Re: Regtest comparison issues for 2.13.31

[Phil Holmes]

Neil Puttock" <address@hidden> wrote in message 
news:address@hidden
On 26 August 2010 13:55, Phil Holmes <address@hidden> wrote:


> display-lily-tests.ly
> >
> > The following output is new to .31:
> >
> > "Test 93 unequal: .
> > in = \time 2/4
> > out = \context Score \context Timing \applyContext #make-time-settings
> > Test 94 unequal: .
> > in = #(set-time-signature 5 8 '(3 2))
> > out = \context Score \context Timing \applyContext #make-time-settings"
> >
> > It looks like a bug, but could a developer confirm?
>
> This isn't a bug.  It's showing a difference between the versions
> where 2.13.30 has the above (broken) output and 2.13.31 doesn't (click
> on the links to each file to see the difference).  It verifies that
> Carl's fix for issue 1198 actually works (the log file output for
> 2.13.31 is correct).

Got it.  Confirmed with Carl, too.

> A log file difference doesn't always mean something's broken; on the
> contrary, for many issues which don't show graphical changes, it may
> mean something's working again (usually indicated by the minus sign in
> front of the warning).

I was using an editor that does "diif".  Obviously I got old and new 
confused :-(  Thanks for pointing out the minus sign - hadn't spotted that. 
I had understood that a change isn't necessarily a bug, but thought these 
worth checking.

> > page-turn-page-breaking-auto-first-page.ly and 
> > page-breaking-page-count3.ly:
> >
> > Both logfiles have "warning: couldn't fit music on page: overflow is" 
> > and
> > then a figure. The figure is new. Is this bug-tracing output or a bug?
>
> It's extra bug-tracing output.  Joe added it here:
>
> http://git.savannah.gnu.org/gitweb/?p=lilypond.git;a=commit;h=76c1654d8e6daceabb68d530729f87194fa49fed

Ta.  Guessed as much but checking to make sure.

> > page-breaking-min-systems-per-page2.ly:
> >
> > Addition of: "warning: couldn't fit music on page" in logfile
>
> The warning is present in 2.13.30, but no longer appears in 2.13.31
> (see the minus sign).


Ta again.  Just the one regression in .31 then :-)

--
Phil Holmes
Bug Squad