[XeTeX] XeTeX bugs in bidirectional typesetting

[Simon Cozens]

On 20/11/2016 12:35, Vafa Khalighi wrote:
> For the past 10 years I have reported numerous issues to the LuaTeX 
> and etex teams without any results but it is more than enough. I am 
> not going to waste time and energy doing useless things.

Well, it's not necessarily useless. If both engines are using the same
model, then you have twice as many communities available to fix bugs.
And it means the bug fixes and bidi expertise can be shared between the
two communities. I would also recommend XeTeX moving to the LuaTeX model
- and then fixing it!

> I put my time and energy into developing an engine that really has a
>  working bidi model. an engine which is developed by a native speaker
>  and meets the needs of people with real documents.

XeTeX has a lot of advantages in terms of opentype support, large set of
packages and mature community. It's a shame the bidi support is not
great; that is a known problem and there are not many people with the
expertise to make it work and do it well. As I understand it, the
problems with bidi were one of the reasons that Khaled stopped working
on XeTeX, which in a sense is a shame - he's exactly the sort of person
you need to get this right...

If you want an engine with a working bidi model, then you might want to
have a look at SILE. (https://github.com/simoncozens/sile) It uses the
Unicode bidi algorithm and so you get multilevel reordering without any
markup required. (See
https://github.com/simoncozens/sile/blob/master/examples/arabic.pdf) But
of course then you don't get the large set of packages and the mature
community...

S