Re: cannot find Voice `uniqueContext0' with \context Staff

[Michael Käppler]

Am 09.09.2020 um 13:12 schrieb David Kastrup:
> Michael Käppler <xmichael-k@web.de> writes:
>
> > Am 08.09.2020 um 22:37 schrieb James Lowe:
> > > On 08/09/2020 20:15, Thomas Morley wrote:
> > > > Hi,
> > > >
> > > > this came up in the german forum:
> > > >
> > > > \context Staff { b1 }
> > > > \addlyrics { foo }
> > > >
> > > > results in not printed lyrics and the warning:
> > > >
> > > > warning: cannot find Voice `uniqueContext0'
> > > > \addlyrics
> > > >              { foo }
> > > >
> > > > It worked up to 2.18.2 but fails for 2.20.0 and upstream
> > > >
> > > > Changing the code to use
> > > > \context Voice or \new Voice or \new Staff
> > > > works.
> > > >
> > > > Cheers,
> > > >     Harm
> > > It seems it occurred somewhere in those builds that we can no longer
> > > compile by default. Also see:
> > > https://lists.gnu.org/archive/html/lilypond-devel/2020-02/msg01024.html
> > >
> > > i.e. 2.19.12 does NOT have the warning and I cannot compile another
> > > build until 2.19.38 where the warning is already manifest.
> > >
> > > I tried to do what David suggested in that email thread above but was
> > > unsuccessful with cherry picking (more my skills than David's
> > > instructions though), so this is going to be tedious to bisect I expect.
> > >
> > > James
> > I can try to hunt this down today, but please, gurus, intervene if you
> > have an immediate guess what's going on.
> commit df3457d85ebfa4bc347a4569241227449f84b901
> Author: David Kastrup <dak@gnu.org>
> Date:   Tue Sep 9 11:14:34 2014 +0200
>
>      Allow \addlyrics to work with arbitrary contexts
>
> would be my guess.  Version 2.19.14.
Almost correct, it was

commit fc9070f6ca7945cf30c991b11874de570d90c9e6
Author: David Kastrup <dak@gnu.org>
Date:   Tue Sep 9 11:04:27 2014 +0200

    Revert "Convert \new Staff ... \addlyrics into \new Staff \new
Voice ... \addlyrics"

    This reverts commit 76f256075f0a4f481d7e11c19dd55ab09030fe5b.

(Btw, using an old LilyDevDebian VM image did the trick and saved me a
lot of time compared to cherry-picking
the compilation fixes IMHO. It is *very* nice to see, how the
compilation logs have been improved in the meantime,
being full of uncomprehensible warnings and with cluttered terminal
output. Many thanks to all that put their effort in this topic)

I can take another look this evening, I don't know, however, if I can
manage to understand the history of
these changes and whether the use case Harm mentioned did only work by
accident.

Michael