On Mon, 7 Nov 2005, Malcolm Rowe wrote:
> On Mon, Nov 07, 2005 at 08:22:46AM +0100, Peter N. Lundblad wrote:
> > Why the
> > post-commit error was written to stdout instead of stderr is another
> > question that should probably be posted separately...)
>
> See svn_cl__print_commit_info().
>
> /* Writing to stdout, as there maybe systems that consider the
> * presence of stderr as an indication of commit failure.
> * OTOH, this is only of informational nature to the user as
> * the commit has succeeded. */
>
Fun. I looked in the file, but didmiss that comment. Sometimes using a
one-line braille display shows up:-)
This was committed recently as part of 443.
Is the above statement really valid? We print other warnings to stderr. If
someone wants to know if the commit failed, then they should use the exit
status IMO. Hmmm, and then we have the proposal that warnings should also
set a non-zero exit status... I still don't think printing this particular
warning to stdout is a good idea. The commit has succeded, but you don't
know how severe the post-commit failure is.
Regards,
//Peter
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Mon Nov 7 11:17:58 2005