"Andrew Bernard" <[email protected]> writes:
> Hi Ming,
>
>
>
> We have been through this before, I am sorry to say. You write
> lilypond code in a way that compiles, yes, but it is pathological in
> the extreme, to put it mildly. You set each and every bar as a
> variable, and when there are several voices, introduce new parallel
> voices in every bar. As an example for the others, this is what you
> do:
>
>
>
> lft.46 = { <<{g2 f2}\\{d1}>> |}
>
> lft.47 = { <<{r4 ef4~ <ef g>2}\\{g,1}>> |}
Those are not "new parallel voices": they still connect. See:
{
<< c''1~ \\ e'1~ >>
<< c''1 \\ e'1 >>
}
<< ... \\ ... >> translates to
<< \context Voice = "1" ... \context Voice = "2" ... >>
so when the parallel voices line up perfectly in time, they still happen
to connect. It still makes a hash of understanding the code and
debugging it and may well trigger rare problems in our context handling.
Without very good reason, choosing such a non-structure definitely looks
like creating more problems than it solves.
--
David Kastrup
_______________________________________________
lilypond-user mailing list
[email protected]
https://lists.gnu.org/mailman/listinfo/lilypond-user
