This patch creates a spanner whose ends are oriented to measure
boundaries, a frequent request from users. The ends of the
spanner may be aligned in various ways to prefatory material.
The spanner may hold text or markup in a centered gap.
The spanner is begun with the command '\startMeasureSpanner'
and ended with '\stopMeasureSpanner'.
Provide two regression tests.
Passes make. make check and a full make doc.
"pkx166h" lilypond-pkx@users.sourceforge.net writes:
I think nestability is sort of a red herring. Much more important would
be heeding spanner-id (in which case one can use this for implementing
nestable spanners in specific cases).
--
David Kastrup
Patch on countdown for Nov 20th (there's a comment about nestability on Reitveld but I don't think this holds anything up). I'd think we'd probably want this in the doc somewhere (Changes and Notation right? that can be a new tracker issue I guess).
Patch counted down - please push.
Patch isn't ready for pushing, yet. I still need to create a patch addressing the concerns of the reviewers.
Last edit: David Nalesnik 2019-11-20
Diff:
Passes make, make check and a full make doc.
There are some 'cell' differences in the reg tests.
See below
I can't imagine what relationship my patch would have with these files. Has there been recent work on musicxml? (The majority of the files listed create musicxml output.)
A quick look in git:
http://git.savannah.gnu.org/gitweb/?p=lilypond.git;a=commit;h=bdd9f912691797b954e9541b03f25dfccc1291a2
I'll leave on review for now in case anyone else wants to chime in.
Last edit: Anonymous 2019-11-23
James, you do a
test-baseline
just before applying each patch, right? In that case I don't see how a recent change might affect this, it must be the applied patch.Perhaps it relates to test-baseline then. None of these regtests has anything to do with my patch. There was no report of cell differences with the initial patch. The changes I made for the revised patch were mostly cosmetic.
I can reproduce the cell differences on my machine - but only with the new regtests added. If I remove the new tests and just check the "old" tests, output-distance only shows the expected test-output-distance.ly test.
So I guess this is yet another non-determinism when you add new tests and can be ignored in this case.
Thank you for checking this!
Patch on countdown for Nov 27th.
Patch counted down - please push.
Patch pushed as commit 0b63a1c5aa1340f08b529133a6fd0b11516ec910.