Re: [XeTeX] problem rendering unicode Devanagari dependent vowel signs and ligatures

[Richard Cobbe]

On Wed, Feb 02, 2011 at 08:07:24PM +0100, Dominik Wujastyk wrote:
> I've done some tests on Leo's first problem (see below), and I also cannot
> find a way to separate, e.g., the visarga sign (similar to a Latin colon)
> from the dotted-circle that is used in Unicode font charts to signify that
> the visarga always follows a vowel.

I'm no expert, but it looks to me as though results will vary with the font
used.  My first clue to this, oddly enough, was the fact that both the
visarga and the anusvara in your email, Dominik, displayed without the
dotted circle.

I tried the following on a Mac OS 10.6 system:

\documentclass{article}

\usepackage{fontspec}
\usepackage{xltxtra}
\usepackage{xunicode}

\newfontfamily{\SktAFont}[Mapping=velthuis-sanskrit]{Devanagari MT}

\newfontfamily{\SktFont}[Mapping=velthuis-sanskrit,Script=Devanagari]{Sanskrit 
2003}

\begin{document}

Devanagari MT: {\SktAFont .h .m atrya}

Sanskrit 2003: {\SktFont .h .m atrya}

\end{document}

Devanagari MT displays both the anusvara and the visarga without dotted
circles; Sanskrit 2003 displays them with.  (I'm guessing from this that
Apple's Terminal.app, in which I run my mail client, uses Devanagari MT for
the relevant range of Unicode code points, but I have no idea how OS X
chooses fonts for non-Ascii characters.)

> On 25 January 2011 11:16, Leo Brouwer  wrote:
>
> > Secondly, it seems Xetex doesn't render the common ligature त्र्य correctly
> > for most freely available Devanagari fonts.

I wonder how much of this is XeTeX and how much of this is an issue of font
encoding and design, although I know very little about the internals of
either.

I can say, Leo, that I was certainly able to reproduce your results:
Sanskrit 2003 doesn't display the trya conjunct correctly, but Devanagari
MT does.  I suspect that font isn't available on non-Mac platforms, though,
so that's of limited utility to you.

Richard


--
Subscriptions, Archive, and List information, etc.:
  http://tug.org/mailman/listinfo/xetex

Re: [XeTeX] How can I use IEEEtran class with XeLaTeX?

[Shrisha Rao]

El ene 31, 2011, a las 9:59 p.m., Maxim Cournoyer escribió:

> Just to be clear, in the present case, I'm _not_ planning to submit a paper 
> for acceptance at the IEEE;
> What I need to do is to write an assignment in the IEEE format for my 
> university. It must be written in french.

That should be easy enough to do with just the article class and two column 
format, with margins and other things set similar to what IEEEtran provides.  
There is no need to start with IEEEtran/bare_jrnl.tex as such.  I know some 
people who write IEEE conference papers this way too, because of package 
conflicts and such with IEEEtran.

Regards,

Shrisha Rao




--
Subscriptions, Archive, and List information, etc.:
  http://tug.org/mailman/listinfo/xetex

Re: [XeTeX] Unaligned combining diacritic

[Pander]

On 2011-02-02 22:21, maxwell wrote:
> On Wed, 02 Feb 2011 16:16:13 -0500, maxwell 
> wrote:
>> On Wed, 02 Feb 2011 15:13:03 +0100, Pander
> 
>> wrote:
>>> For FreeMono I can get combining diacritic mark exactly over the
>>> previous glyph but this does not work with FreeSerif, FreeMono DejaVu
>>> Serif, DejaVu Sans and DejaVu Sans Mono. Please see the attachement.
>>>
>>> Is this a font issue, fontspec issue or (Xe)TeX issue?
>>
>> For DejaVu, I believe this is the bug reports I submitted here:
>> https://bugs.freedesktop.org//show_bug.cgi?id=26940
>> https://bugs.freedesktop.org//show_bug.cgi?id=26941
>> See also
>>https://bugs.freedesktop.org//show_bug.cgi?id=28475
>> marked as a dupe of 26941.  I don't believe anything has been done on
>> either of these bug reports.
> 
> 
> P.S. See also
>http://www.mail-archive.com/xetex@tug.org/msg00016.html
> 
>   Mike Maxwell (who will some day learn not to hit "send" until he's done)

