Re: Coordinating upload of teTeX-3.0 to unstable
Kurt Roeckx [EMAIL PROTECTED] wrote: Those really failed to build: - bison-1.35 dvips bison.dvi -o bison.ps This is dvips(k) 5.95a Copyright 2005 Radical Eye Software (www.radicaleye.com) dvips: ! DVI file can't be opened. I cannot reproduce this here. Do you have a build log available? - doc-debian-es: dvips -o debian-faq-es.ps debian-faq-es.dvi This is dvips(k) 5.95a Copyright 2005 Radical Eye Software (www.radicaleye.com) dvips: ! DVI file can't be opened. - gnuplot: dvips: ! DVI file can't be opened. (And other warnings.) Trying these next... Regards, Frank -- Frank Küster Inst. f. Biochemie der Univ. Zürich Debian Developer
Re: Coordinating upload of teTeX-3.0 to unstable
On 07.08.05 Kurt Roeckx ([EMAIL PROTECTED]) wrote: Hi, - gprolog hevea -O -nosymb -s -exec xxdate.exe -fix custom.hva manual.tex -o manual.html /usr/bin/hevea: unknown option `-nosymb'. I guess hevea has nothing to do with teTeX. - laptop-net: texi2pdf laptop-net.texi make[2]: texi2pdf: Command not found texi2pdf was in tetex-bin. Dunno why it disappeared. - scalable-cyrfonts No file fontinst.rc. ) (./fnstcorr.tex ! Undefined control sequence. l.37 \catcode`\_=\underscorecatcode Probably buggy document. - userv debiandoc2ps: ERROR: output could not be generated properly Need an debiandoc2ps -v and probably -k for that. H. -- If only God would give me some clear sign! Like making a large deposit in my name at a Swiss Bank. - Woody Allen http://hilmarpreusse.forum-rheinland.de/ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Coordinating upload of teTeX-3.0 to unstable
Kurt Roeckx [EMAIL PROTECTED] wrote: This was the list of everything directly build depending on it, and I'll do one with the indirect build dependencies soon. You can stop all build-depending on debiandoc-sgml: It has the classical bug that produces PDF output with teTeX-3.0 even when dvi is desired, causing almost everything to FTBFS. I've not yet submitted a bug, because I'm still testing the patch. Should I file it as important and bump the severity to grave only when teTeX-3.0 is unstable? Or should I file as grave at once? Regards, Frank -- Frank Küster Inst. f. Biochemie der Univ. Zürich Debian Developer
Re: Coordinating upload of teTeX-3.0 to unstable
On Tue, Jul 26, 2005 at 06:49:16PM +0200, Frank Küster wrote: Kurt Roeckx [EMAIL PROTECTED] wrote: I can try to build all those for you if you want, and I'll report those where I think that show problems. That would be great. I think most of the problems will show up in failed builds, some might result in different filelists (pdf file generated instead of dvi file, but somehow the build does not fail), a small number might result in wrong fonts being used. I started with everything build depending on tetex-*. Here is an overview of packages which might have problems. Note that I have no clue what some of those message mean. Those really failed to build: - bison-1.35 dvips bison.dvi -o bison.ps This is dvips(k) 5.95a Copyright 2005 Radical Eye Software (www.radicaleye.com) dvips: ! DVI file can't be opened. - ctie Unmet build dependencies: libkpathsea-dev (= 2.0.2-1) - doc-debian-es: dvips -o debian-faq-es.ps debian-faq-es.dvi This is dvips(k) 5.95a Copyright 2005 Radical Eye Software (www.radicaleye.com) dvips: ! DVI file can't be opened. - gnuplot: dvips: ! DVI file can't be opened. (And other warnings.) - gprolog hevea -O -nosymb -s -exec xxdate.exe -fix custom.hva manual.tex -o manual.html /usr/bin/hevea: unknown option `-nosymb'. - iproute: error post-processing latex2e. - laptop-net: texi2pdf laptop-net.texi make[2]: texi2pdf: Command not found - make: dvips: ! DVI file can't be opened. - ptex-bin syntax error: Last token = 289 (!), error buffer = `_', last id = `O' (parameterless procedure). ./pconvert: conversion of failed, moving dregs: ./pconvert: mv jbibtex.c jbibtex.h jbibd.h ./failure - python-crypto: pdfTeX error (ext1): \pdfinfo used while \pdfoutput is not set. - scalable-cyrfonts No file fontinst.rc. ) (./fnstcorr.tex ! Undefined control sequence. l.37 \catcode`\_=\underscorecatcode ? make[1]: *** [build] Terminated make[2]: *** [latex_mtx] Terminated - userv debiandoc2ps: ERROR: output could not be generated properly - lcdf-typetools Unmet build dependencies: libkpathsea-dev (= 2.0.2-4) - libkpathsea-perl Unmet build dependencies: libkpathsea-dev (= 2.0-1) Those just gave warnings: - am-utils ** Unknown command *** Unknown node in menu entry `FSinfo host netif:' (l. 5664) - blitz++: Doesn't seem to make any documentation at all? - cameleon Warning: Could not find program file: xxdate.exe manual.html:1130: Warning, cannot find anchor: pluginchat - cfitsio ./quick.tex:2: Warning: