`buckettool -c' (cat) now separates buckets by an empty line.

[libucw.git] / lib / url.c

diff --git a/lib/url.c b/lib/url.c
index 73e3b50134b1d3e6a75c7a0385a19be8c99944b6..8475a85f378498835509117b51c5b35362f208eb 100644 (file)
--- a/lib/url.c
+++ b/lib/url.c
@@ -1,8 +1,18 @@
 /*
- *     Sherlock Library -- URL Functions (according to RFC 1738 and 1808)
+ *     Sherlock Library -- URL Functions
  *
- *     (c) 1997--2001 Martin Mares <mj@ucw.cz>
+ *     (c) 1997--2004 Martin Mares <mj@ucw.cz>
  *     (c) 2001 Robert Spalek <robert@ucw.cz>
+ *
+ *     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.
  */
 
 #include "lib/lib.h"
@@ -13,6 +23,7 @@
 #include <string.h>
 #include <stdlib.h>
 #include <stdio.h>
+#include <alloca.h>
 
 /* Configuration */
 
@@ -23,13 +34,13 @@ static uns url_min_repeat_count = 0x7fffffff;
 static uns url_max_repeat_length = 0;
 
 static struct cfitem url_config[] = {
-  { "URL",             CT_SECTION,     NULL },
-  { "IgnoreSpaces",    CT_INT,         &url_ignore_spaces },
-  { "IgnoreUnderflow", CT_INT,         &url_ignore_underflow },
-  { "ComponentSeparators",     CT_STRING,      &url_component_separators },
-  { "MinRepeatCount",          CT_INT,         &url_min_repeat_count },
-  { "MaxRepeatLength",         CT_INT,         &url_max_repeat_length },
-  { NULL,              CT_STOP,        NULL }
+  { "URL",                             CT_SECTION,     NULL },
+  { "IgnoreSpaces",                    CT_INT,         &url_ignore_spaces },
+  { "IgnoreUnderflow",                 CT_INT,         &url_ignore_underflow },
+  { "ComponentSeparators",             CT_STRING,      &url_component_separators },
+  { "MinRepeatCount",                  CT_INT,         &url_min_repeat_count },
+  { "MaxRepeatLength",                 CT_INT,         &url_max_repeat_length },
+  { NULL,                              CT_STOP,        NULL }
 };
 
 static void CONSTRUCTOR url_init_config(void)
@@ -195,7 +206,7 @@ url_split(byte *s, struct url *u, byte *d)
 
          s += 2;
          q = d;
-         while (*s && *s != '/')       /* Copy user:passwd@host:port */
+         while (*s && *s != '/' && *s != '?')  /* Copy user:passwd@host:port */
            *d++ = *s++;
          *d++ = 0;
          w = strchr(q, '@');
@@ -203,6 +214,11 @@ url_split(byte *s, struct url *u, byte *d)
            {
              *w++ = 0;
              u->user = q;
+             if (e = strchr(q, ':'))
+               {
+                 *e++ = 0;
+                 u->pass = e;
+               }
            }
          else
            w = q;
@@ -241,7 +257,7 @@ relpath_merge(struct url *u, struct url *b)
 
   if (a[0] == '/')                     /* Absolute path => OK */
     return 0;
-  if (o[0] != '/')
+  if (o[0] != '/' && o[0] != '?')
     return URL_PATH_UNDERFLOW;
 
   if (!a[0])                           /* Empty URL -> inherit everything */
@@ -352,9 +368,9 @@ url_normalize(struct url *u, struct url *b)
   int err;
 
   /* Basic checks */
-  if (url_proto_path_flags[u->protoid] && !u->host ||
-      u->host && !*u->host ||
+  if (url_proto_path_flags[u->protoid] && (!u->host || !*u->host) ||
       !u->host && u->user ||
+      !u->user && u->pass ||
       !u->rest)
     return URL_SYNTAX_ERROR;
 
@@ -371,12 +387,25 @@ url_normalize(struct url *u, struct url *b)
        {
          u->host = b->host;
          u->user = b->user;
+         u->pass = b->pass;
          u->port = b->port;
          if (err = relpath_merge(u, b))
            return err;
        }
     }
 