:)

Thanks all

Good news for FreeFonts as fixes are in CVS:
  https://savannah.gnu.org/bugs/?32326
  https://savannah.gnu.org/bugs/?32327

> 
> 
> --
> Subscriptions, Archive, and List information, etc.:
>   http://tug.org/mailman/listinfo/xetex



--
Subscriptions, Archive, and List information, etc.:
  http://tug.org/mailman/listinfo/xetex

[XeTeX] Velthuis to Roman translit

[Neal Delmonico]

Hi All,

Is there any easy way to get from Velthuis Devanagari encoding to Roman  
transliteration?  I have lots of documents in Velthuis that I would like  
to switch to Roman transliteration sometimes without having to type them  
in again.  If there is a way to just substitute some LaTex codes, that  
would be tremendous.


Thanks,

Neal

--
Using Opera's revolutionary email client: http://www.opera.com/mail/


--
Subscriptions, Archive, and List information, etc.:
 http://tug.org/mailman/listinfo/xetex

Re: [XeTeX] Unaligned combining diacritic

[maxwell]

On Wed, 02 Feb 2011 16:16:13 -0500, maxwell 
wrote:
> On Wed, 02 Feb 2011 15:13:03 +0100, Pander

> wrote:
>> For FreeMono I can get combining diacritic mark exactly over the
>> previous glyph but this does not work with FreeSerif, FreeMono DejaVu
>> Serif, DejaVu Sans and DejaVu Sans Mono. Please see the attachement.
>> 
>> Is this a font issue, fontspec issue or (Xe)TeX issue?
> 
> For DejaVu, I believe this is the bug reports I submitted here:
> https://bugs.freedesktop.org//show_bug.cgi?id=26940
> https://bugs.freedesktop.org//show_bug.cgi?id=26941
> See also
>https://bugs.freedesktop.org//show_bug.cgi?id=28475
> marked as a dupe of 26941.  I don't believe anything has been done on
> either of these bug reports.


P.S. See also
   http://www.mail-archive.com/xetex@tug.org/msg00016.html

  Mike Maxwell (who will some day learn not to hit "send" until he's done)


--
Subscriptions, Archive, and List information, etc.:
  http://tug.org/mailman/listinfo/xetex

Re: [XeTeX] Unaligned combining diacritic

[maxwell]

On Wed, 02 Feb 2011 15:13:03 +0100, Pander 
wrote:
> For FreeMono I can get combining diacritic mark exactly over the
> previous glyph but this does not work with FreeSerif, FreeMono DejaVu
> Serif, DejaVu Sans and DejaVu Sans Mono. Please see the attachement.
> 
> Is this a font issue, fontspec issue or (Xe)TeX issue?

For DejaVu, I believe this is the bug reports I submitted here:
https://bugs.freedesktop.org//show_bug.cgi?id=26940
https://bugs.freedesktop.org//show_bug.cgi?id=26941
See also
   https://bugs.freedesktop.org//show_bug.cgi?id=28475
marked as a dupe of 26941.  I don't believe anything has been done on
either of these bug reports.

   Mike Maxwell


--
Subscriptions, Archive, and List information, etc.:
  http://tug.org/mailman/listinfo/xetex

[XeTeX] XeTeX's (old) vbox model

[Simon Cozens]

Hello,
This is a bit of an odd question, I think, but please humour me. I read 
from
an old mailing list message
(http://www.tug.org/pipermail/xetex/2005-March/002025.html) that XeTeX
formerly did not retrieve the height and depth of glyphs when building boxes.
I tried the code in that thread, and found that today it *is* certainly
getting those metrics - it was reporting height and depth correctly for a
range of glyphs I tried. So obviously there was a change there at some point.

However, I'm intrigued as to how, in the past, XeTeX put together 
vertical
lists, given that you would need to know height and depth measurements to
apply interline glue: (from the TeXBook)

Here are the exact rules by which TEX calculates the interline glue
between boxes: Assume that a new box of height h (not a rule box) is
about to be appended to the bottom of the current vertical list, and
let \prevdepth = p, \lineskiplimit = l, \baselineskip = (b plus y
minus z). If p <=- -1000 pt, no interline glue is added. Otherwise
if b - p - h >= l, the interline glue `b - p - h plus y minus z)
will be appended just above the new box. Otherwise the \lineskip
glue will be appended. Finally, \prevdepth is set to the depth of
the new box.

