Bug#586549: mktexlsr does not fail on intermittent errors
Hilmar Preusse hill...@web.de wrote: On 10.08.10 Norbert Preining (prein...@logic.at) wrote: On Di, 10 Aug 2010, Hilmar Preusse wrote: Hi Norbert, This is about set -e in basic scripts of TL. However I'm missing the set -e in the scripts. Ahh, you replaced the this morning. Out of luck... Yes, I want to push the changes to unstable/testing and set -e wchange will not pass release manager ... I think *now* is a good time to introduce set -e and break TeX Live in Debian. ;-) I agree. There's already a patch in svn, which seems to work most of the time. However, there is one flaw in it that I just noticed: At least texconfig has lots of constructs like do_something if test $? != 0; then echo $errormessage exit 1 fi Of course the error message will never be shown this way. So quite some more stuff to patch. Regards, Frank -- Frank Küster Sprecher B90/Grüne OV Miltenberg und Umgebung VCD Miltenberg, ADFC Aschaffenburg-Miltenberg Debian Developer (TeXLive) -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#586549: mktexlsr does not fail on intermittent errors
On 10.08.10 Norbert Preining (prein...@logic.at) wrote: On Di, 10 Aug 2010, Hilmar Preusse wrote: Hi Norbert, This is about set -e in basic scripts of TL. However I'm missing the set -e in the scripts. Ahh, you replaced the this morning. Out of luck... Yes, I want to push the changes to unstable/testing and set -e wchange will not pass release manager ... I think *now* is a good time to introduce set -e and break TeX Live in Debian. ;-) Hilmar -- sigmentation fault -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#586549: mktexlsr does not fail on intermittent errors
On 05.08.10 Norbert Preining (prein...@logic.at) wrote: Hi, the change now as potential problems are most likely to be exposed during lenny-squeeze upgrades. And such upgrades will hopefully be tested several times before we finally release. Test packages are on the server: deb http://people.debian.org/~preining/TeX/ tl2009/ but only for amd64, plus deb-src of course. If someone can test updates from lenny it would be great ... Upgrade worked, all format files were created, a minimal LaTeX example runs through. lennyamd64:~# dpkg -l texlive*|grep ^ii ii texlive-base2009-10 TeX Live: Essential programs and files ii texlive-base-bin-doc2007.dfsg.2-4+lenny3TeX Live: Documentation files for texlive-base-bin ii texlive-binaries2009-7 Binaries for TeX Live However I'm missing the set -e in the scripts. Ahh, you replaced the this morning. Out of luck... H. -- sigmentation fault -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#586549: mktexlsr does not fail on intermittent errors
Hi Hilmar, On Di, 10 Aug 2010, Hilmar Preusse wrote: If someone can test updates from lenny it would be great ... Upgrade worked, all format files were created, a minimal LaTeX example runs through. Thanks, looks fine. However I'm missing the set -e in the scripts. Ahh, you replaced the this morning. Out of luck... Yes, I want to push the changes to unstable/testing and set -e wchange will not pass release manager ... I am. The set -e was ignored for years I wouldn't push it now. ;-) Best wishes Norbert Norbert Preiningprein...@{jaist.ac.jp, logic.at, debian.org} JAIST, Japan TeX Live Debian Developer DSA: 0x09C5B094 fp: 14DF 2E6C 0307 BE6D AD76 A9C0 D2BF 4AA3 09C5 B094 NAUGATUCK (n.) A plastic sachet containing shampoo, polyfilla, etc., which is impossible to open except by off the corners. --- Douglas Adams, The Meaning of Liff -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#586549: mktexlsr does not fail on intermittent errors
On Fr, 16 Jul 2010, Frank Küster wrote: Adding set -e to fmtutil, mtexlsr and updmap should be quite straightforward to test. With texconfig it might be harder, since it is I am preparing another round of texlive-binaries update. Should I include the set -e there? Opinions please, but take into respect that we are close to release (well, more or less) of Debian. Best wishes Norbert Norbert Preiningprein...@{jaist.ac.jp, logic.at, debian.org} JAIST, Japan TeX Live Debian Developer DSA: 0x09C5B094 fp: 14DF 2E6C 0307 BE6D AD76 A9C0 D2BF 4AA3 09C5 B094 SHRIVENHAM (n.) One of Germaine Greer's used-up lovers. --- Douglas Adams, The Meaning of Liff -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#586549: mktexlsr does not fail on intermittent errors
