add UTF-8 support.

* glib/gpattern.c: add UTF-8 support.

        * tests/patterntest.c: add UTF-8 and equality tests.

        * docs/reference/glib/Makefile.am (MKDB_OPTIONS): Add --sgml-mode.

        * docs/reference/glib/tmpl/patterns.sgml: Document UTF-8 support.

ChangeLog

@@ -1,3 +1,9 @@
+2001-11-14  Matthias Clasen  <matthiasc@poet.de>
+
+	* glib/gpattern.c: add UTF-8 support.
+
+	* tests/patterntest.c: add UTF-8 and equality tests.
+
 Wed Nov 14 07:34:24 2001  Tim Janik  <timj@gtk.org>
 
 	* glib/galloca.h (g_newa): provide g_newa(ctype, count) on top of

ChangeLog.pre-2-0

@@ -1,3 +1,9 @@
+2001-11-14  Matthias Clasen  <matthiasc@poet.de>
+
+	* glib/gpattern.c: add UTF-8 support.
+
+	* tests/patterntest.c: add UTF-8 and equality tests.
+
 Wed Nov 14 07:34:24 2001  Tim Janik  <timj@gtk.org>
 
 	* glib/galloca.h (g_newa): provide g_newa(ctype, count) on top of

ChangeLog.pre-2-10

@@ -1,3 +1,9 @@
+2001-11-14  Matthias Clasen  <matthiasc@poet.de>
+
+	* glib/gpattern.c: add UTF-8 support.
+
+	* tests/patterntest.c: add UTF-8 and equality tests.
+
 Wed Nov 14 07:34:24 2001  Tim Janik  <timj@gtk.org>
 
 	* glib/galloca.h (g_newa): provide g_newa(ctype, count) on top of

ChangeLog.pre-2-12

@@ -1,3 +1,9 @@
+2001-11-14  Matthias Clasen  <matthiasc@poet.de>
+
+	* glib/gpattern.c: add UTF-8 support.
+
+	* tests/patterntest.c: add UTF-8 and equality tests.
+
 Wed Nov 14 07:34:24 2001  Tim Janik  <timj@gtk.org>
 
 	* glib/galloca.h (g_newa): provide g_newa(ctype, count) on top of

ChangeLog.pre-2-2

@@ -1,3 +1,9 @@
+2001-11-14  Matthias Clasen  <matthiasc@poet.de>
+
+	* glib/gpattern.c: add UTF-8 support.
+
+	* tests/patterntest.c: add UTF-8 and equality tests.
+
 Wed Nov 14 07:34:24 2001  Tim Janik  <timj@gtk.org>
 
 	* glib/galloca.h (g_newa): provide g_newa(ctype, count) on top of

ChangeLog.pre-2-4

@@ -1,3 +1,9 @@
+2001-11-14  Matthias Clasen  <matthiasc@poet.de>
+
+	* glib/gpattern.c: add UTF-8 support.
+
+	* tests/patterntest.c: add UTF-8 and equality tests.
+
 Wed Nov 14 07:34:24 2001  Tim Janik  <timj@gtk.org>
 
 	* glib/galloca.h (g_newa): provide g_newa(ctype, count) on top of

ChangeLog.pre-2-6

@@ -1,3 +1,9 @@
+2001-11-14  Matthias Clasen  <matthiasc@poet.de>
+
+	* glib/gpattern.c: add UTF-8 support.
+
+	* tests/patterntest.c: add UTF-8 and equality tests.
+
 Wed Nov 14 07:34:24 2001  Tim Janik  <timj@gtk.org>
 
 	* glib/galloca.h (g_newa): provide g_newa(ctype, count) on top of

ChangeLog.pre-2-8

@@ -1,3 +1,9 @@
+2001-11-14  Matthias Clasen  <matthiasc@poet.de>
+
+	* glib/gpattern.c: add UTF-8 support.
+
+	* tests/patterntest.c: add UTF-8 and equality tests.
+
 Wed Nov 14 07:34:24 2001  Tim Janik  <timj@gtk.org>
 
