"Trevor Daniels" <t.dani...@treda.co.uk> writes:
> David Kastrup wrote Wednesday, February 18, 2015 6:00 PM
>
>> More seriously, currently \displayLilyMusic deals badly with
>> { c4 c4 8 8 4 } which gets rendered as { c4 c 8 8 4 } and does
>> consequently not recreate its input. It would be rather tricky to fix
>> that since by the time the 8 is printed, c and the following space have
>> already been produced. Instead of making things even more complex here,
>> I consider it saner to become less clever and just print what's in the
>> data, never mind whether it's possibly redundant.
>>
>> Opinions?
>
> Agreed. It would be nice if \displayLilyMusic produced a canonical
> form (using the "preferred notation" meaning) but we're a long way
> from that.
I think that we won't be writing human-readable LilyPond files from
within LilyPond. If we put out a predictable stream, the human-friendly
formatting (linebreaks, linelengths, indentation) from that is more in
the ballpark of Emacs or Frescobaldi.
> In the meantime, producing correct LilyPond input is far more
> important.
I was not suggesting that "producing correct LilyPond input" would be
impossible while keeping the default durations omitted as a rule. It
would just further complicate the printing code, so I am more or less
using the currently not-always-correct output as an opportunity for
stopping to invest the effort for exploiting this redundancy: it's
becoming more tricky (but by no means impossible) to know when it will
work and when not. And actually, it was quite a bit of experimentation
getting the examples in input/regression/display-lily-tests.ly to
combine in a manner where there respective "leave off duration"
decisions check out. Making them work independently from one another is
going to remove at least one source of "Huh?".
I do like the prospect of making input/music/output match closer. That
was one aim of the somewhat repercussion-heavy issue 2240 which made cis
differ from <cis>. Stopping \displayLilyMusic from omitting possibly
redundant durations is much less likely to cause unexpected side
effects.
--
David Kastrup
_______________________________________________
lilypond-devel mailing list
lilypond-devel@gnu.org
https://lists.gnu.org/mailman/listinfo/lilypond-devel
