Luatex (actually Aleph) and Bidi_Mirrored chars in RTL mode

Luatex (actually Aleph) and Bidi_Mirrored chars in RTL mode

[Khaled Hosny]

[-- Attachment #1.1.1: Type: text/plain, Size: 457 bytes --]

Luatex does not mirror characters that has a Bidi_Mirrored property when
the text direction is set to RTL (TRT in Aleph), according to 
http://unicode.org/reports/tr9/#Mirroring, the different types of
parenthesis that has Bidi_Mirrored property should be mirreored in RTL
mode, but this isn't what I get. Is this a bug, feature, or am I missing
some thing?

Thanks,
 Khaled

-- 
 Khaled Hosny
 Arabic localizer and member of Arabeyes.org team

[-- Attachment #1.1.2: sample.tex --]
[-- Type: text/x-tex, Size: 440 bytes --]

% engine=luatex 
\noheaderandfooterlines

\starttext

 \definefontfeature
   [arabic]
   [analyze=yes, clig=yes, curs=yes, dlig=yes,
    fina=yes, init=yes, isol=yes, kern=yes,
    language=dflt, liga=yes, mark=yes, medi=yes,
    mkmk=yes, mode=node, rlig=yes, script=arab]

\font \Arabic = ScheherazadeRegOT*arabic at 24pt

\pardir TRT \textdir TRT

\Arabic
(قوس) و <قوس> و [قوس] و  «قوس»
\stoptext 

[-- Attachment #1.1.3: sample.pdf --]
[-- Type: application/pdf, Size: 10072 bytes --]

[-- Attachment #1.2: Digital signature --]
[-- Type: application/pgp-signature, Size: 189 bytes --]

[-- Attachment #2: Type: text/plain, Size: 487 bytes --]

___________________________________________________________________________________
If your question is of interest to others as well, please add an entry to the Wiki!

maillist : ntg-context@ntg.nl / http://www.ntg.nl/mailman/listinfo/ntg-context
webpage  : http://www.pragma-ade.nl / http://tex.aanhet.net
archive  : https://foundry.supelec.fr/projects/contextrev/
wiki     : http://contextgarden.net
___________________________________________________________________________________

Re: Luatex (actually Aleph) and Bidi_Mirrored chars in RTL mode

[Arthur Reutenauer]

	Hello,

>                                  Is this a bug, feature, or am I missing
> some thing?

  You're not missing anything, it's kind of a misfeature because the
Bidi_Mirrored property is not taken in account by ConTeXt (yet).
Happily enough, it's one of the things I'm sponsored by Google to
implement as a part of my Summer of Code project
(http://code.google.com/soc/2008/tex/appinfo.html?csaid=8BC22C657B7F0D0E)
-- of course the project is more general and there is a part on
bidirectional behaviour, but I didn't get to it yet.  In this case, I
guess the mirroring property should be added to char-def.lua and handled
accordingly; contrary to XeTeX, LuaTeX doesn't know it intrinsically, so
it has to be dealt with at the ConTeXt level.

  Incidentally, you may note that the guillemets are not mirrored
(although I observed some incoherent behaviour with the Geeza font on
the Mac).

	Arthur
___________________________________________________________________________________
If your question is of interest to others as well, please add an entry to the Wiki!

maillist : ntg-context@ntg.nl / http://www.ntg.nl/mailman/listinfo/ntg-context
webpage  : http://www.pragma-ade.nl / http://tex.aanhet.net
archive  : https://foundry.supelec.fr/projects/contextrev/
wiki     : http://contextgarden.net
___________________________________________________________________________________

Re: Luatex (actually Aleph) and Bidi_Mirrored chars in RTL mode

[Hans Hagen]

Arthur Reutenauer wrote:
> 	Hello,
> 
>>                                  Is this a bug, feature, or am I missing
>> some thing?
> 
>   You're not missing anything, it's kind of a misfeature because the
> Bidi_Mirrored property is not taken in account by ConTeXt (yet).
> Happily enough, it's one of the things I'm sponsored by Google to
> implement as a part of my Summer of Code project
> (http://code.google.com/soc/2008/tex/appinfo.html?csaid=8BC22C657B7F0D0E)
> -- of course the project is more general and there is a part on
> bidirectional behaviour, but I didn't get to it yet.  In this case, I
> guess the mirroring property should be added to char-def.lua and handled
> accordingly; contrary to XeTeX, LuaTeX doesn't know it intrinsically, so
> it has to be dealt with at the ConTeXt level.

this morning i added that info to the main table, but i'm still 
pondering about how to use if properly ... in tex normally there are 
explicit mode switches (in the source of the document) which gives 
ultimate control; there's also the issue of these pardir etc changes 
that then need to be injected into the node list;

anyhow, i'll look into it, but whatever solution i come up with, it has 
to be under user control; i don't want hard coded automatisms that then 
are hard to bypass (especially implicid properties can result in messy 
situations, unicode math is a candidate for that)



-----------------------------------------------------------------
                                           Hans Hagen | PRAGMA ADE
               Ridderstraat 27 | 8061 GH Hasselt | The Netherlands
      tel: 038 477 53 69 | fax: 038 477 53 74 | www.pragma-ade.com
                                              | www.pragma-pod.nl
-----------------------------------------------------------------
___________________________________________________________________________________
If your question is of interest to others as well, please add an entry to the Wiki!

maillist : ntg-context@ntg.nl / http://www.ntg.nl/mailman/listinfo/ntg-context
webpage  : http://www.pragma-ade.nl / http://tex.aanhet.net
archive  : https://foundry.supelec.fr/projects/contextrev/
wiki     : http://contextgarden.net
___________________________________________________________________________________

Re: Luatex (actually Aleph) and Bidi_Mirrored chars in RTL mode

[Khaled Hosny]

[-- Attachment #1.1: Type: text/plain, Size: 2673 bytes --]

On Mon, Jun 09, 2008 at 05:20:23PM +0200, Hans Hagen wrote:
> Arthur Reutenauer wrote:
> > 	Hello,
> > 
> >>                                  Is this a bug, feature, or am I missing
> >> some thing?
> > 
> >   You're not missing anything, it's kind of a misfeature because the
> > Bidi_Mirrored property is not taken in account by ConTeXt (yet).
> > Happily enough, it's one of the things I'm sponsored by Google to
> > implement as a part of my Summer of Code project
> > (http://code.google.com/soc/2008/tex/appinfo.html?csaid=8BC22C657B7F0D0E)
> > -- of course the project is more general and there is a part on
> > bidirectional behaviour, but I didn't get to it yet.  In this case, I
> > guess the mirroring property should be added to char-def.lua and handled
> > accordingly; contrary to XeTeX, LuaTeX doesn't know it intrinsically, so
> > it has to be dealt with at the ConTeXt level.
> 
> this morning i added that info to the main table, but i'm still 
> pondering about how to use if properly ... in tex normally there are 
> explicit mode switches (in the source of the document) which gives 
> ultimate control; there's also the issue of these pardir etc changes 
> that then need to be injected into the node list;
> 
> anyhow, i'll look into it, but whatever solution i come up with, it has 
> to be under user control; i don't want hard coded automatisms that then 
> are hard to bypass (especially implicid properties can result in messy 
> situations, unicode math is a candidate for that)

Giving the user more control is good idea, as long as the correct
behaviour is one of the options.