Did it simply use a fixed height and depth per font size, essentially 
using a
fixed leading model? If so, how did it calculate those values? And how does it
do it now - I'm presuming it gets metrics through FreeType; does it use
control boxes (FT_Glyph_Get_CBox) or the exact bounding box? 
(FT_Outline_Get_BBox)

(The reason I want to know is somewhat arcane, and somewhat 
embarrassing.
I've accidentally written a typesetting system, and it's a fun thing to hack
on and I want to keep toying with it. Although it's probably unnecessary, I
believe it's not totally pointless because it's embedded in a high-level
programming language and so it's very easily scriptable. And my goodness I'm
learning a lot about how TrueType/OpenType works. The slightly longer story is
at http://www.simon-cozens.org/content/typesetting-perl)

Thanks,
Simon


--
Subscriptions, Archive, and List information, etc.:
  http://tug.org/mailman/listinfo/xetex

Re: [XeTeX] problem rendering unicode Devanagari dependent vowel signs and ligatures

[Dominik Wujastyk]

I've done some tests on Leo's first problem (see below), and I also cannot
find a way to separate, e.g., the visarga sign (similar to a Latin colon)
from the dotted-circle that is used in Unicode font charts to signify that
the visarga always follows a vowel.

Like this:  ः
If I type the visarga after a vowel, the dotted circle disappears: अः
So, I conjecture, the visarga itself *is* somewhere in the font, and it
should be possible somehow to get at it, for example to discuss it as a
separate character, or teach it to beginners.

It's an identical problem with anusvara:  ं but अं

Similarly, with the Devanagari diacritical accent-like markers that are used
to signify vowels after consonants, the flags on के कै को कौ, or ते तै तो
तौ, etc., apparently can't be typed separately.

I've racked my brains and done several experiments, to no avail.  I think
Leo has put his finger on something here that is probably a stupid error in
the Unicode standard.

Any insights?

Dominik


On 25 January 2011 11:16, Leo Brouwer  wrote:

> Greetings,
>
> I'm working on producing a critical edition of a Sanskrit text using
> XeLaTex. In the edition, I wan to report cancellation marks in the
> manuscript with multiplication signs around the cancelled akshara, e.g. xकx.
> This works fine for consonants, but I can't get XeTex to render dependent
> vowels without the dotted circle. I've tried inputting it in every possible
> combination with a space (U+0020) and/or Zero Width Joiner (U+200D), but
> nothing seems to work. Is there any way to make XeTex display a dependent
> vowel without the circle?
>
> Secondly, it seems Xetex doesn't render the common ligature त्र्य correctly
> for most freely available Devanagari fonts. I can only get shiDeva, Nakula,
> & Sahadeva to display it. Sanskrit 2003 has dedicated characters for
> ligatures in its Private Use Area, so it's possible (but time-consuming) to
> manually input them, but none of the other fonts do. In the attached sample
> I've only included the fonts I find most attractive, but all the others
> available on the Wazu font 
> gallery
>  have the same problem. Ideally, I want to use a font like Surekh or
> Yogesh which includes a real bold font. With these 2 fonts I've noticed an
> additional problem: if ligatures with a superscribed "r" are preceded by a
> dependent vowel "i," the "r" gets moved to the "i." Is it possible to get
> these fonts to display correctly, or are they simply unusable in XeTex?
>
> Here's the tex input for my sample:
>
> \documentclass{article}
>
> \usepackage{fontspec, xunicode, xltxtra}
>
> \usepackage{polyglossia}
>
> \setdefaultlanguage{sanskrit}
>
> \setotherlanguage{english}
>
> \setmainfont{Charis SIL}
>
> \newfontfamily\sanskritfont [Script=Devanagari]{Sanskrit 2003}
>
> \begin{document}
>
> \textenglish{Sanskrit 2003:} स्वातन्त्र्यम्, स्वात\char"E178 म् ,
> त्र्यक्षः, \char"E11Cक्षः $\times$\char"200D\char"093E$\times$
>
>
>  \textenglish {AA Nagari Shree L3:}
> \fontspec[Script=Devanagari]{AA_NAGARI_SHREE_L3}स्वातन्त्र्यम् , त्र्यक्षः,
> $\times$\char"200D\char"093E$\times$
>
>
>  \textenglish{CDAC-GISTYogesh:}
> \fontspec[Script=Devanagari]{CDAC-GISTYogesh}स्वातन्त्र्यम् , त्र्यक्षः,
> विसर्गः, $\times$\char"0020\char"093E$\times$
>
>
>  \textenglish{CDAC-GISTSurekh:}
> \fontspec[Script=Devanagari]{CDAC-GISTSurekh}स्वातन्त्र्यम् , त्र्यक्षः,
> विसर्गः, $\times$\char"0020\char"200D\char"093E$\times$
>
>
>  \textenglish{shiDeva:}
> \fontspec[Script=Devanagari]{shiDeva}स्वातन्त्र्यम् , त्र्यक्षः,
> $\times$\char"200D\char"093E$\times$
>
>
>  \textenglish{Nakula:} \fontspec[Script=Devanagari]{Nakula}स्वातन्त्र्यम्
> , त्र्यक्षः, $\times$\char"200D\char"093E$\times$
>
>
>  \textenglish{Sahadeva:}
> \fontspec[Script=Devanagari]{Sahadeva}स्वातन्त्र्यम् , त्र्यक्षः,
> $\times$\char"200D\char"093E$\times$
>
> \end{document}
>
> Thank you for your help,
>
> Leonard Brouwer
>
>
>
>
>
> --
> Subscriptions, Archive, and List information, etc.:
>  http://tug.org/mailman/listinfo/xetex
>
>


