[tex-live] perl or lua standard library modules in texlive

Michael Palmer mpalmer at uwaterloo.ca
Sun Sep 23 02:04:53 CEST 2012


Hi Karl,

thanks very much for your reply. In the meantime, I have dug a little  
further and am cobbling together a lua version of the thin client that  
would simply delegate all the real work to my server. This lua client  
would need to import a few more lua modules. No Python files would  
need to be installed on the user's computer.

The code generated by mol2chemfig also needs a little package (.sty)  
with custom macros. In addition, I would like to include a sample  
document (.tex) with mol2chemfig-generated examples (also .tex) and  
maybe some molfiles (.mol).

Could you point me to some example of a somewhat similar package, from  
which I could learn how to arrange the files in mine?

Thanks again and best wishes, Michael

Quoting Karl Berry <karl at freefriends.org>:

> Hi Michael,
>
> Thanks for writing.  I looked at your mol2chemfig when it hit CTAN and
> regretted that I could not make it work in TL as-is.
>
> The issue is not Python, per se -- any language is possible,
> dependencies permitting.  There are a couple Python scripts in TL now
> (dviasm, ebong), but they are standalone executables.  I expect it would
> be possible to make mol2chemfig work in Python (if users have whatever
> prerequisite libraries installed, if any).
>
> I haven't written anything beyond the terse notes on pkgcontrib.html
> that you already found.  Here is a more verbose version:
>
> The two principal issues are:
> - The installation instructions said to run "python setup.py", which I
>   know is the standard method for Python packages, but it doesn't work
>   at all with TL.  What I need are files that I can install in TL
>   directly and that do not depend on the "install-time" environment.
>   (Since the TL user who eventually runs the script will have an
>   environment that arbitrarily differs from mine doing the install.)
>
> - The executables have to be "self-locating", which is to say, they have
>   to find their data files relative to their own location, not based on
>   any absolute prefix.  (Since users can install TL anywhere.)  The
>   package structure would end up being something like this:
> bin/*/mol2chemfig symlink to:
> texmf-dist/scripts/mol2chemfig/mol2chemfig.py
> texmf-dist/scripts/mol2chemfig/atom.py
> texmf-dist/scripts/mol2chemfig/common.py
> ..
>
> In practice, when mol2chemfig is run, it needs to use kpsewhich to find
> its files (or likely, in this case, find the directory to add to
> PYTHONPATH or however it works ...):
>
> subpath=`kpsewhich --progname=mol2chemfig --format=texmfscripts  
> mol2chemfig.py`
> .. `dirname $subpath` ...
>
> I know I'm not spelling everything out, but hopefully those clues make a
> little sense?  Please do ask further as needed.
>
>
> Heiko, you probably do more than anyone with finding subsidiary script
> files.  If you have anything to add/correct to the above, please do ...
>
> Best,
> Karl
>



-- 
Michael Palmer
University of Waterloo
Waterloo, Canada
science.uwaterloo.ca/~mpalmer





More information about the tex-live mailing list