[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: run make doc without regtest? (issue 4180)

From: David Kastrup
Subject: Re: run make doc without regtest? (issue 4180)
Date: Sun, 28 Dec 2014 08:00:23 +0100
User-agent: Gnus/5.13 (Gnus v5.13) Emacs/25.0.50 (gnu/linux)

Keith OHara <address@hidden> writes:

> David Kastrup <dak <at> gnu.org> writes:
>> Keith OHara <k-ohara5a5a <at> oco.net> writes:
>> > It seems that we knew about these problems, bug then added the 
>> > regression tests anyway.  
>> >
>> > We have bug-reports, so it seems we lose no information
>> Bug reports or an actual issue in the tracker?
> Well, there is of course the bug report in the issue title,
> but Federico's example and log seem to point more to the bug reported
> in another thread here, now issue 4231
>> > if we avoid these issues in the regression test, but the last time I
>> > tried that I got
>> > <http://lists.gnu.org/archive/html/lilypond-devel/2011-11/msg00559.html>
>> There was no attempt of even making a tracker issue that time as far as
>> I can tell, it was strictly a sweep-under-the-rug approach.
> The tracker issue from 2011 was linked in the email linked above.

Yes, I remember now.  Reported in June.  No attempt of fixing it,
messing up a whole number of regtests and making them unreliable.
Several unrelated regtests were purposefully _complicated_ (adding
stemlets) in order not to trigger the problem.  I called this the wrong
approach in November and started debugging the problem in order to
figure out what was wrong, reporting on my findings.  It was you who
actually found the fix.

Do you want to state that you would have considered it better to keep
obfuscating regtests by adding stemlets and thus avoid looking for the
underlying problem?

With regard to large rationals, the underlying problem is well-known and
separated out and we are not talking about artificial complication of
regtests in ways nobody understands.

If you don't see a difference here, it would seem good for the overall
stability of the code base that other voices get heard as well.

> A new regtest had triggered a previously-tracked issue, and somebody
> with and email starting 'dak' complained on the bug-lilypond. 
> I pointed to the tracker item, and suggested to change the regtest.
> Then a message from 'dak' posted to the tracker item complained that
> the error in the regtest hindered regression testing.  So I put the
> known work-around into the offending regtest, and somebody with the 
> email staring 'dak' reverted the work-around.
> Let's see what happens this time.

Even if you make it appear you'd rather have them silenced.

David Kastrup

reply via email to

[Prev in Thread] Current Thread [Next in Thread]