[Aleph] ovp2ovf, font_install_unicode and the Mac

Marc Wilhelm Küster klists at saphor.de
Mon Jul 18 17:21:51 CEST 2005


Dear Joël,

>Le lundi  4 juillet 2005, Marc Wilhelm Küster a écrit :
>  
>
>>                                                                   Does
>>anybody of you have by any chance access to either the old code or a Mac
>>binary (or, even better, of course, the news that everything is fixed
>>now in some ovp2ovf > 2.0)?
>>    
>>
>
>I also had such trouble with several versions of ovp2ovf. On FreeBSD 5.0,
>it worked fine, then on FreeBSD 5.1, there was a bug that prevented ovp2ovf
>from working at all, the reason was that font utilities had been completely
>rewritten in C whereas previous versions were based on Knuth's vptovf in
>Web. I filled a bug report on Debian since I had the same trouble on my
>Linux box, and proposed a patch that fixed the most visible problems <URL:
>http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=229184 >. This patch and
>some others seem to have been applied in teTeX since that time. But there
>are still some trouble, especially with huge fonts as a friend of mine
>found when dealing with a Chinese font (for the anecdote, we were working
>on the students' directory of our School, any student has a form in it with
>some personal data and some fields are filled with quotations or motto, the
>biggest part of which is in french or in english, the remaining part being
>in some exotic languages : ancient greek, russian, chinese, japanese,
>hindi, elfic... ; this is the reason why we used lambda/lamed). Here is the
>mail that he sent to J. Plaice last year (the patch is attached to my mail) :
>  
>
Many thanks for this patch! It does indeed resolve the issue admirably.
The only remaining restrictions of ovp2ovf are relatively easy to
overcome in my particular context:

  * lack of support for ofmlevel 1: use ofmlevel 0 which, due to your
patches, now works wonderfully. The price are unnecessarily large ofm
files, but that's acceptable

  * the lexer assumes a hex must be composed of 0-9A-F, whereas the old
ovp2ovf supported 0-9A-Fa-f. Again, easy to avoid (or to fix in the
lexer, for that matter).

Your patch is unfortunately not applied in its entirety in the current
texlive source code
(http://www.ctan.org/tex-archive/systems/texlive/Source/source.tar.bz2)
nor does it seem to be active in the current tetex 3.x at least on the
mac using fink. The tetex 3 distribution in the Debian experimental
branch does, for whatever strange reason, not contain ovp2ovf so I
cannot validate the status there. The tetex 3 source code also contains
only a partial patch --- with the result that both the texlive and the
tetex 3 ovp2ovf don't work, not even if tested with omlgc.ovp. With
these experiences I'd encourage every package maintainer to apply your
patch as is.

It would be equally desirable that the distributions contain the old,
web2c-version of the font utilities. While I fully appreciate the
advantages of a lex/yacc-based implementation, it would greatly
facilitate life if the old versions were dropped only when working
replacements are in place.

Again many thanks for your help! Now, let's hope that your full patch
actually makes it quickly into the major distributions.

Best regards,

Marc



More information about the Aleph mailing list