Cannot open file: html.sty ./quick.tex:3: Warning: Command not found: \htmladdtonavigation [...] - dvipdfmx: does not make it's documentation? - ecartis ! pdfTeX warning (ext4): destination with the same identifier (name{page.1}) has been already used, duplicate ignored to be read again \penalty - ecasound2.2: ./eci_doc.latex:15: Warning: File \jobname.htoc not found - freefem: ./freefem.tex:23: Warning: Command not found: \PassOptionsToPackage ./freefem.tex:46: Warning: keyval, uknown key: ``gobble'' ./freefem.tex:46: Warning: keyval, uknown key: ``numbers'' [...] - freefem3d: Package hyperref Warning: Token not allowed in a PDFDocEncoded string, (hyperref)removing `\special' on input line 1. [...] dvips: Unknown keyword (PUSH) in \special will be ignored dvips: Unknown keyword (RED) in \special will be ignored dvips: Could not find figure file COLOR; continuing [...] - freetype1: Does not generate documentation? - gap-ctbllib: Lots of messages like: Reference `CTblLib' undefined. - gnucap: Warning: Could not find program file: xxdate.exe ./gnucap-man.tex:47: Warning: File \jobname.htoc not found Some have the combination of: ** menu entry without previous node: License (l. 98) ** node_prev `Top' for `License' not found ** `Distrib' doesn't appear in menus (am-utils, esh, fdutils) I did not have time to look at all those that appeared to have been built succesful. Can't be build currently (C++ transition, xorg-x11, ...) - acl2 - axiom - boot-floppies - cmucl - crystalspace - darcs - drift - gmsh - ion2 - kdegraphics - lhs2tex - maxima - papaya - evince Have other problems building not related to tetex, and I'll file bugs about those if they don't exist yet: - bibtool - clif - flite - gcl - gclcvs - jlint - lilypond - lyx - pointless - pspp - saoimage - sbcl - sbm - skribe - transfig - xconq - xen - xfree86 - xorg-x11 - tmview Should I worry about messages like: Underfull \hbox (badness 2368) Overfull \hbox (21.3942pt too wide) LaTeX Font Warning: Font shape `OMS/cmtt/m/n' undefined LaTeX Warning: Reference `sec-cblash' on page 7 undefined on input line 452. ** `XXX' doesn't appear in menus: This was the list of everything directly build depending on it, and I'll do one with the indirect build dependencies soon. Kurt -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact
Re: Coordinating upload of teTeX-3.0 to unstable
Kurt Roeckx [EMAIL PROTECTED] wrote: On Tue, Jul 26, 2005 at 06:49:16PM +0200, Frank Küster wrote: Kurt Roeckx [EMAIL PROTECTED] wrote: I can try to build all those for you if you want, and I'll report those where I think that show problems. That would be great. I think most of the problems will show up in failed builds, some might result in different filelists (pdf file generated instead of dvi file, but somehow the build does not fail), a small number might result in wrong fonts being used. I started with everything build depending on tetex-*. Here is an overview of packages which might have problems. Note that I have no clue what some of those message mean. I have, in some cases. Do you have build logs available somewhere? Those really failed to build: - bison-1.35 dvips bison.dvi -o bison.ps This is dvips(k) 5.95a Copyright 2005 Radical Eye Software (www.radicaleye.com) dvips: ! DVI file can't be opened. This is most probably a buggy document (or stylefile) that has a wrong test for pdf output, and produces pdf instead of dvi if the TeX engine understands \pdfoutput, even if it has carefully been set to 0. - ctie Unmet build dependencies: libkpathsea-dev (= 2.0.2-1) Errors like that are delibarate - the library has changed considerably, and is now called libkpathsea4-dev in order to force maintainers to really test their packages. - dvipdfmx: does not make it's documentation? - freetype1: Does not generate documentation? This could point to the same pdf-instead-of-dvi problem, this time as intermediate files, together with a broken build system that doesn't notice that it failed... Some have the combination of: ** menu entry without previous node: License (l. 98) ** node_prev `Top' for `License' not found ** `Distrib' doesn't appear in menus (am-utils, esh, fdutils) That's most probably just a bug in the texinfo sources, nothing that has been triggered by the version change of teTeX. Should I worry about messages like: Underfull \hbox (badness 2368) Overfull \hbox (21.3942pt too wide) Not at all - that's just a warning about bad line breaking, and for sure not dependent on the version. LaTeX Font Warning: Font shape `OMS/cmtt/m/n' undefined Most probably irrelevant and not version-dependent LaTeX Warning: Reference `sec-cblash' on page 7 undefined on input line 452. Minor relevance and not version-dependent ** `XXX' doesn't appear in menus: No idea, but I wouldn't bother too much. This was the list of everything directly build depending on it, and I'll do one with the indirect build dependencies soon. Thank you for your great work! Regards, Frank -- Frank Küster Inst. f. Biochemie der Univ. Zürich Debian Developer