 	* glib/galloca.h (g_newa): provide g_newa(ctype, count) on top of

docs/reference/ChangeLog

@@ -1,10 +1,16 @@
+2001-11-14  Matthias Clasen  <matthiasc@poet.de>
+
+	* glib/Makefile.am (MKDB_OPTIONS): Add --sgml-mode.
+
+	* glib/tmpl/patterns.sgml: Document UTF-8 support.
+
 Wed Nov 14 03:19:49 2001  Tim Janik  <timj@gtk.org>
 
-	* gobject/tmp/param_value_types.sgml: more docs.
+	* gobject/tmpl/param_value_types.sgml: more docs.
 
 Tue Nov 13 21:31:58 2001  Tim Janik  <timj@gtk.org>
 
-	* gobject/tmp/param_value_types.sgml: list parameter and
+	* gobject/tmpl/param_value_types.sgml: list parameter and
 	value types.
 
 	* gobject/tmpl/gparamspec.sgml: more docs for g_param_spec*()

docs/reference/glib/Makefile.am

@@ -13,7 +13,7 @@ DOC_SOURCE_DIR=../../..
 SCAN_OPTIONS=--deprecated-guards="G_DISABLE_DEPRECATED"
 
 # Extra options to supply to gtkdoc-mkdb
-MKDB_OPTIONS=
+MKDB_OPTIONS=--sgml-mode
 
 # Extra options to supply to gtkdoc-fixref
 FIXXREF_OPTIONS=

docs/reference/glib/tmpl/patterns.sgml

@@ -19,26 +19,16 @@ Note that in contrast to <function>glob()</function>, the '/' character
 be escaped to include them literally in a pattern. 
 </para>
 <para>
-The pattern matcher is restricted to ASCII and will not work correctly with
-multibyte UTF-8 characters in the pattern or in the string to match.
-</para>
-<para>
 When multiple strings must be matched against the same pattern, it
 is better to compile the pattern to a #GPatternSpec using 
 g_pattern_spec_new() and use g_pattern_match_string() instead of 
 g_pattern_match_simple().  This avoids the overhead of repeated 
 pattern compilation.
-</para>
-
-<!-- ##### SECTION See_Also ##### -->
-<para>
-g_strreverse()
-</para>
 
 <!-- ##### STRUCT GPatternSpec ##### -->
 <para>
 A <structname>GPatternSpec</structname> is the 'compiled' form of a pattern.
-This structure is opaque and its fields and cannot be accessed.
+This structure is opaque and its fields cannot be accessed directly.
 </para>
 
 
@@ -47,7 +37,7 @@ This structure is opaque and its fields and cannot be accessed.
 Compiles a pattern to a #GPatternSpec.
 </para>
 
-@pattern: a zero terminated string.
+@pattern: a zero-terminated UTF-8 encoded string.
 @Returns: a newly-allocated #GPatternSpec.
 
 
@@ -77,16 +67,24 @@ string given is mandatory. The reversed string can be omitted by passing %NULL,
 this is more efficient if the reversed version of the string to be matched is
 not at hand, as g_pattern_match() will only construct it if the compiled pattern
 requires reverse matches.
-Note that, if the user code will (possibly) match a string against a multitude of
+</para>
-patterns containing wildcards, chances are high that some patterns will require
+<para>
-a reversed string. In this case, it's more efficient to provide the reversed
+Note that, if the user code will (possibly) match a string against a multitude 
-string to avoid multiple constructions thereof in the various calls to
+of patterns containing wildcards, chances are high that some patterns will 
+require a reversed string. In this case, it's more efficient to provide the 
+reversed string to avoid multiple constructions thereof in the various calls to
 g_pattern_match().
 </para>
+<para>
+Note also that the reverse of a UTF-8 encoded string can in general 
+<emphasis>not</emphasis> be obtained by <function>g_strreverse()</function>.
+This works only if the string doesn't contain any multibyte characters.
+Glib doesn't currently offer a function to reverse UTF-8 encoded strings.
+</para>
 
 @pspec: a #GPatternSpec.
 @string_length: the length of @string.
-@string: the string to match.
+@string: the UTF-8 encoded string to match.
 @string_reversed: the reverse of @string or %NULL.
 @Returns: %TRUE if @string matches @pspec.
 
