Re: 2.7.38 bug: proportional duration beaming of 32nds and greater

[Trevor Bača]

On 3/16/06, Trevor Bača <address@hidden> wrote:
> On 3/16/06, Han-Wen Nienhuys <address@hidden> wrote:
> > Trevor Bača wrote:
> > >
> > > Ugh. The exact conditions to induce this bug are very difficult.
> > >
> > > It is possible to come up with an example with somewhat fewer notes
> > > and fewer measures, like this:
> >
> > > ... but cutting the example down even further usually causes the
> > > problem to no longer manifest.
> > >
> >
> > Thanks, this was a particularly hairy bug, but it's fixed now.
> >
> > The problem is that we reorder long grob lists when they cross
> > line-breaks. This is good for efficiency, and harmless for unordered
> > sets of grobs. Of course, stems of a beam are ordered, and it shouldn't
> > happen there.
>
> Really appreciate the work on this particular bug; I couldn't even get
> my bug files to test consistently, much less get a reasonable
> intuition for which parts of the code might be effected.


Works great in 2.7.39.


--
Trevor Bača
address@hidden