[NTG-context] Malayalam conjuncts

Ajith R ajithramayyan at yahoo.co.in
Sat Jan 1 11:00:05 CET 2022

On Sat, 01 Jan 2022 09:48:53 +0530
kauśika <citturs at gmail.com> wrote:

> On Friday, December 31, 2021 6:22:15 PM IST Ajith R via ntg-context
> wrote:
> > The conjuncts that are not formed
> > correctly are those where the second component is ര U+0D30,
> > followed by a symbol that is shown on the right side of the
> > conjunct viz ാ U+0D3E, ി U+0D3F, ീ U+0D40, ു  U+0D41, ൂ U+0D42. If
> > no symbols follow, or symbols follow on the left or on both sides,
> > the conjunct is well formed.In the minimum working example given 5
> > is well formed, while the first conjunct in 6 is well formed and
> > second ill formed conjunct.  
> I have been using ConTeXt to typeset documents in several Indic
> languages and have run into similar issues (in many languages).
> Please see this for a similar issue in some conjuncts for Devanagari:
> https://www.mail-archive.com/ntg-context@ntg.nl/msg99691.html
> For what its worth, I have not had issues with some fonts while
> issues with others persist. Some of these issues we can work around
> as I have pointed out in the above posting.
> In almost all cases I encountered no issues while using Xe(La)Tex.
> Based on some advise from Hans and reading about these OTF features
> and their implementations in Indic fonts, I think these issues might
> be due to differences in implementation. [Not entirely sure since I
> am a novice]. My guess is that Harfbuzz (which is what Xe(La)TeX uses
> by default) uses some heuristics to work out these conjuncts (?!).
> To answer your specific question regarding the conjuncts in the given
> words you have to use some Unicode hacking to get what you want in
> ConTeXt.
> In each of the following ZWS refers to the Unicode character
> (zero-width space U+200B)
> 1. ശ്രീ 
> This is typeset correctly by writing  
> 	ശ്ര + 
> 	ZWS (U+200B) + 
> 2. അശ്രു
> Typeset correctly with
> 	അശ്ര + 
> 	ZWS (U+200B) + 
> 	​ൂ
> 3. ശുശ്രൂഷ
> Typeset correctly with
> 	ശുശ്ര +
> 	ZWS (U+200B) + 
> 	ൂ +
> 4. പ്രാസം
> Typeset correctly with
> 	പ്ര + 
> 	ZWS (U+200B) + 
> 	ാ +
> 	സം
> 5. പ്രേയസി (rendered correctly as entered; no hacks necessary)
> 6. പ്രോഗ്രാം
> Typeset correctly with
> 	പ്രൊ +
> 	ഗ്ര + 
> 	ZWS (U+200B) +
> 	ാ +‌
>> where the last character is the Malayalam Anusvara.
> Consider yet another example:
> സാന്ദ്രാനന്ദാഅവബൊധാത്മകമ്
> Here the 'ന്ദ്രാ' conjunct is not typeset in ConTeXt. To fix this I
> do ന്ദ്ര +
> 	ZWS (U+200B) +
> 	​ാ
> This is what I have been doing to ensure correct typesetting of
> Malayalam and other Indic languages in ConTeXt. Honestly, it is
> inconvenient since the .tex files containing Unicode are no longer
> sanitary. However, ConTeXt has so many remarkable features that the
> very thought of having to go back to (Xe)LaTeX (just for harfbuzz
> rendering) causes me immense pain. As far as I am concerned, in every
> other way ConTeXt simply has no match in the (Xe)LaTeX world. In my
> usage of ConTeXt for my academic work (in English with lots of
> mathematics) I have encountered no issues. Even if I did there was
> always some legitimate (non-hacky) fix for it. For me personally, the
> rendering in Indic languages is the only pain point with ConTeXt
> (which I am willing to live with).
> So I am willing to live with the drawbacks till the day they are
> hopefully fixed. Anyway, I hope you can use these fixes temporarily.
> For example, if your editor supports it, you can replace all glyphs
> with this issue with the corresponding recipe involving ZWS.
> Dear Hans and other developers of ConTeXt, LuaTeX, 
> If you happen to see this please look into the font system (where it
> concerns Indic systems). The present issue is very similar to the one
> I posted about earlier:
> https://www.mail-archive.com/ntg-context@ntg.nl/msg99691.html
> I have described the issue and the hacks to fix it the best I can. In
> case there is any other information that I can provide please let me
> know.
> Best,
> kauśika

Hi Kausika,

Your workaround seems to help, though it entails adding the ZWS
throughout. Anyway, I agree with you in that even with this anomaly,
Context is worth sticking to.
Dear Hans and other developers of ConTeXt, LuaTeX, 
If a novice can help in any way to resolve this issue, please guide us
how we can contribute.


More information about the ntg-context mailing list