[tex-live] textlive seems to ignore top of home tree

hometreetexlive.9.virgilinux at dfgh.net hometreetexlive.9.virgilinux at dfgh.net
Wed Jul 21 15:19:06 CEST 2010



--- On Wed, 7/21/10, Karl Berry - karl at freefriends.org <+hometreetexlive+virgilinux+ba22cd1a74.karl#freefriends.org at spamgourmet.com> wrote:

> From: Karl Berry - karl at freefriends.org <+hometreetexlive+virgilinux+ba22cd1a74.karl#freefriends.org at spamgourmet.com>
> Subject: Re: [tex-live] textlive seems to ignore top of home tree (hometreetexlive: message 2 of 13)
> To: hometreetexlive.9.virgilinux at dfgh.net
> Cc: tex-live at tug.org
> Date: Wednesday, July 21, 2010, 12:19 AM
>     BY DEFAULT it searches
> BOTH the top of TEXMFHOME and the appropriate
>     subdirectory
> 
> As I said before: that would be nice in theory, but as far
> as I can
> tell, it would inflict an unacceptable performance hit for
> some people
> currently using TEXMFHOME.
> 
> The only compatible solution I can think of is to invent a
> new tree that
> doesn't require the subdir structure.
> 
> In any event, neither of these things would be simple to
> do, and require
> considerable thought.  (If you don't understand why
> this is so, please
> just take my word for it.)  Perhaps for TL'11.
> 
> Thanks for the suggestion.
Well, just following Akira's suggestion I achieved the exact functionality that is being proposed with one single line of "code" 
(export TEXINPUTS=/home/rodxxx/texmf: )
With that LaTeX looks for a class first at the top of /home/rodxxx/texmf then in the subdirectory /home/rodxxx/texmf/tex/ which exactly what is being proposed as default behavior (with the obvious benefit that if the user does not put the class/style file in a subdirectory it would still be found).

So the only remotely conceivable "performance hit" that I can imagine is the act of looking for the class/style file at the top of the home tree... and action equivalent to looking for the class/style file in the work directory, which LaTeX currently DOES always, without anyone complaining about it.

I myself cannot notice any performance difference while doing simple tests of this action. Then again, if I had a million items at the top of the home tree then there may be a performance hit... but I cannot imagine a reasonable set of circumstances leading to any user having a huge number of items in the HOME tree... most of what s/he needs should be available at either the distribution tree or the LOCAL tree (both managed by an administrator presumably... otherwise the user could installed any desired package in the main tree, using the standard tools).

But if this was really a problem, it could in principle be solved (I suppose almost as easily) by asking LaTeX to first look inside the appropriate home tree subdirectory and then, ONLY when it could not find what was looking for, at the top of the home tree. In this scenario, whenever the user has put the package in the expected subdirectory, LaTeX would behave in the EXACT same way it does today, so that those with a TEXMFHOME according to the recommended (required?) structure wouls see ZERO effect.

In any case, I continue to be puzzle by the fact that today LaTeX seems to always look (by the fault) in the user work directory (which could conceivably be a huge mess with all sort of files in it) and no one seems bothered by it, yet somehow we are expected to accept that if LaTeX treat the top of the home tree in the exact same way it treats the work directory today some major problems would arise. If this was true, it would be highly counter-intuitive.

Virgil



      




More information about the tex-live mailing list