From: Dirk Mueller:

Only regcomp() can return an error, regexec() only returns match or
no-match. This allows to merge file_regerror into file_regcomp and
thereby simplifies the callers of file_regcomp() as they no longer
need to handle errors. Also this reduces memory overhead slightly
as we can make file_regex_t directly a regex_t, saving ~ 16 bytes per
regex.

4 files changed, 24 insertions, 46 deletions

diff --git a/src/apprentice.c b/src/apprentice.c
index 992102b4..0c422407 100644
--- a/src/apprentice.c
+++ b/src/apprentice.c
@@ -32,7 +32,7 @@
 #include "file.h"
 
 #ifndef	lint
-FILE_RCSID("@(#)$File: apprentice.c,v 1.322 2022/04/18 21:46:43 christos Exp $")
+FILE_RCSID("@(#)$File: apprentice.c,v 1.323 2022/04/18 21:50:49 christos Exp $")
 #endif	/* lint */
 
 #include "magic.h"
@@ -2830,11 +2830,9 @@ getvalue(struct magic_set *ms, struct magic *m, const char **p, int action)
 			file_regex_t rx;
 			int rc = file_regcomp(ms, &rx, m->value.s,
 			    REG_EXTENDED);
-			if (rc) {
-				if (ms->flags & MAGIC_CHECK)
-					file_regerror(&rx, rc, ms);
+			if (rc == 0) {
+				file_regfree(&rx);
 			}
-			file_regfree(&rx);
 			return rc ? -1 : 0;
 		}
 		return 0;
diff --git a/src/file.h b/src/file.h
index 51caa599..aea3f101 100644
--- a/src/file.h
+++ b/src/file.h
@@ -27,7 +27,7 @@
  */
 /*
  * file.h - definitions for file(1) program
- * @(#)$File: file.h,v 1.231 2022/04/11 18:07:38 christos Exp $
+ * @(#)$File: file.h,v 1.232 2022/04/18 21:50:49 christos Exp $
  */
 
 #ifndef __file_h__
@@ -171,11 +171,7 @@
 #define FILE_COMPILE	2
 #define FILE_LIST	3
 
-typedef struct {
-	const char *pat;
-	int rc;
-	regex_t rx;
-} file_regex_t;
+typedef regex_t file_regex_t;
 
 struct buffer {
 	int fd;
@@ -601,7 +597,6 @@ protected int file_regcomp(struct magic_set *, file_regex_t *, const char *,
 protected int file_regexec(struct magic_set *, file_regex_t *, const char *,
     size_t, regmatch_t *, int);
 protected void file_regfree(file_regex_t *);
-protected void file_regerror(file_regex_t *, int, struct magic_set *);
 
 typedef struct {
 	char *buf;
diff --git a/src/funcs.c b/src/funcs.c
index 2c1b2164..aab9c3bd 100644
--- a/src/funcs.c
+++ b/src/funcs.c
@@ -27,7 +27,7 @@
 #include "file.h"
 
 #ifndef	lint
-FILE_RCSID("@(#)$File: funcs.c,v 1.127 2022/04/11 18:07:38 christos Exp $")
+FILE_RCSID("@(#)$File: funcs.c,v 1.128 2022/04/18 21:50:49 christos Exp $")
 #endif	/* lint */
 
 #include "magic.h"
@@ -639,9 +639,7 @@ file_replace(struct magic_set *ms, const char *pat, const char *rep)
 	int rc, rv = -1;
 
 	rc = file_regcomp(ms, &rx, pat, REG_EXTENDED);
-	if (rc) {
-		file_regerror(&rx, rc, ms);
-	} else {
+	if (rc == 0) {
 		regmatch_t rm;
 		int nm = 0;
 		while (file_regexec(ms, &rx, ms->o.buf, 1, &rm, 0) == 0) {
@@ -670,16 +668,22 @@ file_regcomp(struct magic_set *ms file_locale_used, file_regex_t *rx,
 	strlcpy(old, setlocale(LC_CTYPE, NULL), sizeof(old));
 	(void)setlocale(LC_CTYPE, "C");
 #endif
-	rx->pat = pat;
-
-	rx->rc = regcomp(&rx->rx, pat, flags);
+	int rc;
+	rc = regcomp(rx, pat, flags);
 
