Re: Align metronome mark at time signature or first musical element. Fixes #684. (issue1579041)

[n . puttock]

Hi Jan,

I've tested the latest patch, and it looks pretty good so far.

Here are a few thoughts on positioning:

-) A metronome mark at a full-bar rest should be aligned with the
barline instead of the paper column to the left of the rest.

-) If there's a tempo change at a key signature, the metronome mark
shouldn't be aligned with the following note.

Cheers,
Neil


http://codereview.appspot.com/1579041/diff/8001/9001
File lily/metronome-engraver.cc (right):

http://codereview.appspot.com/1579041/diff/8001/9001#newcode88
lily/metronome-engraver.cc:88: Metronome_mark_engraver::acknowledge_grob
(Grob_info info)
Since this always gets called before acknowledge_break_aligned (), you
could fold the code above into this with a check for
break-aligned-interface.

http://codereview.appspot.com/1579041/diff/8001/9001#newcode98
lily/metronome-engraver.cc:98: text_->set_parent (g, X_AXIS);
I find this really weird: do you know why this is necessary for the
tempo mark to appear on the full-bar rest even though
stop_translation_timestep () resets the X-parent to
currentMusicalColumn?

http://codereview.appspot.com/1579041/diff/8001/9002
File scm/define-grobs.scm (right):

http://codereview.appspot.com/1579041/diff/8001/9002#newcode1150
scm/define-grobs.scm:1150: (self-alignment-X . -1)
(self-alignment-X . ,LEFT)

http://codereview.appspot.com/1579041/diff/8001/9002#newcode1157
scm/define-grobs.scm:1157: break-alignable-interface
+ self-alignment-interface

http://codereview.appspot.com/1579041/diff/8001/9002#newcode1162
scm/define-grobs.scm:1162: (break-align-symbol . multi-measure-rest)
This doesn't really work, since a MultiMeasureRest isn't a break-aligned
grob (same for MetronomeMark).

The tempo mark regtests spit out loads of warnings due to the missing
interface for this property, but you can't add break-aligned-interface
to silence them since the rest then gets picked up in the
break-alignment-interface, causing a segfault.

http://codereview.appspot.com/1579041/show