summaryrefslogtreecommitdiffstats
Commit message (Collapse)AuthorAgeFilesLines
...
* doebuild: check distfiles digests less oftenZac Medico2011-08-301-1/+4
| | | | | | If the unpack phase is not going to be executed as a dependency, then we can simply skip the distfiles digest checks if the unpack phase is already marked complete via $PORTAGE_BUILDDIR/.unpacked.
* repoman: discard STDERR output from CVS.Michał Górny2011-08-301-1/+1
| | | | Fixes: https://bugs.gentoo.org/show_bug.cgi?id=373669
* tests/emerge: add USE for installed pkgsZac Medico2011-08-291-0/+2
|
* dblink.treewalk: remove dowgrade checkZac Medico2011-08-291-9/+0
| | | | | This code is unused since commit a37eb8ebd2fad3f8074490a061f067e2c637f05d.
* Fix PORTDIR in emerge tests.Zac Medico2011-08-291-0/+2
|
* Mark some messages for translation.Arfrever Frehtes Taifersar Arahesis2011-08-291-10/+10
|
* env_update: always respect makelinks=FalseZac Medico2011-08-292-8/+5
| | | | | | | | | | | | This ensures that the env-update --no-ldconfig option is always respected. Also, in order to preserve behavior during downgrades for an env_update call in vartree.py, remove the makelinks=False for downgrades since it wasn't respected anyway when lib path mtimes had changed, and we don't want it to behave differently in this case now that makelinks=False is always respected. Note that the unconditional use of the ldconfig -X option (from bug #373341) since commit e1347f9c0dd5ef5ff1a50d6b136b0648efb8a6ca may also come into play for downgrades.
* env_update: fix breakage in last commitZac Medico2011-08-291-1/+1
|
* env_update: fix CHOST/CBUILD ldconfig codeZac Medico2011-08-291-4/+6
| | | | | | This code never worked (since it was added in commit 751893b0272561eb9274110a474d5436a7d2bc76) due to a name collision between the env argument and another variable with the same name.
* Fix env interference in emerge/repoman tests.Zac Medico2011-08-292-15/+26
|
* emerge: add simple unit testsZac Medico2011-08-285-2/+148
| | | | | | | These tests are similar to the repoman tests, using a __PORTAGE_TEST_EPREFIX environment variable to make emerge confine itself to a testing prefix so that things like install and uninstall operations can be performed.
* Send "Recording in world" msg to stdout.Zac Medico2011-08-281-1/+2
|
* ResolverPlayground: fix to make emerge happyZac Medico2011-08-281-1/+15
|
* env_update: add minimal EPREFIX supportZac Medico2011-08-282-10/+16
| | | | | | | This takes EPREFIX from the env argument and uses it when joining all paths. Also, ldconfig calls are disabled when EPREFIX is non-empty, since it's inappropriate to update the global /etc/ld.so.cache in this case.
* python3.2 fixes: ResourceWarning: unclosed fileZac Medico2011-08-281-2/+7
|
* test_poll: fix "ResourceWarning: unclosed file"Zac Medico2011-08-281-1/+9
|
* Sort test cases, for predictable order.Zac Medico2011-08-281-0/+2
|
* spawnebuild: skip previously executed phasesZac Medico2011-08-281-0/+11
| | | | | | | | | | This simply checks of $PORTAGE_BUILDDIR/.${EBUILD_PHASE%e}ed and skips the phase like ebuild.sh would. It preserves a special case for the install phase with FEATURES=noauto, so that dyn_install in ebuild.sh continues to work the same for this case.. Also, note that commit ae9b6cb513c7b29376caecf3b4e52aac452e2b93 preserves the automatic "recreating WORKDIR" behavior that used to be implemented in dyn_unpack.
* doebuild: avoid redundant distfiles checksZac Medico2011-08-282-36/+71
| | | | | | | | | When the unpack phase is already marked as complete, it's wasteful to check distfiles digests. In order to avoid this, we have to migrate the distfiles/workdir timestamp comparisons from ebuild.sh to doebuild.py, so that doebuild always knows when unpack will be triggered. This also allows us to eliminate code in dyn_unpack that duplicated dyn_clean, actually call dyn_clean instead.