Regards,
 Khaled
> 
> 
> 
> -----------------------------------------------------------------
>                                            Hans Hagen | PRAGMA ADE
>                Ridderstraat 27 | 8061 GH Hasselt | The Netherlands
>       tel: 038 477 53 69 | fax: 038 477 53 74 | www.pragma-ade.com
>                                               | www.pragma-pod.nl
> -----------------------------------------------------------------
> ___________________________________________________________________________________
> If your question is of interest to others as well, please add an entry to the Wiki!
> 
> maillist : ntg-context@ntg.nl / http://www.ntg.nl/mailman/listinfo/ntg-context
> webpage  : http://www.pragma-ade.nl / http://tex.aanhet.net
> archive  : https://foundry.supelec.fr/projects/contextrev/
> wiki     : http://contextgarden.net
> ___________________________________________________________________________________

-- 
 Khaled Hosny
 Arabic localizer and member of Arabeyes.org team

[-- Attachment #1.2: Digital signature --]
[-- Type: application/pgp-signature, Size: 189 bytes --]

[-- Attachment #2: Type: text/plain, Size: 487 bytes --]

___________________________________________________________________________________
If your question is of interest to others as well, please add an entry to the Wiki!

maillist : ntg-context@ntg.nl / http://www.ntg.nl/mailman/listinfo/ntg-context
webpage  : http://www.pragma-ade.nl / http://tex.aanhet.net
archive  : https://foundry.supelec.fr/projects/contextrev/
wiki     : http://contextgarden.net
___________________________________________________________________________________

Re: Luatex (actually Aleph) and Bidi_Mirrored chars in RTL mode

[Hans Hagen]

Khaled Hosny wrote:
> Luatex does not mirror characters that has a Bidi_Mirrored property when
> the text direction is set to RTL (TRT in Aleph), according to 
> http://unicode.org/reports/tr9/#Mirroring, the different types of
> parenthesis that has Bidi_Mirrored property should be mirreored in RTL
> mode, but this isn't what I get. Is this a bug, feature, or am I missing
> some thing?

experimental in the beta

\setcharactermirroring[1]

no high level interface yet, i need to think of how to do such things as 
efficient as possible and prevent interference with font features and 
such (currently it's an attribute handler that pops in quite early)


-- 

-----------------------------------------------------------------
                                           Hans Hagen | PRAGMA ADE
               Ridderstraat 27 | 8061 GH Hasselt | The Netherlands
      tel: 038 477 53 69 | fax: 038 477 53 74 | www.pragma-ade.com
                                              | www.pragma-pod.nl
-----------------------------------------------------------------
___________________________________________________________________________________
If your question is of interest to others as well, please add an entry to the Wiki!

maillist : ntg-context@ntg.nl / http://www.ntg.nl/mailman/listinfo/ntg-context
webpage  : http://www.pragma-ade.nl / http://tex.aanhet.net
archive  : https://foundry.supelec.fr/projects/contextrev/
wiki     : http://contextgarden.net
___________________________________________________________________________________

Re: Luatex (actually Aleph) and Bidi_Mirrored chars in RTL mode

[Khaled Hosny]

[-- Attachment #1.1.1: Type: text/plain, Size: 2074 bytes --]

On Mon, Jun 09, 2008 at 11:08:09PM +0200, Hans Hagen wrote:
> Khaled Hosny wrote:
> > Luatex does not mirror characters that has a Bidi_Mirrored property when
> > the text direction is set to RTL (TRT in Aleph), according to 
> > http://unicode.org/reports/tr9/#Mirroring, the different types of
> > parenthesis that has Bidi_Mirrored property should be mirreored in RTL
> > mode, but this isn't what I get. Is this a bug, feature, or am I missing
> > some thing?
> 
> experimental in the beta
> 
> \setcharactermirroring[1]

It does work perfectly with unidirectional texts (RTL or LTR), but when
mixing bi-directional text, like Arabic text between brackets inside
English line, the closing bracket takes the direction of the embedded
text not the main line.
See the attached example.

Regards,
 Khaled

> 
> no high level interface yet, i need to think of how to do such things as 
> efficient as possible and prevent interference with font features and 
> such (currently it's an attribute handler that pops in quite early)
> 
> 
> -- 
> 
> -----------------------------------------------------------------
>                                            Hans Hagen | PRAGMA ADE
>                Ridderstraat 27 | 8061 GH Hasselt | The Netherlands
>       tel: 038 477 53 69 | fax: 038 477 53 74 | www.pragma-ade.com
>                                               | www.pragma-pod.nl
> -----------------------------------------------------------------
> ___________________________________________________________________________________
> If your question is of interest to others as well, please add an entry to the Wiki!
> 
> maillist : ntg-context@ntg.nl / http://www.ntg.nl/mailman/listinfo/ntg-context
> webpage  : http://www.pragma-ade.nl / http://tex.aanhet.net
> archive  : https://foundry.supelec.fr/projects/contextrev/
> wiki     : http://contextgarden.net
> ___________________________________________________________________________________

-- 
 Khaled Hosny
 Arabic localizer and member of Arabeyes.org team

[-- Attachment #1.1.2: mirroring.tex --]
[-- Type: text/x-tex, Size: 1340 bytes --]

% engine=luatex 

% OpenType features needed for Arabic
\definefontfeature
  [arab]
  [mode=node,language=dflt,script=arab,
   init=yes,medi=yes,fina=yes,isol=yes,
   liga=yes,dlig=yes,rlig=yes,clig=yes,
   mark=yes,mkmk=yes,kern=yes,curs=yes]

\font\Arab = arabtype*arab

\setcharactermirroring[1]

\hoffset=0pt

% For inner paragraph control within an LR paragraph

\definestartstop
  [arabictext]
  [commands=%
    {\textdir TRT%
    \Arab}]

\def\ArabicText#1{\startarabictext#1\stoparabictext}

% For separate Arabic-script paragraphs

\def\ArabicDirPar{\textdir TRT\pardir TRT}

\definestartstop
  [arab]
  [commands=%
    {\Arab%
     \ArabicDirPar}]

\showframe[text]

\starttext
\startarab
سلام (قوس) وقوس <قوس> و [قوس]

وهذا قوس حول نص غير عربي \textdir TLT (hello) \textdir TRT ثم عربي 

\stoparab

\blank

Here is some mixed {\em Arabic-} (\ArabicText{عربي}) and
Latin-script. As you can see, Aleph does a very good job mixing
{\em LR} (\ArabicText{يسار-يمين}) and {\em RL}
(\ArabicText{يمين-يسار}) texts. \ArabicText{و
هنا جملة منقطعة في وسط قرينة
لاتينية}. Aleph even does a great job breaking Arabic
phrases across lines.

(bracket) and <bracket> and [bracket]

\stoptext

[-- Attachment #1.1.3: mirroring.pdf --]
[-- Type: application/pdf, Size: 19796 bytes --]

[-- Attachment #1.2: Digital signature --]
[-- Type: application/pgp-signature, Size: 189 bytes --]

[-- Attachment #2: Type: text/plain, Size: 487 bytes --]

___________________________________________________________________________________
If your question is of interest to others as well, please add an entry to the Wiki!

maillist : ntg-context@ntg.nl / http://www.ntg.nl/mailman/listinfo/ntg-context
webpage  : http://www.pragma-ade.nl / http://tex.aanhet.net
archive  : https://foundry.supelec.fr/projects/contextrev/
wiki     : http://contextgarden.net
___________________________________________________________________________________

Re: Luatex (actually Aleph) and Bidi_Mirrored chars in RTL mode

[Idris Samawi Hamid]

[-- Attachment #1: Type: text/plain, Size: 1140 bytes --]

Hi Khaled,

On Fri, 13 Jun 2008 15:52:27 -0600, Khaled Hosny <khaledhosny@eglug.org>  
wrote:

>> experimental in the beta
>>
>> \setcharactermirroring[1]
>
> It does work perfectly with unidirectional texts (RTL or LTR), but when
> mixing bi-directional text, like Arabic text between brackets inside
> English line, the closing bracket takes the direction of the embedded
> text not the main line.
> See the attached example.