+  /* Change path "?" to "/?" because it's the true meaning */
+  if (u->rest[0] == '?')
+    {
+      int l = strlen(u->rest);
+      if (u->bufend - u->buf < l+1)
+       return URL_ERR_TOO_LONG;
+      u->buf[0] = '/';
+      memcpy(u->buf+1, u->rest, l+1);
+      u->rest = u->buf;
+      u->buf += l+2;
+    }
+
   /* Fill in missing info */
   if (u->port == ~0U)
     u->port = std_ports[u->protoid];
@@ -406,8 +435,8 @@ kill_end_dot(byte *b)
   if (b)
     {
       k = b + strlen(b) - 1;
-      if (k > b && *k == '.')
-       *k = 0;
+      while (k > b && *k == '.')
+       *k-- = 0;
     }
 }
 
@@ -458,6 +487,11 @@ url_pack(struct url *u, byte *d)
       if (u->user)
        {
          d = append(d, u->user, e);
+         if (u->pass)
+           {
+             d = append(d, ":", e);
+             d = append(d, u->pass, e);
+           }
          d = append(d, "@", e);
        }
       d = append(d, u->host, e);
@@ -500,7 +534,7 @@ url_error(uns err)
   return errmsg[err];
 }
 
-/* A "macro" for canonical split */
+/* Standard cookbook recipes */
 
 int
 url_canon_split(byte *u, byte *buf1, byte *buf2, struct url *url)
@@ -516,6 +550,19 @@ url_canon_split(byte *u, byte *buf1, byte *buf2, struct url *url)
   return url_canonicalize(url);
 }
 
+int
+url_auto_canonicalize(byte *src, byte *dst)
+{
+  byte buf1[MAX_URL_SIZE], buf2[MAX_URL_SIZE], buf3[MAX_URL_SIZE];
+  int err;
+  struct url ur;
+
+  (void)((err = url_canon_split(src, buf1, buf2, &ur)) ||
+   (err = url_pack(&ur, buf3)) ||
+   (err = url_enescape(buf3, dst)));
+  return err;
+}
+
 /* Testing */
 
 #ifdef TEST
@@ -525,9 +572,12 @@ 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";
 
-  if (argc != 2)
+  if (argc != 2 && argc != 3)
     return 1;
+  if (argc == 3)
+    base = argv[2];
   if (err = url_deescape(argv[1], buf1))
     {
       printf("deesc: error %d\n", err);
@@ -539,8 +589,8 @@ int main(int argc, char **argv)
       printf("split: error %d\n", err);
       return 1;
     }
-  printf("split: @%s@%s@%s@%d@%s\n", url.protocol, url.user, url.host, url.port, url.rest);
-  if (err = url_split("http://mj@www.hell.org/123/sub_dir/index.html;param?query&zzz/subquery#fragment", &url0, buf3))
+  printf("split: @%s@%s@%s@%s@%d@%s\n", url.protocol, url.user, url.pass, url.host, url.port, url.rest);
+  if (err = url_split(base, &url0, buf3))
     {
       printf("split base: error %d\n", err);
       return 1;
@@ -550,19 +600,19 @@ int main(int argc, char **argv)
       printf("normalize base: error %d\n", err);
       return 1;
     }
-  printf("base: @%s@%s@%s@%d@%s\n", url0.protocol, url0.user, url0.host, url0.port, url0.rest);
+  printf("base: @%s@%s@%s@%s@%d@%s\n", url0.protocol, url0.user, url0.pass, url0.host, url0.port, url0.rest);
   if (err = url_normalize(&url, &url0))
     {
       printf("normalize: error %d\n", err);
       return 1;
     }
-  printf("normalize: @%s@%s@%s@%d@%s\n", url.protocol, url.user, url.host, url.port, url.rest);
+  printf("normalize: @%s@%s@%s@%s@%d@%s\n", url.protocol, url.user, url.pass, url.host, url.port, url.rest);
   if (err = url_canonicalize(&url))
     {
       printf("canonicalize: error %d\n", err);
       return 1;
     }
-  printf("canonicalize: @%s@%s@%s@%d@%s\n", url.protocol, url.user, url.host, url.port, url.rest);
+  printf("canonicalize: @%s@%s@%s@%s@%d@%s\n", url.protocol, url.user, url.pass, url.host, url.port, url.rest);
   if (err = url_pack(&url, buf4))
     {
       printf("pack: error %d\n", err);