@@ -99,7 +97,7 @@ g_pattern_match() instead while supplying the reversed string.
 </para>
 
 @pspec: a #GPatternSpec.
-@string: the string to match.
+@string: the UTF-8 encoded string to match.
 @Returns: %TRUE if @string matches @pspec.
 
 
@@ -111,8 +109,8 @@ the pattern once with g_pattern_spec_new() and call g_pattern_match_string()
 repetively.
 </para>
 
-@pattern: the pattern.
+@pattern: the UTF-8 encoded pattern.
-@string: the string to match.
+@string: the UTF-8 encoded string to match.
 @Returns: %TRUE if @string matches @pspec.
 
 

glib/gpattern.c

@@ -21,7 +21,8 @@
 #include "gmacros.h"
 #include "gmessages.h"
 #include "gmem.h"
-#include "gutils.h" /* inline hassle */
+#include "gunicode.h"
+#include "gutils.h" 
 #include <string.h>
 
 /* keep enum and structure of gpattern.c and patterntest.c in sync */
@@ -34,44 +35,51 @@ typedef enum
   G_MATCH_EXACT,     /* "AAAAA" */
   G_MATCH_LAST
 } GMatchType;
+
 struct _GPatternSpec
 {
   GMatchType match_type;
   guint      pattern_length;
+  guint      min_length;
   gchar     *pattern;
 };
 
 
 /* --- functions --- */
-static inline void
+/**
-instring_reverse (guint  length,
+ * g_utf8_reverse:
-		  gchar *str)
+ * string: a UTF-8 string.
+ *
+ * Reverses a UTF-8 string. The @string must be valid UTF-8 encoded text. 
+ * (Use g_utf8_validate() on all text before trying to use UTF-8 
+ * utility functions with it.)
+ *
+ * Note that unlike g_strreverse(), this function returns
+ * newly-allocated memory, which should be freed with g_free() when
+ * no longer needed. 
+ *
+ * Returns: a newly-allocated string which is the reverse of @string.
+ */
+static gchar *
+g_utf8_reverse (guint len, const gchar *string)
 {
-  gchar *f, *l, *b;
+  gchar *result;
+  const gchar *p;
+  gchar *m, *r, skip;
 
-  f = str;
+  result = g_new (gchar, len + 1);
-  l = str + length - 1;
+  r = result + len;
-  b = str + length / 2;
+  p = string;
-  while (f < b)
+  while (*p) 
     {
-      gchar tmp = *l;
+      skip = g_utf8_skip[*(guchar*)p];
-
+      r -= skip;
-      *l-- = *f;
+      for (m = r; skip; skip--)
-      *f++ = tmp;
+        *m++ = *p++;
     }
-}
+  result[len] = 0;
 
-static inline gchar*
+  return result;
-strdup_reverse (guint        length,
-		const gchar *str)
-{
-  gchar *t, *dest = g_new (gchar, length + 1);
-
-  t = dest + length;
-  *t-- = 0;
-  while (t >= dest)
-    *t-- = *str++;
-  return dest;
 }
 
 static inline gboolean
@@ -93,7 +101,7 @@ g_pattern_ph_match (const gchar *match_pattern,
 	case '?':
 	  if (!*string)
 	    return FALSE;
-	  string++;
+	  string = g_utf8_next_char (string);
 	  break;
 
 	case '*':
@@ -105,7 +113,7 @@ g_pattern_ph_match (const gchar *match_pattern,
 		{
 		  if (!*string)
 		    return FALSE;
-		  string++;
+		  string = g_utf8_next_char (string);
 		}
 	    }
 	  while (ch == '*' || ch == '?');
@@ -117,7 +125,7 @@ g_pattern_ph_match (const gchar *match_pattern,
 		{
 		  if (!*string)
 		    return FALSE;
-		  string++;
+		  string = g_utf8_next_char (string);
 		}
 	      string++;
 	      if (g_pattern_ph_match (pattern, string))
@@ -150,10 +158,11 @@ g_pattern_match (GPatternSpec *pspec,
   g_return_val_if_fail (pspec != NULL, FALSE);
   g_return_val_if_fail (string != NULL, FALSE);
 