I just worked with this, and made some symmetrical definitions: see the  
attached modified files.

\setcharactermirroring[0,1] works in both uni- and bi-directional text --  
note that each bracketpair is inside of its respective directional context.

However, for the FIRST occurence of a bracket-pair in a bidi paragraph it  
does not behave as expected.

It may be that the second invocation of \setcharactermirroring does not  
immediately override the last. Maybe we need a \flush or  
\clearsetcharactermirroring command?

Best wishes
Idris

-- 
Professor Idris Samawi Hamid, Editor-in-Chief
International Journal of Shi`i Studies
Department of Philosophy
Colorado State University
Fort Collins, CO 80523

[-- Attachment #2: mirroring.pdf --]
[-- Type: application/pdf, Size: 29836 bytes --]

[-- Attachment #3: mirroring.tex --]
[-- Type: application/x-tex, Size: 2009 bytes --]

[-- Attachment #4: Type: text/plain, Size: 487 bytes --]

___________________________________________________________________________________
If your question is of interest to others as well, please add an entry to the Wiki!

maillist : ntg-context@ntg.nl / http://www.ntg.nl/mailman/listinfo/ntg-context
webpage  : http://www.pragma-ade.nl / http://tex.aanhet.net
archive  : https://foundry.supelec.fr/projects/contextrev/
wiki     : http://contextgarden.net
___________________________________________________________________________________

Re: Luatex (actually Aleph) and Bidi_Mirrored chars in RTL mode

[Khaled Hosny]

[-- Attachment #1.1: Type: text/plain, Size: 2086 bytes --]

On Fri, Jun 13, 2008 at 07:34:03PM -0600, Idris Samawi Hamid wrote:
> Hi Khaled,
>
> On Fri, 13 Jun 2008 15:52:27 -0600, Khaled Hosny <khaledhosny@eglug.org>  
> wrote:
>
>>> experimental in the beta
>>>
>>> \setcharactermirroring[1]
>>
>> It does work perfectly with unidirectional texts (RTL or LTR), but when
>> mixing bi-directional text, like Arabic text between brackets inside
>> English line, the closing bracket takes the direction of the embedded
>> text not the main line.
>> See the attached example.
>
> I just worked with this, and made some symmetrical definitions: see the  
> attached modified files.
>
> \setcharactermirroring[0,1] works in both uni- and bi-directional text -- 
> note that each bracketpair is inside of its respective directional 
> context.

Thanks for the fix, this makes it more organized indeed.

>
> However, for the FIRST occurence of a bracket-pair in a bidi paragraph it 
> does not behave as expected.
>
> It may be that the second invocation of \setcharactermirroring does not  
> immediately override the last. Maybe we need a \flush or  
> \clearsetcharactermirroring command?

I'm new at TeX/ConTeXt, all what I can do is to confirm that I'm getting
the same output as yours.

Regards,
 Khaled

>
> Best wishes
> Idris
>
> -- 
> Professor Idris Samawi Hamid, Editor-in-Chief
> International Journal of Shi`i Studies
> Department of Philosophy
> Colorado State University
> Fort Collins, CO 80523



> ___________________________________________________________________________________
> If your question is of interest to others as well, please add an entry to the Wiki!
> 
> maillist : ntg-context@ntg.nl / http://www.ntg.nl/mailman/listinfo/ntg-context
> webpage  : http://www.pragma-ade.nl / http://tex.aanhet.net
> archive  : https://foundry.supelec.fr/projects/contextrev/
> wiki     : http://contextgarden.net
> ___________________________________________________________________________________


-- 
 Khaled Hosny
 Arabic localizer and member of Arabeyes.org team

[-- Attachment #1.2: Digital signature --]
[-- Type: application/pgp-signature, Size: 189 bytes --]

[-- Attachment #2: Type: text/plain, Size: 487 bytes --]

___________________________________________________________________________________
If your question is of interest to others as well, please add an entry to the Wiki!

maillist : ntg-context@ntg.nl / http://www.ntg.nl/mailman/listinfo/ntg-context
webpage  : http://www.pragma-ade.nl / http://tex.aanhet.net
archive  : https://foundry.supelec.fr/projects/contextrev/
wiki     : http://contextgarden.net
___________________________________________________________________________________

Re: Luatex (actually Aleph) and Bidi_Mirrored chars in RTL mode

[Hans Hagen]

Idris Samawi Hamid wrote:
> Hi Khaled,
> 
> On Fri, 13 Jun 2008 15:52:27 -0600, Khaled Hosny <khaledhosny@eglug.org> 
> wrote:
> 
>>> experimental in the beta
>>>
>>> \setcharactermirroring[1]
>>
>> It does work perfectly with unidirectional texts (RTL or LTR), but when
>> mixing bi-directional text, like Arabic text between brackets inside
>> English line, the closing bracket takes the direction of the embedded
>> text not the main line.
>> See the attached example.
> 
> I just worked with this, and made some symmetrical definitions: see the 
> attached modified files.
> 
> \setcharactermirroring[0,1] works in both uni- and bi-directional text 
> -- note that each bracketpair is inside of its respective directional 
> context.

i wonder what gives you the impression that you can use 0.1 as argument

currently there is no reset command but you can say

\setcharactermirroring[-1]

since -1 resets an attribute

-----------------------------------------------------------------
                                           Hans Hagen | PRAGMA ADE
               Ridderstraat 27 | 8061 GH Hasselt | The Netherlands
      tel: 038 477 53 69 | fax: 038 477 53 74 | www.pragma-ade.com
                                              | www.pragma-pod.nl
-----------------------------------------------------------------
___________________________________________________________________________________
If your question is of interest to others as well, please add an entry to the Wiki!

maillist : ntg-context@ntg.nl / http://www.ntg.nl/mailman/listinfo/ntg-context
webpage  : http://www.pragma-ade.nl / http://tex.aanhet.net
archive  : https://foundry.supelec.fr/projects/contextrev/
wiki     : http://contextgarden.net
___________________________________________________________________________________

Re: Luatex (actually Aleph) and Bidi_Mirrored chars in RTL mode

[Idris Samawi Hamid]

On Sun, 15 Jun 2008 14:32:40 -0600, Hans Hagen <pragma@wxs.nl> wrote:

>> \setcharactermirroring[0,1] works in both uni- and bi-directional text
>> -- note that each bracketpair is inside of its respective directional
>> context.
>
> i wonder what gives you the impression that you can use 0.1 as argument

No, this was an abbreviation for "both \setcharactermirroring[0] and  
\setcharactermirroring[1] work in..."

