X-Git-Url: http://mj.ucw.cz/gitweb/?a=blobdiff_plain;f=lib%2Fregex.c;h=f952333e7e061aabd30dd5013394d32b86017cd1;hb=81a17ab471944888106c70a719e99d6e6ba96f22;hp=b74fb6a290f5abdfe6b0fa5b3f70aec4942056f5;hpb=97a9eb186ea7f6726258e164d75503ebb18fc490;p=libucw.git

diff --git a/lib/regex.c b/lib/regex.c
index b74fb6a2..f952333e 100644
--- a/lib/regex.c
+++ b/lib/regex.c
@@ -16,112 +16,15 @@
 #include <string.h>
 #include <stdlib.h>
 
-#if 1
-
-/* BSD regular expression library */
-
-#include <regex.h>
-
-#define INITIAL_MEM 1024		/* Initial space allocated for each pattern */
-#define CHAR_SET_SIZE 256		/* How many characters in the character set.  */
-
-struct regex {
-  struct re_pattern_buffer buf;
-  struct re_registers regs;		/* Must not change between re_match() calls */
-  int len_cache;
-};
-
-regex *
-rx_compile(byte *p, int icase)
-{
-  regex *r = xmalloc_zero(sizeof(regex));
-  const char *msg;
-
-  r->buf.buffer = xmalloc(INITIAL_MEM);
-  r->buf.allocated = INITIAL_MEM;
-  if (icase)
-    {
-      unsigned i;
-      r->buf.translate = xmalloc (CHAR_SET_SIZE);
-      /* Map uppercase characters to corresponding lowercase ones.  */
-      for (i = 0; i < CHAR_SET_SIZE; i++)
-        r->buf.translate[i] = Cupcase(i);
-    }
-  else
-    r->buf.translate = NULL;
-  re_set_syntax(RE_SYNTAX_POSIX_EXTENDED);
-  msg = re_compile_pattern(p, strlen(p), &r->buf);
-  if (!msg)
-    return r;
-  die("Error parsing pattern `%s': %s", p, msg);
-}
-
-void
-rx_free(regex *r)
-{
-  xfree(r->buf.buffer);
-  if (r->buf.translate)
-    xfree(r->buf.translate);
-  xfree(r);
-}
-
-int
-rx_match(regex *r, byte *s)
-{
-  int len = strlen(s);
-
-  r->len_cache = len;
-  if (re_match(&r->buf, s, len, 0, &r->regs) < 0)
-    return 0;
-  if (r->regs.start[0] || r->regs.end[0] != len) /* XXX: Why regex doesn't enforce implicit "^...$" ? */
-    return 0;
-  return 1;
-}
-
-int
-rx_subst(regex *r, byte *by, byte *src, byte *dest, uns destlen)
-{
-  byte *end = dest + destlen - 1;
-
-  if (!rx_match(r, src))
-    return 0;
-
-  while (*by)
-    {
-      if (*by == '\\')
-	{
-	  by++;
-	  if (*by >= '0' && *by <= '9')	/* \0 gets replaced by entire pattern */
-	    {
-	      uns j = *by++ - '0';
-	      if (j < r->regs.num_regs)
-		{
-		  byte *s = src + r->regs.start[j];
-		  uns i = r->regs.end[j] - r->regs.start[j];
-		  if (r->regs.start[j] > r->len_cache || r->regs.end[j] > r->len_cache)
-		    return -1;
-		  if (dest + i >= end)
-		    return -1;
-		  memcpy(dest, s, i);
-		  dest += i;
-		  continue;
-		}
-	    }
-	}
-      if (dest < end)
-	*dest++ = *by++;
-      else
-	return -1;
-    }
-  *dest = 0;
-  return 1;
-}
-
-#elif 0
+#if defined(CONFIG_OWN_REGEX) || defined(CONFIG_POSIX_REGEX)
 
 /* POSIX regular expression library */
 
+#ifdef CONFIG_OWN_REGEX
+#include "lib/regex/regex-sh.h"
+#else
 #include <regex.h>
+#endif
 
 struct regex {
   regex_t rx;
@@ -205,7 +108,7 @@ rx_subst(regex *r, byte *by, byte *src, byte *dest, uns destlen)
   return 1;
 }
 
