|
From: | Colin Hall |
Subject: | Re: Syntax for make-dynamic-script |
Date: | Sat, 5 Jan 2013 00:23:47 +0000 |
On Dec 26, 2012 10:49 AM, "David Kastrup" <address@hidden> wrote:
>
> Colin Hall <address@hidden> writes:
>
> > David Kastrup <dak <at> gnu.org> writes:
> >
> >>
> >> Richard Shann <richard.shann <at> virgin.net> writes:
> >>
> >> > On Sat, 2012-12-22 at 00:58 +0100, David Kastrup wrote:
> >> > I guess this $ notation is documented somewhere - searching around
> >> > yesterday evening I came across this
> >> > http://lilypond.org/doc/v2.16/Documentation/notation/expressive-marks-
> > attached-to-notes#index-DynamicText
> >> >
> >> > which is where I suspect I got the notation I was using. Should that be
> >> > using $ too
> >>
> >> I repeat: with the exact given examples in the manual, the location data
> >> is correct.
> >
> > Thanks for your help with this, David, it looks like
> > Richard is able to engrave his music.
> >
> > Could you give me some advice on forming a good
> > documentation enhancement tracker for
> > the "$(" syntax? Perhaps there's a description
> > in a commit that I can use.
>
> Well, I have been thinking a bit about it. When using music via #...,
> the music really needs to be constructed for single use and not used
> somewhere else. So it should be feasible to put location data on it
> like it is done for $...
I'm interested to document the existence of the $( syntax. There must have been a motivation. Could you point me to the relevant tracker, or mailing list thread, or provide a fresh explanation. Feel free to be brief.
> So while one current difference in semantics between music entered via
> $... and #... is that the former gets location data (just like \xxx
> does) and the latter doesn't, we probably don't do people a favor by
> having this additional difference between $... and #...
Agreed.
> The question is how we actually want to be dealing with point-and-click
> and errors occuring from within #{ ... #}. For syntax errors, it does
> not seem like anything but the location inside of #{ ... #} itself makes
> sense. But the location data for music might make more sense to be
> associated with the text origin of the problematic construct. Maybe we
> should assign a location only to music that does not yet have one?
A syntax error results from a user error in text-land. So link to the definition.
For other situations, link to the instantiation, I think.
Cheers,
Colin.
[Prev in Thread] | Current Thread | [Next in Thread] |