[tex-live] Re: context / directory structure

[Olaf Weber]

Karl Berry writes:

> (Back on some older mail.)
>     the boundary condition that files are unique is a dangerous one
>     (where did it come from?);

> It has been there since I first implemented subdirectory searching for
> Web2c, what, 15 years ago or so (gadzooks!).  In any given search path,
> if two files have the same name, it's indeterminate which one you'll get.

>     PDFTEX.MAPPATH

> Indeed, since the map files are incompatible, we could need more search
> paths.  However, in practice, I note that right now there are no
> filename duplications within texmf*/fonts/map.  So I guess we have
> muddled through to this point.

> A better solution, as many have pointed out but none have implemented,
> would be to make map files more consistent between programs, with a
> commonly-agreed syntax for extensions so that programs can ignore
> extensions not meant for them or not implemented yet.  Unfortunately I
> cannot take that on.  Is anyone inclined?

If the code is to be common, it might make sense to move the parser
into either web2c/lib/lib.a or into linkpathsea (the latter perhaps in
the longer run).

The main issue to figure out what has to be implemented.  Once that is
known, it shouldn't be hard.

> _______________________________________________
> TeX Live mailing list
> http://tug.org/mailman/listinfo/tex-live

-- 
Olaf Weber

               (This space left blank for technical reasons.)