Re: [I18n]i18n mechanism

2002-02-19 Thread Tomohiro KUBOTA 

Hi,

At Mon, 18 Feb 2002 13:54:56 +0530 (IST),
Aravind Menon wrote:

>   I am working on a project that involves changing Xlib to support Indian
> locales. I am quite new to i18n and would like to get technical information
> about the existing i18n framework, specifically the internal mechanisms and
> architecture.
>   Where can I find the above information.

Sorry, people here seem not to be interested in supporting "complex"
languages such as Arab/Hebrew and Indic in Xlib level or some other
similarly basic level.  Though I am interested in, I don't have enough
skill (I am relatively a newcomer to this list).

Though there are a few projects to support "complex" languages like
Pango (http://www.pango.org/ it is not living now?), I think ST project
(http://stsf.sourceforge.net/) is the one which you may like best.

---
Tomohiro KUBOTA <[EMAIL PROTECTED]>
http://www.debian.or.jp/~kubota/
"Introduction to I18N"  http://www.debian.org/doc/manuals/intro-i18n/
___
I18n mailing list
[EMAIL PROTECTED]
http://XFree86.Org/mailman/listinfo/i18n

Re: [I18n]i18n mechanism

2002-02-19 Thread Aravind Menon 

Hi,
I do not exactly wish to do the entire handling of Indic script. We have
already modified a part of Xlib (the XDrawString and similar functions) to
handle input in a different manner when it is ISCII encoded (ISCII is an
8-bit encoding for Indic scripts and is compatible with ascii). This
modification is, at best, ad hoc. 
We wanted to introduce UTF-8 support for Indic scripts by
modifying lcUTF8.c and introducing appropriate converters in lcUniConv. These
converters convert UTF-8 encoded text into ISCII. The ISCII text is then
appropriately handled by the modified XDrawString.
I introduced a new locale 'DEVANAGARI', and wrote a UTF to ISCII
converter for it. When XmbDrawString is called with UTF encoded DEVANAGARI text,
the conversion is working fine, but the correct font is not getting loaded. In
fact, XCreateFontSet returns DEVANAGARI as a missing charset inspite of the fact
the required font is installed in the system. 
Any suggestions on how to proceed. 

Thanks
Aravind

On Tue, 19 Feb
2002, Tomohiro KUBOTA wrote:

> 
> > I am working on a project that involves changing Xlib to support Indian
> > locales. I am quite new to i18n and would like to get technical information
> > about the existing i18n framework, specifically the internal mechanisms and
> > architecture.
> > Where can I find the above information.
> 
> Sorry, people here seem not to be interested in supporting "complex"
> languages such as Arab/Hebrew and Indic in Xlib level or some other
> similarly basic level.  Though I am interested in, I don't have enough
> skill (I am relatively a newcomer to this list).
> 


___
I18n mailing list
[EMAIL PROTECTED]
http://XFree86.Org/mailman/listinfo/i18n

Re: [I18n]i18n mechanism

2002-02-19 Thread Markus Kuhn 

If you want to use Indic, there are two things you should look at:

Pango
  http://www.pango.org/

ST
  http://stsf.sourceforge.net/

If you want to help with Indic support, I would focus on input methods.
Output methods are now becomming mostly locale independent with UTF-8,
so there is very little Indic-specific to do here.

Markus

-- 
Markus G. Kuhn, Computer Laboratory, University of Cambridge, UK
Email: mkuhn at acm.org,  WWW: 

___
I18n mailing list
[EMAIL PROTECTED]
http://XFree86.Org/mailman/listinfo/i18n

Re: [I18n]i18n mechanism

2002-02-19 Thread Akber Choudhry 

Aravind,
I am working on a similar project.  Where is the font
'installed'?  Also, what type of font is it? what is the OS?

Regards,
Akber

On Tue, 19 Feb 2002, Aravind Menon wrote:

> Hi,
>   I do not exactly wish to do the entire handling of Indic script. We have
> already modified a part of Xlib (the XDrawString and similar functions) to
> handle input in a different manner when it is ISCII encoded (ISCII is an
> 8-bit encoding for Indic scripts and is compatible with ascii). This
> modification is, at best, ad hoc.
>   We wanted to introduce UTF-8 support for Indic scripts by
> modifying lcUTF8.c and introducing appropriate converters in lcUniConv. These
> converters convert UTF-8 encoded text into ISCII. The ISCII text is then
> appropriately handled by the modified XDrawString.
>   I introduced a new locale 'DEVANAGARI', and wrote a UTF to ISCII
> converter for it. When XmbDrawString is called with UTF encoded DEVANAGARI text,
> the conversion is working fine, but the correct font is not getting loaded. In
> fact, XCreateFontSet returns DEVANAGARI as a missing charset inspite of the fact
> the required font is installed in the system.
>   Any suggestions on how to proceed.
>
> Thanks
> Aravind
>
> On Tue, 19 Feb
> 2002, Tomohiro KUBOTA wrote:
>
> >
> > >   I am working on a project that involves changing Xlib to support Indian
> > > locales. I am quite new to i18n and would like to get technical information
> > > about the existing i18n framework, specifically the internal mechanisms and
> > > architecture.
> > >   Where can I find the above information.
> >
> > Sorry, people here seem not to be interested in supporting "complex"
> > languages such as Arab/Hebrew and Indic in Xlib level or some other
> > similarly basic level.  Though I am interested in, I don't have enough
> > skill (I am relatively a newcomer to this list).
> >
>
>
> ___
> I18n mailing list
> [EMAIL PROTECTED]
> http://XFree86.Org/mailman/listinfo/i18n
>