;-)

Best wishes
Idris

-- 
Professor Idris Samawi Hamid, Editor-in-Chief
International Journal of Shi`i Studies
Department of Philosophy
Colorado State University
Fort Collins, CO 80523
___________________________________________________________________________________
If your question is of interest to others as well, please add an entry to the Wiki!

maillist : ntg-context@ntg.nl / http://www.ntg.nl/mailman/listinfo/ntg-context
webpage  : http://www.pragma-ade.nl / http://tex.aanhet.net
archive  : https://foundry.supelec.fr/projects/contextrev/
wiki     : http://contextgarden.net
___________________________________________________________________________________

Re: Luatex (actually Aleph) and Bidi_Mirrored chars in RTL mode

[Hans Hagen]

Idris Samawi Hamid wrote:
> On Sun, 15 Jun 2008 14:32:40 -0600, Hans Hagen <pragma@wxs.nl> wrote:
> 
>>> \setcharactermirroring[0,1] works in both uni- and bi-directional text
>>> -- note that each bracketpair is inside of its respective directional
>>> context.
>> i wonder what gives you the impression that you can use 0.1 as argument
> 
> No, this was an abbreviation for "both \setcharactermirroring[0] and  
> \setcharactermirroring[1] work in..."

ah ... anyhow, you're lucky that 0 is not implementing something then -)

Hans

-----------------------------------------------------------------
                                           Hans Hagen | PRAGMA ADE
               Ridderstraat 27 | 8061 GH Hasselt | The Netherlands
      tel: 038 477 53 69 | fax: 038 477 53 74 | www.pragma-ade.com
                                              | www.pragma-pod.nl
-----------------------------------------------------------------
___________________________________________________________________________________
If your question is of interest to others as well, please add an entry to the Wiki!

maillist : ntg-context@ntg.nl / http://www.ntg.nl/mailman/listinfo/ntg-context
webpage  : http://www.pragma-ade.nl / http://tex.aanhet.net
archive  : https://foundry.supelec.fr/projects/contextrev/
wiki     : http://contextgarden.net
___________________________________________________________________________________

Re: Luatex (actually Aleph) and Bidi_Mirrored chars in RTL mode

[Idris Samawi Hamid]

On Sun, 15 Jun 2008 14:32:40 -0600, Hans Hagen <pragma@wxs.nl> wrote:

> currently there is no reset command but you can say
> \setcharactermirroring[-1]
> since -1 resets an attribute

I tried that, but the result is the same: first occurrence of an RL  
open-paren in an LR paragraph is mis-mirrored. Next occurrences are fine.

The enclosed test file can serve as a benchmark for now.

Best
Idris

=============mirroring.tex==============
% engine=luatex

% OpenType features needed for Arabtype font
\definefontfeature
   [arab]
   [mode=node,language=dflt,script=arab,
    init=yes,medi=yes,fina=yes,isol=yes,
    liga=yes,dlig=yes,rlig=yes,clig=yes,
    mark=yes,mkmk=yes,kern=yes,curs=yes]

\font\Arabic = arabtype*arab   at 19pt

\def\ArabicParDir{\textdir TRT\pardir TRT}
\def\ArabicTextDir{\textdir TRT}
\def\LatinParDir{\textdir TLT\pardir TLT}
\def\LatinTextDir{\textdir TLT}

\definestartstop
   [arabicpar]
   [commands=%
     {\Arabic
	 \ArabicParDir
	 \setcharactermirroring[1]%
	 }]

\definestartstop
   [arabictext]
   [commands=%
     {\Arabic\setcharactermirroring[1]
     \ArabicTextDir
	%
	}]

\definestartstop
   [latinpar]
   [commands=%
     {\Arabic\LatinParDir
     \setcharactermirroring[-1]%
     }]	

\definestartstop
   [latintext]
   [commands={\LatinTextDir
   \setcharactermirroring[-1]%
   }]	

\setupwhitespace[big]

\showframe[text]

\starttext
\startarabicpar
سلام (قوس) وقوس <قوس> و [قوس]

وهذا قوس حول نص غير عربي
\startlatintext(Latin)\stoplatintext\ ثم عربي  \crlf
وهذا قوس حول نص غير عربي
\startlatintext(Latin)\stoplatintext\ ثم عربي  \crlf
وهذا قوس حول نص غير عربي
\startlatintext(Latin)\stoplatintext\ ثم عربي


\stoparabicpar

\blank

\startlatinpar
Peace (paren) \& [paren] \& <paren>

Here is some mixed Arabic
{\startarabictext(عربي)\stoparabictext} \ and Latin script. \crlf
Here is some mixed Arabic
{\startarabictext(عربي)\stoparabictext} \ and Latin script. \crlf
Here is some mixed Arabic
{\startarabictext(عربي)\stoparabictext} \ and Latin script.

As you can see, \LUATEX\ does a very good job mixing LR
\startarabictext(يسار-يمين)\stoparabictext \
and RL
\startarabictext(يمين-يسار)\stoparabictext \
texts. \crlf
As you can see, \LUATEX\ does a very good job mixing LR
\startarabictext(يسار-يمين)\stoparabictext \
and RL
\startarabictext(يمين-يسار)\stoparabictext \
texts.  \crlf
As you can see, \LUATEX\ does a very good job mixing LR
\startarabictext(يسار-يمين)\stoparabictext \
and RL
\startarabictext(يمين-يسار)\stoparabictext \
texts.

% \hfill\break temporary workaround for luatex bug

\LUATEX\ even does a great job breaking Arabic phrases \startarabictext(و  
هنا جملة منقطعة في وسط قرينة لاتينية)\stoparabictext \ across lines.  
\hfill\break
\LUATEX\ even does a great job breaking Arabic phrases \startarabictext(و  
هنا جملة منقطعة في وسط قرينة لاتينية)\stoparabictext \ across lines.   
\hfill\break
\LUATEX\ even does a great job breaking Arabic phrases \startarabictext(و  
هنا جملة منقطعة في وسط قرينة لاتينية)\stoparabictext \ across lines.


(bracket) and <bracket> and [bracket]

\stoplatinpar

\stoptext
========================================

-- 
Professor Idris Samawi Hamid, Editor-in-Chief
International Journal of Shi`i Studies
Department of Philosophy
Colorado State University
Fort Collins, CO 80523
___________________________________________________________________________________
If your question is of interest to others as well, please add an entry to the Wiki!

maillist : ntg-context@ntg.nl / http://www.ntg.nl/mailman/listinfo/ntg-context
webpage  : http://www.pragma-ade.nl / http://tex.aanhet.net
archive  : https://foundry.supelec.fr/projects/contextrev/
wiki     : http://contextgarden.net
___________________________________________________________________________________

Re: Luatex (actually Aleph) and Bidi_Mirrored chars in RTL mode

[Khaled Hosny]

[-- Attachment #1.1: Type: text/plain, Size: 740 bytes --]

On Sun, Jun 15, 2008 at 03:42:30PM -0600, Idris Samawi Hamid wrote:
> On Sun, 15 Jun 2008 14:32:40 -0600, Hans Hagen <pragma@wxs.nl> wrote:
> 
> > currently there is no reset command but you can say
> > \setcharactermirroring[-1]
> > since -1 resets an attribute
> 
> I tried that, but the result is the same: first occurrence of an RL  
> open-paren in an LR paragraph is mis-mirrored. Next occurrences are fine.
> 
> The enclosed test file can serve as a benchmark for now.

I just updated my minimals installation and this example is rendered
correctly now. I think I should try to make a sort of testbed based on
this example.

