[Gnucash-changes] r12189 - gnucash/trunk - utf8_caseequal_len: Fix length-delimited utf8 comparison. Fixes

[Christian Stimming]

Author: cstim
Date: 2005-12-27 16:40:22 -0500 (Tue, 27 Dec 2005)
New Revision: 12189
Trac: http://svn.gnucash.org/trac/changeset/12189

Modified:
   gnucash/trunk/ChangeLog
   gnucash/trunk/src/register/register-core/quickfillcell.c
Log:
utf8_caseequal_len: Fix length-delimited utf8 comparison. Fixes
occasional crashes in register when editing quickfill entries.



Modified: gnucash/trunk/ChangeLog
===================================================================
--- gnucash/trunk/ChangeLog	2005-12-27 19:54:46 UTC (rev 12188)
+++ gnucash/trunk/ChangeLog	2005-12-27 21:40:22 UTC (rev 12189)
@@ -1,3 +1,9 @@
+2005-12-27  Christian Stimming  <stimming at tuhh.de>
+
+	* src/register/register-core/quickfillcell.c (utf8_caseequal_len):
+	Fix length-delimited utf8 comparison. Fixes occasional crashes in
+	register when editing quickfill entries.
+
 2005-12-27  David Hampton  <hampton at employees.org>
 
 	* src/gnome-utils/gnc-recurrence.c:

Modified: gnucash/trunk/src/register/register-core/quickfillcell.c
===================================================================
--- gnucash/trunk/src/register/register-core/quickfillcell.c	2005-12-27 19:54:46 UTC (rev 12188)
+++ gnucash/trunk/src/register/register-core/quickfillcell.c	2005-12-27 21:40:22 UTC (rev 12189)
@@ -103,10 +103,14 @@
 static gboolean
 utf8_caseequal_len (const char *s1, const char *s2, guint len)
 {   
-   char *s1new;
-   char *s2new;
+   gchar *s1new;
+   gchar *s2new;
+   const gchar *s1_offset;
+   const gchar *s2_offset;
    glong s1chars;
    glong s2chars;
+   glong s1_bytes_len;
+   glong s2_bytes_len;
    gboolean equal = FALSE;
     
    if (len == 0)
@@ -118,19 +122,30 @@
    if (!s1 || !s2)
       return FALSE;
 
-   s1chars = g_utf8_strlen (s1, -1);
-   s2chars = g_utf8_strlen (s2, -1);
-   
+   /* Obtain the number of bytes for the given number of characters */
+   s1_offset = g_utf8_offset_to_pointer (s1, len);
+   s2_offset = g_utf8_offset_to_pointer (s2, len);
+   s1_bytes_len = s1_offset - s1;
+   s2_bytes_len = s2_offset - s2;
+
+   /* Test whether the number of characters might be too small anyway
+      (dont need to examine more than bytes_len bytes to check that) */
+   s1chars = g_utf8_strlen (s1, s1_bytes_len);
+   s2chars = g_utf8_strlen (s2, s2_bytes_len);
    if ( (s1chars < len) || (s2chars < len) )
        return FALSE;
-   
-   s1new = g_new0 (gchar, strlen (s1));
-   s2new = g_new0 (gchar, strlen (s2));
 
-   g_utf8_strncpy (s1new, s1, len);
-   g_utf8_strncpy (s2new, s2, len);
+   /* Allocate new strings that are case-independent. */
+   s1new = g_utf8_casefold (s1, s1_bytes_len);
+   s2new = g_utf8_casefold (s2, s2_bytes_len);
 
-   equal = utf8_caseequal (s1new, s2new);
+   /* equal = utf8_caseequal (s1new, s2new); */
+   /* ^^ don't call this to save one string allocation; we used
+      g_utf8_casefold here already. */
+
+   /* Now really compare the two strings */
+   if (g_utf8_collate(s1new, s2new) == 0)
+      equal = TRUE;
    
    g_free (s1new);
    g_free (s2new);