-- 
Akber Choudhry
Dyanet Inc.
http://www.dyanet.com/

___
I18n mailing list
[EMAIL PROTECTED]
http://XFree86.Org/mailman/listinfo/i18n

Re: [I18n]i18n mechanism

2002-02-19 Thread Tomohiro KUBOTA 

Hi,

At Tue, 19 Feb 2002 19:51:54 +0530 (IST),
Aravind Menon wrote:

> 
> Hi,
>   I do not exactly wish to do the entire handling of Indic script. We have
> already modified a part of Xlib (the XDrawString and similar functions) to
> handle input in a different manner when it is ISCII encoded (ISCII is an
> 8-bit encoding for Indic scripts and is compatible with ascii). This
> modification is, at best, ad hoc. 

I imagine your work is related to character -> glyph conversion.
Is that right?

>   We wanted to introduce UTF-8 support for Indic scripts by
> modifying lcUTF8.c and introducing appropriate converters in lcUniConv. These
> converters convert UTF-8 encoded text into ISCII. The ISCII text is then
> appropriately handled by the modified XDrawString.


>   I introduced a new locale 'DEVANAGARI', and wrote a UTF to ISCII
> converter for it. When XmbDrawString is called with UTF encoded DEVANAGARI text,
> the conversion is working fine, but the correct font is not getting loaded. In
> fact, XCreateFontSet returns DEVANAGARI as a missing charset inspite of the fact
> the required font is installed in the system. 

What font name you use?  I imagine Devanagari will need more glyph
codepoints than character codepoints.  Since these X11 functions
assume the character encoding to be *glyph* index, I think some
radical redesign would be needed for X font mechanism to support
Devanagari.  It is because X's font mechanism including XFontStruct,
XFontSet, and Xft all assume that character and glyph correspond
one-to-one.

I once have seen Indic ISCII fonts which, in reality, have more glyphs
than ISCII codepoints.  We need to have some orthodox codepoint for
such expression forms.  And, I remember these fonts have '-*-iso8859-1'
names, which is a bad idea against the idea of FontSet.


---
Tomohiro KUBOTA <[EMAIL PROTECTED]>
http://www.debian.or.jp/~kubota/
"Introduction to I18N"  http://www.debian.org/doc/manuals/intro-i18n/
___
I18n mailing list
[EMAIL PROTECTED]
http://XFree86.Org/mailman/listinfo/i18n

Re: [I18n]i18n mechanism

2002-02-19 Thread Keith Packard 


Around 12 o'clock on Feb 20, Tomohiro KUBOTA wrote:

> What font name you use?  I imagine Devanagari will need more glyph
> codepoints than character codepoints.  Since these X11 functions
> assume the character encoding to be *glyph* index, I think some
> radical redesign would be needed for X font mechanism to support
> Devanagari.  It is because X's font mechanism including XFontStruct,
> XFontSet, and Xft all assume that character and glyph correspond
> one-to-one.

Xft exposes the underlying FreeType file so that apps can access the 
underlying font information as needed to perform the N unicode chars -> M 
glyphs needed for Devanagari layout. Pango uses this in it's 
implementation of arabic support, but it doesn't yet have Xft support for 
indic scripts.  I don't know how hard adding that will be; indic scripts 
aren't the easiest to typeset.

Xft also provides a simple Unicode layer on top of the glyph layer, but 
that is only a convenience for locales where a 1->1 mapping from Unicode 
char to glyph exists and where layout is straightforward.

Keith PackardXFree86 Core TeamCompaq Cambridge Research Lab


___
I18n mailing list
[EMAIL PROTECTED]
http://XFree86.Org/mailman/listinfo/i18n

Re: [I18n]i18n mechanism

2002-02-19 Thread Aravind Menon 

Hi,

On Wed, 20 Feb 2002, Tomohiro KUBOTA wrote:

> I imagine your work is related to character -> glyph conversion.
> Is that right?

Yes, we have written some parsing routines for conversion from ISCII to glyph

> > converter for it. When XmbDrawString is called with UTF encoded DEVANAGARI text,
> > the conversion is working fine, but the correct font is not getting loaded. In
> > fact, XCreateFontSet returns DEVANAGARI as a missing charset inspite of the fact
> > the required font is installed in the system. 
> 
> What font name you use?  I imagine Devanagari will need more glyph

I am using the font -mf-devanagari10-0-0-300-300--0-- and similar fonts.

The correct font is not getting loaded because Xlib looks for a font which has
the same XLC_FONTSET.font name specified in the XLC_LOCALE file for the locale.
ie if I specify XLC_FONTSET.font.primary as ISO8859-1, Xlib tries to load a font
with the extension iso8859-1.

So, how can I modify -mf-devanagari10-0-0-300-300--0-- so that its encoding
name is, say DEVANAGARI or something like that.

Regards

Aravind


___
I18n mailing list
[EMAIL PROTECTED]
http://XFree86.Org/mailman/listinfo/i18n