Avenue for a fix

[Richard Stallman]

This may be the right approach for a fix, but I don't think
it is safe to change current_buffer directly this way.
But now that the bug is probably identified, would someone
like to write a clean fix?


Date: Tue, 20 Jan 2004 17:33:34 +0300
From: Dmitry Antipov <address@hidden>
X-Accept-Language: en-us, en
To: address@hidden
Subject: Probably fixed [Re: Crash calling md5 for a list of buffers]
Sender: address@hidden

Index: fns.c
===================================================================
RCS file: /cvsroot/emacs/emacs/src/fns.c,v
retrieving revision 1.354
diff -u -r1.354 fns.c
--- fns.c       29 Dec 2003 13:51:54 -0000      1.354
+++ fns.c       20 Jan 2004 13:33:17 -0000
@@ -5386,7 +5386,7 @@
   int start_char = 0, end_char = 0;
   int start_byte = 0, end_byte = 0;
   register int b, e;
-  register struct buffer *bp;
+  register struct buffer *bp, *tem;
   int temp;
 
   if (STRINGP (object))
@@ -5542,8 +5542,10 @@
                  Fsignal (Qcoding_system_error, Fcons (coding_system, 
Qnil));
            }
        }
-
+      /* At this moment, current buffer may be not equal to bp.  */
+      tem = current_buffer, current_buffer = bp;
       object = make_buffer_string (b, e, 0);
+      current_buffer = tem;
 
       if (STRING_MULTIBYTE (object))
        object = code_convert_string1 (object, coding_system, Qnil, 1);



_______________________________________________
Bug-gnu-emacs mailing list
address@hidden
http://mail.gnu.org/mailman/listinfo/bug-gnu-emacs