>> Pure relevant is used all over the code. > > Good point. There is even a predicate 'pure-relevant?' > defining which grobs qualify. >
Exactly. This is how the Pure_from_neighbor_engraver evaluates which grobs have pure height functions (line 49).
http://codereview.appspot.com/5323062/diff/29001/lily/pure-from-neighbor-engraver.cc File lily/pure-from-neighbor-engraver.cc (right): http://codereview.appspot.com/5323062/diff/29001/lily/pure-from-neighbor-engraver.cc#newcode49 lily/pure-from-neighbor-engraver.cc:49: SCM pure_relevant_p = ly_lily_module_constant ("pure-relevant?"); Now, this fills 'items_' with things that pass the 'pure-relevant' test. (Since you say Item as opposed to Grob, do you exclude slurs and such?) http://codereview.appspot.com/5323062/diff/29001/lily/pure-from-neighbor-engraver.cc#newcode60 lily/pure-from-neighbor-engraver.cc:60: pure_relevants_.push_back (i.item ()); But 'pure_relevants_' is something different from simply things that satisfy the predicate 'pure-relevant?'. It looks like it includes only items in the immediately-adjacent columns. This would seems to let accidentals on the second-neightbor-column cross the extended bar lines. How would this (untested) work? \new PianoStaff << \new Staff { R1*4 } \new Staff { e'1 | \acciaccatura e'16 <e'' cis''' dis'''>4 r2. | \slurUp \acciaccatura c''16 <c' fis gis>4 r2. <c' fis gis>4 r2. } >> http://codereview.appspot.com/5323062/ _______________________________________________ lilypond-devel mailing list lilypond-devel@gnu.org https://lists.gnu.org/mailman/listinfo/lilypond-devel