Hi Norbert, On Fr, 16 Jul 2010, Frank Küster wrote: Adding set -e to fmtutil, mtexlsr and updmap should be quite straightforward to test. With texconfig it might be harder, since it is I am preparing another round of texlive-binaries update. Should I include the set -e there? Opinions please, but take into respect that we are close to release (well, more or less) of Debian. I'm not sure whether you want to count on my opinion as any bugs caused by such a change probably won't bother me that much. Anyway, I believe it's worth doing the change now as potential problems are most likely to be exposed during lenny-squeeze upgrades. And such upgrades will hopefully be tested several times before we finally release. Best, Michael pgp8s30TguTTc.pgp Description: PGP signature
Bug#586549: mktexlsr does not fail on intermittent errors
the change now as potential problems are most likely to be exposed during lenny-squeeze upgrades. And such upgrades will hopefully be tested several times before we finally release. Test packages are on the server: deb http://people.debian.org/~preining/TeX/ tl2009/ but only for amd64, plus deb-src of course. If someone can test updates from lenny it would be great ... I installed them on my system and it worked, up to now. Best wishes Norbert Norbert Preiningprein...@{jaist.ac.jp, logic.at, debian.org} JAIST, Japan TeX Live Debian Developer DSA: 0x09C5B094 fp: 14DF 2E6C 0307 BE6D AD76 A9C0 D2BF 4AA3 09C5 B094 VENTNOR (n.) One who, having been visited as a child by a mysterious gypsy lady, is gifted with the strange power of being able to operate the air-nozzles above aeroplane seats. --- Douglas Adams, The Meaning of Liff -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#586549: mktexlsr does not fail on intermittent errors
On 05.08.10 Norbert Preining (prein...@logic.at) wrote: Hi, the change now as potential problems are most likely to be exposed during lenny-squeeze upgrades. And such upgrades will hopefully be tested several times before we finally release. Test packages are on the server: deb http://people.debian.org/~preining/TeX/ tl2009/ but only for amd64, plus deb-src of course. If someone can test updates from lenny it would be great ... I installed them on my system and it worked, up to now. I'll try to set up an AMD64 (stable) ASAP and test upgrades. H. -- The sun never sets on those who ride into it. -- RKO http://www.hilmar-preusse.de.vu/ -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#586549: mktexlsr does not fail on intermittent errors
Norbert Preining prein...@logic.at wrote: As tested by the OP, *no* changes were necessary for fmtutil ... that is a good sign. I'm quite sure that some years ago, one of the main teTeX scripts (fmtutil, mktexlsr, texconfig, updmap) just wouldn't run with it, and the fix was either non-trivial or would only show up the next instance. Maybe there's even a bug report in the teTeX packages about it? I tried to find a discussion in the teTeX lists which I think did happen, but failed. Adding set -e to fmtutil, mtexlsr and updmap should be quite straightforward to test. With texconfig it might be harder, since it is interactive and more complex. (and ugly spagetti code). Regards, Frank -- Dr. Frank Küster Debian Developer (TeXLive) VCD Aschaffenburg-Miltenberg, ADFC Miltenberg B90/Grüne KV Miltenberg -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#586549: mktexlsr does not fail on intermittent errors
Well, I definitely do not want to add set -e to our fundamental scripts a week before our final release. But if you want to add it for Debian, I don't know of any specific reason not to. You can experience the pain, or lack thereof. As I've written before: - I am not a fan of set -e in general, because I've seen too much otherwise-unnecessary code added to placate it, and confusing silent aborts result. - but I do not categorically refuse to add it in the future. - After the release. Do you know of any reason why set -e was not used from the very begining? I don't know of any specific technical reason. I expect Thomas was no fan of set -e either. karl -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#586549: mktexlsr does not fail on intermittent errors
