand_args(char args)
2677	# endif	2677	# endif
2678	#endif /* FEATURE_VI_COLON */	2678	#endif /* FEATURE_VI_COLON */
2679		2679
		2680	#if ENABLE_FEATURE_VI_REGEX_SEARCH
		2681	# define MAX_SUBPATTERN 10 // subpatterns \0 .. \9
		2682
		2683	// If the return value is not NULL the caller should free R
		2684	static char regex_search(char q, regex_t preg, const char Rorig,
		2685	size_t len_F, size_t len_R, char **R)
		2686	{
		2687	regmatch_t regmatch[MAX_SUBPATTERN], *cur_match;
		2688	char *found = NULL;
		2689	const char *t;
		2690	char *r;
		2691
		2692	regmatch[0].rm_so = 0;
		2693	regmatch[0].rm_eo = end_line(q) - q;
		2694	if (regexec(preg, q, MAX_SUBPATTERN, regmatch, REG_STARTEND) != 0)
		2695	return found;
		2696
		2697	found = q + regmatch[0].rm_so;
		2698	*len_F = regmatch[0].rm_eo - regmatch[0].rm_so;
		2699	*R = NULL;
		2700
		2701	fill_result:
		2702	// first pass calculates len_R, second fills R
		2703	*len_R = 0;
		2704	for (t = Rorig, r = R; t; t++) {
		2705	size_t len = 1; // default is to copy one char from replace pattern
		2706	const char *from = t;
		2707	if (*t == '\\') {
		2708	from = ++t; // skip backslash
		2709	if (t >= '0' && t < '0' + MAX_SUBPATTERN) {
		2710	cur_match = regmatch + (*t - '0');
		2711	if (cur_match->rm_so >= 0) {
		2712	len = cur_match->rm_eo - cur_match->rm_so;
		2713	from = q + cur_match->rm_so;
		2714	}
		2715	}
		2716	}
		2717	*len_R += len;
		2718	if (*R) {
		2719	memcpy(r, from, len);
		2720	r += len;
		2721	/* r = '\0'; - xzalloc did it /
		2722	}
		2723	}
		2724	if (*R == NULL) {
		2725	R = xzalloc(len_R + 1);
		2726	goto fill_result;
		2727	}
		2728
		2729	return found;
		2730	}
		2731	#endif /* ENABLE_FEATURE_VI_REGEX_SEARCH */
		2732
2680	// buf must be no longer than MAX_INPUT_LEN!	2733	// buf must be no longer than MAX_INPUT_LEN!
2681	static void colon(char *buf)	2734	static void colon(char *buf)
2682	{	2735	{
@@ -3084,6 +3137,14 @@ static void colon(char *buf)
3084	# if ENABLE_FEATURE_VI_VERBOSE_STATUS	3137	# if ENABLE_FEATURE_VI_VERBOSE_STATUS
3085	int last_line = 0, lines = 0;	3138	int last_line = 0, lines = 0;
3086	# endif	3139	# endif
		3140	# if ENABLE_FEATURE_VI_REGEX_SEARCH
		3141	regex_t preg;
		3142	int cflags;
		3143	char *Rorig;
		3144	# if ENABLE_FEATURE_VI_UNDO
		3145	int undo = 0;
		3146	# endif
		3147	# endif
3087		3148
3088	// F points to the "find" pattern	3149	// F points to the "find" pattern
3089	// R points to the "replace" pattern	3150	// R points to the "replace" pattern
@@ -3100,7 +3161,6 @@ static void colon(char *buf)
3100	*flags++ = '\0'; // terminate "replace"	3161	*flags++ = '\0'; // terminate "replace"
3101	gflag = *flags;	3162	gflag = *flags;
3102	}	3163	}
3103	len_R = strlen(R);
3104		3164
3105	if (len_F) { // save "find" as last search pattern	3165	if (len_F) { // save "find" as last search pattern
3106	free(last_search_pattern);	3166	free(last_search_pattern);
@@ -3122,31 +3182,68 @@ static void colon(char *buf)
3122	b = e;	3182	b = e;
3123	}	3183	}
3124		3184
		3185	# if ENABLE_FEATURE_VI_REGEX_SEARCH
		3186	Rorig = R;
		3187	cflags = 0;
		3188	if (ignorecase)
		3189	cflags = REG_ICASE;
		3190	memset(&preg, 0, sizeof(preg));
		3191	if (regcomp(&preg, F, cflags) != 0) {
		3192	status_line(":s bad search pattern");
		3193	goto regex_search_end;
		3194	}
		3195	# else
