[svn] Add FLAGS argument to map_html_tags.

[wget] / src / html-url.c

diff --git a/src/html-url.c b/src/html-url.c
index 756bf2abb128180e539bade30815027430cf9c9f..09962eddc9286a0f6a6569c697660caf9f2e96a8 100644 (file)
--- a/src/html-url.c
+++ b/src/html-url.c
@@ -270,30 +270,36 @@ init_interesting (void)
   }
 }
 
+/* Find tag with name TAG_NAME in KNOWN_TAGS and return its index.  */
+
 static int
 find_tag (const char *tag_name)
 {
-  int i;
+  /* Originally implemented as linear search.  In Wget 1.9 known_tags
+     contains 21 elements, for which binary search requires max. 5
+     comparisons, whereas linear search performs 10 on average.  */
 
-  /* This is linear search; if the number of tags grow, we can switch
-     to binary search.  */
+  int lo = 0, hi = countof (known_tags) - 1;
 
-  for (i = 0; i < countof (known_tags); i++)
+  while (lo <= hi)
     {
-      int cmp = strcasecmp (known_tags[i].name, tag_name);
-      /* known_tags are sorted alphabetically, so we can
-         micro-optimize.  */
-      if (cmp > 0)
-       break;
-      else if (cmp == 0)
-       return i;
+      int mid = (lo + hi) >> 1;
+      int cmp = strcasecmp (tag_name, known_tags[mid].name);
+      if (cmp < 0)
+       hi = mid - 1;
+      else if (cmp > 0)
+       lo = mid + 1;
+      else
+       return mid;
     }
+
   return -1;
 }
 
 /* Find the value of attribute named NAME in the taginfo TAG.  If the
    attribute is not present, return NULL.  If ATTRIND is non-NULL, the
    index of the attribute in TAG will be stored there.  */
+
 static char *
 find_attr (struct taginfo *tag, const char *name, int *attrind)
 {
@@ -637,6 +643,7 @@ get_urls_html (const char *file, const char *url, int *meta_disallow_follow)
 {
   struct file_memory *fm;
   struct map_context ctx;
+  int flags;
 
   /* Load the file. */
   fm = read_file (file);
@@ -657,8 +664,16 @@ get_urls_html (const char *file, const char *url, int *meta_disallow_follow)
   if (!interesting_tags)
     init_interesting ();
 
-  map_html_tags (fm->content, fm->length, interesting_tags,
-                interesting_attributes, collect_tags_mapper, &ctx);
+  /* Specify MHT_TRIM_VALUES because of buggy HTML generators that
+     generate <a href=" foo"> instead of <a href="foo"> (Netscape
+     ignores spaces as well.)  If you really mean space, use &32; or
+     %20.  */
+  flags = MHT_TRIM_VALUES;
+  if (opt.strict_comments)
+    flags |= MHT_STRICT_COMMENTS;
+
+  map_html_tags (fm->content, fm->length, collect_tags_mapper, &ctx, flags,
+                interesting_tags, interesting_attributes);
 
   DEBUGP (("no-follow in %s: %d\n", file, ctx.nofollow));
   if (meta_disallow_follow)