Hi Karl, On So, 11 Jul 2010, Karl Berry wrote: Well, I definitely do not want to add set -e to our fundamental scripts a week before our final release. Sorry if I was unclear, I didn't suggest to do it in TL now, only your general comment or whether there are some background infos I don't know. But if you want to add it for Debian, I don't know of any specific reason not to. You can experience the pain, or lack thereof. We will do that and see how it works out. After some testing here and hopefully not too many bug reporits, and after release of TL 2010, we can do the same for upstream TL, too. - I am not a fan of set -e in general, because I've seen too much otherwise-unnecessary code added to placate it, and confusing silent aborts result. As tested by the OP, *no* changes were necessary for fmtutil ... that is a good sign. Best wishes Norbert Norbert Preiningprein...@{jaist.ac.jp, logic.at, debian.org} JAIST, Japan TeX Live Debian Developer DSA: 0x09C5B094 fp: 14DF 2E6C 0307 BE6D AD76 A9C0 D2BF 4AA3 09C5 B094 DUNGENESS (n.) The uneasy feeling that the plastic handles of the overloaded supermarket carrier bag you are carrying are getting steadily longer. --- Douglas Adams, The Meaning of Liff -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#586549: mktexlsr does not fail on intermittent errors
On 04.07.10 Michael Tautschnig (m...@debian.org) wrote: Hi, Thanks for your help! If I understood the mktexlsr code correctly, something as simple as adding set -e could suffice to make mktexlsr exit/fail early, and the existing ls-R file would not get overwritten. I you're willing to work an this we'd be happy. I assume that not only mktexlsr and fmtutil are affected. If somebody can give me a few usage scenarios for fmtutil (i.e., proper command lines) I'll happily try out that as well, fmtutil-sys --all (run as root) regenerates all format files, fmtutil-sys --byfmt $format just (re-)generates a single format file, where $format is tex, amstex, pdftex etc. Also you should try running fmtutil -all fmtutil --byfmt $format, they generate user-specific formats. but at the moment this doesn't seem to be as painful as expected!? Maybe we should simply introduce these -e switches and then wait for the side effects. Norbert, Frank? H. -- sigmentation fault signature.asc Description: Digital signature
Bug#586549: mktexlsr does not fail on intermittent errors
Hi! [...] If somebody can give me a few usage scenarios for fmtutil (i.e., proper command lines) I'll happily try out that as well, fmtutil-sys --all (run as root) regenerates all format files, fmtutil-sys --byfmt $format just (re-)generates a single format file, where $format is tex, amstex, pdftex etc. Also you should try running fmtutil -all fmtutil --byfmt $format, they generate user-specific formats. Done for all of the above, no errors. but at the moment this doesn't seem to be as painful as expected!? Maybe we should simply introduce these -e switches and then wait for the side effects. Norbert, Frank? I really wouldn't claim to have done exhaustive testing, but I'd hope that I have at least covered the very basic use cases, and none of them causes an error. Therefore I'd also opt for adding set -e to the scripts and see what happens. Best, Michael pgpq3fY0oEzPm.pgp Description: PGP signature
Bug#586549: mktexlsr does not fail on intermittent errors
