Fix #1205.

Fix #1205.

Reinstate an event for \tempo, instead of relying on context property changes.

[Neil Puttock, 2010-09-12 21:42:50]

Hi,

This patch changes the internal representation of \tempo so it uses an event to
trigger creation of a metronome mark.

Please review.

Thanks,
Neil

[Valentin Villenave, 2010-11-27 00:27:27]

On 2010/09/12 21:42:50, Neil Puttock wrote:
> This patch changes the internal representation of \tempo so it uses an event
to
> trigger creation of a metronome mark.

Wow, I can understand why my patch would wreak havoc through your work!
I only had a quick look, but it certainly does look elegant. If you don't mind
my asking, what makes it better to have a tempo-change event rather than just
waiting for some properties to change?

There were some undocumented "features" of Reinhold's implementation that your
code might not break compatibility with (for example, manually setting
Score.tempoUnitCount would reprint the tempoText, etc.).

It's unfortunate that whilst you were trying to make less things hardcoded in
parser.yy, I've been doing the exact opposite... :-)

Cheers,
Valentin.

[Neil Puttock, 2010-11-27 00:57:24]

On 2010/11/27 00:27:27, Valentin Villenave wrote:

> Wow, I can understand why my patch would wreak havoc through your work!
> I only had a quick look, but it certainly does look elegant. If you don't mind
> my asking, what makes it better to have a tempo-change event rather than just
> waiting for some properties to change?

http://code.google.com/p/lilypond/issues/detail?id=1205

There are too many context properties associated with \tempo (unlike say, \clef,
where they're really needed, for example, to set a clef inside an instrument
change definition)
 
> There were some undocumented "features" of Reinhold's implementation that your
> code might not break compatibility with (for example, manually setting
> Score.tempoUnitCount would reprint the tempoText, etc.).

Hmm, interesting, though I'm not sure how useful that is.

> It's unfortunate that whilst you were trying to make less things hardcoded in
> parser.yy, I've been doing the exact opposite... :-)

Actually, I got a bit carried away, since using `scalar' allows silly things
like \tempo <digit>, which obviously isn't desirable.  :)

Cheers,
Neil

[Graham Percival (old account), 2011-01-22 20:34:05]

Unfortunately this patch no longer applies cleanly to git master.

[Neil Puttock, 2011-02-04 10:31:40]

On 2011/01/22 20:34:05, Graham Percival wrote:
> Unfortunately this patch no longer applies cleanly to git master.

It should work fine now.

Cheers,
Neil

[Graham Percival (old account), 2011-02-04 13:15:51]

LGTM

[Graham Percival (old account), 2011-02-06 20:47:32]

argh, I just sent the 48-hour notice before checking the regtests.  Minor issue:

input/regression/display-lily-tests.log
no longer prints a warning for "test 90 unequal".

I know that you know \displayMusic better than anybody else, so I'm not worried
about this.  :)

[nicolas.sceaux, 2011-02-06 21:13:34]

http://codereview.appspot.com/2145047/diff/6001/scm/define-music-display-meth...
File scm/define-music-display-methods.scm (right):

http://codereview.appspot.com/2145047/diff/6001/scm/define-music-display-meth...
scm/define-music-display-methods.scm:927: (format #f "\\tempo ~{~a~a~}~a = ~a~a"
The indentation is wrong.
The usual indentation with this kind of macro is:
(with-xxx something
  body) ; two spaces
or:
(with-xxx
    something ; four spaces
  body) ; two spaces
so that the special 'something' argument is not confused with the regular body.

For the same reason, the new indentation of "\\melismaEnd" on line 915 is
incorrect.

[Neil Puttock, 2011-02-07 21:27:33]

On 2011/02/06 21:13:34, nicolas.sceaux wrote:
>
http://codereview.appspot.com/2145047/diff/6001/scm/define-music-display-meth...
> File scm/define-music-display-methods.scm (right):
> 
>
http://codereview.appspot.com/2145047/diff/6001/scm/define-music-display-meth...
> scm/define-music-display-methods.scm:927: (format #f "\\tempo ~{~a~a~}~a =
~a~a"
> The indentation is wrong.
> The usual indentation with this kind of macro is:
> (with-xxx something
>   body) ; two spaces
> or:
> (with-xxx
>     something ; four spaces
>   body) ; two spaces
> so that the special 'something' argument is not confused with the regular
body.
> 
> For the same reason, the new indentation of "\\melismaEnd" on line 915 is
> incorrect.

Thanks, fixed.

I relied on the default indentation provided by emacs, but I've added the
following to my .emacs file to make sure:

(put 'with-music-match 'scheme-indent-function 1)

Cheers,
Neil