Re: osx intel broken?

[Sean Reed]

> > > > > after unpacking the archives, they appear to be PPC programs that
> > > > > won't run on my new iMac. I know I've used this software on this

> > > > The frontend of the intel version claims to be ppc, but the actual
> > > > processing is done in native intel code.  It works fine here.

> > > however, the macIntel 2.11.x versions will run a few times, but  
> > > after
> > > a certain number of measures it doesn't seem to complete the
> > > compilation process. (i haven't found a pattern to it yet, the

> > > the console stops producing any feedback suddenly and opens the .pdf
> > > in Preview immediately after the "Preprocessing graphical  
> > > objects..."
> > > statement with no changes to the output. the "date modified" of

> > I run into this exact situation every so often. Almost every time
> > there winds up being something quite spectacularly off in my rhythmic
> > input.

> I think you're hitting on crashes that are triggered due to your  
> unorthodox
> rhythms. The process log window in the macos X version is not so  
> sophisticated,
> so it doesn't properly print error messages for core dumps and  
> similar.
>
> Still, if you can manage to reproduce any of this, please send me  
> an offending
> input.

Well, I found the offending input, and it appears to be something  
completely different all together:

It appears to be unterminated slurs that are the offending input, but  
only as of lilypond 2.11.x, and only at bar lines.

Here's an example:

Using 2.11.5, lilypond stops at the "Preprocessing graphical  
objects..." and outputs no .ps/.pdf when compiling this expression:
{d'4 cis' d' cis'(}

However, THIS expression (slur beginning moved back one note) results  
in only a warning in the console about an unterminated slur, but  
the .ps/.pdf is created and displayed:
{d'4 cis' d'( cis'}

Using 2.10.5, both of these expressions are compiled to completion  
and the .pdf is opened. Both return the "unterminated slur" warning,  
but finish anyway. In the case where the slur is begun on the last  
note of the measure using 2.10.5, the slur is merely not drawn on the  
output.

This issue made itself evident to me because my scores are often  
hundreds of measures long, and when tweaking them I often comment out  
all but 5 or 10 bars at a time.

Hope this is of some significance for you.

-Sean
---------------
Sean Reed
Dublin
www.seanreed.ie