FAQ Search Today's Posts Mark Forums Read
» Video Reviews

» Linux Archive

Linux-archive is a website aiming to archive linux email lists and to make them easily accessible for linux users/developers.


» Sponsor

» Partners

» Sponsor

Go Back   Linux Archive > Redhat > Fedora Development

 
 
LinkBack Thread Tools
 
Old 01-06-2008, 11:24 AM
Caolan McNamara
 
Default mktexlsr/texhash/texconfig-sys rehash, what's the canonical %post/%postun for a TeX thing ?

On Sat, 2008-01-05 at 16:24 -0600, Jason L Tibbitts III wrote:
> >>>>> "CM" == Caolan McNamara <caolanm@redhat.com> writes:
>
> CM> So, what exactly is the current problem with these two ? Is it
> CM> just the mysterious failure of TeX to execute some stuff, i.e.
>
> Yes, I debugged this the other day. The problem for bacula is that
> latex2html somehow does not properly run texhash in %post, so nothing
> can find html.sty. If you force a texhash it works. Changing the
> post scriptlet for latex2html to not use "/usr/bin/env -
> /usr/bin/texhash" but to just call texhash directly works. I don't
> know why.

I see that another tex package e.g. jadetex has a comment about
"use texconfig-sys rehash instead of texhash", and the change was...

%post
-/usr/bin/env - PATH=$PATH:%{TeXdir}/bin texhash > /dev/null 2>&1
+[ -x %{_bindir}/texconfig-sys ] && %{_bindir}/texconfig-sys rehash
2> /dev/null || :

+%postun
+[ -x %{_bindir}/texconfig-sys ] && %{_bindir}/texconfig-sys rehash
2> /dev/null || :
+

So maybe a determination of the right way to rehash is called for and
consistent usage might clear this up ?

C.

--
fedora-devel-list mailing list
fedora-devel-list@redhat.com
https://www.redhat.com/mailman/listinfo/fedora-devel-list
 
Old 01-06-2008, 01:30 PM
Jindrich Novy
 
Default mktexlsr/texhash/texconfig-sys rehash, what's the canonical %post/%postun for a TeX thing ?

On Sun, Jan 06, 2008 at 12:24:51PM +0000, Caolan McNamara wrote:
>
> On Sat, 2008-01-05 at 16:24 -0600, Jason L Tibbitts III wrote:
> > >>>>> "CM" == Caolan McNamara <caolanm@redhat.com> writes:
> >
> > CM> So, what exactly is the current problem with these two ? Is it
> > CM> just the mysterious failure of TeX to execute some stuff, i.e.
> >
> > Yes, I debugged this the other day. The problem for bacula is that
> > latex2html somehow does not properly run texhash in %post, so nothing
> > can find html.sty. If you force a texhash it works. Changing the
> > post scriptlet for latex2html to not use "/usr/bin/env -
> > /usr/bin/texhash" but to just call texhash directly works. I don't
> > know why.
>
> I see that another tex package e.g. jadetex has a comment about
> "use texconfig-sys rehash instead of texhash", and the change was...
>
> %post
> -/usr/bin/env - PATH=$PATH:%{TeXdir}/bin texhash > /dev/null 2>&1
> +[ -x %{_bindir}/texconfig-sys ] && %{_bindir}/texconfig-sys rehash
> 2> /dev/null || :
>
> +%postun
> +[ -x %{_bindir}/texconfig-sys ] && %{_bindir}/texconfig-sys rehash
> 2> /dev/null || :
> +
>
> So maybe a determination of the right way to rehash is called for and
> consistent usage might clear this up ?
>

Yes, running rehashing of the kpathsea's ls-R files via texconfig-sys
as noted above shouldn't break anything and should correctly call
mktexlsr to regenerate ls-Rs. I actually suggested it to Ondrej while
fixing the jadetex and iproute build.

Jindrich

--
Jindrich Novy <jnovy@redhat.com> http://people.redhat.com/jnovy/

--
fedora-devel-list mailing list
fedora-devel-list@redhat.com
https://www.redhat.com/mailman/listinfo/fedora-devel-list
 
Old 01-06-2008, 02:17 PM
Caolan McNamara
 
Default mktexlsr/texhash/texconfig-sys rehash, what's the canonical %post/%postun for a TeX thing ?

On Sun, 2008-01-06 at 15:30 +0100, Jindrich Novy wrote:
> Yes, running rehashing of the kpathsea's ls-R files via texconfig-sys
> as noted above shouldn't break anything and should correctly call
> mktexlsr to regenerate ls-Rs. I actually suggested it to Ondrej while
> fixing the jadetex and iproute build.

Sounds good, can we give this a whirl then ? Probably has to be you to
do it as there's ACLs for latex2html.

C.

--
fedora-devel-list mailing list
fedora-devel-list@redhat.com
https://www.redhat.com/mailman/listinfo/fedora-devel-list
 
Old 01-06-2008, 02:29 PM
Patrice Dumas
 
Default mktexlsr/texhash/texconfig-sys rehash, what's the canonical %post/%postun for a TeX thing ?

