Re: Seg fault in 1.12.5

[Larry Jones]

Steve McIntyre writes:
> 
> Simply replacing the outer check with the old (1.12.1) code:
[...]
> _seems_ to work, but I don't pretend to be able to follow this code at
> 1.30AM.

It fixes the crash but reintroduces the bug that the code was originally
changed to fix.  The correct fix should be:

Index: update.c
===================================================================
RCS file: /cvs/ccvs/src/update.c,v
retrieving revision 1.202.4.13
diff -u -r1.202.4.13 update.c
--- update.c    20 Feb 2004 21:21:06 -0000      1.202.4.13
+++ update.c    26 Feb 2004 02:35:09 -0000
@@ -2352,7 +2352,7 @@
      * revision.  i.e. we know that diff3(file2,file1,file2) will produce
      * file2.
      */
-    if (vers->ts_user
+    if (vers->vn_user != NULL && vers->ts_user != NULL
         && strcmp (vers->ts_user, vers->ts_rcs) == 0
         && strcmp (rev2, vers->vn_user) == 0)
     {

I'd be very interested to know an easy way to reproduce the problem,
though.  It would occur when there's a file in the current directory
(ts_user not NULL) that did not come from the repository and has not
been added (vn_user NULL), but nonetheless has a corresponding
repository file (vn_rcs and ts_rcs not NULL or we wouldn't have gotten
to this code) -- I haven't been able to come up with a scenario where
that occurs.

I've checked the above fix in on the stable branch, but cvs.cvshome.org
went catatonic immediately afterwards so I haven't been able to update
the trunk.

-Larry Jones

Isn't it sad how some people's grip on their lives is so precarious
that they'll embrace any preposterous delusion rather than face an
occasional bleak truth? -- Calvin