aboutsummaryrefslogtreecommitdiff
path: root/pbuilder-buildpackage
Commit message (Collapse)AuthorAge
* fixup a bunch of things for binNMU supportMattia Rizzolo2015-11-23
|
* buildpackage: umount as the very last option, due keeping the bindmounts ↵Mattia Rizzolo2015-11-22
| | | | | | while executing the hooks Closes: #805552
* use log.e() instead of log()Mattia Rizzolo2015-11-15
|
* use log.w() instead of log()Mattia Rizzolo2015-11-15
|
* use log.i() instead of log()Mattia Rizzolo2015-11-15
|
* use relative sources where possibleMattia Rizzolo2015-11-01
| | | | | | | | | | This way it's a lot easier to test stuff without actually installing it Use BASH_SOURCE, which contains the path of the current script being executed. BASH_SOURCE is clearly a bashism, but it's cheaper than using dirname(1). Thanks to Gianfranco Costamagna for noticing how this bit could have be improved. Gbp-Dch: Short
* no need to load runhooks anymore now that it's all handled in -modulesMattia Rizzolo2015-10-28
| | | | Gbp-Dch: ignore
* pbuilder-buildpackage: only override HOME during dpkg-buildpackage invocationMattia Rizzolo2015-10-14
|
* New config: BUILD_HOME, to be able to set the value of HOME while building. ↵Mattia Rizzolo2015-10-14
| | | | | | | | Default it to /nonexistent to prevent builds from writing to /home. This reverts what was done for #170762 Closes: #441052
* add hook type I: it runs after successfully build a package, after copying ↵Mattia Rizzolo2015-10-13
| | | | | | back the build results Closes: #652527
* always load hooks just after extracting the build place, and unload them ↵Mattia Rizzolo2015-10-13
| | | | | | | | when cleaning up So, no need to load&run&unload everywhere as it is now. Gbp-Dch: Short
* pbuilder-buildpackage: some more whitespaces normalizationMattia Rizzolo2015-10-03
| | | | Gbp-Dch: Ignore
* fix a whole bunch of warning from shellcheckMattia Rizzolo2015-09-18
| | | | | | | There are still quite some, this is a first chunk based on a given patch applied where i felt confident enough. Thanks: Herbert Parentes Fortes Neto <hpfn@ig.com.br> for the initial patch-set
* buildpackage: Unset DISPLAY just before the build, othewise there is ↵Mattia Rizzolo2015-08-29
| | | | | | software that tries to connect to a X server LP: #816556
* parametrize the build directory using the conf option BUILDDIRMattia Rizzolo2015-08-08
|
* pbuilder-buildpackage: remove the -e parameter from the dpkg-buildpackage ↵Mattia Rizzolo2015-06-24
| | | | | | invocation Closes: #567484
* Deprecate the PKGNAME_LOGFILE_EXTENTION conf entry in favour of ↵Mattia Rizzolo2015-06-23
| | | | | | PKGNAME_LOGFILE_EXTENSION to fix the typo Closes: #693458
* set the USER env variable to BUILDUSERNAME, to fix FTBFS of some package ↵Mattia Rizzolo2015-06-17
| | | | | | | which check for not being run as root using that variable Closes: #677666 Thanks: Tim Landscheidt <tim@tim-landscheidt.de> for the initial patch
* import 0.125+nmu3Mattia Rizzolo2015-06-17
|
* import 0.125+nmu2Mattia Rizzolo2015-06-17
|
* import 0.125+nmu1Mattia Rizzolo2015-06-17
|
* fix logging message and conditionality to make it less lines.Junichi Uekawa2012-03-13
| | | | | | I think we don't need the if. I'm not quite convinced ||true to kill the error message is a good idea
* fix space handling in additional-buildresults.Junichi Uekawa2012-03-13
|
* Bug#652634: Allow to copy additional files out of the build directoryGuido Günther2012-03-13
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | On Sat, Dec 24, 2011 at 05:22:05PM +0900, Junichi Uekawa wrote: > Hi, > > I'd use a shell array for specifying the values; but yes I think I > too wanted this feature at one point. Attached is a patch that uses a shell array and adds some documentation. Please apply if you see fit. > Would list of filename (I will imagine there will be requests for all > sorts of wildcards after release, which is kind of a pain) be useful > or a directory that a post-build script can write to after build be > useful? (I think qemubuilder needed something to do with a specific > directory that you can write to, but I will need to check how I did > it). I'd be great to have this in first and see how others use it. It wouldn't be a problem to introduce an ADDITIONAL_BUILD_RESULTS_DIR later. Cheers, -- Guido >From ad3569e07a9cc64cd5b126193cddf311e48180a6 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Guido=20G=C3=BCnther?= <agx@sigxcpu.org> Date: Mon, 19 Dec 2011 13:31:59 +0100 Subject: [PATCH] Add ADDITIONAL_BUILDRESULTS variable this can be used to copy additional build results out of the build directory. It's useful to e.g. preserve a xml testresult file when using a build system like Jenkins.