Re: Coordinating upload of teTeX-3.0 to unstable
On Tue, Jul 26, 2005 at 03:05:58PM +0200, Andreas Barth wrote: * Frank Küster ([EMAIL PROTECTED]) [050726 15:04]: Remember, to compile tetex-bin in experimental, the build machine has to install tetex-base from experimental. Oh yes. Bad. So, either you need to make sure it works on all machines, or I have to go onto all buildd chroots by hand (or I have to finally install Wouters patch :). All need some time. Ok, I'd say that I make sure after the final upload to experimental that it works on all archs. Are you interested in sparc, mips and hppa builds? In that case, please tell me, I'll build them on my machines. Bye, Aurelien -- .''`. Aurelien Jarno | GPG: 1024D/F1BCDB73 : :' : Debian GNU/Linux developer | Electrical Engineer `. `' [EMAIL PROTECTED] | [EMAIL PROTECTED] `-people.debian.org/~aurel32 | www.aurel32.net -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Coordinating upload of teTeX-3.0 to unstable
Frank Küster a écrit : Aurelien Jarno [EMAIL PROTECTED] wrote: On Tue, Jul 26, 2005 at 03:05:58PM +0200, Andreas Barth wrote: * Frank Küster ([EMAIL PROTECTED]) [050726 15:04]: Remember, to compile tetex-bin in experimental, the build machine has to install tetex-base from experimental. Oh yes. Bad. So, either you need to make sure it works on all machines, or I have to go onto all buildd chroots by hand (or I have to finally install Wouters patch :). All need some time. Ok, I'd say that I make sure after the final upload to experimental that it works on all archs. Are you interested in sparc, mips and hppa builds? In that case, please tell me, I'll build them on my machines. That would be great. Ideally, you'd take the latest versions from deb http://people.debian.org/~frank/teTeX-3.0 experimental main I have build them successfully for sparc and mips. For hppa, I am waiting for a package part of the libgcc2 transition to finish to build before starting tetex-3.0. Now that ftp-master is back online, are you planning to upload version 3.0-5 to experimental? In that case, I'll upload my builds directly to experimental. In the contrary I'll put them somewhere on the web. Bye, Aurelien -- .''`. Aurelien Jarno | GPG: 1024D/F1BCDB73 : :' : Debian GNU/Linux developer | Electrical Engineer `. `' [EMAIL PROTECTED] | [EMAIL PROTECTED] `-people.debian.org/~aurel32 | www.aurel32.net -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Coordinating upload of teTeX-3.0 to unstable
Dear release team, with the next upload (scheduled for today) of teTeX-3.0 to experimental, I think we have only cosmetic issues left before we can upload it to unstable. Before doing this, however, I'd like to notify you, so that you have a chance to yell out no!. There are two problem areas where tetex might interfere with other packages: a) libkpathsea: It has a new soname, and I have not checked at all whether this causes problems in compiling other packages. I do not expect, however, big problems with that because of two reasons: First of all, only a limited number of packages (and no libraries if you don't count the libkpathsea-perl perl module) with a limited scope on TeX depend on it. Second, without looking it up, I assume that most or even all packages in Debian that depend on it are also part of TeX-live, where compilation and runtime problems should have revealed themselves. b) More severely, tetex-bin is a build-dependency of many packages. While I think that our packages do not contain major bugs, I would be surprised if the upgrade did not reveal bugs in packages that use it, causing some to FTBFS. We've had such cases a year ago when we changed internals of tetex-bin and found that certain packages messed with them during build or installation. If nobody objects, I think we will be able to do an upload next week. TIA, Frank -- Frank Küster Inst. f. Biochemie der Univ. Zürich Debian Developer
Re: Coordinating upload of teTeX-3.0 to unstable