--
Subscriptions, Archive, and List information, etc.:
  http://tug.org/mailman/listinfo/xetex

Re: [XeTeX] Unaligned combining diacritic

[Dominik Wujastyk]

> Junicode is one open-source font that does support combining marks well, if
> you need to use them.  Charis SIL is another.
>

This is my experience too.  These two fonts have been carefully tuned for
accent positioning.

The same is true of John Smith's IndUni fonts, at
http://bombay.indology.info.

Dominik


--
Subscriptions, Archive, and List information, etc.:
  http://tug.org/mailman/listinfo/xetex

Re: [XeTeX] substitutions for missing or wrong glyphs

[Susan Dittmar]

Dear Michael,

thanks a lot for your help. It works partially. The line 
"\XeTeXcharclass `\-- 4" results in an error, perhaps because of the use of
-- instead of the unicode endash. But I think I can work on from there.
At least I already managed to make it work for the german quotation marks.
Now I need to find out what those commands do exactly, but I feel confident
to find further information in the 'net now that I know what to look out
for.

Thanks again,

Susan


--
Subscriptions, Archive, and List information, etc.:
  http://tug.org/mailman/listinfo/xetex

Re: [XeTeX] substitutions for missing or wrong glyphs

[Fr. Michael Gilmary, mma]

On Feb 2 AD 2011, at 4:40 AM, Susan Dittmar wrote:



I would like to use a font (several in fact, some of which are true  
type,
others open type) which does not include all glyphs I need.  For  
example
the endash and emdash are missing, as are the german quotation  
marks. Is
there a way, other than hacking the font file itself, to tell xetex  
to take
those glyphs from another font? I know that might look ugly, but  
it's still
less ugly than the 'missing glyph' glyph or another completely  
wrong glyph.




Hi Susan:

This seems to be a perfect case for XeTeX interchartoks!

This works, but you'll want to change the Scale for the \tgb font  
I've used here (I made an exaggerated example so it could be seen to  
work).




