Re: CueVoice issue fixed

[Mike Solomon]

Unfortunately, I cannot automate the clef change stuff because of the way the 
engraver works - unless beams are announced earlier, the engraver can't catch 
them in time.  That means that I'd have to keep every grob alive in case there 
were a beam that extended over the whole duration of the piece through 
auto-beaming and not through explicit signaling (this is admittedly insanely 
unlikely, but it is also exactly the kinda thing I'd do).

Perhaps we should just say in the docs that explicit beaming is mandatory 
around stuff like key changes, time signatures, and clef changes if people want 
collision avoidance.  It is kinda buggy, but it would require a radical change 
to lilypond's engraver system to have some sorta retroactive event triggering 
(unless I am misunderstanding how this thing works, which wouldn't be the first 
time...).

Everything else is implemented / fixed.  Please add to Rietveld.

Cheers,
MS

0006-Better-reflects-what-is-actually-under-the-beam.patch
Description: Binary data

0005-Attempt-to-deal-with-smaller-font-sizes-like-cue-not.patch
Description: Binary data

0004-Changes-suggested-by-Carl.patch
Description: Binary data

0003-Intermediary-37-patch.patch
Description: Binary data

0002-Fixes-formatting-problems.patch
Description: Binary data

0001-Fixing-issue-37-with-extra-position-callback.patch
Description: Binary data

better.pdf
Description: Adobe PDF document

better.ly
Description: Binary data

On Jan 11, 2011, at 1:32 PM, Neil Puttock wrote:

> On 9 January 2011 19:48, Mike Solomon <address@hidden> wrote:
>> The patch set now fixes Issue 400 as well.  It also fixes an unreported 
>> issue of time signature collisions.
> 
> Any chance of making the clef avoidance automatic?  I'm not too
> concerned about automating the time signature case since it usually
> occurs across a barline (though there are of course exceptions to this
> rule. :)
> 
> You might consider adding key signatures too.
> 
>> Not much difference in the regtests - I've attached all offenders.
> 
> Great! They look pretty innocuous apart from slur-scoring.ly (assuming
> all the changes are the result of accidental avoidance).
> 
> As for slur-scoring.ly, the raised beam in the second bar is
> unfortunate, since the accidental in the upper voice isn't colliding
> with the beamlet.
> 
> Have you considered what happens when beams start at the same
> timestep?  Here's an example which looks better without your changes:
> 
> \relative c'' {
>  << { g8[ fis] } \\ { c''[ dis,,] } >>
> }
> 
> Cheers,
> Neil