Re: Odd things with \repeat percent and R1

[David Kastrup]

Noeck <address@hidden> writes:

> Am 26.01.2013 06:15, schrieb Jim Long:
>> On FreeBSD 8.3 amd64, this works for R1, R1*1, or R1*2, but
>> crashes lilypond with R1*3 (or 4 ...).  Puzzling.
>> 
>> … \repeat percent 4 R1*2
>> 
>
> Hi Jim,
>
> I think, this is because there are commonly used symbols in the percent
> repeat style for repeating exactly one measure (°/.) and for repeating
> exactly two measures (°//.). But not for three or more measures. It is
> usually used for short patterns and not for parts longer than 2
> measures. So there is no sign for that. One could invent °///. and so
> on, but at some point a volta repeat would make it more readable.
>
> By the way \repeat percent 4 R2. doesn't work either with a 4/4 metre,
> because it is not a full measure.
>
> Is that an answer to your question?

Well, there may or may not be a musical rationale behind what can
sensibly supported.

However,

Interpreting 
music.../usr/local/share/lilypond/2.16.1/scm/music-functions.scm:318:21: In 
procedure car in expression (car durs):
/usr/local/share/lilypond/2.16.1/scm/music-functions.scm:318:21: Wrong type 
(expecting pair): ()

is not exactly proof that the developers have carefully weighed the
evidence in all directions and came to the conclusion that they can't
cater for that case sensibly and instead chose to inform the user of it.

I don't know whether LilyPond can do better regarding the typesetting,
but it certainly should be able to do better regarding the error
message.

-- 
David Kastrup