This becomes apparent when using \voiceTwo
as opposed to simply overriding LigatureBracket.direction
.
\version "2.19.82" << \new Staff { \override LigatureBracket.direction = #DOWN %\override Stem.direction = #DOWN \[ c'1 d' \] } \new Staff { \voiceTwo \[ c'1 d' \] } >>
It’s the same for TupletBracket and for longer durations (breve, longa, maxima). I’ll have a look at
ly:tuplet-bracket::calc-x-positions
.According to Gould, a tuplet bracket should start at the left edge of the first and end at the right edge of the last note head; this would mean that aligning to stems, whether visible or not, is always wrong.
Edit: Gould doesn’t say anything about ligature brackets.
Last edit: Malte Meyn 2018-07-19
If we wanted to follow Gould here, IIUC this means we could just delete lines 69 to 73 from tuplet-bracket.cc.
Edit: Gould’s example for testing and comparing.
Last edit: Malte Meyn 2018-07-19
Issue 5387: align brackets only to visible Stems
TupletBrackets and LigatureBrackets now don’t take into account stems of
rests, whole and longer notes and stems that are omitted.
Contains regtests.
http://codereview.appspot.com/346130043
Passes make, make check and a full make doc.
Reg test diffs here
https://cloud.indie.host/s/965JHw2QEDFLs5F ~ 300KB
I had a close look and can now say that all of these regressions are expected :)
Patch on countdown for July 25th.
Patch counted down - please push
commit 1cc61d1a0851d2f35d0579f4d05a3bffbddfa0c6
Author: Malte Meyn lilypond@maltemeyn.de
Date: Thu Jul 19 16:34:12 2018 +0200
commit a69a18ff931c7405e6473e9a564c11276f3c9f9b
Author: Malte Meyn lilypond@maltemeyn.de
Date: Thu Jul 19 15:31:47 2018 +0200
(For the record: This doesn’t follow Gould but modifies
LilyPond’s current behaviour: If there is a stem on the same side/direction,
align to that. Else align to the rest/note head.)