This goes against the documentation, which reads:
-Q Causes the command to be really quiet; the command will
generate output only for serious problems.
-q Causes the command to be somewhat quiet; informational
messages, such as reports of recursion through sub-
directories, are suppressed.
Redirecting stdout and stderr to /dev/null can serve as a bandaid, but
that throws out legitimate errors too.
Enclosed is my patch to make it honor -q/-Q. With -q, it now prints:
Checking in foo.txt
Checking in bar.txt
and with -Q it prints nothing if there are no errors.
M.E.O.
Enc.
Derek
M.E.O'Neill wrote:
>------------------------------------------------------------------------
>
>_______________________________________________
>Bug-cvs mailing list
>Bug...@gnu.org
>http://mail.gnu.org/mailman/listinfo/bug-cvs
>
Derek R. Price <de...@ximbiot.com> writes:
> Would anyone mind if I checked in a patch like this one on stable
> since it does make CVS's behavior match its documentation?
It is a reasonable request and a mostly reasonable patch (having some
sanity.sh tests would have been an even better patch... :-). I have no
objections to making a cvs commit more quiet given the -Q and -q flags.
> I'll probably expand the -q output a little bit, though I will likely
> keep it to one line.
Yes, I'd like for it to mention the new version number if possible.
I suggest that messages like
cvs commit: Rebuilding administrative file database
should also be retained with -q, but possibly be supressed with -Q if
the CVSROOT directory is being updated.
Thanks!
-- Mark
> Derek
>
> M.E.O'Neill wrote:
>
> > cvs commit/checkin doesn't honor -q or -Q, and is extraordinarily
> > chatty, which can be annoying in scripts that run cvs commit (with
> > the -m option for the log message).
> >
> > This goes against the documentation, which reads:
> >
> > -Q Causes the command to be really quiet; the command will
> > generate output only for serious problems.
> >
> > -q Causes the command to be somewhat quiet; informational
> > messages, such as reports of recursion through sub-
> > directories, are suppressed.
> >
> > Redirecting stdout and stderr to /dev/null can serve as a bandaid,
> > but that throws out legitimate errors too.
> >
> > Enclosed is my patch to make it honor -q/-Q. With -q, it now prints:
> >
> > Checking in foo.txt
> > Checking in bar.txt
> >
> > and with -Q it prints nothing if there are no errors.
> >
> > M.E.O.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.3 (FreeBSD)
iD8DBQFAPa5N3x41pRYZE/gRArJvAJ9FP+nB7npeYG8lF3MbCoBU4Xd1xQCfa+Cs
7rGCaQYeypgjVm39ZHuk5V0=
=sUuH
-----END PGP SIGNATURE-----
I agree.
-Larry Jones
No one can prove I did that!! -- Calvin
>-----BEGIN PGP SIGNED MESSAGE-----
>Hash: SHA1
>
>Derek R. Price <de...@ximbiot.com> writes:
>
>
>
>>Would anyone mind if I checked in a patch like this one on stable
>>since it does make CVS's behavior match its documentation?
>>
>>
>
>It is a reasonable request and a mostly reasonable patch (having some
>sanity.sh tests would have been an even better patch... :-).
>
That's a good point. Mr. (Ms.?) O'Neill, would you mind updating the
sanity.sh test cases to take account of your changes, possibly answer
Mark's other concerns below, and resubmit your patch?
>Yes, I'd like for it to mention the new version number if possible.
>I suggest that messages like
>
> cvs commit: Rebuilding administrative file database
>
>should also be retained with -q, but possibly be supressed with -Q if
>the CVSROOT directory is being updated.
>
Derek
--
Get CVS support at <http://ximbiot.com>!