* Bug#659703: pbuilder: Typo in error messageMaarten Bezemer2012-03-09
| | | | | | | | | | | | | | | | | | | | | | | | | | | Package: pbuilder Version: 0.206 Severity: minor Tags: patch User: ubuntu-devel@lists.ubuntu.com Usertags: origin-ubuntu precise ubuntu-patch There is a typo in an error message, introduced in http://anonscm.debian.org/gitweb/?p=pbuilder/pbuilder.git;a=commit;h=163e0751e2066cd24ec4de61f30489b5a21161e8 *** /tmp/tmpQkEY6_ This patch corrects the typo. Thanks for considering the patch. -- System Information: Debian Release: wheezy/sid APT prefers oneiric-updates APT policy: (500, 'oneiric-updates'), (500, 'oneiric-security'), (500, 'oneiric'), (100, 'oneiric-backports') Architecture: amd64 (x86_64) Kernel: Linux 3.0.0-15-generic (SMP w/2 CPU cores) Locale: LANG=en_US.UTF-8, LC_CTYPE=en_US.UTF-8 (charmap=UTF-8) Shell: /bin/sh linked to /bin/dash
* I think I need to quote PATHJunichi Uekawa2012-01-27
|
* Bug#644507: pbuilder produces a confusing error message about the dcs ↵Junichi Uekawa2011-12-31
| | | | | | | filename if it can't read the dsc file. Make the message say that the file is not found instead of complaining about the validity of file name format, which is irrelevant.
* Bug#606687: ccache support fails in the face of su PATH manglingMatthew Palmer2011-12-31
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Package: pbuilder Version: 0.199 Severity: normal Tags: patch pbuilder's ccache support relies on being able to modify PATH when building packages. Unfortunately, su's documented behaviour is to reset PATH to a known safe setting (in login.defs), which breaks this (along with any other attempts to manually set PATH in pbuilderrc). Apparently (see irc-log.txt) su has had a bug in it that has failed to mangle PATH, which has kept this bug "hidden". However, the version of login in squeeze definitely resets PATH. The attached patch adjusts the way that commands are called in the su session, preserving the environment and making ccache work. -- System Information: Debian Release: 5.0.3 APT prefers stable APT policy: (500, 'stable'), (1, 'experimental'), (1, 'unstable'), (1, 'testing') Architecture: i386 (i686) Kernel: Linux 2.6.26-2-686 (SMP w/2 CPU cores) Locale: LANG=en_AU, LC_CTYPE=en_AU (charmap=ISO-8859-1) Shell: /bin/sh linked to /bin/bash Versions of packages pbuilder depends on: ii cdebootstrap 0.5.4 Bootstrap a Debian system ii coreutils 6.10-6 The GNU core utilities ii debconf [debconf-2.0] 1.5.24 Debian configuration management sy ii debianutils 2.30 Miscellaneous utilities specific t ii debootstrap 1.0.10lenny1 Bootstrap a basic Debian system ii wget 1.11.4-2+lenny2 retrieves files from the web Versions of packages pbuilder recommends: ii devscripts 2.10.35lenny7 scripts to make the life of a Debi ii fakeroot 1.11 Gives a fake root environment ii sudo 1.6.9p17-3 Provide limited super user privile Versions of packages pbuilder suggests: ii cowdancer 0.47 Copy-on-write directory tree utili pn gdebi-core <none> (no description available) pn pbuilder-uml <none> (no description available) -- debconf information excluded
* Ignore LD_PRELOAD when inside 'pbuilder build' building something.Junichi Uekawa2011-09-21
| | | | | | | | This should alleviate problems encountered when fakeroot / cowdancer interacts badly with whatever that's going in. This will probably break fakechroot, so it needs some kind of configuration, but fixes problem with lighttpd.
* lighttpd test fails with cowdancer loaded. (closes: #640684)Junichi Uekawa2011-09-21
| | | | | Do not load cowdancer when su to normal user privilege; cowdancer protection isn't needed there.
