Timing translator difference between version 2.22 and above with 2.19.84

[Karim Haddad]

Hi,

I am experiencing trouble with timing translation of a big polymetric and 
polytempic score using latest stable and dev version of lilypond. This doesn't 
happen with version 2.19.84. I suspect some changing in handling breaks ??

here is the output:

%%%%%%%%%%%%%%%%%%%%%%%%%%
/home/karim/Work/quatuor_2/scores/quatuor2/Real_2nd/testings/1/test1.ly:3013:1: 
warning: mid-measure time signature without \partial

\time 5/4
Preprocessing graphical objects...
programming error: insane spring distance requested, ignoring it
continuing, cross fingers
programming error: insane spring min_distance requested, ignoring it
continuing, cross fingers
programming error: insane spring constant
continuing, cross fingers
Finding the ideal number of pages...
Fitting music on 9 or 10 pages...
Drawing systems...
/home/karim/Work/quatuor_2/scores/quatuor2/Real_2nd/testings/1/test1.ly:2869:5: 
programming error: bounds of this piece aren't breakable.
b'16
    [
    
/home/karim/Work/quatuor_2/scores/quatuor2/Real_2nd/testings/1/test1.ly:2869:5: 
continuing, cross fingers
    Converting to `test1.pdf'...
    Success: compilation successfully completed

Compilation finished at Sun Nov  7 14:40:46
%%%%%%%%%%%%%%%%%%%%%%%%%%

Although the compilation doesn't fail, but i have an incomplete measure with 
recent versions where the 2.19.84 version render it ok

I am sorry not to post the .ly score since it  is far than an MWE. I tried 
isolating the problem, but this seems an accumulative calculation somewhere 
(?). Tried removing measures before and after, it happens when the score is 
complete.

I can post however a screenshot of the differecne betweene version rendering, 
but i don't think this is pertinent? if yes, i will do this.


Thanks for any insight on this matter.

Best to you all

-- 
Karim Haddad

Music Representations Team, IRCAM
Research and development manager.
email   : karim.haddad@ircam.fr
webpage : http://karim.haddad.free.fr