Re: Guile-memory-leek in 1.5 and span bars bug

[Mats Bengtsson]

> IIRC we never was able to reproduce the guile-1.4-memory-leek in 1.5.
> Well, I just stumbled over it in the middle of all the other problems 
> with the 48 handclappers score.
> I'll attach the first 14 measures of the piece. latex CAN handle the 
> first measures because there doesn't happen so much in the start.
>
>
> The problem is in voice 28 - where an extra measure is inserted.
> If you comment out most of the other voices, the problem disappear.

Please specify exactly what Lilypond version and Guile version
you use, the file processes nicely over here with 1.5.59 and
Guile 1.5.6. If you wish, I could give it a try with Guile 1.4
as long as I know what Lilypond version I should try.

Your story about how you produced the score reminds me about
a project I worked on several years ago, where we made a 
program that automatically generated C++ programs for a 
specific purpose. The generated programs didn't really
look like anything you'd write by hand and they could easily 
contain several thousand lines of code within a single
function. As a result, we managed to find all kind of bugs in 
the compiler (unfortunately we had to stick to the specific
compiler, but they got long lists of bug reports because of 
our unintended stress test).

   /Mats