Regards,
 Khaled

-- 
 Khaled Hosny
 Arabic localizer and member of Arabeyes.org team

[-- Attachment #1.2: Digital signature --]
[-- Type: application/pgp-signature, Size: 189 bytes --]

[-- Attachment #2: Type: text/plain, Size: 487 bytes --]

___________________________________________________________________________________
If your question is of interest to others as well, please add an entry to the Wiki!

maillist : ntg-context@ntg.nl / http://www.ntg.nl/mailman/listinfo/ntg-context
webpage  : http://www.pragma-ade.nl / http://tex.aanhet.net
archive  : https://foundry.supelec.fr/projects/contextrev/
wiki     : http://contextgarden.net
___________________________________________________________________________________

Re: Luatex (actually Aleph) and Bidi_Mirrored chars in RTL mode

[Hans Hagen]

Idris Samawi Hamid wrote:
> Hi Khaled,
> 
> On Fri, 13 Jun 2008 15:52:27 -0600, Khaled Hosny <khaledhosny@eglug.org> 
> wrote:
> 
>>> experimental in the beta
>>>
>>> \setcharactermirroring[1]
>>
>> It does work perfectly with unidirectional texts (RTL or LTR), but when
>> mixing bi-directional text, like Arabic text between brackets inside
>> English line, the closing bracket takes the direction of the embedded
>> text not the main line.
>> See the attached example.
> 
> I just worked with this, and made some symmetrical definitions: see the 
> attached modified files.

i need to look into the luatex source to see what kind of direction 
nodes are possible (i just check for one now) or maybe tac has such a 
list at hand

Hans


-----------------------------------------------------------------
                                           Hans Hagen | PRAGMA ADE
               Ridderstraat 27 | 8061 GH Hasselt | The Netherlands
      tel: 038 477 53 69 | fax: 038 477 53 74 | www.pragma-ade.com
                                              | www.pragma-pod.nl
-----------------------------------------------------------------
___________________________________________________________________________________
If your question is of interest to others as well, please add an entry to the Wiki!

maillist : ntg-context@ntg.nl / http://www.ntg.nl/mailman/listinfo/ntg-context
webpage  : http://www.pragma-ade.nl / http://tex.aanhet.net
archive  : https://foundry.supelec.fr/projects/contextrev/
wiki     : http://contextgarden.net
___________________________________________________________________________________

Re: Luatex (actually Aleph) and Bidi_Mirrored chars in RTL mode

[Hans Hagen]

Khaled Hosny wrote:
> On Mon, Jun 09, 2008 at 11:08:09PM +0200, Hans Hagen wrote:
>> Khaled Hosny wrote:
>>> Luatex does not mirror characters that has a Bidi_Mirrored property when
>>> the text direction is set to RTL (TRT in Aleph), according to 
>>> http://unicode.org/reports/tr9/#Mirroring, the different types of
>>> parenthesis that has Bidi_Mirrored property should be mirreored in RTL
>>> mode, but this isn't what I get. Is this a bug, feature, or am I missing
>>> some thing?
>> experimental in the beta
>>
>> \setcharactermirroring[1]
> 
> It does work perfectly with unidirectional texts (RTL or LTR), but when
> mixing bi-directional text, like Arabic text between brackets inside
> English line, the closing bracket takes the direction of the embedded
> text not the main line.
> See the attached example.

i need to figure out what the subtypes are of the direction nodes

i suggest that you and idris cook up a set of test files that can serve 
as benchmark, preferably something organized (arabtest-whatever) and 
such; eventually these can go into the regression test machinery

Hans


-----------------------------------------------------------------
                                           Hans Hagen | PRAGMA ADE
               Ridderstraat 27 | 8061 GH Hasselt | The Netherlands
      tel: 038 477 53 69 | fax: 038 477 53 74 | www.pragma-ade.com
                                              | www.pragma-pod.nl
-----------------------------------------------------------------
___________________________________________________________________________________
If your question is of interest to others as well, please add an entry to the Wiki!

maillist : ntg-context@ntg.nl / http://www.ntg.nl/mailman/listinfo/ntg-context
webpage  : http://www.pragma-ade.nl / http://tex.aanhet.net
archive  : https://foundry.supelec.fr/projects/contextrev/
wiki     : http://contextgarden.net
___________________________________________________________________________________

Re: Luatex (actually Aleph) and Bidi_Mirrored chars in RTL mode

[Idris Samawi Hamid]

Hi Khaled,

On Sun, 08 Jun 2008 19:21:28 -0600, Khaled Hosny <khaledhosny@eglug.org>  
wrote:

> Luatex does not mirror characters that has a Bidi_Mirrored property when
> the text direction is set to RTL (TRT in Aleph), according to
> http://unicode.org/reports/tr9/#Mirroring, the different types of
> parenthesis that has Bidi_Mirrored property should be mirreored in RTL
> mode, but this isn't what I get. Is this a bug, feature, or am I missing
> some thing?

In addition to the comments by Arthur, Hans:

This issue needs to be looked at in the larger context of bidi in luatex.  
For example, we need a flexible model to implement Bidirectional Character  
Types. For example, we need an option that will recognize certain  
character-types as RL and switch directions accordingly -- that is,  
automatically switch \textdir, within a given paragraph, in accordance  
with the directionality of the word-string. This needs a well thought-out  
model so that we can map the Unicode specification to our \texdir \pardir  
functionality...

So the whole directionality issue needs to be looked at carefully, in  
addition to temporary patches like \setcharactermirroring[1]

OTOH: SC Unipad gets this right is perhaps the ideal to strive for...  
Indeed, it implements unicode better than any text application of which I  
am aware...

Also, we need not be too slavish: The Yudit author has pointed out areas  
where the bidi algorithm makes no sense or is deficient:

http://www.yudit.org/bidi/surprise.html

So this all needs careful thought. On my list...

Best wishes
Idris

-- 
Professor Idris Samawi Hamid, Editor-in-Chief
International Journal of Shi`i Studies
Department of Philosophy
Colorado State University
Fort Collins, CO 80523
___________________________________________________________________________________
If your question is of interest to others as well, please add an entry to the Wiki!

maillist : ntg-context@ntg.nl / http://www.ntg.nl/mailman/listinfo/ntg-context
webpage  : http://www.pragma-ade.nl / http://tex.aanhet.net
archive  : https://foundry.supelec.fr/projects/contextrev/
wiki     : http://contextgarden.net
___________________________________________________________________________________

Re: Luatex (actually Aleph) and Bidi_Mirrored chars in RTL mode

[Arthur Reutenauer]

> Also, we need not be too slavish: The Yudit author has pointed out areas  
> where the bidi algorithm makes no sense or is deficient:
> 
> http://www.yudit.org/bidi/surprise.html

  To be honest, this page dates back to half a dozen years ago and
reflects Gáspár Sinai's positions at that time; there are hints that he
has changed his mind at least slightly, as he started to make quite a
scandal about possible "security problems" in the bidi algorithm and
later backed off (http://yudit.org/security/).  This is of course not to
say that the bidi algorithm is perfect, because it's not, but I don't
think that the link you quote makes a really strong point against it.
In particular, I find that some of the recommandations Sinai criticizes
are amazingly close to what we do in the TeX world ("formatting codes
should be inserted" => mark up the text with direction-switching
commands, etc.)

	Arthur
___________________________________________________________________________________
If your question is of interest to others as well, please add an entry to the Wiki!

maillist : ntg-context@ntg.nl / http://www.ntg.nl/mailman/listinfo/ntg-context
webpage  : http://www.pragma-ade.nl / http://tex.aanhet.net
archive  : https://foundry.supelec.fr/projects/contextrev/
wiki     : http://contextgarden.net
___________________________________________________________________________________

Re: Luatex (actually Aleph) and Bidi_Mirrored chars in RTL mode

[Idris Samawi Hamid]

Hi,

On Mon, 09 Jun 2008 19:36:00 -0600, Arthur Reutenauer  
<arthur.reutenauer@normalesup.org> wrote:

>> Also, we need not be too slavish: The Yudit author has pointed out areas
>> where the bidi algorithm makes no sense or is deficient:
>>
>> http://www.yudit.org/bidi/surprise.html
>
>   To be honest, this page dates back to half a dozen years ago and
> reflects Gáspár Sinai's positions at that time; there are hints that he
> has changed his mind at least slightly, as he started to make quite a
> scandal about possible "security problems" in the bidi algorithm and
> later backed off (http://yudit.org/security/).  This is of course not to
> say that the bidi algorithm is perfect, because it's not, but I don't
> think that the link you quote makes a really strong point against it.
> In particular, I find that some of the recommandations Sinai criticizes
> are amazingly close to what we do in the TeX world ("formatting codes
> should be inserted" => mark up the text with direction-switching
> commands, etc.)

