Dear teTeX maintainers!
I have a strange problem with texinfo and autobuilders:
On Son, 12 Mär 2006, Bastian Blank wrote:
> Package: texinfo
> Version: 4.8-6
> Severity: serious
>
> > Automatic build of gmsh_1.64.0-1 on debian01 by sbuild/s390 85
> [...]
> > Setting up texinfo (4.8-6) ...
> > Running mktexlsr. This may take some time. ... done.
> > dpkg: error processing texinfo (--configure):
> > subprocess post-installation script returned error exit status 1
> > Setting up html2text (1.3.2a-3) ...
On Son, 12 Mär 2006, Bastian Blank wrote:
> On Sun, Mar 12, 2006 at 01:53:58PM +0100, Norbert Preining wrote:
> > Huuu, could you please run
> > sh -ex /var/lib/dpkg/info/texinfo.postinst configure
> > and send me the output. This can only happen in the kpsewhich call or in
> > the install-info call, and I need to know where.
>
> As it only fails in the autobuilder, I can't provide a sh output, but strace
> shows the problem:
>
> | execve("/usr/bin/kpsewhich", ["kpsewhich", "-var-value", "TEXMFSYSVAR"],
> [/* 26 vars */]) = 0
> | [...]
> | access("/usr/share/texmf/aliases", R_OK) = -1 ENOENT (No such file or
> directory)
> | fstat64(1, {st_mode=S_IFIFO|0600, st_size=0, ...}) = 0
> | mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) =
> 0x77fe6000
> | write(1, "\n", 1) = 1
> | munmap(0x77fe6000, 4096) = 0
> | exit_group(1)
>
> At this point, the postinstall script bails out.
>
> The configuration succeds if I do it by hand after it failed one time, so I
> think it may some sort of recursive dependency, where the order of
> configuration is undefined.
For your information, the texinfo.postinst script contains:
# update-fmtutil checks for /etc/texmf/fmt.d/00tex.cnf which
# may not be present, also the update-fmtutil is already
# present (config files are installed at config time!)
# So we have to check for the existence of this file
# before we can call update-fmtutil
if [ -r /etc/texmf/fmt.d/00tex.cnf ] ; then
if which update-fmtutil > /dev/null ; then update-fmtutil ;
fi
fi
if which mktexlsr >/dev/null; then update_lsr_files; fi
# tetex might be unpacked but not configured, we have to check
# wether kpsewhich can be run!
#if which fmtutil-sys > /dev/null ; then
if kpsewhich --version > /dev/null 2>&1 ; then
build_format_if_format_exists --byfmt texinfo
build_format_if_format_exists --byfmt cyrtexinfo
fi
Where the call to kpsewhich -var-value TEXMFSYSVAR occurs in
build_format_if_format_exists
This seems to be a problem that tetex is not already configured.
Although I thought that this should NOT happen, namely that only with
tex-common installed a call to kpsewhich should succeed. But it seems
that it doesn't. Do you have an idea why?
And how can I check this or circumvent this? I assume that I have to
check whether tetex-base and or -bin is configured. But I don't
understand why the above call to kpsewhich fails. It is strange.
Best wishes
Norbert
-------------------------------------------------------------------------------
Dr. Norbert Preining <preining AT logic DOT at> Università di Siena
gpg DSA: 0x09C5B094 fp: 14DF 2E6C 0307 BE6D AD76 A9C0 D2BF 4AA3 09C5 B094
-------------------------------------------------------------------------------
LIFF (n.)
A book, the contents of which are totally belied by its cover. For
instance, any book the dust jacket of which bears the words. 'This
book will change your life'.
--- Douglas Adams, The Meaning of Liff
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