Hi Frank, * Frank K?ster ([EMAIL PROTECTED]) [050726 10:30]: with the next upload (scheduled for today) of teTeX-3.0 to experimental, I think we have only cosmetic issues left before we can upload it to unstable. That sounds good, thanks. b) More severely, tetex-bin is a build-dependency of many packages. While I think that our packages do not contain major bugs, I would be surprised if the upgrade did not reveal bugs in packages that use it, causing some to FTBFS. We've had such cases a year ago when we changed internals of tetex-bin and found that certain packages messed with them during build or installation. Currently, we have (at least) two large transitions at our hands: 1. gcc-4.0/glibc 2. xorg I really would be happy if we can avoid that tetex becomes part of any of these transitions. That means for me, in doubt don't upload tetex to unstable right now. Cheers, Andi -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Coordinating upload of teTeX-3.0 to unstable
On Tue, Jul 26, 2005 at 10:14:15AM +0200, Frank Küster wrote: b) More severely, tetex-bin is a build-dependency of many packages. While I think that our packages do not contain major bugs, I would be surprised if the upgrade did not reveal bugs in packages that use it, causing some to FTBFS. We've had such cases a year ago when we changed internals of tetex-bin and found that certain packages messed with them during build or installation. Then can you please notify debian-devel of this before you upload? Thanks, -- Steve Langasek postmodern programmer signature.asc Description: Digital signature
Re: Coordinating upload of teTeX-3.0 to unstable
Andreas Barth [EMAIL PROTECTED] wrote: That sounds good, thanks. b) More severely, tetex-bin is a build-dependency of many packages. While I think that our packages do not contain major bugs, I would be surprised if the upgrade did not reveal bugs in packages that use it, causing some to FTBFS. We've had such cases a year ago when we changed internals of tetex-bin and found that certain packages messed with them during build or installation. Currently, we have (at least) two large transitions at our hands: 1. gcc-4.0/glibc 2. xorg I really would be happy if we can avoid that tetex becomes part of any of these transitions. That means for me, in doubt don't upload tetex to unstable right now. How long do you expect the transitions to take, and what would you think is appropriate to replace doubt by knowlegde? The number of packages build-depending on tetex (directly or indirectly via a build-dep on some conversion tool like debiandoc-sgml) is really big, and we will not be able to check this all by hand. Writing to -devel and writing to the maintainers of those packages might help, but we will probably not even be able to keep track of the responses and be sure whether all maintainers of important packages have answered. Regards, Frank -- Frank Küster Inst. f. Biochemie der Univ. Zürich Debian Developer
Re: Coordinating upload of teTeX-3.0 to unstable
* Frank K?ster ([EMAIL PROTECTED]) [050726 12:34]: Andreas Barth [EMAIL PROTECTED] wrote: Currently, we have (at least) two large transitions at our hands: 1. gcc-4.0/glibc 2. xorg I really would be happy if we can avoid that tetex becomes part of any of these transitions. That means for me, in doubt don't upload tetex to unstable right now. How long do you expect the transitions to take, and what would you think is appropriate to replace doubt by knowlegde? What is appropriate for the second: Bring up ftp-master again :) Well, xorg seems to depend only on some small issues. Without being able to investigate further right now, we should be able to solve that in less than a weeks time after ftp-master is available again - or at least, we know much more. When that is done, I don't see a hard stopper for uploading tetex to unstable, except of course you should notify the maintainers again (e.g. via debian-devel). Cheers, Andi -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Coordinating upload of teTeX-3.0 to unstable
Andreas Barth writes: * Frank K?ster ([EMAIL PROTECTED]) [050726 12:34]: Andreas Barth [EMAIL PROTECTED] wrote: Currently, we have (at least) two large transitions at our hands: 1. gcc-4.0/glibc 2. xorg I really would be happy if we can avoid that tetex becomes part of any of these transitions. That means for me, in doubt don't upload tetex to unstable right now. How long do you expect the transitions to take, and what would you think is appropriate to replace doubt by knowlegde? What is appropriate for the second: Bring up ftp-master again :) Well, xorg seems to depend only on some small issues. Without being able to investigate further right now, we should be able to solve that in less than a weeks time after ftp-master is available again - or at least, we know much more. When that is done, I don't see a hard stopper for uploading tetex to unstable, except of course you should notify the maintainers again (e.g. via debian-devel). please make sure, that tetex-bin builds on all architectures. avoid having new arch-indep packages in unstable and unbuildable arch packages. Matthias -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]