X-Git-Url: http://mj.ucw.cz/gitweb/?a=blobdiff_plain;f=ucw%2Furl.c;h=39e52946e2bd086cd3cbdaefa4985fe217b56fde;hb=342e0c3edeacf4eecd03da36c879ca817c64a0f3;hp=8f5073a99721aadef2b1b29526940ee5d58bc8db;hpb=689b60e4c61d300b576fdb7007de51189d2f0f7e;p=libucw.git

diff --git a/ucw/url.c b/ucw/url.c
index 8f5073a9..39e52946 100644
--- a/ucw/url.c
+++ b/ucw/url.c
@@ -7,11 +7,6 @@
  *	This software may be freely distributed and used according to the terms
  *	of the GNU Lesser General Public License.
  *
- *	The URL syntax corresponds to RFC 2396 with several exceptions:
- *
- *	   o  Escaping of special characters still follows RFC 1738.
- *	   o  Interpretation of path parameters follows RFC 1808.
- *
  *	XXX: The buffer handling in this module is really horrible, but it works.
  */
 
@@ -35,6 +30,7 @@ static uns url_min_repeat_count = 0x7fffffff;
 static uns url_max_repeat_length = 0;
 static uns url_max_occurences = ~0U;
 
+#ifndef TEST
 static struct cf_section url_config = {
   CF_ITEMS {
     CF_UNS("IgnoreSpaces", &url_ignore_spaces),
@@ -51,6 +47,7 @@ static void CONSTRUCTOR url_init_config(void)
 {
   cf_declare_section("URL", &url_config, 0);
 }
+#endif
 
 /* Escaping and de-escaping */
 
@@ -95,6 +92,14 @@ url_deescape(const char *s, char *d)
 	      val = NCC_AND; break;
 	    case '#':
 	      val = NCC_HASH; break;
+#ifndef CONFIG_URL_ESCAPE_COMPAT
+	    case '$':
+	      val = NCC_DOLLAR; break;
+	    case '+':
+	      val = NCC_PLUS; break;
+	    case ',':
+	      val = NCC_COMMA; break;
+#endif
 	    }
 	  *d++ = val;
 	  s += 3;
@@ -133,16 +138,19 @@ url_enescape(const char *s, char *d)
     {
       if (d >= end)
 	return URL_ERR_TOO_LONG;
-      if (Calnum(c) ||							/* RFC 1738(2.2): Only alphanumerics ... */
-	  c == '$' || c == '-' || c == '_' || c == '.' || c == '+' ||	/* ... and several other exceptions ... */
-	  c == '!' || c == '*' || c == '\'' || c == '(' || c == ')' ||
-	  c == ',' ||
-	  c == '/' || c == '?' || c == ':' || c == '@' ||		/* ... and reserved chars used for reserved purpose */
-	  c == '=' || c == '&' || c == '#' || c == ';')
+      if (Calnum(c) ||							/* RFC 2396 (2.1-2.3): Only alphanumerics ... */
+	  c == '!' || c == '*' || c == '\'' || c == '(' || c == ')' ||	/* ... and some exceptions and reserved chars */
+	  c == '$' || c == '-' || c == '_' || c == '.' || c == '+' ||
+	  c == ',' || c == '=' || c == '&' || c == '#' || c == ';' ||
+	  c == '/' || c == '?' || c == ':' || c == '@'
+#ifndef CONFIG_URL_ESCAPE_COMPAT
+	  || c == '~'
+#endif
+	)
 	*d++ = *s++;
       else
 	{
-	  uns val = ((byte)*s < NCC_MAX) ? NCC_CHARS[(byte)*s] : *s;
+	  uns val = (byte)(((byte)*s < NCC_MAX) ? NCC_CHARS[(byte)*s] : *s);
 	  *d++ = '%';
 	  *d++ = enhex(val >> 4);
 	  *d++ = enhex(val & 0x0f);
@@ -162,14 +170,14 @@ url_enescape_friendly(const char *src, char *dest)
     {
       if (dest >= end)
 	return URL_ERR_TOO_LONG;
-      if (*srcb < NCC_MAX)
+      if ((byte)*srcb < NCC_MAX)
 	*dest++ = NCC_CHARS[*srcb++];
       else if (*srcb >= 0x20 && *srcb < 0x7f)
 	*dest++ = *srcb++;
       else
 	{
 	  *dest++ = '%';
-	  *dest++ = enhex(*srcb >> 4);
+	  *dest++ = enhex((byte)*srcb >> 4);
 	  *dest++ = enhex(*srcb++ & 0x0f);
 	}
     }
@@ -183,7 +191,7 @@ char *url_proto_names[URL_PROTO_MAX] = URL_PNAMES;
 static int url_proto_path_flags[URL_PROTO_MAX] = URL_PATH_FLAGS;
 
 uns
-identify_protocol(const char *p)
+url_identify_protocol(const char *p)
 {
   uns i;
 
@@ -211,7 +219,7 @@ url_split(char *s, struct url *u, char *d)
 	  while (s < p)
 	    *d++ = *s++;
 	  *d++ = 0;
-	  u->protoid = identify_protocol(u->protocol);
+	  u->protoid = url_identify_protocol(u->protocol);
 	  s++;
 	  if (url_proto_path_flags[u->protoid] && (s[0] != '/' || s[1] != '/'))
 	    {
@@ -318,15 +326,9 @@ relpath_merge(struct url *u, struct url *b)
 	;
       goto copy;
     }
-  if (a[0] == ';')			/* Change parameters */
-    {
-      for(p=o; *p && *p != ';' && *p != '?' && *p != '#'; p++)
-	;
-      goto copy;
-    }
 
   p = NULL;				/* Copy original path and find the last slash */
-  while (*o && *o != ';' && *o != '?' && *o != '#')
+  while (*o && *o != '?' && *o != '#')
     {
       if (d >= e)
 	return URL_ERR_TOO_LONG;
@@ -517,7 +519,7 @@ url_pack(struct url *u, char *d)
     {
       d = append(d, u->protocol, e);
       d = append(d, ":", e);
-      u->protoid = identify_protocol(u->protocol);
+      u->protoid = url_identify_protocol(u->protocol);
     }
   if (u->host)
     {
@@ -610,7 +612,7 @@ int main(int argc, char **argv)
   char buf1[MAX_URL_SIZE], buf2[MAX_URL_SIZE], buf3[MAX_URL_SIZE], buf4[MAX_URL_SIZE];
   int err;
   struct url url, url0;
-  char *base = "http://mj@www.hell.org/123/sub_dir/index.html;param?query&zzz/subquery#fragment";
+  char *base = "http://mj@www.hell.org/123/sub_dir;param/index.html;param?query&zzz/sub;query+#fragment?";
 
   if (argc != 2 && argc != 3)
     return 1;