%
\documentclass[11pt]{article}

\XeTeXinterchartokenstate = 1

\usepackage{fontspec, realscripts}
\usepackage{polyglossia}
\setdefaultlanguage[variant=us]{english}
\defaultfontfeatures{Ligatures=TeX}
\setmainfont[Numbers={OldStyle,Proportional}, Contextuals=NoAlternate] 
{Garamond Premier Pro}


\newfontfamily\tgb[Scale=2]{TeX Gyre Bonum}

\XeTeXinterchartoks 0 4 = {\tgb}
\XeTeXinterchartoks 4 0 = {\normalfont}
\XeTeXcharclass `\– 4
\XeTeXcharclass `\— 4
\XeTeXinterchartoks 255 0 = {\normalfont}
\XeTeXinterchartoks 0 255 = {\normalfont}
\XeTeXinterchartoks 255 4 = {\tgb}
\XeTeXinterchartoks 4 255 = {\normalfont}


\begin{document}
Try this — I mean does it really work from 2–3 times?
\end{document}
%



HTH.



United in adoration of Jesus,



fr. michael gilmary, mma

Most Holy Trinity Monastery
67 Dugway Road
Petersham, MA 01366

www.MaroniteMonks.org







--
Subscriptions, Archive, and List information, etc.:
 http://tug.org/mailman/listinfo/xetex

Re: [XeTeX] substitutions for missing or wrong glyphs

[Fr. Michael Gilmary, mma]

This works, but you'll want to change the Scale for the \tgb font  
I've used here (I made an exaggerated example so it could be seen  
to work).



What I mean to say was: change the scale to [Scale=MatchLowercase] so  
it will be close to your main font.





United in adoration of Jesus,



fr. michael gilmary, mma

Most Holy Trinity Monastery
67 Dugway Road
Petersham, MA 01366

www.MaroniteMonks.org






--
Subscriptions, Archive, and List information, etc.:
 http://tug.org/mailman/listinfo/xetex

Re: [XeTeX] substitutions for missing or wrong glyphs

[Peter Dyballa]

Am 02.02.2011 um 10:40 schrieb Susan Dittmar:

Is there a way, other than hacking the font file itself, to tell  
xetex to take those glyphs from another font?



Yes. Make these characters active and bind them to some font(s).  
Examples where given on this list during the last few months.


--
Mit friedvollen Grüßen

  Pete

The box said "Use Windows 95 or better," so I got a Macintosh.




--
Subscriptions, Archive, and List information, etc.:
 http://tug.org/mailman/listinfo/xetex

Re: [XeTeX] Unaligned combining diacritic

[Peter Dyballa]

Am 02.02.2011 um 15:13 schrieb Pander:


Is this a font issue, fontspec issue or (Xe)TeX issue?



Certainly! When the same algorithm produces with different input fonts  
so different output results, then it's likely the fault of the fonts.  
(Just because the members of a font family share one name it's not  
guaranteed that they also share the same font coding basis.)


--
Greetings

  Pete
You can never know too little of what is not worth knowing at all.
– Anon.




--
Subscriptions, Archive, and List information, etc.:
 http://tug.org/mailman/listinfo/xetex

Re: [XeTeX] Unaligned combining diacritic

[David J. Perry]

It's a font issue.  Unless a font maker specifically sets up each combining 
mark to fit correctly over all the possible base characters through the 
OpenType mark to base feature, they won't work right.  A partial exception 
might be a monospaced font; since all the characters are the same width, 
things will work better than in a proportionally spaced font.  But even in a 
monspaced font, you won't get perfection without some manual adjustments on 
the font maker's part; e.g., an acute or grave that is mathematically 
centered over a character probably won't look right, and there is also the 
need to replace a dotted i with the dotless version before applying accents.


Junicode is one open-source font that does support combining marks well, if 
you need to use them.  Charis SIL is another.


David

- Original Message - 
From: "Pander" 

To: "xetex" 
Sent: Wednesday, February 02, 2011 9:13 AM
Subject: [XeTeX] Unaligned combining diacritic



Hi all,

For FreeMono I can get combining diacritic mark exactly over the
previous glyph but this does not work with FreeSerif, FreeMono DejaVu
Serif, DejaVu Sans and DejaVu Sans Mono. Please see the attachement.

Is this a font issue, fontspec issue or (Xe)TeX issue?

Thanks,

Pander










--
Subscriptions, Archive, and List information, etc.:
 http://tug.org/mailman/listinfo/xetex





--
Subscriptions, Archive, and List information, etc.:
 http://tug.org/mailman/listinfo/xetex

[XeTeX] substitutions for missing or wrong glyphs

[Susan Dittmar]

Hi folks,

I guess this is a FAQ, but my search foo seems to have left me today...

