Hi Deri,
thank you for taking the iniative. I spent a whole weekend without computer.
Unfortunately, the font I suggested does not work either.
The problem is exemplified by the ligature rgy-. on the second line of
the original text, quite at the beginning. The base letter here is ག g,
with a superscript ར r- and a subscript ཡ -y- . (I showed the base
forms). The correct composition is: རྒྱ As you can see, the r is reduced
to it top stroke, and the subscript -y- is reduced to a swoosh-like
angle which is also available at code point U+0FB1 TIBETAN SUBJOINED
LETTER YA.
The complete ligatures are *not* part of the Unicode standard for
Tibetan since it is not the aim to encode the appearance; the semantics
gets encoded. In the late 1990s, I participated in the standard work for
ISO/Unicode of several Asian languages and scripts.
The problem present in Tom's non-ok files is the same as in Deri's
example. I'll have to dig into the ligature tables which should make the
right selection of glyphs why they are not triggered as expected.
In the flawed files, it is visible that the base forms of these latters
are stacked in one place, but without selection of proper glyph shape.
Tibetan script vertical stacks are built in a quite regular fashion. It
is possible to completely rely on Latin (ASCII) input and have the
ligature mechanism of the font perform the necessary acrobatics.
Long before Unicode was widely accepted and in use, I had written such a
system for TeX/LaTeX: https://ctan.org/pkg/ctib4tex
The ligature in question appears on line no. 2 of page 8 of the pdf
documentation. §3. 2 Known transliteration problems" on p. 5 of said
document demonstrates the problem of consonant clusters which may lead
to incorrect ligatures.
At that time, I still relied on Metafont sources. I never reworked the
package for Unicode fonts, but on the other hand this is an incentive to
create a Latin-input driven Tibetan macro package for groff which also
accepts native Tibetan as input.
May I kindly ask you to post the source code of your trial file? Then I
can dig deeper into the ligature problem.
Best regards,
Oliver.
On 21/01/2024 18:48, Deri wrote:
On Sunday, 21 January 2024 00:28:42 GMT Oliver Corff via wrote:
Hi Tom,
བཀྲ་ཤིས་བདེ་ལེགས།
(sorry, I forgot the shad)
have you tried the font TibMachUni-1.901b.ttf? It is available via
package manager e.g. in Fedora 39.
Can post a source file for your examples, please? I may try in tomorrow.
Best regards,
Oliver.
On 21/01/2024 00:09, Tom wrote:
Hi,
I did typeset few books in Heirloom Troff with quite good outcome.
For next book I need to use Tibetan font and unfortunately I can't
make work in Heirloom Troff and Groff, only Neatroff does work.
Right now I'm eager to completely switch to Groff. I can make any font
work in Groff but not Tibetan. To my basic knowledge, I guess it is
all about blwm and blws not accessible. There are only few Tibetan
fonts with complex glyph composition which works in Neatroff. In groff
I have managed only BabelStoneTibetan to display but several glyphs
doesn't compose.
For viewing I have attached Groff and Neatroff pdfs.
YagpoTibetanUni: 100% composing in Neatroff but 100% failed in Groff
and Heirloom troff.
BabelStoneTibetan: 100% failed in Groff and Neatroff.
I have tried various tibetan fonts, as well NotoSerifTibetan present
in linux repositories. All those fonts don't work and groff yields:
troff:<standard input>:20: warning: special character 'u0F04' not defined
troff:<standard input>:20: warning: special character 'u0F05' not defined
troff:<standard input>:20: warning: special character 'u0F0D' not defined
...
If I understand enough, the warnings are about missing glyph mappings.
Would you mind to have a look and check if Tibetan fonts are possible
to make work in Groff ? I appreciate for any hints, and direction I
can follow to make it happen.
Regards,
Tom
Hi Tom,
Using the font suggested by Oliver (after installing it with Peter Schafter's
install-font.sh script) and using the latest unicode aware pdf generation
version of groff (awaiting release by Branden), with this script:-
.ft TibetanR \" What I called Oliver's suggested font
.ps 24
.vs 28
.pdfbookmark 1 བཀྲ་ཤིས་བདེ་ལེགས།
བཀྲ་ཤིས་བདེ་ལེགས།
.br
.ad l
༄༅། །དབུ་མ་རྩ་བའི་ཚིག་ལེའུར་བྱས་པ་ཤེས་རབ་ཅེས་བྱ་བ་བཞུགས་སོ།
ཧཱུ༔ ཨོ་རྒྱན་ཡུལ་གྱི་ནུབ་བྱང་མཚམས༔
It produced the attached.
Cheers
Deri
--
Dr. Oliver Corff
Wittelsbacherstr. 5A
10707 Berlin
GERMANY
Tel.: +49-30-85727260
mailto:oliver.co...@email.de