+  if (pspec->min_length > string_length)
+    return FALSE;
+
   switch (pspec->match_type)
     {
-      gboolean result;
-      gchar *tmp;
     case G_MATCH_ALL:
       return g_pattern_ph_match (pspec->pattern, string);
     case G_MATCH_ALL_TAIL:
@@ -161,47 +170,23 @@ g_pattern_match (GPatternSpec *pspec,
 	return g_pattern_ph_match (pspec->pattern, string_reversed);
       else
 	{
-	  tmp = strdup_reverse (string_length, string);
+          gboolean result;
+          gchar *tmp;
+	  tmp = g_utf8_reverse (string_length, string);
 	  result = g_pattern_ph_match (pspec->pattern, tmp);
 	  g_free (tmp);
 	  return result;
 	}
     case G_MATCH_HEAD:
-      if (pspec->pattern_length > string_length)
+      if (pspec->pattern_length == string_length)
-	return FALSE;
-      else if (pspec->pattern_length == string_length)
 	return strcmp (pspec->pattern, string) == 0;
       else if (pspec->pattern_length)
 	return strncmp (pspec->pattern, string, pspec->pattern_length) == 0;
       else
 	return TRUE;
     case G_MATCH_TAIL:
-      if (pspec->pattern_length > string_length)
+      if (pspec->pattern_length)
-	return FALSE;
+        return strcmp (pspec->pattern, string + (string_length - pspec->pattern_length)) == 0;
-      else if (pspec->pattern_length == string_length)
-	{
-	  if (string_reversed)
-	    return strcmp (pspec->pattern, string_reversed) == 0;
-	  else
-	    {
-	      tmp = strdup_reverse (string_length, string);
-	      result = strcmp (pspec->pattern, tmp) == 0;
-	      g_free (tmp);
-	      return result;
-	    }
-	}
-      else if (pspec->pattern_length)
-	{
-	  if (string_reversed)
-	    return strncmp (pspec->pattern, string_reversed, pspec->pattern_length) == 0;
-	  else
-	    {
-	      tmp = strdup_reverse (string_length, string);
-	      result = strncmp (pspec->pattern, tmp, pspec->pattern_length) == 0;
-	      g_free (tmp);
-	      return result;
-	    }
-	}
       else
 	return TRUE;
     case G_MATCH_EXACT:
@@ -222,6 +207,7 @@ g_pattern_spec_new (const gchar *pattern)
   gboolean seen_joker = FALSE, seen_wildcard = FALSE, more_wildcards = FALSE;
   gint hw_pos = -1, tw_pos = -1, hj_pos = -1, tj_pos = -1;
   gboolean follows_wildcard = FALSE;
+  guint pending_jokers = 0;
   const gchar *s;
   gchar *d;
   guint i;
@@ -231,6 +217,7 @@ g_pattern_spec_new (const gchar *pattern)
   /* canonicalize pattern and collect necessary stats */
   pspec = g_new (GPatternSpec, 1);
   pspec->pattern_length = strlen (pattern);
+  pspec->min_length = 0;
   pspec->pattern = g_new (gchar, pspec->pattern_length + 1);
   d = pspec->pattern;
   for (i = 0, s = pattern; *s != 0; s++)
@@ -249,17 +236,29 @@ g_pattern_spec_new (const gchar *pattern)
 	  tw_pos = i;
 	  break;
 	case '?':
-	  if (hj_pos < 0)
+	  pending_jokers++;
-	    hj_pos = i;
+	  pspec->min_length++;
-	  tj_pos = i;
+	  continue;
-	  /* fall through */
 	default:
+	  for (; pending_jokers; pending_jokers--, i++) {
+	    *d++ = '?';
+  	    if (hj_pos < 0)
+	     hj_pos = i;
+	    tj_pos = i;
+	  }
 	  follows_wildcard = FALSE;
+	  pspec->min_length++;
 	  break;
 	}
       *d++ = *s;
       i++;
     }
+  for (; pending_jokers; pending_jokers--) {
+    *d++ = '?';
+    if (hj_pos < 0)
+      hj_pos = i;
+    tj_pos = i;
+  }
   *d++ = 0;
   seen_joker = hj_pos >= 0;
   seen_wildcard = hw_pos >= 0;
@@ -271,8 +270,8 @@ g_pattern_spec_new (const gchar *pattern)
       if (pspec->pattern[0] == '*')
 	{
 	  pspec->match_type = G_MATCH_TAIL;
-	  instring_reverse (pspec->pattern_length, pspec->pattern);
+          memmove (pspec->pattern, pspec->pattern + 1, --pspec->pattern_length);
-	  pspec->pattern[--pspec->pattern_length] = 0;
+	  pspec->pattern[pspec->pattern_length] = 0;
 	  return pspec;
 	}
       if (pspec->pattern[pspec->pattern_length - 1] == '*')
@@ -295,8 +294,11 @@ g_pattern_spec_new (const gchar *pattern)
     pspec->match_type = tw_pos > hw_pos ? G_MATCH_ALL_TAIL : G_MATCH_ALL;
   else /* seen_joker */
     pspec->match_type = tj_pos > hj_pos ? G_MATCH_ALL_TAIL : G_MATCH_ALL;
-  if (pspec->match_type == G_MATCH_ALL_TAIL)
+  if (pspec->match_type == G_MATCH_ALL_TAIL) {
-    instring_reverse (pspec->pattern_length, pspec->pattern);
+    gchar *tmp = pspec->pattern;
+    pspec->pattern = g_utf8_reverse (pspec->pattern_length, pspec->pattern);
+    g_free (tmp);
+  }
   return pspec;
 }
 

