Re: regression tests

[Rune Zedeler]

Kieren MacMillan skrev:

> > development of lilypond will run faster than any effort to try
> > and keep up with such extended documentation of properties.
>
> Possibly... but an *automated* page -- containing a listing exactly as 
> Trevor outlined BUT WITH ONLY THE DEFAULT SETTING/VALUE  -- would keep up
> *precisely* with development, since it would be scripted into the main 
> compilation process.

Yep listing default values is easy - I agree - this was not what I was 
commenting. I was commenting the possibility of listing all POSSIBLE values.
I don't understand what you mean - we already list the default values 
today: 
http://lilypond.org/doc/v2.11/Documentation/user/lilypond-internals/Clef

Though, default values are expanded before they are listed.
If some property is set to ,begin-of-line-invisible it will show in the 
docs as #(#t #f #f) - not as ,begin-of-line-invisible - which makes the 
whole thing fairly unusable. So we have to do "something" with the 
scheme representation in order to make it more feasible for the docs.

-Rune