Granted. The main point is that we have to reinterpret bidi in way that  
fits with TeX's/ConTeXt's needs, idiosyncracies, etc.... I don't believe  
we need to treat the unicode bidi algorithm as canonical.

But I'm still studying the matter.

Best
Idris

-- 
Professor Idris Samawi Hamid, Editor-in-Chief
International Journal of Shi`i Studies
Department of Philosophy
Colorado State University
Fort Collins, CO 80523
___________________________________________________________________________________
If your question is of interest to others as well, please add an entry to the Wiki!

maillist : ntg-context@ntg.nl / http://www.ntg.nl/mailman/listinfo/ntg-context
webpage  : http://www.pragma-ade.nl / http://tex.aanhet.net
archive  : https://foundry.supelec.fr/projects/contextrev/
wiki     : http://contextgarden.net
___________________________________________________________________________________

Re: Luatex (actually Aleph) and Bidi_Mirrored chars in RTL mode

[Hans Hagen]

Idris Samawi Hamid wrote:

> Granted. The main point is that we have to reinterpret bidi in way that  
> fits with TeX's/ConTeXt's needs, idiosyncracies, etc.... I don't believe  
> we need to treat the unicode bidi algorithm as canonical.

i think that we should stick to things that make sense; with tex we're 
often talking of tagged sources and anything ambiguous should be tagged; 
in a sense this is not even related to arab at all, take an url ... i 
can imagine an url-algorithm, but it could never be perfect (just see 
what some programs that try to do it sometimes make of it)

if for instance ( ) are officially not symbols but open/close thingies, 
then we need to deal with them (although i then wonder why we have no 
proper open/close code point for them instead of reusing the ascii () 
which have for users some expected visual appearance, but in that 
respect unicode puzzles me on a daily basis)

Hans

-----------------------------------------------------------------
                                           Hans Hagen | PRAGMA ADE
               Ridderstraat 27 | 8061 GH Hasselt | The Netherlands
      tel: 038 477 53 69 | fax: 038 477 53 74 | www.pragma-ade.com
                                              | www.pragma-pod.nl
-----------------------------------------------------------------
___________________________________________________________________________________
If your question is of interest to others as well, please add an entry to the Wiki!

maillist : ntg-context@ntg.nl / http://www.ntg.nl/mailman/listinfo/ntg-context
webpage  : http://www.pragma-ade.nl / http://tex.aanhet.net
archive  : https://foundry.supelec.fr/projects/contextrev/
wiki     : http://contextgarden.net
___________________________________________________________________________________

Re: Luatex (actually Aleph) and Bidi_Mirrored chars in RTL mode

[Arthur Reutenauer]

> if for instance ( ) are officially not symbols but open/close thingies, 
> then we need to deal with them

  Yes, they are.  That's what the Bidi_Mirrored property is for.

>                                (although i then wonder why we have no 
> proper open/close code point for them instead of reusing the ascii () 

  If anything, for backward compatibility.

	Arthur
___________________________________________________________________________________
If your question is of interest to others as well, please add an entry to the Wiki!

maillist : ntg-context@ntg.nl / http://www.ntg.nl/mailman/listinfo/ntg-context
webpage  : http://www.pragma-ade.nl / http://tex.aanhet.net
archive  : https://foundry.supelec.fr/projects/contextrev/
wiki     : http://contextgarden.net
___________________________________________________________________________________

Re: Luatex (actually Aleph) and Bidi_Mirrored chars in RTL mode

[Hans Hagen]

Arthur Reutenauer wrote:
>> if for instance ( ) are officially not symbols but open/close thingies, 
>> then we need to deal with them
> 
>   Yes, they are.  That's what the Bidi_Mirrored property is for.
> 
>>                                (although i then wonder why we have no 
>> proper open/close code point for them instead of reusing the ascii () 
> 
>   If anything, for backward compatibility.

sure but originally they were in ascii just ( and ) and not some generic 
opener and closer

Hans


-----------------------------------------------------------------
                                           Hans Hagen | PRAGMA ADE
               Ridderstraat 27 | 8061 GH Hasselt | The Netherlands
      tel: 038 477 53 69 | fax: 038 477 53 74 | www.pragma-ade.com
                                              | www.pragma-pod.nl
-----------------------------------------------------------------
___________________________________________________________________________________
If your question is of interest to others as well, please add an entry to the Wiki!

maillist : ntg-context@ntg.nl / http://www.ntg.nl/mailman/listinfo/ntg-context
webpage  : http://www.pragma-ade.nl / http://tex.aanhet.net
archive  : https://foundry.supelec.fr/projects/contextrev/
wiki     : http://contextgarden.net
___________________________________________________________________________________

Re: Luatex (actually Aleph) and Bidi_Mirrored chars in RTL mode

[Arthur Reutenauer]

> sure but originally they were in ascii just ( and ) and not some generic 
> opener and closer

  I did not mean ASCII ... in the 8-bit encodings for Arabic that built
over ASCII (ISO 8859-6, Windows-1256, etc.) there also was a single
couple of parentheses, and I don't think they had a different meaning as
in Unicode ... I couldn't find the exact reference (the ISO 8859-6 text
sells for 70 Swiss Francs on the ISO site, and the only specification
Microsoft provides on his site is the character table --
http://www.microsoft.com/globaldev/reference/sbcs/1256.mspx), but HTML
pages I find on the Web tend to indicate that in Arabic text encoded in
Windows-1256, the character '(' (byte 0x28) was really used as an
opening bracket, and ')' as a closing one.  Maybe people on the list
from Arabic-speaking countries, Israel or Iran can tell us more (in
particular, what do one type on a standard keyboard to input an opening
bracket?).

	Arthur
___________________________________________________________________________________
If your question is of interest to others as well, please add an entry to the Wiki!

maillist : ntg-context@ntg.nl / http://www.ntg.nl/mailman/listinfo/ntg-context
webpage  : http://www.pragma-ade.nl / http://tex.aanhet.net
archive  : https://foundry.supelec.fr/projects/contextrev/
wiki     : http://contextgarden.net
___________________________________________________________________________________

Re: Luatex (actually Aleph) and Bidi_Mirrored chars in RTL mode

[Otared Kavian]

On 10 juin 08, at 23:30, Arthur Reutenauer wrote:

> […]
> Maybe people on the list from Arabic-speaking countries, Israel or  
> Iran can tell us more (in
> particular, what do one type on a standard keyboard to input an  
> opening
> bracket?).
>
> 	Arthur

Hi Arthur,

Sorry for answering to this message so late (this thread is far beyond  
my knowledge and I read it just by curiosity…).
Since I am Iranian and familiar with Persian and a little bit of  
Arabic, I wanted to let you know that one types bracket and such as  
follows:
--- when writing a Persian text enclosed in parentheses, typing on a  
Persian keyboard, one types a "right parenthesis" that is U+0029,  
followed by the text intended to be enclosed in brackets (running  
right to left), and then a "left parenthesis" that is U+0028. The same  
applies when the text is enclosed within "guillemets".
(For numbers there is no difference with roman script).
However on my Mac's Persian or Arabic AZERTY keyboards, the mapping is  
not correct (that is typing a right parenthesis results in U+0028 in  
the text one is typing).

