lilypond-devel
[Top][All Lists]
Advanced

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

Re: Regtest 2.17.13


From: David Kastrup
Subject: Re: Regtest 2.17.13
Date: Mon, 25 Feb 2013 14:56:06 +0100
User-agent: Gnus/5.13 (Gnus v5.13) Emacs/24.3.50 (gnu/linux)

James <address@hidden> writes:

> Hello,
>
> On 25 February 2013 12:00, David Kastrup <address@hidden> wrote:
>
>     "Phil Holmes" <address@hidden> writes:
>     
>     > I've run my pixel comparator on 2.17.13 versus 2.17.12 (note -
>     the
>     > output shows this as 2.17.11 since the windows exe contained the
>     wrong
>     > version number). Output is at
>     >
>     > http://philholmes.net/lilypond/regtestresults/2.17.13/
>     >
>     > As might be supposed, most of the differences are connecting bar
>     > lines.
>     
>     
>     Good grief. Yes. A real pity that our default regtest comparisons
>     blanked those out.
>     
>
> Well it could always be a case of me missing that assuming I did the
> check. Although I have been known to miss the odd change (see Werner's
> dot issue) I thought i would have spotted that change.
>
> It's part of the reason though that if I do see reg tests appear that
> aren't just strings or cell diffs, that I zip and upload the
> test-results for others to look at.
>
> Do you know when the before and after would have been between the two
> as I could re-run a test patchy perhaps by reverting master and
> staging on my local server and then just explicitly testing the patch
> that might have changed this.

Well, it was

commit 0218d1c504463d1ea5b534f23ab4d45c1bc00d6e
Author: David Kastrup <address@hidden>
Date:   Sun Feb 17 18:22:57 2013 +0100

    Issue 3192: Clean up bar-line.scm some more

It has been reverted for now, but of course you can cherry-pick it again
for testing.  There is a comment in the patch saying

+              ;; end of loop
+              ;; model-bar is the last bar found in the elts list
+              ;; (former version had the first here).

And it is quite possible that this is the problem and "model-bar" is
more than just an arbitrary model bar.  I'll try my hand on this one
first (it slightly complicates the code), but I am somewhat at a loss
about how to test it with good confidence.

-- 
David Kastrup



reply via email to

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