Re: various Lilypond 2.9 Win-uninstaller problems

2006-10-06 Thread Han-Wen Nienhuys

Uwe Stöhr schreef:

Han-Wen Nienhuys schrieb:

I tested the latest Lilypond 2.9.16 for Windows and found out that it 
is not correctly uninstalled. After uninstalling it the following 
folders remain:


\usr\etc
\usr\lib

and its subfolders. The installation directory should completely be 
removed.


With Lilypond 2.9.18 also the folder

\usr\var remains.f

I've switched on logging for the Mingw installer. Can you (un)install 
2.9.18 with /s when it comes out? This will generate an installation 
log, which will hopefully provide some clues.


I did this. The logfile is attached. The problem is that the folder

\usr\etc can't be deleted because the file fontconfig.reloc is in use. 
I get an eror message that a program is using it and that the file could 
therefore not be deleted.

The same is for the folder \usr\lb where aifc.py is in use.
Of course there are more files in use.
In \usr\var it is the folder fontconfig that could for an unknown 
reason not be deleted.


This is really strange. The only program using that should be 
lilypond.exe itself. Please investigate.


Btw. I noticed that you are using Python 2.4. Perhaps an update to 
python 2.5 could be an advantage. Then you need at least less files to 
ship with the installer so it becomes a bit smaller.


usually newer releases get bigger, I'd expect the contrary.

--

Han-Wen Nienhuys - [EMAIL PROTECTED] - http://www.xs4all.nl/~hanwen

LilyPond Software Design
 -- Code for Music Notation
http://www.lilypond-design.com



___
lilypond-devel mailing list
lilypond-devel@gnu.org
http://lists.gnu.org/mailman/listinfo/lilypond-devel


Re: Longa notes

2006-10-06 Thread Juergen Reuter

Hi all,

the longa notes problem reduces to this piece of code in method 
internal_print (Grob *me, string *font_char) in file lily/note-head.cc:



  if (!scm_is_symbol (style))
style = ly_symbol2scm (default);

  [... snip ...]

  if (style != ly_symbol2scm (default))
{
   SCM gn = me-get_property (glyph-name);
   if (scm_is_string (gn))
   suffix = ly_scm2string (gn);
}

The problem here is that 'me-get_property (glyph-name)' is called only 
if 'style != ly_symbol2scm (default)'.  That is, the code in 
output-lib.scm for default style is dead.  That is, why longa notes are 
not working for default style.


If I remove the 'if (style != ly_symbol2scm (default))' condition _and_
if I explicitly say: \override NoteHead #'style = #'default
then the longa notes show correctly.  However, without explicitly setting 
#'style to #'default, I get:


/home/reuter/project/lilypond-2.9/share/lilypond/2.9.22/scm/output-lib.scm:144:58: 
In procedure symbol-string in expression (symbol-string style):
/home/reuter/project/lilypond-2.9/share/lilypond/2.9.22/scm/output-lib.scm:144:58: 
Wrong type argument in position 1 (expecting SYMBOLP): ()


Hence, obviously, the above 'if' condition was introduced to suppress this 
error, but this fix is bad, since it makes the code for the default style 
in output-lib.scm dead.


The real problem is the case style switch statement in line 114 of 
output-lib.  If style is not set, it erroneously selects the else case 
(rather than the default case) and dies when trying to execute 
symbol-string on the undefined style.


Does some scm guru know how to test _before_ the case style switch, if 
style is undefined, and if so, set it to default?  This would be 
the real fix, I guess.


Greetings,
Juergen


___
lilypond-devel mailing list
lilypond-devel@gnu.org
http://lists.gnu.org/mailman/listinfo/lilypond-devel


Re: Longa notes

2006-10-06 Thread Han-Wen Nienhuys

Juergen Reuter schreef:

Hi all,

the longa notes problem reduces to this piece of code in method 
internal_print (Grob *me, string *font_char) in file lily/note-head.cc:





The problem here is that 'me-get_property (glyph-name)' is called 
only if 'style != ly_symbol2scm (default)'.  That is, the code in 
output-lib.scm for default style is dead.  That is, why longa notes 
are not working for default style.



Hence, obviously, the above 'if' condition was introduced to suppress 
this error, but this fix is bad, since it makes the code for the default 
style in output-lib.scm dead.


Actually, IIRC, it was there to skip the potentially expensive 
calculation of glyph-name. Be sure to do some timings after removing the 
code too.


--

Han-Wen Nienhuys - [EMAIL PROTECTED] - http://www.xs4all.nl/~hanwen

LilyPond Software Design
 -- Code for Music Notation
http://www.lilypond-design.com



___
lilypond-devel mailing list
lilypond-devel@gnu.org
http://lists.gnu.org/mailman/listinfo/lilypond-devel