On 2012/10/30 06:17:30, Keith wrote:
On 2012/10/30 06:05:57, dak wrote:
> Actually, I would be perfectly fine with binning both \violin.1
> as well as \"violin1".
That should be fine. No-one has indicated they would actually use
either of
these. (But I do use "vn1mvt2" with $vn1mvt2 )
$vn1mvt2 has the disadvantage of running in lexical closure. That means
that it does not work for something like
#{ \layout { \context { \Voice ... $vn1mvt2 ... } } #}
since it inherits the lexical scope from outside and does not look
inside of the LilyPond scope.
The scoping is a difference between \... and $... that may be relevant
at times, so there is a bit of an incentive to have a \-like operator.
Rethinking this carefully, a native \violin2 (as an actual
array/vector/alist) should not really be away more than a few months
hopefully.
This is not really the same as freeform identifiers (you need to declare
the structure of \violin) but if we find that people can be satisfied
with that solution (which is more flexible in some manners), not opening
the somewhat quirky can of worms with weirder identifier syntax forms
might prove acceptable.
If we see that the vector/struct solution will not make people happy,
there will still be enough time to resuscitate one of the existing
proposals or come up with yet another one. We've gotten along without
this for a dozen years or so, so we won't die from waiting a bit more.
This code/patch is not likely to go stale anytime soon.
http://codereview.appspot.com/6778055/
_______________________________________________
lilypond-devel mailing list
lilypond-devel@gnu.org
https://lists.gnu.org/mailman/listinfo/lilypond-devel
