summaryrefslogtreecommitdiffstats
path: root/pym
Commit message (Collapse)AuthorAgeFilesLines
* Enable EbuildIpcDaemon support by default.Zac Medico2010-08-152-14/+10
|
* Add support for /etc/portage/package.accept_keywords. If bothZac Medico2010-08-151-1/+7
| | | | | | | | | | | package.accept_keywords and package.keywords are present, both of them will be used, and values from package.accept_keywords will override values from package.keywords. The package.accept_keywords file is intended to replace the package.keywords file, since profiles support a different form of package.keywords which modifies effective KEYWORDS (rather than ACCEPT_KEYWORDS). Note that bug #292083 requests support for package.accept_keywords in profiles.
* Implement ExtendedAtomDict.iteritems(), so that it doesn't invokeZac Medico2010-08-151-0/+7
| | | | the special __getitem__ behavior.
* Fix make.conf code to handle None return value from getconfig().Zac Medico2010-08-151-4/+6
|
* Add support for /etc/portage/make.profile as an alternative toZac Medico2010-08-151-1/+6
| | | | | | /etc/make.profile. If both /etc/make.profile/ and /etc/portage/make.profile/ exist, then /etc/make.profile/ will be preferred.
* Add support for /etc/portage/make.conf. If present, this file willZac Medico2010-08-151-5/+10
| | | | override settings from /etc/make.conf.
* Remove PollScheduler._schedule_waitpid() since this case isn't triggeredZac Medico2010-08-154-33/+3
| | | | | | anymore, now that PollScheduler._unregister() has be fixed to discard unhandled events (so that reallocation of file descriptors no longer triggers erroneous delivery of stale events).
* Don't handle KeyError when looking up event handlers, sinceZac Medico2010-08-151-31/+12
| | | | _unregister() automatically discards stale events now.
* Fix PollScheduler._unregister() to discard any unhandled eventsZac Medico2010-08-151-0/+13
| | | | | | | | | that belong to the unregistered file, in order to prevent these events from being erroneously delivered to a future handler that is using a reallocated file descriptor of the same numeric value (causing extremely confusing bugs). Note that this is was the cause of the intermittent bug that forced me to disable EbuildIpcDaemon by default.
* Fix SubProcess instances to call _set_returncode() when appropriate,Zac Medico2010-08-154-8/+8
| | | | instead of setting self.returncode directly.
* Remove questionable lockfile removal code in dbapi.invalidentry().Zac Medico2010-08-151-9/+1
|
* Dispose of lock object after calling unlockfile(), in order to ensureZac Medico2010-08-151-0/+1
| | | | that we only try to unlock it once.
* Evaluate SRC_URI and return early if there's nothing to fetch.Zac Medico2010-08-151-26/+52
|
* Make DoebuildSpawnTestCase instantiate a minimal Package instance, andZac Medico2010-08-153-5/+11
| | | | fix code to tolerate missing metadata values in various places.
* Remove unuser return values from PollScheduler event handlers.Zac Medico2010-08-154-6/+0
|
* Remove redundant os.path.join() call in _init_ipc_fifos().Zac Medico2010-08-151-3/+1
|
* Add the "unexpected bash exit" case to the docstring.Zac Medico2010-08-151-1/+3
|
* Add some ebuild metadata for realism.Zac Medico2010-08-151-1/+9
|
* Pass along PORTAGE_USERNAME and PORTAGE_GRPNAME since theyZac Medico2010-08-151-0/+9
| | | | need to be inherited by ebuild subprocesses.
* Make create_trees() pass along PORTAGE_GRPNAME and PORTAGE_USERNAMEZac Medico2010-08-151-1/+1
| | | | to all config instances.
* Avoid KeyError if PORTAGE_USERNAME or PORTAGE_GRPNAME are unset.Zac Medico2010-08-151-4/+7
|
* Pass along PORTAGE_USERNAME and PORTAGE_GRPNAME inside _load_config()Zac Medico2010-08-151-1/+5
| | | | since since they need to be inherited by ebuild subprocesses.
* Add some more info to the 'Cannot chown a lockfile' message, sinceZac Medico2010-08-151-1/+5
| | | | it's being triggered on the buildbot.