* cannot cross build i386 only package on amd64 host (closes: #624530)YOSHINO Yoshihito2011-08-07
| | | | | | | | | | | | | | | | | | | | | | | | | | Followup-For: Bug #624530 Package: pbuilder Version: 0.200 found 624530 0.199+nmu2 severity 624530 important tags 624530 patch thanks Hi, The changes introduced in 0.199+nmu2 (arch-wildcard-part2.patch in Message #159 of Bug#363193) breaks the cross building of packages (not only i386 packages on amd64, but also amd64 ones on i386 running on amd64 kernel). The checkarchitecture must be done inside the build place. That patch does not actually check anything, and what's worse, breaks the cross building. So I propose reverting back the changes. Thanks, -- YOSHINO Yoshihito <yy.y.ja.jp@gmail.com>
* Bug#606699: pbuilder NMUEmilio Pozuelo Monfort2011-06-24
| | | | | | | | | | Hi, I talked to Junichi and he told me it was fine if I NMU'ed pbuilder to fix these bugs. This is the debdiff for my NMU. Regards, Emilio
* Move ccache setup to pbuilder-buildpackage-funcsLoïc Minier2010-01-03
|
* Add builtin ccache support, enabled by defaultLoïc Minier2010-01-02
| | | | | | | | | Add builtin support for using ccache in pbuilder and enable it by default. Ship a new /var/cache/pbuilder/ccache dir and bind-mount and chown it to BUILDUSERID at build time. Install/remove ccache automatically on create/update if CCACHEDIR is set/unset. Update docs and remove old ccache config example. Add a NEWS entry featuring the change.
* Avoid appending extra spaces in DEBBUILDOPTSLoïc Minier2009-12-14
|
* Factor "exec >" + "exec 2>&1" togetherLoïc Minier2009-12-13
|
* Log dpkg-buildpackage command linesLoïc Minier2009-12-12
|
* Add support for building packages twice; #493538Loïc Minier2009-12-11
| | | | | Add support for building packages twice; based on a patch by Nicolas Valcárcel; closes: #493538
* Make error message visible when copying out to buildresult (closes: #505695)Junichi Uekawa2009-12-05
| | | | | | | | NFS directories can't be written to by root processes, and they will fail quietly. 1. Only attempt to copy the files (so we won't attempt to copy the directories). 2. Expose error messages so that users will see them.
* quote-escape debemail so that email with space can be handled properly.Junichi Uekawa2009-11-02
|
* pbuilder-buildpackage: pass -e to dpkg-bp, not -mLoïc Minier2009-09-21
| | | | | pbuilder-buildpackage: pass -e to dpkg-buildpackage instead of -m; thanks Russ Allbery; closes: #546531
* Bug#520251: pbuilder: Pbuilder does not umount bindmounts when ssh ↵Stefan Fritsch2009-03-18
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | connection dies Package: pbuilder Version: 0.181 Severity: important When pbuilder is executed over ssh and the ssh connection dies, it won't umount bind mounts and won't clean up the build dir. Unfortunately, cowbuilder WILL cleanup the build dir, wiping all bind mounts. The execution order is like this: ssh -> sudo -> cowbuilder -> pbuilder Because of the sudo, the SIGHUP will not be delivered to pbuilder. However, pbuilder will get SIGPIPE when writing to stdout or stderr. In addition, even when SIGPIPE is ignored, a write to stdout will give EPIPE, which causes echo to return with an error. Therefore, you - need to catch sigpipe in addition to sighup - need to ignore sigpipe and sighup in the cleanup handler - need to set +x in the cleanup handler to avoid exiting just because echo returns an error. The attached patch works for me but I have only edited pbuilder-buildpackage and not the other commands.
* implement --inputfile option.Junichi Uekawa2009-03-07
| | | | Copies extra files to inside chroot.
* refactor to use 'log' function rather than using 'echo' directly.Junichi Uekawa2009-02-26
| | | | First cut into doing this, hopefully we're not breaking anything.
* 398865: apply patch for LOGNAME, and implement also shadow group fileJunichi Uekawa2007-05-28
|
* fix always ending with "Aborting with error"Junichi Uekawa2007-05-27
|
* fix thinko: trap exit -> trap - exitJunichi Uekawa2007-05-27
|
* change "trap" handling so that all trap function calls are called _trap, and ↵Junichi Uekawa2007-05-27
| | | | they will exit pbuilder after receiving a trap.
* trap SIGHUP as well as EXIT.Junichi Uekawa2007-05-27
|
* use tee instead of not outputting anything for logging mode.Junichi Uekawa2007-05-17
|
generated by cgit v1.2.3 (git 2.25.4) at 2024-05-07 21:47:16 +0000