> [EMAIL PROTECTED] writes:
> >
> > Context ranges are defined by various sorts of brackets depending on
> > the context level, but the most consistent context delimiter is curly
> > brackets {}. However, at present, the following
> >
> > property 1 {
> > {property 2 ...}
> > { section 3 }
> > }
> >
> > results in property 2 being applied to section 3 in many (perhaps
> > all?) cases.
>
> Yes. There is some confusion: a \property is something like (for
> example) the value of "current clef". If you change the clef in one
> voice, the clef for all voices on the staff change.
John's idea is basically equivalent to the property scope rule which I
discussed in my email from Mon, 19 Apr.
Unfortunately, the original subject of the mail was slightly different;
so, maybe some of you missed those lines.
> I understand
> that this mechanism is not good enough for what you want, but I still
> have to figure out what is needed.
>From the perspective of a lilypond user, the suggested scope rule is
* more general (i.e. you can mimmick the current scope model just by
using properties only outside of any {}, and you can use local
properties which is difficult with the current model)
* and more 'natural' (i.e. less unexpected behaviour caused
by side-effects).
I think, the user would not lose anything, but profit from better
comfort. Hence, to reply to your statement, it really would give to the
user what is needed. However, I also notice (as John already indicated),
that substantial parts of lilypond's code might have to be rewritten
for such property scope handling...
Bye,
Juergen