* Fix possible NameError in finally block.Zac Medico2011-08-281-1/+3
|
* Fix "ResourceWarning: unclosed file" with Python 3.2.Arfrever Frehtes Taifersar Arahesis2011-08-281-2/+5
|
* Fix "ResourceWarning: unclosed file" with Python 3.2.Arfrever Frehtes Taifersar Arahesis2011-08-271-1/+1
|
* egencache: don't require xml in global scopeZac Medico2011-08-271-8/+8
|
* egencache: implement XMLParser.doctype()Zac Medico2011-08-271-1/+11
| | | | | | | This is prevents "DeprecationWarning: This method of XMLParser is deprecated. Define doctype() method on the TreeBuilder target." warnings. The code is copied from repoman, commit 5a5e51a2737a0855bb562683f50c57cc31587460.
* python3.2 fixes: ResourceWarning: unclosed fileZac Medico2011-08-263-6/+14
|
* Update --rebuild-if-* flags to rebuild when build dependencies are changed.David James2011-08-264-137/+110
| | | | | | | | | | | | | | Right now, the --rebuild-if-* flags only rebuild packages that are used at both run-time and build-time. This doesn't help for packages that are used only at build-time (for example, static libaries). Rebuilding packages whenever a build-time dependency is changed is easier to understand and explain, and it handles all cases correctly. BUG=chromium-os:15517 TEST=Run emerge test suite. Change-Id: Iae8dab24e8acb6625bc1a0ce41862e90b232eb84
* Fix EPREFIX/EROOT typo in repoman test.Zac Medico2011-08-261-1/+1
|
* python3.2 fixes: ResourceWarning: unclosed fileZac Medico2011-08-263-2/+6
|
* Add __test__ for repoman tests.Zac Medico2011-08-261-0/+0
|
* repoman: add a simple unit testZac Medico2011-08-254-5/+180
|
* Add skipped test support (for python 2.6 compat)Zac Medico2011-08-251-1/+17
| | | | | | Python supports skipped tests since 2.7, but we'll implement support ourselves for python 2.6 compatibility. This is a simple extension of the todo support that we have already implemented.
* testLazyImport: fix PYTHONPATH settingZac Medico2011-08-251-1/+1
|
* Add a copy of metadata.dtd, for repoman tests.Zac Medico2011-08-251-0/+99
| | | | | Since repoman uses metadata.dtd for metadata.xml checks, our sources need to include a copy in order to test it.
* python3.2 fixes: ResourceWarning: unclosed fileZac Medico2011-08-251-2/+4
|
* Enable -Wd for tests, so DeprecationWarning worksZac Medico2011-08-252-2/+2
| | | | | DeprecationWarning is suppressed by default since Python 2.7, so it's a good idea to explicitly enable it for tests.
* python3.2 fixes: ResourceWarning: unclosed fileZac Medico2011-08-251-0/+13
|
* TestCase: remove python 2.4 compat constructorZac Medico2011-08-251-9/+5
|
* python3.2 fixes: ResourceWarning: unclosed fileZac Medico2011-08-257-7/+33
|
* Improve a comment.Arfrever Frehtes Taifersar Arahesis2011-08-251-1/+1
|
* Add some comments, which will allow to easier find code, which should beArfrever Frehtes Taifersar Arahesis2011-08-256-0/+7
| | | | removed when support for older versions of Python is removed.
* die: defer die hooks for EAPI 4 helpers_dieZac Medico2011-08-251-1/+6
| | | | This will fix bug #380625.
* python3.2 fixes: ResourceWarning: unclosed fileZac Medico2011-08-245-13/+34
|
* python3.2 fixes: use array.tobytes()Zac Medico2011-08-245-11/+27
|
* python3.2 fixes: ResourceWarning: unclosed fileZac Medico2011-08-246-17/+38
|
* ElementTree: use iter if available (bug #380565)Zac Medico2011-08-241-2/+11
|
* SafeConfigParser: renamed to ConfigParser in 3.2Zac Medico2011-08-242-3/+13
|
* make.conf.5: document EMERGE_WARNING_DELAYZac Medico2011-08-221-0/+6
|
* dosym: QA warn if target is existing dirZac Medico2011-08-221-1/+2
|
* dosym: QA warn if basename omitted (bug #379899)Zac Medico2011-08-221-0/+5
|
* ebuild(1): show merge/noauto msg only if neededZac Medico2011-08-201-4/+4
|