Re: page breaking changed in 2.13 ???

[Joe Neeman]

On Sun, 2009-05-24 at 11:53 -0700, ArnoWaschk wrote:
> 
> 
> Werner LEMBERG wrote:
> > 
> > 
> >> i am using a 2.13 lilypond compiled from today's git repo, and i am
> >> observing - in comparison to a version from around 10th of April -
> >> that the page breaking strategy seems to have changed significantly.
> >> Which leads me in one case of a before 47page piano reduction score
> >> (1 to 4 staves per page, staff lines number irregular) becoming a
> >> 123page score only having one line per page, even if just being a
> >> piano solo line...
> > 
> > This looks like a bug.  Can you reduce it so that Joe can check it?
> > 
> > Thank you for your reply!
> > No, i do not have a clue how to reduce a huge score, which seems to fail
> > because of a certain hugeness...
> > I only see that shorter scores get less extreme results...
> > And anyway i am far too short of time for experiments. Which is why i am
> > *most urgently* seeking for a way back!
> > Going to 2.12 would be quite a step back in quality...

I would suggest avoiding the current git code (and probably the current
development branch altogether) if you really need things to work
urgently. You can try

git checkout b0ab8d44ad4c8232f672da9775bd8a564771e1c4

since this particular commit occurred before I started messing with
page-breaking stuff. I would appreciate it, however, if you could
produce a small example when you have time, as that would help me fix
the bug.

> > 
> >> P. S. I would also be grateful for instructions to roll back my
> >> source tree to the April version, if that is the easiest...
> > 
> > Just use `gitk' to select a certain revision (or tag) `foo'.  Then
> > say, on the command line,
> > 
> >   git checkout foo
> > 
> > Yes, but how to translate 10th of April 2009 into foo? I cannot see a tag
> > in that period...?

If there is no tag, you can replace "foo" by the SHA1 ID of a commit.

Joe