Re: Disappearing barlines with skipBars - possible bug?

[Mark Knoop]

At 15:43 on 24 Mar 2015, Kevin Barry wrote:
>On Tue, Mar 24, 2015 at 2:38 PM, Mark Knoop <address@hidden> wrote:
>
>> { \time 2/4 c'4 c'1 c'4 } % 3 bars, 2nd of which is empty
>>
>> { \time 2/4 c'4 c'2 c'4 } % 2 bars, neither of which is empty
>>
>
>At best I would consider these to be non-standard notation. At worst
>I'd say it's incorrect. It's possible to write a note value longer
>than a bar, but I don't think there is an accepted standard for
>notating that, except for connecting smaller note values with ties.
>And as I said before I have never seen it in a score (do you have any
>examples?).

I agree these are contrived examples and am not claiming them to be
standard notation. However tuplets spanning barlines are common in
contemporary music and that is the context where this bug appears.

>> skipBars is introduced in the Learning Manual (3.4.5 Scores and
>> parts) as a way of condensing multi-measure rests. Its only
>> references in the Manuals are regarding this function. I suspect its
>> impact on notes crossing barlines is either unintended or at least
>> not thought through
>
>It's just used there as an example of a context property. 

Um, no: "When printing the part, multi-rests must be condensed. This is
done by setting a run-time variable \set Score.skipBars = ##t"

It is in Notation Reference 5.3.2 that it's used as an example of a
context property, although still with reference to its use for
combining multimeasure rests.

>I agree that its behaviour for notes rather than rests might not be
>thought through (but it /is/ mentioned in the IR so it may not be
>unintended). Normally full-bar rests are condensed with
>\compressFullBarRests (which probably just sets skipBars to #t, but I
>didn't check).

Yes, it's just a shortcut:

ly/property-init.ly:156:compressFullBarRests = \set Score.skipBars = ##t

>I can't imagine a situation where the current behaviour would be
>> desirable (silently hiding a barline thus changing the length of a
>> bar) and it would certainly seem to be a very different use case than
>> condensing multi-measure rests.
>
>I think the default behaviour (without touching skipBars) is ok, and
>agree that there's no apparent need for it to affect notes (the
>possible scenario with long notes and short empty bars seems
>implausible). So perhaps skipBars could be changed to only affect
>rests, or at the very least a different context property could be used
>as an example in the learning manual (or both).
>
>I verified that \compressFullBarRests does indeed affect notes as
>well, so if you want to file a bug report then perhaps proceed on that
>basis.

cced to bug-lilypond

-- 
Mark Knoop