[glib/glib-2-22] Don't give up too early when collecting mime types

[Matthias Clasen <matthiasc src gnome org>]

commit 36df8c9b9c36973794cd32f5076f15eadd142db0
Author: Matthias Clasen <mclasen redhat com>
Date:   Sun Oct 25 02:47:23 2009 -0400

    Don't give up too early when collecting mime types
    
    Since returning exactly one match has special significance, don't
    give up matching before we've found at least 2 types. Also, make
    sure that we don't return the same mime type more than once.
    Bug 541236.

 gio/xdgmime/xdgmimecache.c |   67 ++++++++++++++++++++++++++++++++------------
 gio/xdgmime/xdgmimeglob.c  |   43 +++++++++++++++++++++++++---
 2 files changed, 87 insertions(+), 23 deletions(-)
---
diff --git a/gio/xdgmime/xdgmimecache.c b/gio/xdgmime/xdgmimecache.c
index 97fe975..43cda89 100644
--- a/gio/xdgmime/xdgmimecache.c
+++ b/gio/xdgmime/xdgmimecache.c
@@ -444,11 +444,11 @@ cache_glob_lookup_fnmatch (const char *file_name,
 	    }
 	}
 
-      if (n > 0)
-	return n;
+      if (n == n_mime_types)
+	break;
     }
-  
-  return 0;
+
+  return n;
 }
 
 static int
@@ -539,6 +539,7 @@ cache_glob_lookup_suffix (const char *file_name,
 {
   int i, n;
 
+  n = 0;
   for (i = 0; _caches[i]; i++)
     {
       XdgMimeCache *cache = _caches[i];
@@ -547,17 +548,17 @@ cache_glob_lookup_suffix (const char *file_name,
       xdg_uint32_t n_entries = GET_UINT32 (cache->buffer, list_offset);
       xdg_uint32_t offset = GET_UINT32 (cache->buffer, list_offset + 4);
 
-      n = cache_glob_node_lookup_suffix (cache, 
-					 n_entries, offset, 
-					 file_name, len,
-					 ignore_case,
-					 mime_types,
-					 n_mime_types);
-      if (n > 0)
-	return n;
+      n += cache_glob_node_lookup_suffix (cache, 
+	 				  n_entries, offset, 
+					  file_name, len,
+					  ignore_case,
+					  mime_types + n,
+					  n_mime_types - n);
+      if (n == n_mime_types)
+	break;
     }
 
-  return 0;
+  return n;
 }
 
 static int compare_mime_weight (const void *a, const void *b)
@@ -585,7 +586,35 @@ ascii_tolower (const char *str)
 }
 
 static int
-cache_glob_lookup_file_name (const char *file_name, 
+filter_out_dupes (MimeWeight mimes[], int n_mimes)
+{
+  int last;
+  int i, j;
+
+  last = n_mimes;
+
+  for (i = 0; i < last; i++)
+    {
+      j = i + 1;
+      while (j < last)
+        {
+          if (strcmp (mimes[i].mime, mimes[j].mime) == 0)
+            {
+              mimes[i].weight = MAX (mimes[i].weight, mimes[j].weight);
+              last--;
+              mimes[j].mime = mimes[last].mime;
+              mimes[j].weight = mimes[last].weight;
+            }
+          else
+            j++;
+        }
+    }
+
+  return last;
+}
+
+static int
+cache_glob_lookup_file_name (const char *file_name,
 			     const char *mime_types[],
 			     int         n_mime_types)
 {
@@ -619,14 +648,16 @@ cache_glob_lookup_file_name (const char *file_name,
 
   len = strlen (file_name);
   n = cache_glob_lookup_suffix (lower_case, len, FALSE, mimes, n_mimes);
-  if (n == 0)
-    n = cache_glob_lookup_suffix (file_name, len, TRUE, mimes, n_mimes);
+  if (n < 2)
+    n += cache_glob_lookup_suffix (file_name, len, TRUE, mimes + n, n_mimes - n);
 
   free (lower_case);
 
   /* Last, try fnmatch */
-  if (n == 0)
-    n = cache_glob_lookup_fnmatch (file_name, mimes, n_mimes);
+  if (n < 2)
+    n += cache_glob_lookup_fnmatch (file_name, mimes + n, n_mimes - n);
+
+  n = filter_out_dupes (mimes, n);
 
   qsort (mimes, n, sizeof (MimeWeight), compare_mime_weight);
 
diff --git a/gio/xdgmime/xdgmimeglob.c b/gio/xdgmime/xdgmimeglob.c
index c82bf22..daea606 100644
--- a/gio/xdgmime/xdgmimeglob.c
+++ b/gio/xdgmime/xdgmimeglob.c
@@ -37,6 +37,10 @@
 #include <string.h>
 #include <fnmatch.h>
 
+#ifndef MAX
+#define MAX(a,b) ((a) > (b) ? (a) : (b))
+#endif
+
 #ifndef	FALSE
 #define	FALSE	(0)
 #endif
@@ -400,6 +404,34 @@ ascii_tolower (const char *str)
   return lower;
 }
 
+static int
+filter_out_dupes (MimeWeight mimes[], int n_mimes)
+{
+  int last;
+  int i, j;
+
+  last = n_mimes;
+
+  for (i = 0; i < last; i++)
+    {
+      j = i + 1;
+      while (j < last)
+        {
+          if (strcmp (mimes[i].mime, mimes[j].mime) == 0)
+            {
+              mimes[i].weight = MAX (mimes[i].weight, mimes[j].weight);
+              last--;
+              mimes[j].mime = mimes[last].mime;
+              mimes[j].weight = mimes[last].weight;
+            }
+          else
+            j++;
+        }
+    }
+
+  return last;
+}
+
 int
 _xdg_glob_hash_lookup_file_name (XdgGlobHash *glob_hash,
 				 const char  *file_name,
@@ -412,7 +444,6 @@ _xdg_glob_hash_lookup_file_name (XdgGlobHash *glob_hash,
   int n_mimes = 10;
   int len;
   char *lower_case;
-  int try_lower_case;
 
   /* First, check the literals */
 
@@ -447,11 +478,11 @@ _xdg_glob_hash_lookup_file_name (XdgGlobHash *glob_hash,
   len = strlen (file_name);
   n = _xdg_glob_hash_node_lookup_file_name (glob_hash->simple_node, lower_case, len, FALSE,
 					    mimes, n_mimes);
-  if (n == 0)
-    n = _xdg_glob_hash_node_lookup_file_name (glob_hash->simple_node, file_name, len, TRUE,
-					      mimes, n_mimes);
+  if (n < 2)
+    n += _xdg_glob_hash_node_lookup_file_name (glob_hash->simple_node, file_name, len, TRUE,
+					       mimes + n, n_mimes - n);
 
-  if (n == 0)
+  if (n < 2)
     {
       for (list = glob_hash->full_list; list && n < n_mime_types; list = list->next)
         {
@@ -465,6 +496,8 @@ _xdg_glob_hash_lookup_file_name (XdgGlobHash *glob_hash,
     }
   free (lower_case);
 
+  n = filter_out_dupes (mimes, n);
+
   qsort (mimes, n, sizeof (MimeWeight), compare_mime_weight);
 
   if (n_mime_types < n)