Re: extenders over bar lines in 2.19.55 [was: Automatic Lyric Extenders]

[Alexander Kobel]

Hi David, hi all,

On 2017-02-19 15:58, David Kastrup wrote:
> Michael Gerdau <address@hidden> writes:
> 
>> Am 18.02.2017 um 21:15 schrieb Alexander Kobel:
>>> [Bug report summary:]
>>>
>>> Extenders are not drawn anymore for melismata that include notes that
>>> are not bar-aligned, starting somewhere between 2.19.50 and 2.19.55.
>>> M(N)WE attached - the first and second score should have extenders until
>>> the last note.
>>
>> I've tried this on 2.19.50 - 2.19.55 and it seems as if it used to work
>> until 2.19.54, i.e. apparently it is a 2.19.55 regression (which kind of
>> explains why I had not seen this problem before :) )
>>
>> The first example has an extender only because the default minimum
>> length is 1.5. if that is reduced to 0 that extender vanishes altogether
>> in 2.19.55.
> 
> This is a consequence of
> 
> commit 6c6d1f6ac9e6a7a9aba760dcbb41b4fbbc8f0536
> Author: David Kastrup <address@hidden>
> Date:   Sat Feb 4 14:43:47 2017 +0100
> 
>     Issue 5053/2: Fix extendersOverRests property
>     
>     This previously behaved as always-on.

Argh, sure.  Should have thought about that after the earlier, similar report 
from 2017-02-15...

> This program part now works as intended.  Unfortunately,
> extendersOverRests appears to be a misnomed property, so the resulting
> effective change from extendersOverRest being interpreted as ##t
> regardless of its setting to having it default to ##f affects more than
> extenders over rests.

Not sure about the original intention.  I see the point for having the choice 
to stop extenders over rests (as the documentation suggests).  The "side 
effects" are hardly what I expect, and I don't immediately see a use for that.

> Maybe one should let the setting default to ##t for now and try matching
> its documentation to its behavior before changing the default back to
> ##f.

+1.  Right now, the cure seems to do more harm than good, at least from a 
user's point of view...
Concerning "matching doc to behavior": do you intend to change rather docs, 
behavior, or both?


On a somewhat, but not quite, unrelated note, concerning 
https://sourceforge.net/p/testlilyissues/issues/4509/ and 
https://codereview.appspot.com/313240043:
Do you want/need any input there, or are you merely keeping a log of your 
pondering?  I feel somewhat involved as Knut's and your "Rietveld proxy", but 
I'm not sure whether or how I can assist.  (Note: I don't intend to push by any 
means; I've got a wagonload of way more important things to deal with these 
days...)


Cheers,
Alexander