Re: score-level headers made global

[Ishizaki]

Hi, 
by seeing various codes which call ly_make_anonymous_module, 
I simply thought it was safer to give "false" to this function.

But writing "SCM nid = ly_make_anonymous_module (true);" was OK on my computer.
Its output was same as giving "false".

//

Regarding this problem, I feel it is better to be re-considered how
Lilypond must handle these header definitions by core developers.

As Mr. Bengtsson pointed out, when \paper{printallheaders=##t} is set,
its output gets more strange (but at least understandable) with my patch
applyed.

Here, pleae let me explain why I got concerned with this.
When I write small ensemble score and parts, I want to divide headers
from each source ly files to avoid many copy&paste-ing.

For example, a "inst1.ly" is like this;
\include "headers.ly"
\header {
  instrument = "instr1"
}

\score{
  ...
}

and "headers.ly" is like this; 
\header {
  composer = "composer"
  title = "title"
  opus = "opus"
}

But originally this did not output as I wanted, the latter header def in
"inst1.ly" override headers in "headers.ly" and all contents in
"headers.ly" was ignored.
So, I made patch (and it made another problem arouse).

Recently I noticed that I can live without my patch by writing ly files like 
these;
inst1.ly:
\header {
  \include "headers.ly"
  instrument = "instr1"
}

\score{
  ...
}

headers.ly:
composer = "composer"
title = "title"
opus = "opus"

So, all my patches should be removed again, then printallheaders works as 
intended ?
Any thoughts?

Thank you,
------------------------------------------------------------------------
|                         Yoshinobu Ishizaki                           |
|                http://www2c.biglobe.ne.jp/~isizaki/                  |
------------------------------------------------------------------------

On Sun, 14 Aug 2005 15:19:31 -0700
Graham Percival <address@hidden> wrote:

> 
> On 14-Aug-05, at 3:03 PM, Han-Wen Nienhuys wrote:
> 
> > Graham Percival wrote:
> >> There's a header patch that's waiting for Han-Wen to look at it.  I
> >
> >> don't
> >> know if it fixes this, but you could try it.
> >> --- lilypond-org/lily/lily-parser.cc    2005-08-06
> >22:30:23.000000000  > +0900
> >> +++ lilypond/lily/lily-parser.cc    2005-08-06 21:48:29.000000000 
> >> +0900
> >> @@ -220,7 +220,13 @@
> >>  {
> >>    SCM id = parser->lexer_->lookup_identifier("$defaultheader");
> >>    if (!ly_is_module (id))
> >> -    id = ly_make_anonymous_module (be_safe_global);
> >> -
> >> -  return id;
> >> +    {
> >> +      return ly_make_anonymous_module (be_safe_global);
> >> +    }
> >> +  else
> >> +    {
> >> +      SCM nid = ly_make_anonymous_module (false);
> >
> > generally looks ok, but why is nid non-safe?
> 
> I don't know; this patch was sent by Yoshinobu Ishizaki.  I don't
> understand how the parser works.  Yoshinobu-san?
> 
> - Graham
>