Re: (de)cresendi syntax

[Valentin Villenave]

Le 16 mars 2009 16:31, Graham Percival <address@hidden> a écrit :
> I don't think anybody else has replied, so I'll take a stab at it.

I've been working at a draft reply for a few days, but never found
what I was looking for.

> What about reserving \< \> \!  **only** for hairpins, and \cr
> \endCr \dim \endDim  **only** for text?  ok, we'd need extra
> commands for decresc and decr, but this way we could avoid the
> \crescTextXXX stuff entirely.

Yes. Yes. Yes. We certainly need that.

> Again, I'd rather use text for text dynamics.  What about \cresc
> #"cresc. poco a poco" ?  i.e. if \cresc is used by itself, it
> prints "cresc.", but if you provide the optional argument, that's
> used instead of the default text?

Yes, but I'd really like to use this \cresc command of yours *after*
the note (not before, as it is now).

> This is the sticking point in my mind.  Are you able to modify it
> yourself (including Carl's help in "yourself")?   If so, then we
> can have a long discussion about the new dynamic syntax.  I'd
> encourage you to experiment with changing the parser before we
> get into lots of details about the new syntax, though.

The parser can certainly be rewritten (like Reinhold did for \tempo).

However, I think we can take advantage of the (make-music
('DynamicEvent ('tweaks ('text "")))) trick that you've been using
(and so have I).

The only drawback for now is that you can't specify a string as an
argument without using a music function :

instead of a'\cresc #"my text"

you have to first define a make-dyn-from-this-text function, and then
predefine your text:

crescMyText =
#(make-dyn-from-this-text "my text")

and only then use

a'\crescMyText

Whereas with a music function, you can simply do:

\cresc #"mytext" a'

... but then again, the command has to be used *before* the note and
this has many drawbacks (plus a big deal of inconsistency through the
code, that i can tell: see for instance
http://repo.or.cz/w/opera_libre.git?a=blob;f=definitions/text-functions.ly#l178
-- or, line 300, the awful mess with text spanners...)

Which is why, rather than a C++ parser change, what I'd really like to
have is a (define-music-tweak) function, that could act like
define-music-function but that would *not* take the music as an
argument, that would just take the string and cons' it as a 'tweak
property.

Regards,
Valentin