[luatex] ini files for TL 2009

[Manuel Pégourié-Gonnard]

Karl Berry a écrit :
>     > Can I object, please? Why pretend that luatex is nothing more than
>     > pdftex?
> 
> I agree.  Sorry, I didn't realize the implication.
> 
Now, my turn to object. I thought the whole point of hiding primitives in INI
mode and providing tex.enableprimitive() with the ability to activate primitive
with a prefix, was to be able not to activate them and let the macro package decide.

Using a *lua*.ini file which activates only pdfetex primitives does exactly
this: avoid name conflicts (since one can assume most packages to be compatible
with pdftex, and otoh doing so allows to reuse existing .ini files). Then the
user willing to have a full luatex experience loads sdome package such as
luatextra for LaTeX.

The package tries to take care of possible conflicts, possibly applying a
coherent naming scheme (well, probably not in LaTeX2e, but we can hope for
LaTeX3), and more generally provides an additional abstraction layers between
the engine and the user.

This abstraction layer frees luatex developers of the burden of compatibility
questions, allowing them to name the new primitives as they wish, and change
their syntax when the want to, like for \directlua, since the macro package is
responsible of this questions.

So my opinion is: only pdfetex primitives ( + \directlua) in the format, and
then macro packages such as ifluatex.sty and luatextra.sty.

> It doesn't make sense to me for "luatex" to not provide anything extra.
> It seems to me that (pdf)lua(la)tex should provide the luatex
> primitives, just as pdf(la)tex provides pdftex primitives, etc.
> 
So what exactly is the point of tex.enableprimitives() and hiding primitives in
the first place?

Manuel.