As long as all of your commits build successfully, it's not a big deal if there
is a regression problem in the middle, IMO. I'd just add a commit at the end
and call it good.
Carl
On 8/29/20, 5:50 PM, "lilypond-devel on behalf of Owen Lamb"
<lilypond-devel-bounces+carl.d.sorensen+digest=gmail....@gnu.org on behalf of
owendl...@gmail.com> wrote:
On Sat, Aug 29, 2020 at 2:39 AM Werner LEMBERG <w...@gnu.org> wrote:
>
> > Ah! It looks like the issue is in line 12 of palm-mute.ly:
> >
> > e8^\markup { \musicglyph "noteheads.u2do" = palm mute }
> >
> > Here, \musicglyph is looking for noteheads.u2do, which I combined
> > with .d2do to make .s2do. Changing the string to "noteheads.s2do"
> > makes this regression test work as expected again.
> >
> > (Is it permissible to simply update this regtest as part of the main
> > commit, or is there a better way to do it?)
>
> If you mean you should simply add another commit to your branch to fix
> this, then yes :-)
>
Ah. No, I meant re-committing my first commit with the fix, then rebasing
the dev/lamb/GSoC-2020-final branch on top of it to keep my history pretty.
Would you rather I just made a new commit?
--Owen
