Interface to our own regex library.

diff --git a/lib/Makefile b/lib/Makefile
index 8791d2406b2cb4975bb22dbf472d717241260b9a..c3b93e8387331671fd9a64314145225e7d31ed6f 100644 (file)
--- a/lib/Makefile
+++ b/lib/Makefile
@@ -11,6 +11,11 @@ LIBSH_MODS=alloc alloc_str ctmatch db fastbuf fb-file fb-mem lists \
        finger proctitle ipaccess profile bitsig randomkey \
        hashfunc base64 base224 fb-temp fb-mmap fb-printf urlkey \
        partmap fb-limfd fb-buffer
+
+ifdef CONFIG_OWN_REGEX
+include lib/regex/Makefile
+endif
+
 LIBSH_MOD_PATHS=$(addprefix obj/lib/,$(LIBSH_MODS)) $(CUSTOM_LIB_MODULES)
 
 obj/lib/libsh.a: $(addsuffix .o,$(LIBSH_MOD_PATHS))

diff --git a/lib/regex.c b/lib/regex.c
index a80ed74185f68b8c5cef60b174a5dbff2574f06e..f952333e7e061aabd30dd5013394d32b86017cd1 100644 (file)
--- a/lib/regex.c
+++ b/lib/regex.c
@@ -16,112 +16,15 @@
 #include <string.h>
 #include <stdlib.h>
 
-#if !defined(CONFIG_PCRE) && !defined(CONFIG_POSIX_RE)
-
-/* 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 defined(CONFIG_POSIX_RE)
+#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

diff --git a/lib/regex.t b/lib/regex.t
index 4cf0d08cec0626f05f6b7bebfe92f428f5b49706..9b47cd99fdbf5ce98ddd9668eeb746b74a651f44 100644 (file)
--- a/lib/regex.t
+++ b/lib/regex.t
@@ -50,4 +50,16 @@ Out: a<aabb>b
 
 Run:   obj/lib/regex-t '.*\?(.*&)*([a-z_]*sess[a-z_]*|random|sid|S_ID|rnd|timestamp|referer)=.*'
 In:    /nemecky/ubytovani/hotel.php?sort=&cislo=26&mena=EUR&typ=Hotel&luz1=ANO&luz2=ANO&luz3=&luz4=&luz5=&maxp1=99999&maxp2=99999&maxp3=99999&maxp4=99999&maxp5=99999&apart=&rada=8,9,10,11,19,22,26,27,28,29,3&cislo=26&mena=EUR&typ=Hotel&luz1=ANO&luz2=ANO&luz3=&luz4=&luz5=&maxp1=99999&maxp2=99999&maxp3=99999&maxp4=99999&maxp5=99999&apart=&rada=8,9,10,11,19,22,26,27,28,29,3&cislo=26&mena=EUR&typ=Hotel&luz1=ANO&luz2=ANO&luz3=&luz4=&luz5=&maxp1=99999&maxp2=99999&maxp3=99999&maxp4=99999&maxp5=99999&apart=&rada=8,9,10,11,19,22,26,27,28,29,3
+       /test...?f=1&s=3&sid=123&q=3&
 Out:   NO MATCH
+       MATCH
+
+Run:   obj/lib/regex-t '.*[0-9a-f]{8,16}.*'
+In:    abcdabcdabcd
+       aaaaaaaaaaaaaaaaaaaaaaaaaaaa
+       asddajlkdkajlqwepoiequwiouio
+       000001111p101010101010q12032
+Out:   MATCH
+       MATCH
+       NO MATCH
+       MATCH