* Set PORTAGE_PYTHON for ebuild.sh subprocesses.Zac Medico2010-08-151-0/+2
|
* DoebuildSpawnTestCase: Invoke portage.package.ebuild.doebuild.spawn()Zac Medico2010-08-151-0/+47
| | | | | | with a minimal environment. This gives coverage to some of the ebuild execution internals, like ebuild.sh, EbuildSpawnProcess, and EbuildIpcDaemon.
* Disable EbuildIpcDaemon for now, since it seems to be triggeringZac Medico2010-08-143-40/+82
| | | | | intermittent build failures in my stage builds. For testing purposes set PORTAGE_IPC_DAEMON_ENABLE=1 to enable EbuildIpcDaemon.
* portage.dep: Integrate the functionality of flatten into use_reduceSebastian Luther2010-08-147-44/+142
|
* Merge branch 'master' of git+ssh://git.overlays.gentoo.org/proj/portageZac Medico2010-08-141-2/+15
|\
| * Use more correct paths to byte-compiled Python modules with Python 3.2Arfrever Frehtes Taifersar Arahesis2010-08-141-2/+15
| | | | | | | | and remove empty __pycache__ directories.
* | Pass $USE via IPC in order to ensure that we have the correct value forZac Medico2010-08-141-2/+8
|/ | | | | | built/installed packages, since the config class doesn't currently provide a way to access built/installed $USE that would work in all possible scenarios.
* Fix the ACCEPT_LICENSE setting that's generated for eutils check_license()Zac Medico2010-08-141-15/+26
| | | | to account for package.license.
* When in background mode, set NOCOLOR=true in order to prevent colorZac Medico2010-08-141-0/+5
| | | | | codes from showing up in logs, since we're not displaying to a terminal anyway. Thanks to Brian Harring for the suggestion.
* Add PORTAGE_PYTHON to _environ_whitelist.Zac Medico2010-08-141-1/+1
|
* Move PORTAGE_PYTHON setting to doebuild_environment() so that itZac Medico2010-08-144-6/+5
| | | | doesn't pollute os.environ.
* Move 'phase' attribute to AbstractEbuildProcess from subclasses, soZac Medico2010-08-143-11/+11
| | | | the _get_phase() method isn't needed.
* Add a QueryCommand.settings attribute and use it so that $USEZac Medico2010-08-142-5/+6
| | | | | doesn't have to be passed to the daemon in has_version and best_version calls.
* Fix EbuildSpawnProcess to use self.settings.environ() for the spawnZac Medico2010-08-143-22/+12
| | | | | environment, so all AbstractEbuildProcess subclasses are consistent in this way.
* portage.dep: Integrate the functionality of dep_opconvert into use_reduceSebastian Luther2010-08-145-27/+92
|
* Make sure PORTAGE_IPC_DAEMON is only in the environment forZac Medico2010-08-141-3/+6
| | | | | appropriate phases (it was leaking into clean/cleanrm from previous phases.
* Fix _schedule spelling in _schedule_waitpid().Zac Medico2010-08-141-1/+1
|
* Add a PollScheduler._schedule_waitpid() method for SubProcess._wait()Zac Medico2010-08-144-4/+49
| | | | | to use as an alternative to hanging up the scheduler in a blocking waitpid call.
* Atom.violated_conditionals: Use iuse.is_valid_flag instead of iuse.allSebastian Luther2010-08-134-24/+37
|
* Add a SIGUSR1 handling to trigger pdb.set_trace().Zac Medico2010-08-131-0/+6
|
* Add back the POLLHUP handling for now since PipeReaderTestCase breaksZac Medico2010-08-131-1/+5
| | | | without it.
* Stop handling POLLHUP events, since we're only polling for input andZac Medico2010-08-131-5/+1
| | | | | | POLLHUP is really intendend for purposes of output polling. Hopefully this solves a problem with POLLHUP events causing premature unregistration of event handlers (and subsequent hang in waitpid).
* Define basestring for python3.Zac Medico2010-08-131-0/+4
|
* Implement use dependency defaultsSebastian Luther2010-08-1310-152/+541
|
* Set PORTAGE_PYTHON in processes executed by portage.process._exec().Arfrever Frehtes Taifersar Arahesis2010-08-141-1/+4
|
* portage.os is apparently imported, so os doesn't need to be imported.Arfrever Frehtes Taifersar Arahesis2010-08-141-1/+0
|
* Ensure that all Python processes use the same version of Python.Arfrever Frehtes Taifersar Arahesis2010-08-142-0/+4
|