path-searching again

[Paul Vojta]

On Sat Oct 28 13:56:43 1995, mackay@cs.washington.edu (Pierre MacKay) wrote:

>    >   This improves the maintainability of the font tree,
>    >   but, unless all the programs that search this tree employ some form of
>    >   caching, there are serious performance concerns.  For example, in order
>    >   to find \path|cmr10.tfm|, {\TeX} would potentially have to search
>    >                       ^^^
>    >   through all the directories that contain \path|pk| files
>    >                                                  ^^
>    >   in all modes and at all resolutions.
>    > 
>    > I don't understand this. Is this a mix-up or something?
> 
>    No, but implementing texmf/fonts/supplier/typeface/type/... would be a
>    mix-up for the above-mentioned reason.  The problem is:  how does TeX
>    know that the dvi [I meant tfm] file isn't located in some weird place like
> 
> 	   texmf/fonts/foo/bar/pk/dpi300/tfm/cmr10.tfm ?
> 
> 
> Was this ever fully answered?

I believe that I fully answered the first question.  I intended the
second question to be rhetorical.

> I am rushing through back mail and
> it may be redundant to comment at this time.  But---
> 
> With a rationally constructed texmf.cnf, or its equivalent
> in another environment, a tfm file placed in
> 
> 	   texmf/fonts/foo/bar/pk/dpi300/tfm/cmr10.tfm 
> 
> would simply not be found, and why should it?

How would you "rationally" set up a texmf.cnf without creating all the
problems that recursive searches were intended to avoid?  The most
straightforward path in texmf.cnf would be

	texmf/fonts//tfm

and that most certainly would have to find the above .tfm file.

--Paul Vojta, vojta@math.berkeley.edu