On Thu, Jun 26, 2008 at 21:59 +0200, Frank Küster wrote:
> This code
>
> #!/bin/sh
>
> IFS=':'
> kpsepath=$(kpsewhich --expand-path='$TEXMF')
> for path in $kpsepath
> do
> eval FOO=\$path echo $FOO
> done
>
> Just echo's nothing, and therefore, if we replace "echo $FOO" by a
> kpsewhich call
reopen 488038
thanks
[Resending, instead of moving to the Cc field to remove the -done part,
I sent the mail]
Ralf Stubner <[EMAIL PROTECTED]> wrote:
> [EMAIL PROTECTED]:~$ kpsewhere --format='web2c files' fmtutil.cnf
> /home/rfstubne/texmf/web2c/fmtutil.cnf
> /var/lib/texmf/web2c/fmtutil.cnf
>
On Wed, Jun 25, 2008 at 21:38 +0200, Frank Küster wrote:
> $ for path in $kpsepath; do eval TEXMF=\$path kpsewhich texmf.cnf; done
> /usr/share/texmf/web2c/texmf.cnf
> /usr/share/texmf/web2c/texmf.cnf
> /usr/share/texmf/web2c/texmf.cnf
> /usr/share/texmf/web2c/texmf.cnf
> /usr/share/texmf/web2c/te
kpsewhere acts very strangely here. It looks as if the whole concept of
kpsewhere flawed, maybe only in newer web2c/libkpathsea versions?
I've never understood/maintained/looked at kpsewhere.
This is what I get in native TL (development):
$ kpsewhere texmf.cnf
/home/texlive/karl/Master/t
Package: texlive-base-bin
Version: 2007.dfsg.1-4.np.etch.1
Severity: normal
[By the way, Norbert, your etch backport of the package has a higher
version number than the lenny version. I've switched to lenny meanwhile,
but I still have the backport installed until 2007.dfsg.2-1 reaches
testing. Bet
5 matches
Mail list logo