On Sun, Jan 06, 2008 at 03:30:45PM +0100, Jindrich Novy wrote:
> On Sun, Jan 06, 2008 at 12:24:51PM +0000, Caolan McNamara wrote:
>
> Yes, running rehashing of the kpathsea's ls-R files via texconfig-sys
> as noted above shouldn't break anything and should correctly call
> mktexlsr to regenerate ls-Rs. I actually suggested it to Ondrej while
> fixing the jadetex and iproute build.

Calling
/usr/bin/env - /usr/bin/texhash
should also do the right thing. Currently there is a bug in texhash
when $HOME is empty, introduced in texlive-mktexlsr_fixes.patch.

I'll see what I can do.

--
Pat

--
fedora-devel-list mailing list
fedora-devel-list@redhat.com
https://www.redhat.com/mailman/listinfo/fedora-devel-list
 
Old 01-06-2008, 02:43 PM
Patrice Dumas
 
Default mktexlsr/texhash/texconfig-sys rehash, what's the canonical %post/%postun for a TeX thing ?

On Sun, Jan 06, 2008 at 12:24:51PM +0000, Caolan McNamara wrote:
>
> On Sat, 2008-01-05 at 16:24 -0600, Jason L Tibbitts III wrote:
> > >>>>> "CM" == Caolan McNamara <caolanm@redhat.com> writes:
> >
> > CM> So, what exactly is the current problem with these two ? Is it
> > CM> just the mysterious failure of TeX to execute some stuff, i.e.
> >
> > Yes, I debugged this the other day. The problem for bacula is that
> > latex2html somehow does not properly run texhash in %post, so nothing
> > can find html.sty. If you force a texhash it works. Changing the
> > post scriptlet for latex2html to not use "/usr/bin/env -
> > /usr/bin/texhash" but to just call texhash directly works. I don't
> > know why.

I think that the atached patch should fix that issue.

--
Pat
? .build-2007-6.fc9.log
? .texlive.spec.swp
? fmtutil-ptex.cnf
? mendexk2.6e
? ptex-src-3.1.10
? texlive-2007
? texlive-mktexlsr_fixes.patch-save
? texlive.spec-font_dep
? texlive.spec-obsfonts
Index: texlive-mktexlsr_fixes.patch
================================================== =================
RCS file: /cvs/pkgs/rpms/texlive/devel/texlive-mktexlsr_fixes.patch,v
retrieving revision 1.1
diff -u -3 -p -r1.1 texlive-mktexlsr_fixes.patch
--- texlive-mktexlsr_fixes.patch 2 Dec 2007 08:03:19 -0000 1.1
+++ texlive-mktexlsr_fixes.patch 6 Jan 2008 15:43:18 -0000
@@ -1,18 +1,23 @@
-## 10_mktexlsr_fixes by <frank@debian.org> etc etc
-##
-## DP: Fixes wrong paths in various scripts to make lintian shut up.
-## DP: Fix creation of ls-R in root's homedir
-## DP: Also add a note to the man page of mktexlsr about the above fix
-
- build/source/texk/kpathsea/mktexlsr | 24 ++++++++++++------------
- texmf/doc/man/man1/mktexlsr.1 | 9 +++++++++
- 2 files changed, 21 insertions(+), 12 deletions(-)
-
-Index: texlive-bin-2006.svn3816/build/source/texk/kpathsea/mktexlsr
-================================================== =================
---- texlive-bin-2006.svn3816.orig/build/source/texk/kpathsea/mktexlsr 2006-12-25 19:44:43.000000000 +0100
-+++ texlive-bin-2006.svn3816/build/source/texk/kpathsea/mktexlsr 2007-01-26 03:55:05.000000000 +0100
-@@ -82,6 +82,9 @@
+diff -up texlive-2007/texk/kpathsea/mktexlsr.man.mktexlsr_fixes texlive-2007/texk/kpathsea/mktexlsr.man
+--- texlive-2007/texk/kpathsea/mktexlsr.man.mktexlsr_fixes 2006-01-17 22:41:51.000000000 +0100
++++ texlive-2007/texk/kpathsea/mktexlsr.man 2008-01-06 16:13:33.000000000 +0100
+@@ -44,3 +44,12 @@ Print help message and exit.
+ .B --version
+ .rb
+ Print version information and exit.
++."============================================== =======================
++.SH NOTES
++When called by root with no arguments, fBmktexlsrfP in Debian ignores
++TEXMF trees under fI$HOMEfP. This is to avoid creating undesirable files
++such as fI/root/texmf/ls-RfP when doing usual maintainance (it is generally
++a bad idea to work with TeX as root, therefore having a file such as
++fI/root/texmf/ls-RfP in the first place is rather pointless). If you really
++want to update the ls-R databases for such TEXMF trees, simply list them
++explicitely on the command-line.
+diff -up texlive-2007/texk/kpathsea/mktexlsr.mktexlsr_fixes texlive-2007/texk/kpathsea/mktexlsr
+--- texlive-2007/texk/kpathsea/mktexlsr.mktexlsr_fixes 2006-12-25 19:44:43.000000000 +0100
++++ texlive-2007/texk/kpathsea/mktexlsr 2008-01-06 16:37:22.000000000 +0100
+@@ -82,10 +82,21 @@ test $# = 0 && {
'
set x `kpsewhich --show-path=ls-R | tr : '
' | sort | uniq`; shift
@@ -22,18 +27,19 @@ Index: texlive-bin-2006.svn3816/build/so
IFS=$OIFS
}

