Re: Proposal: Changing tremolo beam gap implementation

[Thomas Morley]

Am Fr., 27. März 2020 um 18:35 Uhr schrieb Torsten Hämmerle
<address@hidden>:
>
> Hello all,
>
> Looking more deeply into Harm's strange whole-note tremolo beam gap
> behaviour, I've stumbled over the current gap implementation:
>
> It's probably rather academic, but my understanding of "gap size" is the
> actual gap size (white space) between two objects.
>
> Looking at the current implementation and playing around with stem
> thickness, however, it becomes obvious that the gap is, other than expected,
> not the free space between stem and beam:  It it is measured from the outer
> side of the stems and not from the inner side of the stems.
>
> The following illustration will compare the current behaviour to what I
> would expect, varying the beam thicknesses from 0 via standard up to
> exaggerated values in order to make the effect clearly visible:
>
> <http://lilypond.1069038.n5.nabble.com/file/t3887/tremolo-gap-comp.png>
>
> *What do you think?*
> Wouldn't it be better to have "gap" actually mean "free space"?
> Before opening an issue and uploading a patch, I'd rather ask for opinions.
>
> Thanks in advance,
> Torsten
>
> tremolo-gap-comp.png
> <http://lilypond.1069038.n5.nabble.com/file/t3887/tremolo-gap-comp.png>

Hi Torsten,

for whole-note-tremolo there is taken care for not colliding with
left/right NoteHeads, even for 'gap = 0
If 'gap <= 2 then there is white space added, exactly as specified
(half of gap-size to left/right), measured from NoteHeads to Beam.

(1) Why should it be different if 'gap is greater than 2.0?
(2) Why should it be different for other note-durations?

Thus I think 'gap should always specify the _white_ space.


Btw, I stumbled across it while trying to code a custom-function
attempting to make whole-note-tremolo-beams avoid possible Dots to the
left and possible Accidentals to the right.

Any chance for a fix, while you're on it?


Many thanks,
  Harm

P.S. I crossposted to devel, not all are subscribed to the user-list