		3196	len_R = strlen(R);
		3197	# endif
		3198
3125	for (i = b; i <= e; i++) { // so, :20,23 s \0 find \0 replace \0	3199	for (i = b; i <= e; i++) { // so, :20,23 s \0 find \0 replace \0
3126	char *ls = q; // orig line start	3200	char *ls = q; // orig line start
3127	char *found;	3201	char *found;
3128	vc4:	3202	vc4:
		3203	# if ENABLE_FEATURE_VI_REGEX_SEARCH
		3204	found = regex_search(q, &preg, Rorig, &len_F, &len_R, &R);
		3205	# else
3129	found = char_search(q, F, (FORWARD << 1) \| LIMITED); // search cur line only for "find"	3206	found = char_search(q, F, (FORWARD << 1) \| LIMITED); // search cur line only for "find"
		3207	# endif
3130	if (found) {	3208	if (found) {
3131	uintptr_t bias;	3209	uintptr_t bias;
3132	// we found the "find" pattern - delete it	3210	// we found the "find" pattern - delete it
3133	// For undo support, the first item should not be chained	3211	// For undo support, the first item should not be chained
3134	text_hole_delete(found, found + len_F - 1,	3212	// This needs to be handled differently depending on
3135	subs ? ALLOW_UNDO_CHAIN: ALLOW_UNDO);	3213	// whether or not regex support is enabled.
3136	// can't do this above, no undo => no third argument	3214	# if ENABLE_FEATURE_VI_REGEX_SEARCH
3137	subs++;	3215	# define TEST_LEN_F len_F // len_F may be zero
3138	# if ENABLE_FEATURE_VI_VERBOSE_STATUS	3216	# define TEST_UNDO1 undo++
3139	if (last_line != i) {	3217	# define TEST_UNDO2 undo++
3140	last_line = i;	3218	# else
3141	++lines;	3219	# define TEST_LEN_F 1 // len_F is never zero
		3220	# define TEST_UNDO1 subs
		3221	# define TEST_UNDO2 1
		3222	# endif
		3223	if (TEST_LEN_F) // match can be empty, no delete needed
		3224	text_hole_delete(found, found + len_F - 1,
		3225	TEST_UNDO1 ? ALLOW_UNDO_CHAIN: ALLOW_UNDO);
		3226	if (len_R) { // insert the "replace" pattern, if required
		3227	bias = string_insert(found, R,
		3228	TEST_UNDO2 ? ALLOW_UNDO_CHAIN: ALLOW_UNDO);
		3229	found += bias;
		3230	ls += bias;
		3231	dot = ls;
		3232	//q += bias; - recalculated anyway
3142	}	3233	}
		3234	# if ENABLE_FEATURE_VI_REGEX_SEARCH
		3235	free(R);
3143	# endif	3236	# endif
3144	// insert the "replace" patern	3237	if (TEST_LEN_F \|\| len_R) {
3145	bias = string_insert(found, R, ALLOW_UNDO_CHAIN);	3238	dot = ls;
3146	found += bias;	3239	subs++;
3147	ls += bias;	3240	# if ENABLE_FEATURE_VI_VERBOSE_STATUS
3148	dot = ls;	3241	if (last_line != i) {
3149	//q += bias; - recalculated anyway	3242	last_line = i;
		3243	++lines;
		3244	}
		3245	# endif
		3246	}
3150	// check for "global" :s/foo/bar/g	3247	// check for "global" :s/foo/bar/g
3151	if (gflag == 'g') {	3248	if (gflag == 'g') {
3152	if ((found + len_R) < end_line(ls)) {	3249	if ((found + len_R) < end_line(ls)) {
@@ -3166,6 +3263,10 @@ static void colon(char *buf)
3166	status_line("%d substitutions on %d lines", subs, lines);	3263	status_line("%d substitutions on %d lines", subs, lines);
3167	# endif	3264	# endif
3168	}	3265	}
		3266	# if ENABLE_FEATURE_VI_REGEX_SEARCH
		3267	regex_search_end:
		3268	regfree(&preg);
		3269	# endif
3169	# endif /* FEATURE_VI_SEARCH */	3270	# endif /* FEATURE_VI_SEARCH */
3170	} else if (strncmp(cmd, "version", i) == 0) { // show software version	3271	} else if (strncmp(cmd, "version", i) == 0) { // show software version
3171	status_line(BB_VER);	3272	status_line(BB_VER);