[tex-live] [luatex] [lltx] Location of recorder file

[Philipp Stephani]

Am 17.05.2011 um 22:23 schrieb Reinhard Kotucha:

> The gap between stock Lua and LuaTeX is already quite large.  What I
> miss in stock Lua is TeX, for example. :)
> 
> Well, I don't think that the gap has to be kept small in the first
> place.  Lua is designed to be embedded in programs on a particular
> platform and is not aware of the operating system at all.  It just
> inherits behavior from the operating system it runs on, for instance
> line endings.  I'd like to use texlua as a general, platform
> independent scripting language without the need of conditional code
> for different platforms in the scripts.  As a first step into this
> direction, Taco made io.read('*lines') line ending neutral.  There are
> not many differences between Lua and LuaTeX, mostly extensions, which
> are described in the manual.
> 
> Lua provides os.execute() as the only function for calling external
> programs.  Taco added os.exec(), which was supposed to replace the
> current process image by another one.  Since this doesn't work as
> expected on Windows, he added os.spawn().  All these functions
> expected a string as an argument, which means that file names with
> spaces have to be quoted in each script, which is quite annoying.
> Also, string concatenation is quite ugly.  With table support we can
> simply pass arg{} to the new process.

Additions don't cause problems, it's more the changes I'm a bit worried about. I think that changed functionality should generally be reflected in an interface addition. What happens if I run a script that expects stock Lua behavior through TeXLua?

> 
> Therefore I asked Taco to add support for Lua tables to
> the new functions.  They now accept either a string or a table.
> 
> If you pass arguments as a table, you don't have to quote, except on
> Windows.  I looked at the description of CreateProcess() and it seems
> that it accepts a string only.  However, I suppose that
> 
>  http://msdn.microsoft.com/en-us/library/20y988d2.aspx
> 
> is the specification Taco was missing.

Yes, that is just the conversion from the argument vector to the command-line string that CreateProcessW expects.
But then you still have to write wrapper code for CreateProcessW, convert the UTF-8 arguments to UTF-16, etc. Not hard to do, but adds even more platform-specific code to LuaTeX.

> 
>> The best thing would be to write an extension module that wraps
>> CreateProcessW directly.  If it's only about calculating the
>> command line from an argument vector (i.e. the list2cmdline)
>> function, then that can be done in pure Lua.
> 
> Sure, this is quite similar to what the fixwin() function does, though
> the latter doesn't create a string, but it would be much nicer if we
> can avoid it completely.

The problem remains that unless you wrap CreateProcessW or at least wspawn* directly, you never get Unicode support on Windows. Therefore you have to add platform-specific code anyway. It's a bit less if you call spawn/wspawn, but it's still there.
If you really want to make Lua truly platform-independent, you have to add those Unicode wrapper functions for Windows. No Unicode support on a Unicode operating system is almost unacceptable.