Re: cvs cores when no log message

[Derek Robert Price]

Mark D. Baushke wrote:

> The following has a revised sanity.sh that works with remotecheck.
>
> It might also be useful to write a test that does a 
>
>  rm -f empty.msg
>  touch empty.msg
>  cvs commit -F empty.msg
>
> to exercise that code path as well...
>
>         Enjoy!
>         -- Mark
>
> ChangeLog Entry:
> 2003-02-16  Mark D. Baushke  <mdb@cvshome.org>
>
>         * logmsg.c (logfile_write): Do not pass a NULL pointer to
>         fprintf() when we have an empty log message.
>         (Reported by Piotr KUCHARSKI <chopin@sgh.waw.pl>.)
>         * sanity.sh (editor): Add new tests to verify correct behavior of
>         empty log messages.
>  

Committed, without the extra codepath test you mention, though it does 
sound like a good idea.

Derek

--
               *8^)

Email: derek@ximbiot.com

Get CVS support at <http://ximbiot.com>!
--
Doesn't expecting the unexpected make the unexpected become the expected?