Re: cvs cores when no log message

[Piotr KUCHARSKI]

On Sat, Feb 15, 2003 at 05:06:47PM -0800, Mark D. Baushke wrote:
> >      setup_tmpfile (pipefp, "", changes);
> > +    if (message)
> >      (void) fprintf (pipefp, "Log Message:\n%s\n", message);

Oh well, that was quick hack not to have cores and lock files
waiting around. Of course I thought about (message) ? message : "(null)",
but I guessed you'd do fine after I signal it only ;)

> Actually, a number of stdio implementations replace %s of a NULL
> pointer with "(null)" 

This is bad behaviour, IMO. :)

> so that might be a better patch... more than
> one commit script might be looking for the "Log Message:" text for
> whatever reasons of its own...

Yes, of course.

> However, it would be even better to find the root cause where a NULL
> pointer is getting into message rather than an xstrdup of "" ... do you
> have a test case? Is this in import or commit ? Is it the result of a
> verifymsg script rewriting the message or not?

I just do commit, don't change log message, after quitting editor:
Log message unchanged or not specified
a)bort, c)ontinue, e)dit, !)reuse this message unchanged for remaining dirs
Action: (continue)

I hit 'c' and there goes the core.

> More details please. Could you give a test case that shows this problem?

Simple file, simple commit.

> Could you also include information on what system you have?

Shouldn't matter, but Solaris 2.6.

p.

-- 
Beware of he who would deny you access to information, for in his
heart he dreams himself your master.   -- Commissioner Pravin Lal
http://nerdquiz.sgh.waw.pl/  -- polska wersja quizu dla nerdów ;)