tests/patterntest.c

@@ -16,10 +16,9 @@
  * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
  * Boston, MA 02111-1307, USA.
  */
-
-#include <string.h>
-
 #include "glib.h"
+#include "glib/gpattern.h"
+
 
 /* keep enum and structure of gpattern.c and patterntest.c in sync */
 typedef enum
@@ -31,10 +30,12 @@ typedef enum
   G_MATCH_EXACT,     /* "AAAAA" */
   G_MATCH_LAST
 } GMatchType;
+
 struct _GPatternSpec
 {
   GMatchType match_type;
   guint      pattern_length;
+  guint      min_length;
   gchar     *pattern;
 };
 
@@ -65,16 +66,22 @@ match_type_name (GMatchType match_type)
     }
 }
 
+/* this leakes memory, but we don't care */
+
+#define utf8(str) g_convert (str, -1, "Latin1", "UTF-8", NULL, NULL, NULL)
+#define latin1(str) g_convert (str, -1, "UTF-8", "Latin1", NULL, NULL, NULL)
+
 static gboolean
 test_compilation (gchar *src, 
 		  GMatchType match_type, 
-		  gchar *pattern)
+		  gchar *pattern,
+		  guint min)
 {
   GPatternSpec *spec; 
 
-  g_print ("compiling \"%s\" \t", src);
+  g_print ("compiling \"%s\" \t", utf8(src));
   spec = g_pattern_spec_new (src);
-  
+
   if (spec->match_type != match_type)
     {
       g_print ("failed \t(match_type: %s, expected %s)\n",
@@ -86,8 +93,8 @@ test_compilation (gchar *src,
   if (strcmp (spec->pattern, pattern) != 0)
     {
       g_print ("failed \t(pattern: \"%s\", expected \"%s\")\n",
-	       spec->pattern,
+	       utf8(spec->pattern),
-	       pattern);
+	       utf8(pattern));
       return FALSE;
     }
   