Hi Karl, over here on Debian we have a discussion concerning adding set -e to fmtutil, mktexlsr, updmap etc. It seems that at least for mktexlsr and fmtutil that is deffinitely possible: On Sa, 10 Jul 2010, Michael Tautschnig wrote: If somebody can give me a few usage scenarios for fmtutil (i.e., proper command lines) I'll happily try out that as well, fmtutil-sys --all (run as root) regenerates all format files, fmtutil-sys --byfmt $format just (re-)generates a single format file, where $format is tex, amstex, pdftex etc. Also you should try running fmtutil -all fmtutil --byfmt $format, they generate user-specific formats. Done for all of the above, no errors. but at the moment this doesn't seem to be as painful as expected!? Maybe we should simply introduce these -e switches and then wait for the side effects. Norbert, Frank? I really wouldn't claim to have done exhaustive testing, but I'd hope that I have at least covered the very basic use cases, and none of them causes an error. Therefore I'd also opt for adding set -e to the scripts and see what happens. What is your opinion on that? Do you know of any reason why set -e was not used from the very begining? For full emails text, please see bugs.debian.org/586549 Best Norbert -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#586549: mktexlsr does not fail on intermittent errors
[...] If I understood the mktexlsr code correctly, something as simple as adding set -e could suffice to make mktexlsr exit/fail early, and the existing ls-R file would not get overwritten. I you're willing to work an this we'd be happy. I assume that not only mktexlsr and fmtutil are affected. Hmm, I just did a very quick try on mktexlsr and just added set -e at the very beginning, expecting a longer session of debugging -- but it just works!? If somebody can give me a few usage scenarios for fmtutil (i.e., proper command lines) I'll happily try out that as well, but at the moment this doesn't seem to be as painful as expected!? Best, Michael pgpCFIBE9pw2z.pgp Description: PGP signature
Bug#586549: mktexlsr does not fail on intermittent errors
Package: texlive-base-bin Version: 2007.dfsg.2-4+lenny2 Severity: minor It seems that mktexlsr does not exit if some error occurs in executing the commands. This possibly results in corrupted/incomplete ls-R files, as seen more than once on alioth.debian.org (hence CC'ed Tollef, alioth admin), e.g., if you're running out of disk space (on /tmp, most probably): /etc/cron.daily/tetex-bin: sed: couldn't write 6 items to stdout: No space left on device and then m...@alioth:~$ latex bla.tex This is pdfTeXk, Version 3.141592-1.40.3 (Web2C 7.5.6) %-line parsing enabled. entering extended mode (./bla.tex LaTeX2e 2005/12/01 Babel v3.8h and hyphenation patterns for english, usenglishmax, dumylang, noh yphenation, croatian, ukrainian, russian, bulgarian, czech, slovak, danish, dut ch, finnish, basque, french, german, ngerman, ibycus, greek, monogreek, ancient greek, hungarian, italian, latin, mongolian, norsk, icelandic, interlingua, tur kish, coptic, romanian, welsh, serbian, slovenian, estonian, esperanto, upperso rbian, indonesian, polish, portuguese, spanish, catalan, galician, swedish, loa ded. ! LaTeX Error: File `article.cls' not found. Type X to quit or RETURN to proceed, or enter new name. (Default extension: cls) Enter file name: x (for some very minimal bla.tex) If I understood the mktexlsr code correctly, something as simple as adding set -e could suffice to make mktexlsr exit/fail early, and the existing ls-R file would not get overwritten. Best, Michael pgpVl0GW9BpSk.pgp Description: PGP signature
Bug#586549: mktexlsr does not fail on intermittent errors
severity 586549 normal reassign 586549 texlive-binaries merge 136051 586549 retitle 136051 Please make upstream TL's scripts set -e (at least fmtutil and mktexlsr) stop On 20.06.10 Michael Tautschnig (m...@debian.org) wrote: Package: texlive-base-bin Version: 2007.dfsg.2-4+lenny2 Severity: minor It seems that mktexlsr does not exit if some error occurs in executing the commands. This possibly results in corrupted/incomplete ls-R files, as seen more than once on alioth.debian.org (hence CC'ed Tollef, alioth admin), e.g., if you're running out of disk space (on /tmp, most probably): Unfortunately the problem is not new and nobody really did care about this. Reassigning and merging. If I understood the mktexlsr code correctly, something as simple as adding set -e could suffice to make mktexlsr exit/fail early, and the existing ls-R file would not get overwritten. I you're willing to work an this we'd be happy. I assume that not only mktexlsr and fmtutil are affected. H. -- sigmentation fault signature.asc Description: Digital signature