[tex-live] texdoc error

[Manuel Pégourié-Gonnard]

Hi Reinhard,

Reinhard Kotucha a écrit :
> [...]
> With all this said, the wrapper script
> (bin/win32/tl-w32-wrapper.texlua) should be comprehensible.  The
> function fixwin() is supposed to solve the quoting problem.
> 
Thanks for this interesting informations. To me, the script looked
rather understandable except ths need for fixwin() which is indeed
unintuitive as you say. Anyway, by a funny coincidence, I was just
beginning to think the os.spawn may well be the culprit just before I
read your mail. More about this in a few lines.

> If you replace "--[[" by "---[[", i.e. prepending an extra hyphen, the
> wrapper script doesn't invoke a particular script but it prints to
> screen *how* it would invoke it.
> 
Well, we did it and the results were ok, but it didn't explain us
anything: it just told us where the problem isn't...

Now, here is what I believe is the really minimal example of this
problem. This example assumes you have acrord32 in your path and it's
adobe reader 9, and that presntation.pdf exists. It uses two files:

% wrapper.tlu
kpse.set_program_name("texlua")
os.spawn('texlua d:/mpg/program.tlu')
-- os.execute('texlua d:/mpg/program.tlu)

% program.tlu
os.execute ('acrord32 "d:\\mpg\\presentation.pdf"')
-- os.spawn ('acrord32 "d:\\mpg\\presentation.pdf"')

Now this is the only sequence (spawn then execute) that makes the
problem happen. Every other sequence does work (spawn -> spawn, execute
-> execute, execute -> spawn). For some reason kpse needs to be
initilised to exhibit the problem but I believe (without being able to
prove is) that it could be reproduce with something else in lieu of kpse.

Now I think it's really time to ask the luatex people what they think of it.

Manuel.