Hi folks,
I've come across a problem in lily 2.11.26 which didn't exist in
2.11.25. When I compile a file under certain conditions the
compilation proceeds to the 'Fitting music on 2 or 3 pages...' message
at which point my machine begins to swap heavily and the CPU usage
goes through the roof. I left the machine running on a job for about
twenty minutes yesterday before I gave up. Here are the conditions
under which the problem shows itself:
First, I need to be using the 'optimal' page breaking algorithm. The
problem happens at the 'Fitting music on 2 or 3 pages..' message. If I
use the new 'page-turning' algorithm there is no issue.
Second, the problem only seems to occur when I have many notes with
ledger lines. In my example here it can be triggered by replacing
\relative c'' with \relative c' .
Third, I need to have many scores in my file. In my example, I can
repeat the bar 500 times in one or two scores with no problems.
However, I have used many \scores and the problem arises.
Here's my problem displaying sample:
%%%%%%%%%%%%%%%%%%%%%%%%%%%%
\version "2.11.26"
notes = {
\relative c' { %if c' is changed to c'' the problem won't occur
\repeat unfold 16 {a b c d}
}
}
\score{
\notes
}
\score{
\notes
}
\score{
\notes
}
\score{
\notes
}
\score{
\notes
}
\score{
\notes
}
\score{
\notes
}
\score{
\notes
}
\score{
\notes
}
\score{
\notes
}
\score{
\notes
}
\score{
\notes
}
\score{
\notes
}
\score{
\notes
}
\score{
\notes
}
\score{
\notes
}
\score{
\notes
}
\score{
\notes
}
\score{
\notes
}
\score{
\notes
}
%\paper{ %The problem doesn't occur if this block is uncommented.
% #(define page-breaking ly:page-turn-breaking)
%}
--
=============================================
Cameron Horsburgh
=============================================
_______________________________________________
bug-lilypond mailing list
[email protected]
http://lists.gnu.org/mailman/listinfo/bug-lilypond
