[Pan-devel] patch for minor utf8 issue

[K. Haley]

There is a potential problem in handling bold,italic,underline text when 
there are multibyte characters in the text.  This patch fixes the 
problem by properly iterating over characters instead of bytes.  The 
actuall tests remain the same, only the iteration is changed to prevent 
a false positive.  Besides, does anyone know what would happen if you 
set a tag in the middle of a multibyte character?

--- old/pan/gui/body-pane.cc    Fri Apr 14 18:17:33 2006
+++ new/pan/gui/body-pane.cc    Sat Apr 15 23:13:40 2006
@@ -456,18 +456,20 @@
       const char * e (v_all.end());
       enum { NORMAL, UNDERLINE, BOLD, ITALIC };
       int mode = NORMAL;
-      for (const char * pch(b); pch!=e; ++pch) {
-        if (*pch=='_' && mode==UNDERLINE && (pch+1==e || ::isspace(pch[1]) || 
::ispunct(pch[1])))
-          set_section_tag (buffer, &start, v_all, 
StringView(mode_begin,pch+1-mode_begin), "underline");
-        if (*pch=='/' && mode==ITALIC && (pch+1==e || ::isspace(pch[1]) || 
::ispunct(pch[1])))
-          set_section_tag (buffer, &start, v_all, 
StringView(mode_begin,pch+1-mode_begin), "italic");
-        if (*pch=='*' && mode==BOLD && (pch+1==e || ::isspace(pch[1]) || 
::ispunct(pch[1])))
-          set_section_tag (buffer, &start, v_all, 
StringView(mode_begin,pch+1-mode_begin), "bold");
-        else if (*pch=='_' && (pch==b ||isspace(pch[-1])))
+      for (const char * pch(b),*p(b),*n(b); pch!=e;p=pch,pch=n) {
+               if(n!=e)
+                       n=g_utf8_next_char(n);
+        if (*pch=='_' && mode==UNDERLINE && (n==e || ::isspace(*n) || 
::ispunct(*n)))
+          set_section_tag (buffer, &start, v_all, 
StringView(mode_begin,n-mode_begin), "underline");
+        if (*pch=='/' && mode==ITALIC && (n==e || ::isspace(*n) || 
::ispunct(*n)))
+          set_section_tag (buffer, &start, v_all, 
StringView(mode_begin,n-mode_begin), "italic");
+        if (*pch=='*' && mode==BOLD && (n==e || ::isspace(*n) || 
::ispunct(*n)))
+          set_section_tag (buffer, &start, v_all, 
StringView(mode_begin,n-mode_begin), "bold");
+        else if (*pch=='_' && (pch==b ||isspace(*p)))
           { mode = UNDERLINE; mode_begin=pch; }
-        else if (*pch=='*' && (pch==b ||isspace(pch[-1])))
+        else if (*pch=='*' && (pch==b ||isspace(*p)))
           { mode = BOLD; mode_begin=pch; }
-        else if (*pch=='/' && (pch==b ||isspace(pch[-1])))
+        else if (*pch=='/' && (pch==b ||isspace(*p)))
           { mode = ITALIC; mode_begin=pch; }
         else if (isspace(*pch))
           { mode = NORMAL; mode_begin=pch; }

signature.asc
Description: OpenPGP digital signature