 #ifdef USE_C_LOCALE
 	uselocale(old);
 #else
 	(void)setlocale(LC_CTYPE, old);
 #endif
-	return rx->rc;
+	if (rc > 0 && (ms->flags & MAGIC_CHECK)) {
+		char errmsg[512];
+
+		(void)regerror(rc, rx, errmsg, sizeof(errmsg));
+		file_magerror(ms, "regex error %d for `%s', (%s)", rc, pat,
+		    errmsg);
+	}
+	return rc;
 }
 
 protected int
@@ -695,11 +699,10 @@ file_regexec(struct magic_set *ms file_locale_used, file_regex_t *rx,
 	(void)setlocale(LC_CTYPE, "C");
 #endif
 	int rc;
-	assert(rx->rc == 0);
 	/* XXX: force initialization because glibc does not always do this */
 	if (nmatch != 0)
 		memset(pmatch, 0, nmatch * sizeof(*pmatch));
-	rc = regexec(&rx->rx, str, nmatch, pmatch, eflags);
+	rc = regexec(rx, str, nmatch, pmatch, eflags);
 #ifdef USE_C_LOCALE
 	uselocale(old);
 #else
@@ -711,18 +714,7 @@ file_regexec(struct magic_set *ms file_locale_used, file_regex_t *rx,
 protected void
 file_regfree(file_regex_t *rx)
 {
-	if (rx->rc == 0)
-		regfree(&rx->rx);
-}
-
-protected void
-file_regerror(file_regex_t *rx, int rc, struct magic_set *ms)
-{
-	char errmsg[512];
-
-	(void)regerror(rc, &rx->rx, errmsg, sizeof(errmsg));
-	file_magerror(ms, "regex error %d for `%s', (%s)", rc, rx->pat,
-	    errmsg);
+	regfree(rx);
 }
 
 protected file_pushbuf_t *
diff --git a/src/softmagic.c b/src/softmagic.c
index 99010da3..064196d7 100644
--- a/src/softmagic.c
+++ b/src/softmagic.c
@@ -32,7 +32,7 @@
 #include "file.h"
 
 #ifndef	lint
-FILE_RCSID("@(#)$File: softmagic.c,v 1.321 2022/03/19 19:52:09 christos Exp $")
+FILE_RCSID("@(#)$File: softmagic.c,v 1.322 2022/04/18 21:50:49 christos Exp $")
 #endif	/* lint */
 
 #include "magic.h"
@@ -474,14 +474,13 @@ check_fmt(struct magic_set *ms, const char *fmt)
 {
 	file_regex_t rx;
 	int rc, rv = -1;
+        const char* pat = "%[-0-9\\.]*s";
 
 	if (strchr(fmt, '%') == NULL)
 		return 0;
 
-	rc = file_regcomp(ms, &rx, "%[-0-9\\.]*s", REG_EXTENDED|REG_NOSUB);
-	if (rc) {
-		file_regerror(&rx, rc, ms);
-	} else {
+	rc = file_regcomp(ms, &rx, pat, REG_EXTENDED|REG_NOSUB);
+	if (rc == 0) {
 		rc = file_regexec(ms, &rx, fmt, 0, 0, 0);
 		rv = !rc;
 	}
@@ -2023,9 +2022,6 @@ alloc_regex(struct magic_set *ms, struct magic *m)
 	if (rc == 0)
 		return rx;
 
-fprintf(stderr, "regcomp %s %d\n", m->value.s, rc);
-	file_regerror(rx, rc, ms);
-	file_regfree(rx);
 	free(rx);
 	return NULL;
 }
@@ -2268,12 +2264,9 @@ magiccheck(struct magic_set *ms, struct magic *m, file_regex_t **m_cache)
 			break;
 
 		default:
-			file_regerror(rx, rc, ms);
-			v = CAST(uint64_t, -1);
+			return -1;
 			break;
 		}
-		if (v == CAST(uint64_t, -1))
-			return -1;
 		break;
 	}
 	case FILE_USE: