Dear Talal, Salaam. Your tex file linertl.tex appears to be corrupted, could you either repost it or send it to me off list? I was also going to write something similar to Pablo. My suggestion is to move forward with the rest of your project and this will almost certainly be solved before you're done. Unless Hans says that there is no chance of a fix, which I *highly* doubt. Send a reminder to the list every few days or so, as you did today. Proceeding this way produces an uneasy feeling; believe me, I know. But I've been in your shoes more times than I can count and it always worked out in the end. Best wishes Idris On Sat, 01 Aug 2015 09:37:57 -0600, talazem@fastmail.fm wrote:

Sorry to pester, but this problem is barring me from moving my book project into ConTeXt this weekend.

The use of linenumbers seems to throw off the R2L mechanism. Example of the problem attached.

...

On 30 Jul 2015, at 16:30, talazem@fastmail.fm wrote:

There seems to be some limitation (or bug) in right-to-left (r2l) alignment, when used alongside linenumbering. Consider the following:

...

\setuppapersize[A6] \starttext \showframe

\section{Incorrect} \startlinenumbering \startalignment[r2l] Lorem ipsum dolor sit amet, consectetur adipiscing elit. Pellentesque molestie dictum quam non congue. Integer ut dolor vel mi vestibulum molestie vel et nulla. \stopalignment \stoplinenumbering

\section{Correct} \startalignment[r2l] Lorem ipsum dolor sit amet, consectetur adipiscing elit. Pellentesque molestie dictum quam non congue. Integer ut dolor vel mi vestibulum molestie vel et nulla. \stopalignment

\stoptext

The RTL alignment works correctly on its own. However, when you introduce another factor — here, linenumbers — the RTL alignment is ignored.

Perhaps this linked to the problem with of RTL footnotes being discussed here: gmane.comp.tex.context/92113 http://article.gmane.org/gmane.comp.tex.context/92113?

Best wishes, Talal ___________________________________________________________________________________ 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 : http://foundry.supelec.fr/projects/contextrev/ wiki : http://contextgarden.net ___________________________________________________________________________________

-- Idris Samawi Hamid Professor of Philosophy Colorado State University Fort Collins, CO 80523

Thanks for that, Wolfgang. I can only hope that, as the culprit is known, it can be apprehended! Idris: here is the document that I had sent (and many thanks to you and to Pablo for the more general advice):

\unexpanded\def\doVariant#1#2#3% {\startlinenote[#1]{#2] #3}#2\stoplinenote[#1]}

\newcounter\countvariants \unexpanded\def\variant {\doglobal\increment\countvariants \normalexpanded{\doVariant{Varia:\countvariants}}}

\setupalign[r2l]

\showframe \starttext

\section{Correct} Lorem ipsum dolor sit amet, consectetur adipiscing elit. Pellentesque molestie dictum quam non congue. Integer ut dolor vel mi vestibulum molestie vel et nulla. Morbi id justo vitae quam rhoncus tincidunt ut adipiscing nisi.

\section{Incorrect} \startlinenumbering Lorem ipsum dolor sit amet, consectetur adipiscing elit. Pellentesque molestie dictum quam non congue. Integer ut dolor vel mi vestibulum molestie vel et nulla. Morbi id justo vitae quam rhoncus tincidunt ut adipiscing nisi. \stoplinenumbering

\stoptext

All the best, Talal

On 1 Aug 2015, at 19:47, Wolfgang Schuster wrote:

...

talazem@fastmail.fm mailto:talazem@fastmail.fm 1. August 2015 17:37 Sorry to pester, but this problem is barring me from moving my book project into ConTeXt this weekend.

The use of linenumbers seems to throw off the R2L mechanism. Example of the problem attached.

The culprit is the function "boxed.stage_two" in page.lua which output each text line in the line numbering environment in left to right mode. When you have justified lines everything is fine but when you have a ragged line (e.g. the last line in a paragraph) ConTeXt aligns the context with the left margin.

\starttext \showframe

\startalignment[righttoleft,flushleft]

\input ward

\blank

\startlinenumbering \input ward \stoplinenumbering

\stopalignment

\stoptext

Wolfgang ___________________________________________________________________________________ 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 : http://foundry.supelec.fr/projects/contextrev/ wiki : http://contextgarden.net ___________________________________________________________________________________

Thank you very much for this, Idris. It works nicely in the sample file you sent. But it goes off again once you add \setupdirections[bidi=global]. Consider the following: ========== \setuppapersize[A6] \definefontfamily[mainface][serif][ALMFixed][% features=arabic, range=arabic, ] \setupbodyfont [mainface] \starttext \showframe \setupdirections[bidi=global] \setupthinrules[color=white] \setupwhitespace[big] \section{Incorrect} \startlinenumbering \startalignment[r2l] هذه رسالة لطيفة ألفها العالم الكبير الفهامة النحرير. هذه رسالة جامعة مانعة ألفها الشيخ الفاضل... \thinrule Lorem ipsum dolor sit amet, consectetur adipiscing elit. Pellentesque molestie dictum quam non congue. Integer ut dolor vel mi vestibulum molestie vel et nulla. \stopalignment \stoplinenumbering \section{Correct} \startalignment[r2l] Lorem ipsum dolor sit amet, consectetur adipiscing elit. Pellentesque molestie dictum quam non congue. Integer ut dolor vel mi vestibulum molestie vel et nulla. \stopalignment \stoptext ============= With many thanks, and salam, Talal

On 2 Aug 2015, at 12:51, Idris Samawi Hamid ادريس سماوي حامد wrote:

On Sun, 02 Aug 2015 03:24:12 -0600, talazem@fastmail.fm wrote:

...

Thanks for that, Wolfgang. I can only hope that, as the culprit is known, it can be apprehended!

Talal, here is a barbarous workaround using \thinrule:

====================== \setuppapersize[A6] \starttext \showframe

\setupthinrules[color=white] \setupwhitespace[big]

\section{Incorrect} \startlinenumbering \startalignment[r2l] Lorem ipsum dolor sit amet, consectetur adipiscing elit. Pellentesque molestie dictum quam non congue. Integer ut dolor vel mi vestibulum molestie vel et nulla.\thinrule

Lorem ipsum dolor sit amet, consectetur adipiscing elit. Pellentesque molestie dictum quam non congue. Integer ut dolor vel mi vestibulum molestie vel et nulla. \stopalignment \stoplinenumbering

\section{Correct} \startalignment[r2l] Lorem ipsum dolor sit amet, consectetur adipiscing elit. Pellentesque molestie dictum quam non congue. Integer ut dolor vel mi vestibulum molestie vel et nulla. \stopalignment

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

This effectively hides the problem. Unfortunately it is apparently not straightforward to invoke the \thinrule automatically at the end of every paragraph [see the other thread with Alan].

Best wishes Idris -- Idris Samawi Hamid Professor 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 : http://foundry.supelec.fr/projects/contextrev/ wiki : http://contextgarden.net ___________________________________________________________________________________

Thank you very much for this, Idris. What with the other bugs and problems I’ve faced with R2L (in footnotes, etc.), I was starting to wonder whether a more explicit, rather than global, use of RTL might be a better approach. I’m going to experiment with them, at least to learn how to use them with greater focus. The good news is that apparently Hans has already patched the page-lin.lua file that Wolfgang had identified as the culprit, and it will be included in the next beta. Many thanks to you, Pablo, Wolfgang, and of course Hans so much for your help in getting these right-to-left issues solved. There are a couple of small bidi issues remaining, including that mentioned in the “footnotes when using RtL (Arabic)” thread. The one or two other ones I will post shortly. But, in any case, I have now completely moved my project over into ConTeXt, confident that it (and the team behind it) can handle all of the needs of a multilingual academic monograph and of a bi-direction, Arabic critical edition’s need With many thanks, Talal

On 2 Aug 2015, at 18:15, Idris Samawi Hamid ادريس سماوي حامد wrote:

Dear Talal, Salaam. See below:

On Sun, 02 Aug 2015 07:31:37 -0600, talazem@fastmail.fm wrote:

...

Thank you very much for this, Idris. It works nicely in the sample file you sent. But it goes off again once you add \setupdirections[bidi=global]. Consider the following:

========== <snip>

...

\setupdirections[bidi=global]

\setupthinrules[color=white] \setupwhitespace[big]

\section{Incorrect} \startlinenumbering \startalignment[r2l] هذه رسالة لطيفة ألفها العالم الكبير الفهامة النحرير. هذه رسالة جامعة مانعة ألفها الشيخ الفاضل... \thinrule ===========

You're right. \thinrule has LR unicode characters which is probably why it gets thrown off. So if you turn off global bidi and then either 1) just go with either global RL or 2) invoke RL inside the linenumbering it will work:

% \setupdirections[bidi=global] \righttoleft %% global RL

or just

\startlinenumbering \righttoleft %% local RL, not needed if global RL invoked \startalignment[r2l]

Again, it's a barbarous (and temporary) workaround. See attached.

Wassalaam Idris

PS Unless there is a lot of Latin in the critical edition file, I would probably depend more on explicit dir commands (\righttolet, \lefttoright) and less on bidi, which is still buggy in a number of ways. Beyond the ConTeXt implementation: The unicode bidi algorithm was not designed with advanced typesetting and typography in mind, works best for text-editing etc. per se. so local bidi *may perhaps* be more useful than global bidi for the critical-edition portion your project.

I'm not at all suggesting that this particular bug should not be a priority; printing paragraph lines in a linenumbering environment *must* be direction-agnostic. Just a matter of general strategy to consider. -- Idris Samawi Hamid Professor 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 : http://foundry.supelec.fr/projects/contextrev/ wiki : http://contextgarden.net ___________________________________________________________________________________

what would be the proper name for such a conversion? simplearabic? arabicdigits?

[myconversion] [١,٢,٣,٤,٥,٦,٧,٨,٩,١٠, ١١,١٢,١٣,١٤,١٥,١٦,١٧,١٨,١٩,٢٠, ۲۱,۲۲,۲۳,۲٤,۲٥,۲٦,۲۷,۲۸,۲۹,۳۰, another million entries

Hans ----------------------------------------------------------------- Hans Hagen | PRAGMA ADE Ridderstraat 27 | 8061 GH Hasselt | The Netherlands tel: 038 477 53 69 | voip: 087 875 68 74 | www.pragma-ade.com | www.pragma-pod.nl -----------------------------------------------------------------

Hi Hans, On Mon, 03 Aug 2015 09:13:34 -0600, Hans Hagen wrote:

On 8/3/2015 4:56 PM, Hans Hagen wrote:

...

what would be the proper name for such a conversion? simplearabic? arabicdigits?

...

[myconversion] [١,٢,٣,٤,٥,٦,٧,٨,٩,١٠, ١١,١٢,١٣,١٤,١٥,١٦,١٧,١٨,١٩,٢٠, ۲۱,۲۲,۲۳,۲٤,۲٥,۲٦,۲۷,۲۸,۲۹,۳۰, another million entries

the next beta will have:

\setuplinenumbering[conversion=arabicdecimals]

What's wrong with 'persiannumerals' and 'arabicnumerals' which we had before? They are not supposed to by synonymous for abjadnumerals... In any case we need U+: [0030-0039] -> [0660-0669]: arabicdecimals U+: [0030-0039] -> [06F0-06F9]: persiandecimals Best wishes Idris -- Idris Samawi Hamid Professor of Philosophy Colorado State University Fort Collins, CO 80523

On Mon, 03 Aug 2015 09:56:33 -0600, Hans Hagen wrote:

On 8/3/2015 5:42 PM, Idris Samawi Hamid ادريس سماوي حامد wrote:

...

Hi Hans,

On Mon, 03 Aug 2015 09:13:34 -0600, Hans Hagen wrote:

...

On 8/3/2015 4:56 PM, Hans Hagen wrote:

...

what would be the proper name for such a conversion? simplearabic? arabicdigits?

...

[myconversion] [١,٢,٣,٤,٥,٦,٧,٨,٩,١٠, ١١,١٢,١٣,١٤,١٥,١٦,١٧,١٨,١٩,٢٠, ۲۱,۲۲,۲۳,۲٤,۲٥,۲٦,۲۷,۲۸,۲۹,۳۰, another million entries

the next beta will have:

\setuplinenumbering[conversion=arabicdecimals]

What's wrong with 'persiannumerals' and 'arabicnumerals' which we had before? They are not supposed to by synonymous for abjadnumerals... In any case we need

U+: [0030-0039] -> [0660-0669]: arabicdecimals U+: [0030-0039] -> [06F0-06F9]: persiandecimals

don't confuse decimals with alphabetic ones

a b c ... z aa ab ac ... az ....

vs

1 2 3 ... 10 11 12

so decimals is mapping chars 0-9 onto whatever else, in principle we can have 15 => ae and 21 => ba

Sure, but that's what abjadnumerals are for. Right now we have abjadnumerals=persiannumerals=arabicnumerals If that's intentional, then no problem, arabicdecimals and persiandecimals are fine nomenclature Ah, it just hit me: You're making a literal analogy with romannumerals.... Best wishes Idris -- Idris Samawi Hamid Professor of Philosophy Colorado State University Fort Collins, CO 80523

Hi Hans, Idris, Talal, Some years ago we had a discussion on the issue, but I don’t remember where we ended up… In my opinion, I think that the numberings should be maned as follows: arabicnumerals: should be ١,٢,٣,٤,٥,٦,٧,٨,٩,٠ persiannumerals: should be ۱ ۲ ۳ ۴ ۵ ۶ ۷ ۸ ۹ ۰ abjadnumerals: should be الف، ب، ج، د، ه، و، ز،….. As of today it seems that « persiannumerals » results in what should be named abjadnuemrals, and this is not correct. Best regards: OK

On 03 Aug 2015, at 16:56, Hans Hagen wrote:

what would be the proper name for such a conversion? simplearabic? arabicdigits?

...

[myconversion] [١,٢,٣,٤,٥,٦,٧,٨,٩,١٠, ١١,١٢,١٣,١٤,١٥,١٦,١٧,١٨,١٩,٢٠, ۲۱,۲۲,۲۳,۲٤,۲٥,۲٦,۲۷,۲۸,۲۹,۳۰, another million entries

Hans

----------------------------------------------------------------- Hans Hagen | PRAGMA ADE Ridderstraat 27 | 8061 GH Hasselt | The Netherlands tel: 038 477 53 69 | voip: 087 875 68 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 : http://foundry.supelec.fr/projects/contextrev/ wiki : http://contextgarden.net ___________________________________________________________________________________

Hi Hans, Thanks for the new beta. As Idris pointed out, the logic you follow for naming greeknumerals, romannumerals, persiannumerals, arabicnumerals, and all others in core-con.lua, makes sense but I didn’t notice this logic before… Which means one may learn each and evry day :-) As far as I know, at least in the Persian use of the abjad system of counting, the first case corresponding to the number 1, is usually written الف in order to avoid the confusion with the digit ۱ I don’t know if this can be fixed in the table you have in core-con.lua. A further remark is that, don’t you think that the pagenumbering, as well as the numbers for chapters, sections, items, equations, etc, should be switched by default to arabicdecimals, or persiandecimals, as soon as \mainlanguage has been set to arabic or persian? Or maybe one could have a command like \setupdefaultnumberings[arabicdecimals] which sets a default for conversion of numbers for all the counters. Again many thanks for the features you included in ConTeXt for bidi writing. Best regards: OK

On 04 Aug 2015, at 01:30, Hans Hagen wrote:

On 8/3/2015 9:25 PM, Otared Kavian wrote:

...

Hi Hans, Idris, Talal,

Some years ago we had a discussion on the issue, but I don’t remember where we ended up… In my opinion, I think that the numberings should be maned as follows:

arabicnumerals: should be ١,٢,٣,٤,٥,٦,٧,٨,٩,٠

persiannumerals: should be ۱ ۲ ۳ ۴ ۵ ۶ ۷ ۸ ۹ ۰

abjadnumerals: should be الف، ب، ج، د، ه، و، ز،…..

As of today it seems that « persiannumerals » results in what should be named abjadnuemrals, and this is not correct.

so you and idris and tatal has to come to some agreement on that

we have these alphabetic ones

arabicnumerals persiannumerals arabicexnumerals

and some special ones

abjadnumerals abjadnodotnumerals

(these use small, medium, large chars and do special things with 10 100 1000)

and now

arabicdecimals persiandecimals

see core-con.lua

Hans

----------------------------------------------------------------- Hans Hagen | PRAGMA ADE Ridderstraat 27 | 8061 GH Hasselt | The Netherlands tel: 038 477 53 69 | voip: 087 875 68 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 : http://foundry.supelec.fr/projects/contextrev/ wiki : http://contextgarden.net ___________________________________________________________________________________

Just to note: all numbers in the body text are having their digits placed individually right-to-left. Notice, in the attached PDF, how the ١٤ is displayed as ٤١, and how ١٨٧٢ is ٢٧٨١ … but also the footnote marker is ٣١ instead of ١٣ as it correctly is below in the footnote itself. Is there any temporary work around for this, until we sort out the underlying causes? Talal

On 3 Aug 2015, at 16:19, Hans Hagen wrote:

On 8/3/2015 3:54 PM, talazem@fastmail.fm wrote:

...

I am having problems with the directionality of numbers and mirrored characters (like parentheses). When using bidi, the numbers are inverted (5102 instead of 2015); with the explicit commands, the mirrored characters are inverted as well.

we have several approaches:

the oldest method is based on some heuristics and targets at tex input in the sense that we assume tagging and directional switches

the other two are unicode bidi based (successive versions and there might be more as that evolves)

in any case, everything bidi / rl has to fit into some system and that's something idris and i are working on (well, we did at some point but we need to pick up that thread) ... so maybe you can trigger idris into onthology mode and help him

Hans

----------------------------------------------------------------- Hans Hagen | PRAGMA ADE Ridderstraat 27 | 8061 GH Hasselt | The Netherlands tel: 038 477 53 69 | voip: 087 875 68 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 : http://foundry.supelec.fr/projects/contextrev/ wiki : http://contextgarden.net ___________________________________________________________________________________

Hans, the beta is a solid step forward regarding bidi and Arabic support, and not just in footnotes. Thank you. However, it appears that my earlier declarations of ‘problem solved’ were a bit premature. Footnotes display everything correctly, but not linenumber notes. Using the three bidi methods of \setupdirections[bidi]: - ‘default’ orders the linenumber notes and their texts right-to-left, but any number therein is reversed (4102 instead of 2014). - ‘one’ does not render either the note block nor the individual notes' texts right-to-left. - ‘two’ is identical to method ‘one’. Compare these to regular footnotes, where everything is rendered correctly as of this new beta. Below is a minimal example with resulting PDF output. Talal ========= %%% LANGUAGE FONTS BIDI \mainlanguage[arabic] \definefontfamily [mainface] [rm] [ALM Fixed] [range=arabic,features=arabic] \setupbodyfont[mainface,11pt] \setupdirections[bidi=global,method=two,fences=no] \setupalign[r2l] %% HEADINGS \setuppagenumber[numberconversion=arabicdecimals] %%% NOTES \definenote[footnote][% Used for my comments in English. setups=note:footnote, align=righttoleft, rule={on,right}, paragraph=yes, ] \setupnotations[footnote][% align=righttoleft, alternative=serried, width=broad, distance=.5em, display=no, numberconversion=arabicdecimals, headstyle=\lefttoright, ] \setuplinenumbering[% style=\tfxx, referencing=on, step=1, location=outer, method=page, align=left, distance=1em, width=0.4em, conversion=arabicdecimals, ] \definelinenote[linenotedefault][% setups=note:linenotedefault, align=righttoleft, rule={on,right}, paragraph=yes, ] \setupnotations[linenotedefault][% align=righttoleft, alternative=serried, width=broad, distance=.5em, display=no, numberconversion=arabicdecimals, headstyle=\lefttoright, ] \definelinenote[linenoteone][% setups=note:linenoteone, align=righttoleft, rule={on,right}, paragraph=yes, ] \setupnotations[linenoteone][% align=righttoleft, alternative=serried, width=broad, distance=.5em, display=no, numberconversion=arabicdecimals, headstyle=\lefttoright, ] \definelinenote[linenotetwo][% setups=note:linenotetwo, align=righttoleft, rule={on,right}, paragraph=yes, ] \setupnotations[linenotetwo][% align=righttoleft, alternative=serried, width=broad, distance=.5em, display=no, numberconversion=arabicdecimals, headstyle=\lefttoright, ] \startsetups[note:footnote] \setupdirections[bidi=on,method=two,fences=no] \stopsetups \startsetups[note:linenotedefault] \setupdirections[bidi=on,method=default] \stopsetups \startsetups[note:linenoteone] \setupdirections[bidi=on,method=one] \stopsetups \startsetups[note:linenotetwo] \setupdirections[bidi=on,method=two] \stopsetups %%% DOCUMENT \starttext \showframe \startlinenumbering \input ward\footnote{هذه تعليقة أولى كتبت عام ٢٠١٤.} \input ward\footnote{هذه تعليقة ثانية (٢٠١٤).} \input ward\linenotedefault{هذه تعليقة أولى كتبت عام ٢٠١٤.} \input ward\linenotedefault{هذه تعليقة ثانية (٢٠١٤).} \input ward\linenoteone{هذه تعليقة أولى كتبت عام ٢٠١٤.} \input ward\linenoteone{هذه تعليقة ثانية (٢٠١٤).} \input ward\linenotetwo{هذه تعليقة أولى كتبت عام ٢٠١٤.} \input ward\linenotetwo{هذه تعليقة ثانية (٢٠١٤).} \stoplinenumbering \stoptext \stoptext =========

On 3 Aug 2015, at 17:52, talazem@fastmail.fm wrote:

Well, I don’t know if ‘solves’ the underlying questions that Hans wrote about, but after some more fiddling, I now have all the numbers appearing as I expect them to.

It required specifying `method=two` as part of `\setupdirections[bidi=global,method=two]`. That, along with the `numbercommand=\righttoleft` in \setuphead[section], means that it now looks as I expect it to.

Unless someone identifies a problem with any of this, I think I’ll write up a summary on the Wiki. Without trolling through the mailing list for anything on RTL and then experimenting with the various combinations, it would be hard for one to divine how to get this to work.

For posterity’s sake, here is a full MWE example:

====== %%% LANGUAGE AND FONTS \mainlanguage[arabic] \definefontfamily [mainface] [rm] [ALM Fixed][range=arabic,features=arabic] \setupbodyfont[mainface,12pt]

\setupdirections[bidi=global,method=two] \setupalign[r2l] %\righttoleft

%% HEADINGS

\setuphead[section][ conversion=myconversion, numbercommand=\righttoleft, % <-- toggle ] \setuplinenumbering[conversion=myconversion] \setupnotations[footnote][numberconversion=myconversion] \setupnotations[linenote][numberconversion=myconversion] \setuppagenumber[numberconversion=myconversion]

\defineconversion [myconversion] [١,٢,٣,٤,٥,٦,٧,٨,٩,١٠, ١١,١٢,١٣,١٤,١٥,١٦,١٧,١٨,١٩,٢٠, ۲۱,۲۲,۲۳,۲٤,۲٥,۲٦,۲۷,۲۸,۲۹,۳۰, ٣١,٣٢,٣٣,٣٤,٣٥,٣٦,٣٧,٣٨,٣٩,٤٠, ٤١,٤٢,٤٣,٤٤,٤٥,٤٦,٤٧,٤٨,٤٩,٥٠, ٥١,٥٢,٥٣,٥٤,٥٥,٥٦,٥٧,٥٨,٥٩,٦٠, ٦١,٦٢,٦٣,٦٤,٦٥,٦٦,٦٧,٦٨,٦٩,٧٠, ٧١,٧٢,٧٣,٧٤,٧٥,٧٦,٧٧,٧٨,٧٩,٨٠, ٨١,٨٢,٨٣,٨٤,٨٥,٨٦,٨٧,٨٨,٨٩,٩٠, ٩١,٩٢,٩٣,٩٤,٩٥,٩٦,٩٧,٩٨,٩٩,١٠٠, ١٠١,١٠٢,١٠٣,١٠٤,١٠٥,١٠٦,١٠٧,١٠٨,١٠٩,١١٠, ١١١,١١٢,١١٣,١١٤,١١٥,١١٦,١١٧,١١٨,١١٩,١٢٠, ١٢١,١٢٢,١٢٣,١٢٤,١٢٥,١٢٦,١٢٧,١٢٨,١٢٩,١٣٠, ١٣١,١٣٢,١٣٣,١٣٤,١٣٥,١٣٦,١٣٧,١٣٨,١٣٩,١١٣٠, ١٤١,١٤٢,١٤٣,١٤٤,١٤٥,١٤٦,١٤٧,١٤٨,١٤٩,١٥٠, ١٥١,١٥٢,١٥٣,١٥٤,١٥٥,١٥٦,١٥٧,١٥٨,١٥٩,١٦٠, ١٦١,١٦٢,١٦٣,١٦٤,١٦٥,١٦٦,١٦٧,١٦٨,١٦٩,١٧٠, ١٧١,١٧٢,١٧٣,١٧٤,١٧٥,١٧٦,١٧٧,١٧٨,١٧٩,١٨٠, ١٨١,١٨٢,١٨٣,١٨٤,١٨٥,١٨٦,١٨٧,١٨٨,١٨٩,١٩٠, ١٩١,١٩٢,١٩٣,١٩٤,١٩٥,١٩٦,١٩٧,١٩٨,١٩٩,٢٠٠, ٢٠١,٢٠٢,٢٠٣,٢٠٤,٢٠٥,٢٠٦,٢٠٧,٢٠٨,٢٠٩,٢١٠, ٢١١,٢١٢,٢١٣,٢١٤,٢١٥,٢١٦,٢١٧,٢١٨,٢١٩,٢٢٠, ٢٢١,٢٢٢,٢٢٣,٢٢٤,٢٢٥,٢٢٦,٢٢٧,٢٢٨,٢٢٩,٢٣٠, ٢٣١,٢٣٢,٢٣٣,٢٣٤,٢٣٥,٢٣٦,٢٣٧,٢٣٨,٢٣٩,٢٢٣٠, ٢٤١,٢٤٢,٢٤٣,٢٤٤,٢٤٥,٢٤٦,٢٤٧,٢٤٨,٢٤٩,٢٥٠, ٢٥١,٢٥٢,٢٥٣,٢٥٤,٢٥٥,٢٥٦,٢٥٧,٢٥٨,٢٥٩,٢٦٠, ٢٦١,٢٦٢,٢٦٣,٢٦٤,٢٦٥,٢٦٦,٢٦٧,٢٦٨,٢٦٩,٢٧٠, ٢٧١,٢٧٢,٢٧٣,٢٧٤,٢٧٥,٢٧٦,٢٧٧,٢٧٨,٢٧٩,٢٨٠, ٢٨١,٢٨٢,٢٨٣,٢٨٤,٢٨٥,٢٨٦,٢٨٧,٢٨٨,٢٨٩,٢٩٠, ٢٩١,٢٩٢,٢٩٣,٢٩٤,٢٩٥,٢٩٦,٢٩٧,٢٩٨,٢٩٩,٣٠٠, ]

%%% DOCUMENT

\starttext \showframe

\section{section one}

\startlinenumbering \input tufte \input tufte \stoplinenumbering

\dorecurse{20}{ \section{محمد الجمل (١٨٧٢—١٩٢٣)} مرحباً باعالم . ١٢٣٤٥٦٧٨٩٠ مرحباً بالعالم. ولد عام ١٩٣٤. \footnote{تعليقة لطيفة.} \pagebreak }

\stoptext ======

Best wishes, Talal

...

On 3 Aug 2015, at 17:29, talazem@fastmail.fm wrote:

Just to note: all numbers in the body text are having their digits placed individually right-to-left. Notice, in the attached PDF, how the ١٤ is displayed as ٤١, and how ١٨٧٢ is ٢٧٨١ … but also the footnote marker is ٣١ instead of ١٣ as it correctly is below in the footnote itself.

Is there any temporary work around for this, until we sort out the underlying causes?

Talal

...

On 3 Aug 2015, at 16:19, Hans Hagen wrote:

On 8/3/2015 3:54 PM, talazem@fastmail.fm wrote:

...

I am having problems with the directionality of numbers and mirrored characters (like parentheses). When using bidi, the numbers are inverted (5102 instead of 2015); with the explicit commands, the mirrored characters are inverted as well.

we have several approaches:

the oldest method is based on some heuristics and targets at tex input in the sense that we assume tagging and directional switches

the other two are unicode bidi based (successive versions and there might be more as that evolves)

in any case, everything bidi / rl has to fit into some system and that's something idris and i are working on (well, we did at some point but we need to pick up that thread) ... so maybe you can trigger idris into onthology mode and help him

Hans

----------------------------------------------------------------- Hans Hagen | PRAGMA ADE Ridderstraat 27 | 8061 GH Hasselt | The Netherlands tel: 038 477 53 69 | voip: 087 875 68 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 : http://foundry.supelec.fr/projects/contextrev/ wiki : http://contextgarden.net ___________________________________________________________________________________

___________________________________________________________________________________ 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 : http://foundry.supelec.fr/projects/contextrev/ wiki : http://contextgarden.net ___________________________________________________________________________________

___________________________________________________________________________________ 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 : http://foundry.supelec.fr/projects/contextrev/ wiki : http://contextgarden.net ___________________________________________________________________________________