Best regards: OK
___________________________________________________________________________________
If your question is of interest to others as well, please add an entry to the Wiki!

maillist : ntg-context@ntg.nl / http://www.ntg.nl/mailman/listinfo/ntg-context
webpage  : http://www.pragma-ade.nl / http://tex.aanhet.net
archive  : https://foundry.supelec.fr/projects/contextrev/
wiki     : http://contextgarden.net
___________________________________________________________________________________

Re: Luatex (actually Aleph) and Bidi_Mirrored chars in RTL mode

[Arthur Reutenauer]

	Hello Otared,

  Thanks for the answer (I should have known the results depended on the
actual configuration :-)

	Arthur
___________________________________________________________________________________
If your question is of interest to others as well, please add an entry to the Wiki!

maillist : ntg-context@ntg.nl / http://www.ntg.nl/mailman/listinfo/ntg-context
webpage  : http://www.pragma-ade.nl / http://tex.aanhet.net
archive  : https://foundry.supelec.fr/projects/contextrev/
wiki     : http://contextgarden.net
___________________________________________________________________________________

Re: Luatex (actually Aleph) and Bidi_Mirrored chars in RTL mode

[Idris Samawi Hamid]

Hi Arthur,

On Wed, 18 Jun 2008 02:29:00 -0600, Arthur Reutenauer  
<arthur.reutenauer@normalesup.org> wrote:

>   Thanks for the answer (I should have known the results depended on the
> actual configuration :-)

Get SC unipad and type identical logical input in Arabic and Farsi modes.  
The makers of Unipad are Iranian and they implement both Arabic and Farsi  
rules correctly, have keyboards etc.  AFAIK, no other editor implements  
the bidi alg as well.

I'll forward you something in this regard.

Best wishes
Idris

-- 
Professor Idris Samawi Hamid, Editor-in-Chief
International Journal of Shi`i Studies
Department of Philosophy
Colorado State University
Fort Collins, CO 80523
___________________________________________________________________________________
If your question is of interest to others as well, please add an entry to the Wiki!

maillist : ntg-context@ntg.nl / http://www.ntg.nl/mailman/listinfo/ntg-context
webpage  : http://www.pragma-ade.nl / http://tex.aanhet.net
archive  : https://foundry.supelec.fr/projects/contextrev/
wiki     : http://contextgarden.net
___________________________________________________________________________________

Re: Luatex (actually Aleph) and Bidi_Mirrored chars in RTL mode

[Khaled Hosny]

[-- Attachment #1.1: Type: text/plain, Size: 1650 bytes --]

On Wed, Jun 18, 2008 at 09:22:58AM -0600, Idris Samawi Hamid wrote:
> Hi Arthur,
> 
> On Wed, 18 Jun 2008 02:29:00 -0600, Arthur Reutenauer  
> <arthur.reutenauer@normalesup.org> wrote:
> 
> >   Thanks for the answer (I should have known the results depended on the
> > actual configuration :-)
> 
> Get SC unipad and type identical logical input in Arabic and Farsi modes.  
> The makers of Unipad are Iranian and they implement both Arabic and Farsi  
> rules correctly, have keyboards etc.  AFAIK, no other editor implements  
> the bidi alg as well.

I didn't use Unipad that much, but I think Gedit (GTK pased) does a
perfect job, I'm not aware of any windows port of it though. I'd like to
know what features in Unipad (regarding Arabic) that Gedit hasn't.

Regards,
 Khaled

> 
> I'll forward you something in this regard.
> 
> Best wishes
> Idris
> 
> -- 
> Professor Idris Samawi Hamid, Editor-in-Chief
> International Journal of Shi`i Studies
> Department of Philosophy
> Colorado State University
> Fort Collins, CO 80523
> ___________________________________________________________________________________
> If your question is of interest to others as well, please add an entry to the Wiki!
> 
> maillist : ntg-context@ntg.nl / http://www.ntg.nl/mailman/listinfo/ntg-context
> webpage  : http://www.pragma-ade.nl / http://tex.aanhet.net
> archive  : https://foundry.supelec.fr/projects/contextrev/
> wiki     : http://contextgarden.net
> ___________________________________________________________________________________

-- 
 Khaled Hosny
 Arabic localizer and member of Arabeyes.org team

[-- Attachment #1.2: Digital signature --]
[-- Type: application/pgp-signature, Size: 189 bytes --]

[-- Attachment #2: Type: text/plain, Size: 487 bytes --]

___________________________________________________________________________________
If your question is of interest to others as well, please add an entry to the Wiki!

maillist : ntg-context@ntg.nl / http://www.ntg.nl/mailman/listinfo/ntg-context
webpage  : http://www.pragma-ade.nl / http://tex.aanhet.net
archive  : https://foundry.supelec.fr/projects/contextrev/
wiki     : http://contextgarden.net
___________________________________________________________________________________

Re: Luatex (actually Aleph) and Bidi_Mirrored chars in RTL mode

[Idris Samawi Hamid]

[-- Attachment #1: Type: text/plain, Size: 2087 bytes --]

On Wed, 18 Jun 2008 09:44:19 -0600, Khaled Hosny <khaledhosny@eglug.org>  
wrote:

>> Get SC unipad and type identical logical input in Arabic and Farsi  
>> modes.
>> The makers of Unipad are Iranian and they implement both Arabic and  
>> Farsi
>> rules correctly, have keyboards etc.  AFAIK, no other editor implements
>> the bidi alg as well.
>
> I didn't use Unipad that much, but I think Gedit (GTK pased) does a
> perfect job,

No, it does not ;-)

> I'm not aware of any windows port of it though.

andLinux on W32 is wonderful, no more cygwin etc.

> I'd like to
> know what features in Unipad (regarding Arabic) that Gedit hasn't.

One example:

The following comes out correctly in Unipad, also in Opera's email  
composer (don't know how it will look in your client). It does not come  
out right in Windows Notepad or in gedit. Kate does get it right (and  
other kde apps as well), as well as OOo:

==============
European:
18%, 1--3

Arabic:
١٨٪، ١--٣

Farsi:
۱۸٪، ۱--۳
==============

See attached as well.

Unipad is designed as a complete unicode utility, and has lots of other  
nifty features as well, especially for editing Arabic script. Its  
Arabic-script editing features are truly unique and useful. Eg, I  
select/edit vowels horizontally not vertically.

Unipad has:

Char Info bar:
character code value (U+0000)
character name (if assigned)
character category (letter, decimal digit, etc.)
character block (script, collection, etc.)

