Re: Issue 1881 in lilypond: Multiple cyclic dependency errors for Beam/Stem

[address@hidden]

On Sep 11, 2011, at 11:30 PM, address@hidden wrote:

> Status: Accepted
> Owner: ----
> Labels: Type-Critical
> 
> New issue 1881 by address@hidden: Multiple cyclic dependency errors for 
> Beam/Stem
> http://code.google.com/p/lilypond/issues/detail?id=1881
> 
> Following commit 6465274e66a851cccd4cd32a521abc853f3e79dd, `Restores stem 
> length and stem-begin-position.', even the simplest snippets with beaming 
> spit out multiple cyclic dependency errors:
> 
> \version "2.15.12"
> 
> \relative c' {
> c8 d e f
> }
> 
> 
> /tmp/tmpcPppzK/document.ly:2:2: programming error: cyclic dependency: 
> calculation-in-progress encountered for #'Y-extent (Stem)
> 
> c8 d e f
> /tmp/tmpcPppzK/document.ly:2:2: continuing, cross fingers
> 
> 
> c8 d e f
> /tmp/tmpcPppzK/document.ly:2:2: programming error: cyclic dependency: 
> calculation-in-progress encountered for #'quantized-positions (Beam)
> 
> c8 d e f
> /tmp/tmpcPppzK/document.ly:2:2: continuing, cross fingers
> 
> 
> c8 d e f
> /tmp/tmpcPppzK/document.ly:2:2: programming error: cyclic dependency: 
> calculation-in-progress encountered for #'quantized-positions (Beam)
> 
> c8 d e f
> /tmp/tmpcPppzK/document.ly:2:2: continuing, cross fingers
> 
> 
> c8 d e f
> /tmp/tmpcPppzK/document.ly:2:2: programming error: cyclic dependency: 
> calculation-in-progress encountered for #'quantized-positions (Beam)
> 
> c8 d e f
> /tmp/tmpcPppzK/document.ly:2:2: continuing, cross fingers
> 
> 
> c8 d e f
> 
> This is a serious problem for regression testing, since it obscures genuine 
> changes.
> 

Hey all,

I'm well on my way to getting a fix worked out for this.  I think that the 
problem only comes from lines that used the convention:

(void) me->get_property ("foo")

to trigger calculations if they hadn't happened yet.  The values were actually 
never used.

It may be useful to come up with a sort of "void_get_property" function that 
triggers a callback without returning a property.  It, in turn, would not throw 
an error for cyclic dependencies.  Thoughts?

Cheers,
MS