I would like to use a font (several in fact, some of which are true type,
others open type) which does not include all glyphs I need.  For example
the endash and emdash are missing, as are the german quotation marks. Is
there a way, other than hacking the font file itself, to tell xetex to take
those glyphs from another font? I know that might look ugly, but it's still
less ugly than the 'missing glyph' glyph or another completely wrong glyph.

Thanks for any hints,

Susan


--
Subscriptions, Archive, and List information, etc.:
  http://tug.org/mailman/listinfo/xetex

[XeTeX] Unaligned combining diacritic

[Pander]

Hi all,

For FreeMono I can get combining diacritic mark exactly over the
previous glyph but this does not work with FreeSerif, FreeMono DejaVu
Serif, DejaVu Sans and DejaVu Sans Mono. Please see the attachement.

Is this a font issue, fontspec issue or (Xe)TeX issue?

Thanks,

Pander
\documentclass{article}
\usepackage{xltxtra}
\begin{document}
\noindent\fontspec{FreeMono}◌̀ à\\%U+25CC U+0300
◌́\\%U+25CC U+0301
◌̂ â\\%U+25CC U+0302
◌̃\\%U+25CC U+0303
◌̄\\%U+25CC U+0304
◌̅\\%U+25CC U+0305
\noindent\fontspec{FreeSerif}◌̀ à\\%U+25CC U+0300
◌́\\%U+25CC U+0301
◌̂ â\\%U+25CC U+0302
◌̃\\%U+25CC U+0303
◌̄\\%U+25CC U+0304
◌̅\\%U+25CC U+0305
\noindent\fontspec{FreeSans}◌̀ à\\%U+25CC U+0300
◌́\\%U+25CC U+0301
◌̂ â\\%U+25CC U+0302
◌̃\\%U+25CC U+0303
◌̄\\%U+25CC U+0304
◌̅\\%U+25CC U+0305
\noindent\fontspec{DejaVu Sans}◌̀ à\\%U+25CC U+0300
◌́\\%U+25CC U+0301
◌̂ â\\%U+25CC U+0302
◌̃\\%U+25CC U+0303
◌̄\\%U+25CC U+0304
◌̅\\%U+25CC U+0305
\noindent\fontspec{DejaVu Serif}◌̀ à\\%U+25CC U+0300
◌́\\%U+25CC U+0301
◌̂ â\\%U+25CC U+0302
◌̃\\%U+25CC U+0303
◌̄\\%U+25CC U+0304
◌̅\\%U+25CC U+0305
\noindent\fontspec{DejaVu Sans Mono}◌̀ à\\%U+25CC U+0300
◌́\\%U+25CC U+0301
◌̂ â\\%U+25CC U+0302
◌̃\\%U+25CC U+0303
◌̄\\%U+25CC U+0304
◌̅\\%U+25CC U+0305
\end{document}


--
Subscriptions, Archive, and List information, etc.:
  http://tug.org/mailman/listinfo/xetex

Re: [XeTeX] How to load pstricks?

[Sven Siegmund]

Actually I am not sure what helped. Before I tried to uninstall and
reinstall them I copied them manually in their respected location from
CTAN from the path you mentioned. They were there, but I thought I
must rebuild the FNDB and other stuff, so I also uninstalled the
package and installed it again. then it worked -- but maybe because of
the manual copying.

S.


--
Subscriptions, Archive, and List information, etc.:
  http://tug.org/mailman/listinfo/xetex

Re: [XeTeX] How to load pstricks?

[Ulrike Fischer]

Am Tue, 1 Feb 2011 21:12:20 +0100 schrieb Sven Siegmund:

> As Ulrike pointed out, this has nothing to do with XeTeX, it is a
> Miktex-specific packaging bug and can be currently solved by
> uninstalling and reinstaling the pstricks package.

? Does that mean that the problem was solved by reinstalling
pstricks with the miktex package manager? I can't reproduce this. I
still have to install the missing config files manually. 


-- 
Ulrike Fischer 



--
Subscriptions, Archive, and List information, etc.:
  http://tug.org/mailman/listinfo/xetex