Extended Char Info bar:
decimal character code value
encoded byte sequence (octals)
bidirectional character type
decimal digit value (only if the character represents a decimal digit)

Unipad is not a good choice for a programmer's or TeX editor, but in  
overall unicode editing it is unrivalled, although it has not been updated  
since Unicode 4.1.0 :-(

Best wishes
Idris

-- 
Professor Idris Samawi Hamid, Editor-in-Chief
International Journal of Shi`i Studies
Department of Philosophy
Colorado State University
Fort Collins, CO 80523

[-- Attachment #2: unipad-bidi.pdf --]
[-- Type: application/pdf, Size: 10891 bytes --]

[-- Attachment #3: Type: text/plain, Size: 487 bytes --]

___________________________________________________________________________________
If your question is of interest to others as well, please add an entry to the Wiki!

maillist : ntg-context@ntg.nl / http://www.ntg.nl/mailman/listinfo/ntg-context
webpage  : http://www.pragma-ade.nl / http://tex.aanhet.net
archive  : https://foundry.supelec.fr/projects/contextrev/
wiki     : http://contextgarden.net
___________________________________________________________________________________

Re: Luatex (actually Aleph) and Bidi_Mirrored chars in RTL mode

[Khaled Hosny]

[-- Attachment #1.1: Type: text/plain, Size: 2389 bytes --]

On Wed, Jun 18, 2008 at 01:50:35PM -0600, Idris Samawi Hamid wrote:
> On Wed, 18 Jun 2008 09:44:19 -0600, Khaled Hosny <khaledhosny@eglug.org>  
> wrote:
>
>>> Get SC unipad and type identical logical input in Arabic and Farsi  
>>> modes.
>>> The makers of Unipad are Iranian and they implement both Arabic and  
>>> Farsi
>>> rules correctly, have keyboards etc.  AFAIK, no other editor implements
>>> the bidi alg as well.
>>
>> I didn't use Unipad that much, but I think Gedit (GTK pased) does a
>> perfect job,
>
> No, it does not ;-)
>
>> I'm not aware of any windows port of it though.
>
> andLinux on W32 is wonderful, no more cygwin etc.
>
>> I'd like to
>> know what features in Unipad (regarding Arabic) that Gedit hasn't.
>
> One example:
>
> The following comes out correctly in Unipad, also in Opera's email  
> composer (don't know how it will look in your client). It does not come  
> out right in Windows Notepad or in gedit. Kate does get it right (and  
> other kde apps as well), as well as OOo:
>
> ==============
> European:
> 18%, 1--3
>
> Arabic:
> ١٨٪، ١--٣
>
> Farsi:
> ۱۸٪، ۱--۳
> ==============

Oh, it is shame that even my terminal (mlterm) renders it right! I think
I'm going to report this :)

>
> See attached as well.
>
> Unipad is designed as a complete unicode utility, and has lots of other  
> nifty features as well, especially for editing Arabic script. Its  
> Arabic-script editing features are truly unique and useful. Eg, I  
> select/edit vowels horizontally not vertically.
>
> Unipad has:
>
> Char Info bar:
> character code value (U+0000)
> character name (if assigned)
> character category (letter, decimal digit, etc.)
> character block (script, collection, etc.)
>
> Extended Char Info bar:
> decimal character code value
> encoded byte sequence (octals)
> bidirectional character type
> decimal digit value (only if the character represents a decimal digit)
>
> Unipad is not a good choice for a programmer's or TeX editor, but in  
> overall unicode editing it is unrivalled, although it has not been 
> updated since Unicode 4.1.0 :-(

Fortunately, thanks to WINE, it does run under Linux, even their web
site lists WINE as one of the supported platforms.

Regards,
 Khaled

-- 
 Khaled Hosny
 Arabic localizer and member of Arabeyes.org team

[-- Attachment #1.2: Digital signature --]
[-- Type: application/pgp-signature, Size: 189 bytes --]

[-- Attachment #2: Type: text/plain, Size: 487 bytes --]

___________________________________________________________________________________
If your question is of interest to others as well, please add an entry to the Wiki!

maillist : ntg-context@ntg.nl / http://www.ntg.nl/mailman/listinfo/ntg-context
webpage  : http://www.pragma-ade.nl / http://tex.aanhet.net
archive  : https://foundry.supelec.fr/projects/contextrev/
wiki     : http://contextgarden.net
___________________________________________________________________________________

Re: Luatex (actually Aleph) and Bidi_Mirrored chars in RTL mode

[Hans Hagen]

Idris Samawi Hamid wrote:

> This issue needs to be looked at in the larger context of bidi in luatex.  
> For example, we need a flexible model to implement Bidirectional Character  
> Types. For example, we need an option that will recognize certain  
> character-types as RL and switch directions accordingly -- that is,  
> automatically switch \textdir, within a given paragraph, in accordance  
> with the directionality of the word-string. This needs a well thought-out  
> model so that we can map the Unicode specification to our \texdir \pardir  
> functionality...

it's not hard to do but indeed we need a clean definition; it may even 
be that we need to change some of this \*dir stuff; (it's no problem for 
me to inject dir nodes in the node list); especially messy seems the 
number stuff

btw is there a reason why in bidi arabic r->l is tagged 'al' and not 'r'?

Hans

-----------------------------------------------------------------
                                           Hans Hagen | PRAGMA ADE
               Ridderstraat 27 | 8061 GH Hasselt | The Netherlands
      tel: 038 477 53 69 | fax: 038 477 53 74 | www.pragma-ade.com
                                              | www.pragma-pod.nl
-----------------------------------------------------------------
___________________________________________________________________________________
If your question is of interest to others as well, please add an entry to the Wiki!

maillist : ntg-context@ntg.nl / http://www.ntg.nl/mailman/listinfo/ntg-context
webpage  : http://www.pragma-ade.nl / http://tex.aanhet.net
archive  : https://foundry.supelec.fr/projects/contextrev/
wiki     : http://contextgarden.net
___________________________________________________________________________________

Re: Luatex (actually Aleph) and Bidi_Mirrored chars in RTL mode

[Arthur Reutenauer]

> btw is there a reason why in bidi arabic r->l is tagged 'al' and not 'r'?

  Arabic Letter.  Basic right-to-left characters have type R.

	Arthur
___________________________________________________________________________________
If your question is of interest to others as well, please add an entry to the Wiki!

maillist : ntg-context@ntg.nl / http://www.ntg.nl/mailman/listinfo/ntg-context
webpage  : http://www.pragma-ade.nl / http://tex.aanhet.net
archive  : https://foundry.supelec.fr/projects/contextrev/
wiki     : http://contextgarden.net
___________________________________________________________________________________

Re: Luatex (actually Aleph) and Bidi_Mirrored chars in RTL mode

[Hans Hagen]

Arthur Reutenauer wrote:
>> btw is there a reason why in bidi arabic r->l is tagged 'al' and not 'r'?
> 
>   Arabic Letter.  Basic right-to-left characters have type R.

do you have any idea why is arab treated special?

-----------------------------------------------------------------
                                           Hans Hagen | PRAGMA ADE
               Ridderstraat 27 | 8061 GH Hasselt | The Netherlands
      tel: 038 477 53 69 | fax: 038 477 53 74 | www.pragma-ade.com
                                              | www.pragma-pod.nl
-----------------------------------------------------------------
___________________________________________________________________________________
If your question is of interest to others as well, please add an entry to the Wiki!

maillist : ntg-context@ntg.nl / http://www.ntg.nl/mailman/listinfo/ntg-context
webpage  : http://www.pragma-ade.nl / http://tex.aanhet.net
archive  : https://foundry.supelec.fr/projects/contextrev/
wiki     : http://contextgarden.net
___________________________________________________________________________________

Re: Luatex (actually Aleph) and Bidi_Mirrored chars in RTL mode

[Arthur Reutenauer]

> do you have any idea why is arab treated special?

  Arabic letters behave differently from Hebrew, Thaana, etc. in
particular in the vicinity of European digits.  But the difference is
quite minute.

	Arthur
___________________________________________________________________________________
If your question is of interest to others as well, please add an entry to the Wiki!

maillist : ntg-context@ntg.nl / http://www.ntg.nl/mailman/listinfo/ntg-context
webpage  : http://www.pragma-ade.nl / http://tex.aanhet.net
archive  : https://foundry.supelec.fr/projects/contextrev/
wiki     : http://contextgarden.net
___________________________________________________________________________________