@@ -99,6 +106,14 @@ test_compilation (gchar *src,
       return FALSE;
     }
   
+  if (spec->min_length != min)
+    {
+      g_print ("failed \t(min_length: %d, expected %d)\n",
+	       spec->min_length,
+	       min);
+      return FALSE;
+    }
+  
   g_print ("passed (%s: \"%s\")\n",
 	   match_type_name (spec->match_type),
 	   spec->pattern);
@@ -111,7 +126,7 @@ test_match (gchar *pattern,
 	    gchar *string, 
 	    gboolean match)
 {
-  g_print ("matching \"%s\" against \"%s\" \t", string, pattern);
+  g_print ("matching \"%s\" against \"%s\" \t", utf8(string), utf8(pattern));
   
   if (g_pattern_match_simple (pattern, string) != match)
     {
@@ -132,14 +147,14 @@ test_equal (gchar *pattern1,
   GPatternSpec *p2 = g_pattern_spec_new (pattern2);
   gboolean equal = g_pattern_spec_equal (p1, p2);
 
-  g_print ("comparing \"%s\" with \"%s\" \t", pattern1, pattern2);
+  g_print ("comparing \"%s\" with \"%s\" \t", utf8(pattern1), utf8(pattern2));
 
   if (expected != equal)
     {
       g_print ("failed \t{%s, %u, \"%s\"} %s {%s, %u, \"%s\"}\n",
-	       match_type_name (p1->match_type), p1->pattern_length, p1->pattern,
+	       match_type_name (p1->match_type), p1->pattern_length, utf8(p1->pattern),
 	       expected ? "!=" : "==",
-	       match_type_name (p2->match_type), p2->pattern_length, p2->pattern);
+	       match_type_name (p2->match_type), p2->pattern_length, utf8(p2->pattern));
     }
   else
     g_print ("passed (%s)\n", equal ? "equal" : "unequal");
@@ -150,9 +165,9 @@ test_equal (gchar *pattern1,
   return expected == equal;
 }
 
-#define TEST_COMPILATION(src, type, pattern) { \
+#define TEST_COMPILATION(src, type, pattern, min) { \
   total++; \
-  if (test_compilation (src, type, pattern)) \
+  if (test_compilation (latin1(src), type, latin1(pattern), min)) \
     passed++; \
   else \
     failed++; \
@@ -160,7 +175,7 @@ test_equal (gchar *pattern1,
 
 #define TEST_MATCH(pattern, string, match) { \
   total++; \
-  if (test_match (pattern, string, match)) \
+  if (test_match (latin1(pattern), latin1(string), match)) \
     passed++; \
   else \
     failed++; \
@@ -168,7 +183,7 @@ test_equal (gchar *pattern1,
 
 #define TEST_EQUAL(pattern1, pattern2, match) { \
   total++; \
-  if (test_equal (pattern1, pattern2, match)) \
+  if (test_equal (latin1(pattern1), latin1(pattern2), match)) \
     passed++; \
   else \
     failed++; \
@@ -180,16 +195,21 @@ main (int argc, char** argv)
   gint total = 0;
   gint passed = 0;
   gint failed = 0;
-  gchar *string;
+  gchar *string, *pattern;
 
