summaryrefslogtreecommitdiffstats
Commit message (Collapse)AuthorAgeFilesLines
...
* portage(5): add more pointers to make.confMike Frysinger2013-01-151-2/+4
| | | | Signed-off-by: Mike Frysinger <vapier@gentoo.org>
* emerge: add reference to the portage(5) man page when failingMike Frysinger2013-01-151-8/+9
| | | | | | | | | | | | | | | | | | | For example, the current licensing error message looks like: The following license changes (package.license) are necessary to proceed: #required by quake3-bin (argument) >=games-fps/quake3-bin-1.32c-r1 GPL-2 Q3AEULA If you don't know much about licensing issues, this error message doesn't help. Instead, give references to the man page so people can easily delve further. Now it looks like: The following license changes are necessary to proceed: (see package.license in the portage(5) man page for more details) #required by quake3-bin (argument) >=games-fps/quake3-bin-1.32c-r1 GPL-2 Q3AEULA Signed-off-by: Mike Frysinger <vapier@gentoo.org>
* make.conf.5: document @preserved-rebuildZac Medico2013-01-151-0/+3
|
* doebuild: tweak handling of _unused paramZac Medico2013-01-152-5/+5
|
* EbuildMetadataPhase: use dynamic pipe fdZac Medico2013-01-153-41/+12
| | | | Also, remove deprecated code from doebuild (with hardcoded pipe fd).
* SpawnProcess: improve dummy pipe allocation logicZac Medico2013-01-158-30/+57
| | | | | | | By using allocated file descriptors for keys in fd_pipes, we naturally avoid interference with callers such as FileDigester and MergeProcess. See the _setup_pipes docstring for more benefits of this allocation approach.
* Warn if old make.globals found, bug #452176v2.2.0_alpha153Zac Medico2013-01-151-2/+16
|
* Warn if 2 make.profile dirs found, bug #452176Zac Medico2013-01-151-3/+15
|
* test-fail-continue: mark complete, bug #452030Zac Medico2013-01-151-1/+8
|
* SpawnProcess: avoid fd_pipes interferenceZac Medico2013-01-151-2/+8
| | | | | This fixes a random collision that can occur with FileDigester, triggering malfunction and a KeyError.
* ebuild.1: update obsolete md5 referencesZac Medico2013-01-151-6/+6
|
* Respect LINGUAS in 'make install'.Arfrever Frehtes Taifersar Arahesis2013-01-151-7/+10
|
* Makefile: install man/ru/*v2.2.0_alpha152Zac Medico2013-01-141-4/+10
|
* ebuild.1: add Russian translation, bug #450372Vladimir Romanov2013-01-141-0/+242
|
* selinux: python3 unicode paths, bug #430488Arfrever Frehtes Taifersar Arahesis2013-01-142-27/+30
|
* Tweak init.d qa check for bug #451386.Zac Medico2013-01-111-0/+4
|
* make.conf.5: tweak escape format for apostrophyv2.2.0_alpha151Zac Medico2013-01-111-1/+1
|
* DoebuildSpawnTestCase: also call spawn_nofetchZac Medico2013-01-101-12/+27
|
* spawn_nofetch: import portage, bug #451344Zac Medico2013-01-101-0/+1
|
* Revert "_ctypes: don't cache library, bug #448858"v2.2.0_alpha150Zac Medico2013-01-101-4/+6
| | | | | | | | This reverts commit 9e37cca4f54260bd8c45a3041fcee00938c71649. As noted in bug #448858, comment #14, dlclose is not called automatically, so we may as well cache our library handles. In order to protect ourselves, we use a fork since commit 7ebb2f54877edb28621c33e380f8777b1b1dc201.
* Use EventLoop, no SchedulerInterface if possible.Zac Medico2013-01-101-2/+2
|
* emerge/test_simple.py: override PATH moreZac Medico2013-01-102-3/+9
| | | | | Override things that may be unavailable, or may have portability issues when running tests in exotic environments.
* SpawnProcess: handle fcntl ENOTTY for FreeBSDZac Medico2013-01-101-3/+8
|
* _emirrordist/main.py: remove extra Config importZac Medico2013-01-101-1/+0
|
* FetchTask: use umask for copied filesZac Medico2013-01-101-12/+2
|
* FetchTask: avoid possible duplicate URIZac Medico2013-01-101-2/+4
|
* FetchTask: randomize thirdpartymirrorsZac Medico2013-01-101-1/+7
|
* FetchTask: log_failure for distfiles digesterZac Medico2013-01-101-0/+7
|
* FetchIterator: support PORTAGE_CHECKSUM_FILTERZac Medico2013-01-091-0/+13
|
* make.conf.5: escape apostropy quote, bug #451118Zac Medico2013-01-091-2/+2
|
* FetchIterator: use RepoConfig.load_manifest()Zac Medico2013-01-091-3/+5
|
* _fetch_digester_exit: fix NoneType AttributeErrorZac Medico2013-01-091-2/+2
|
* Makefile: install emirrordistZac Medico2013-01-091-1/+1
|
* Add emirrordist, a tool for mirroring distfiles.Zac Medico2013-01-0912-0/+1967
| | | | | Special thanks to Brian Harring, author of the mirror-dist program from which emirrordist is derived.
* Manifest: make distdir argument optionalZac Medico2013-01-091-2/+2
|
* autounmask: tweak --autounmask-write suggestionZac Medico2013-01-081-3/+4
|
* autounmask: show file names for suggested changesBen Kohler2013-01-081-5/+7
| | | | | | | Also, advise the user to examine the proposed config changes carefully. X-Gentoo-Bug: 450960 X-Gentoo-Bug-URL: https://bugs.gentoo.org/show_bug.cgi?id=450960
* SpawnProcess: stdout_fd FD_CLOEXECZac Medico2013-01-071-0/+15
|
* ManifestTask: use PipeLogger for monitoringZac Medico2013-01-071-2/+8
|
* Declare regex literal as such.Zac Medico2013-01-071-1/+1
|
* parseManifest2: handle space in name, bug 450736Zac Medico2013-01-071-14/+18
|
* PipeLogger: handle file object for input_fdZac Medico2013-01-072-9/+14
|
* If 2 make.conf files found, then warn.Zac Medico2013-01-071-3/+12
| | | | | This may give a clue to those who unintentionally end up with 2 make.conf files somehow.
* MetadataRegen: simplify _iter_every_cpZac Medico2013-01-062-14/+6
|
* MetadataRegen: call cp_all for each categoryZac Medico2013-01-062-16/+21
| | | | | List categories individually, in order to start yielding quicker, and in order to reduce latency in case of a signal interrupt.
* portdbapi.cp_all(): add reverse sort parameterZac Medico2013-01-061-2/+3
|
* _keep_scheduling: check _terminated.is_set()Zac Medico2013-01-062-4/+4
| | | | | This ensures that scheduling loops terminate as soon as possible after an interrupt is received.
* ManifestScheduler: fix signal interrupt handlingZac Medico2013-01-061-3/+5
|
* MetadataRegen: portdbapi.cp_all() already sortedZac Medico2013-01-061-1/+1
|
* MetadataRegen: fix signal interrupt handlingZac Medico2013-01-061-3/+4
| | | | | We have to force it out of the scheduling loop before _termination_check can have an effect.