-#else
+#elif defined(CONFIG_PCRE)
 
 /* PCRE library */
 
@@ -305,6 +208,111 @@ rx_subst(regex *r, byte *by, byte *src, byte *dest, uns destlen)
   return 1;
 }
 
+#else
+
+/* BSD regular expression library */
+
+#ifdef CONFIG_OWN_BSD_REGEX
+#include "lib/regex/regex-sh.h"
+#else
+#include <regex.h>
+#endif
+
+#define INITIAL_MEM 1024		/* Initial space allocated for each pattern */
+#define CHAR_SET_SIZE 256		/* How many characters in the character set.  */
+
+struct regex {
+  struct re_pattern_buffer buf;
+  struct re_registers regs;		/* Must not change between re_match() calls */
+  int len_cache;
+};
+
+regex *
+rx_compile(byte *p, int icase)
+{
+  regex *r = xmalloc_zero(sizeof(regex));
+  const char *msg;
+
+  r->buf.buffer = xmalloc(INITIAL_MEM);
+  r->buf.allocated = INITIAL_MEM;
+  if (icase)
+    {
+      unsigned i;
+      r->buf.translate = xmalloc (CHAR_SET_SIZE);
+      /* Map uppercase characters to corresponding lowercase ones.  */
+      for (i = 0; i < CHAR_SET_SIZE; i++)
+        r->buf.translate[i] = Cupcase(i);
+    }
+  else
+    r->buf.translate = NULL;
+  re_set_syntax(RE_SYNTAX_POSIX_EXTENDED);
+  msg = re_compile_pattern(p, strlen(p), &r->buf);
+  if (!msg)
+    return r;
+  die("Error parsing pattern `%s': %s", p, msg);
+}
+
+void
+rx_free(regex *r)
+{
+  xfree(r->buf.buffer);
+  if (r->buf.translate)
+    xfree(r->buf.translate);
+  xfree(r);
+}
+
+int
+rx_match(regex *r, byte *s)
+{
+  int len = strlen(s);
+
+  r->len_cache = len;
+  if (re_match(&r->buf, s, len, 0, &r->regs) < 0)
+    return 0;
+  if (r->regs.start[0] || r->regs.end[0] != len) /* XXX: Why regex doesn't enforce implicit "^...$" ? */
+    return 0;
+  return 1;
+}
+
+int
+rx_subst(regex *r, byte *by, byte *src, byte *dest, uns destlen)
+{
+  byte *end = dest + destlen - 1;
+
+  if (!rx_match(r, src))
+    return 0;
+
+  while (*by)
+    {
+      if (*by == '\\')
+	{
+	  by++;
+	  if (*by >= '0' && *by <= '9')	/* \0 gets replaced by entire pattern */
+	    {
+	      uns j = *by++ - '0';
+	      if (j < r->regs.num_regs)
+		{
+		  byte *s = src + r->regs.start[j];
+		  uns i = r->regs.end[j] - r->regs.start[j];
+		  if (r->regs.start[j] > r->len_cache || r->regs.end[j] > r->len_cache)
+		    return -1;
+		  if (dest + i >= end)
+		    return -1;
+		  memcpy(dest, s, i);
+		  dest += i;
+		  continue;
+		}
+	    }
+	}
+      if (dest < end)
+	*dest++ = *by++;
+      else
+	return -1;
+    }
+  *dest = 0;
+  return 1;
+}
+
 #endif
 
 #ifdef TEST
@@ -312,9 +320,16 @@ rx_subst(regex *r, byte *by, byte *src, byte *dest, uns destlen)
 int main(int argc, char **argv)
 {
   regex *r;
-  byte buf1[256], buf2[256];
+  byte buf1[4096], buf2[4096];
+  int opt_i = 0;
 
-  r = rx_compile(argv[1], 0);
+  if (!strcmp(argv[1], "-i"))
+    {
+      opt_i = 1;
+      argv++;
+      argc--;
+    }
+  r = rx_compile(argv[1], opt_i);
   while (fgets(buf1, sizeof(buf1), stdin))
     {
       char *p = strchr(buf1, '\n');