-  TEST_COMPILATION("*A?B*", G_MATCH_ALL, "*A?B*");
+  TEST_COMPILATION("*A?B*", G_MATCH_ALL, "*A?B*", 3);
-  TEST_COMPILATION("ABC*DEFGH", G_MATCH_ALL_TAIL, "HGFED*CBA");
+  TEST_COMPILATION("ABC*DEFGH", G_MATCH_ALL_TAIL, "HGFED*CBA", 8);
-  TEST_COMPILATION("ABCDEF*GH", G_MATCH_ALL, "ABCDEF*GH");
+  TEST_COMPILATION("ABCDEF*GH", G_MATCH_ALL, "ABCDEF*GH", 8);
-  TEST_COMPILATION("ABC**?***??**DEF*GH", G_MATCH_ALL, "ABC*?*??*DEF*GH");
+  TEST_COMPILATION("ABC**?***??**DEF*GH", G_MATCH_ALL, "ABC*???DEF*GH", 11);
-  TEST_COMPILATION("*A?AA", G_MATCH_ALL_TAIL, "AA?A*");
+  TEST_COMPILATION("*A?AA", G_MATCH_ALL_TAIL, "AA?A*", 4);
-  TEST_COMPILATION("ABCD*", G_MATCH_HEAD, "ABCD");
+  TEST_COMPILATION("ABCD*", G_MATCH_HEAD, "ABCD", 4);
-  TEST_COMPILATION("*ABCD", G_MATCH_TAIL, "DCBA");
+  TEST_COMPILATION("*ABCD", G_MATCH_TAIL, "ABCD", 4);
-  TEST_COMPILATION("ABCDE", G_MATCH_EXACT, "ABCDE");
+  TEST_COMPILATION("ABCDE", G_MATCH_EXACT, "ABCDE", 5);
+  TEST_COMPILATION("A?C?E", G_MATCH_ALL, "A?C?E", 5);
+  TEST_COMPILATION("*?x", G_MATCH_ALL_TAIL, "x?*", 2);
+  TEST_COMPILATION("?*x", G_MATCH_ALL_TAIL, "x?*", 2);
+  TEST_COMPILATION("*?*x", G_MATCH_ALL_TAIL, "x?*", 2);
+  TEST_COMPILATION("x*??", G_MATCH_ALL_TAIL, "??*x", 3);
 
   TEST_EQUAL ("*A?B*", "*A?B*", TRUE);
   TEST_EQUAL ("A*BCD", "A*BCD", TRUE);
@@ -198,6 +218,8 @@ main (int argc, char** argv)
   TEST_EQUAL ("*YZ", "*YZ", TRUE);
   TEST_EQUAL ("A1x", "A1x", TRUE);
   TEST_EQUAL ("AB*CD", "AB**CD", TRUE);
+  TEST_EQUAL ("AB*?*CD", "AB*?CD", TRUE);
+  TEST_EQUAL ("AB*?CD", "AB?*CD", TRUE);
   TEST_EQUAL ("AB*CD", "AB*?*CD", FALSE);
   TEST_EQUAL ("ABC*", "ABC?", FALSE);
 
@@ -215,11 +237,20 @@ main (int argc, char** argv)
   TEST_MATCH("?*x", "x", FALSE);
   TEST_MATCH("*?*x", "yx", TRUE);
   TEST_MATCH("*?*x", "xxxx", TRUE);
-
+  TEST_MATCH("x*??", "xyzw", TRUE);
-  string = g_convert ("<EFBFBD>x", -1, "UTF-8", "Latin1", NULL, NULL, NULL);
+  TEST_MATCH("*x", "<EFBFBD>x", TRUE);
-  TEST_MATCH("*x", string, TRUE);
+  TEST_MATCH("?x", "<EFBFBD>x", TRUE);
-  TEST_MATCH("?x", string, TRUE);
+  TEST_MATCH("??x", "<EFBFBD>x", FALSE);
-  TEST_MATCH("??x", string, FALSE);
+  TEST_MATCH("ab<EFBFBD><EFBFBD>", "ab<EFBFBD><EFBFBD>", TRUE);
+  TEST_MATCH("ab<EFBFBD><EFBFBD>", "abao", FALSE);
+  TEST_MATCH("ab?<3F>", "ab<EFBFBD><EFBFBD>", TRUE);
+  TEST_MATCH("ab?<3F>", "abao", FALSE);
+  TEST_MATCH("ab<EFBFBD>?", "ab<EFBFBD><EFBFBD>", TRUE);
+  TEST_MATCH("ab<EFBFBD>?", "abao", FALSE);
+  TEST_MATCH("ab??", "ab<EFBFBD><EFBFBD>", TRUE);
+  TEST_MATCH("ab*", "ab<EFBFBD><EFBFBD>", TRUE);
+  TEST_MATCH("ab*<2A>", "ab<EFBFBD><EFBFBD>", TRUE);
+  TEST_MATCH("ab*<2A>", "aba<EFBFBD>x<EFBFBD>", TRUE);
   
   g_print ("\n%u tests passed, %u failed\n", passed, failed);