-@@ -89,6 +92,10 @@
- # Prepend cwd if the directory was relative.
- case "$TEXMFLS_R" in
- "") continue ;; # Strictly speaking, it is an error if this case is taken.
-+ $HOME/*) if test -n "$NOROOTHOME"; then
+ for TEXMFLS_R in "$@"; do
++ if [ "z$HOME" != 'z' ]; then
++ case "$TEXMFLS_R" in
++ $HOME/*) if test -n "$NOROOTHOME"; then
+ tty -s && echo "$progname: Skipping $TEXMFLS_R" >&2
+ continue
+ fi ;;
- /* | [A-z]:/*) ;;
- *) TEXMFLS_R="`pwd`/$TEXMFLS_R"
- esac
-@@ -112,9 +119,9 @@
++ esac
++ fi
+ # Prepend cwd if the directory was relative.
+ case "$TEXMFLS_R" in
+ "") continue ;; # Strictly speaking, it is an error if this case is taken.
+@@ -112,9 +123,9 @@ for TEXMFLS_R in "$@"; do
db_dir=`echo "$db_file" | sed 's%/[^/][^/]*$%%'` # can't rely on dirname

test -d "$db_dir" || continue
@@ -44,7 +50,7 @@ Index: texlive-bin-2006.svn3816/build/so
cp /dev/null "$db_file"
# Use same permissions as parent directory, minus x,s, or t bits.
chmod `kpsestat -xst "$db_dir"` "$db_file"
-@@ -128,11 +135,8 @@
+@@ -128,11 +139,8 @@ for TEXMFLS_R in "$@"; do
# Skip if we cannot write the file:
kpseaccess -w "$db_file" || { echo "$progname: $db_file: no write permission. Skipping..." >&2; continue; }

@@ -58,7 +64,7 @@ Index: texlive-bin-2006.svn3816/build/so

$verbose && echo "$progname: Updating $db_file... " >&2
echo "$ls_R_magic" >"$db_file_tmp"
-@@ -152,12 +156,8 @@
+@@ -152,12 +160,8 @@ for TEXMFLS_R in "$@"; do
| sed -e '/.svn.*:$/,/^$/d'
>>"$db_file_tmp"

@@ -73,21 +79,3 @@ Index: texlive-bin-2006.svn3816/build/so
done
$verbose && echo "$progname: Done." >&2
exit 0
-Index: texlive-bin-2006.svn3816/texmf/doc/man/man1/mktexlsr.1
-================================================== =================
---- texlive-bin-2006.svn3816.orig/build/source/texk/kpathsea/mktexlsr.man 2007-01-14 19:01:06.000000000 +0100
-+++ texlive-bin-2006.svn3816/build/source/texk/kpathsea/mktexlsr.man 2007-01-26 03:55:05.000000000 +0100
-@@ -44,3 +44,12 @@
- .B --version
- .rb
- Print version information and exit.
-+."=============================================== ======================
-+.SH NOTES
-+When called by root with no arguments, fBmktexlsrfP in Debian ignores
-+TEXMF trees under fI$HOMEfP. This is to avoid creating undesirable files
-+such as fI/root/texmf/ls-RfP when doing usual maintainance (it is generally
-+a bad idea to work with TeX as root, therefore having a file such as
-+fI/root/texmf/ls-RfP in the first place is rather pointless). If you really
-+want to update the ls-R databases for such TEXMF trees, simply list them
-+explicitely on the command-line.
-
Index: texlive.spec
================================================== =================
RCS file: /cvs/pkgs/rpms/texlive/devel/texlive.spec,v
retrieving revision 1.12
diff -u -3 -p -r1.12 texlive.spec
--- texlive.spec 4 Jan 2008 13:24:51 -0000 1.12
+++ texlive.spec 6 Jan 2008 15:43:18 -0000
@@ -432,7 +432,7 @@ chmod -x texk/dvipdfm/encodings.c
%patch43 -p1 -b .perl

%patch100 -p3
-%patch101 -p3
+%patch101 -p1 -b .mktexlsr_fixes
%patch102 -p3
%patch104 -p3
%patch105 -p3
--
fedora-devel-list mailing list
fedora-devel-list@redhat.com
https://www.redhat.com/mailman/listinfo/fedora-devel-list
 

Thread Tools




All times are GMT. The time now is 04:48 PM.

VBulletin, Copyright ©2000 - 2014, Jelsoft Enterprises Ltd.
Content Relevant URLs by vBSEO ©2007, Crawlability, Inc.
Copyright 2007 - 2008, www.linux-archive.org