summaryrefslogtreecommitdiffstats
Commit message (Collapse)AuthorAgeFilesLines
* Bug #337178 - Fix slot conflict backtracking to mask the higherZac Medico2010-09-271-10/+36
| | | | | | | | | | version if appropriate. Generally, we mask the higher version since this solves common cases in which a lower version is needed so that all dependencies will be satisfied (bug #337178). However, if existing_node happens to be installed then we mask that since this is a common case that is triggered when --update is not enabled.
* Remove --depclean-lib-check docs from the man/emerge.1 since it's not ↵Zac Medico2010-09-271-6/+0
| | | | supported in the stable branch.
* Adjust package sets documentation to match behavior in the stable branch.Zac Medico2010-09-273-20/+4
|
* Add warnings advising use of revdep-rebuild in --depclean documentation ↵Zac Medico2010-09-271-1/+4
| | | | since _ENABLE_DYN_LINK_MAP is False in the stable branch.
* Remove documentation for FEATURES=preserve-libs since it's not supported in ↵Zac Medico2010-09-271-4/+0
| | | | the stable branch.
* Comment out the FEATURES=preserve-libs setting in make.globals since it's ↵Zac Medico2010-09-271-1/+1
| | | | not supported in the stable branch.
* Toggle private constants for stable branch:Zac Medico2010-09-271-5/+5
| | | | | | | | _ENABLE_DYN_LINK_MAP = False _ENABLE_PRESERVE_LIBS = False _ENABLE_REPO_NAME_WARN = False _ENABLE_SET_CONFIG = False _SANDBOX_COMPAT_LEVEL = "16"
* Revert one more change from bug #337178: parent_atoms is supposedZac Medico2010-09-261-1/+1
| | | | to refer to the new pkg, not the one to_be_masked.
* Revert slot conflict behavior backtracking behavior changes fromZac Medico2010-09-261-5/+6
| | | | | | | | bug #337178 and comment about reasoning for behavior. We always mask existing_node since _select_package tries to avoid slot conflicts when possible and therefore a conflict typically means that existing_node was a poor choice.
* Fix confusing messages shown byZac Medico2010-09-261-6/+6
| | | | | | | | | | | depgraph._show_missed_update_slot_conflicts() since the fix from bug #337178. When depgraph._add_pkg() was fixed to mask the higher version in 4bc78ab0b563697b98527eebcdfe474863383cf0, we didn't adjust the parent_atoms variable so that it would contain parent atoms from the versions being masked. This resulted in confusing messages from depgraph._show_missed_update_slot_conflicts().
* Add PORTAGE_BINHOST sftp protocol support.v2.2_rc87Zac Medico2010-09-253-5/+27
|
* Fix a typo in a comment.Zac Medico2010-09-251-1/+1
|
* Add non-essential backslashes for clarity in GetConfigTestCase.Zac Medico2010-09-251-1/+1
|
* Add test cases for getconfig() bash compatibility with variablesZac Medico2010-09-251-0/+28
| | | | from make.globals.
* Add default values for FETCHCOMMAND_SSH and RESUMECOMMAND_SSHZac Medico2010-09-251-0/+3
| | | | that call rsync.
* Fix varexpand() so that it doesn't do redundant quote removal thatZac Medico2010-09-251-1/+7
| | | | | | | | is already handled by shlex. This fixes a bug in getconfig() which caused it to remove backslash characters it front of quote characters in cases where bash would have preserved them when sourcing the same input.
* Fix BinpkgVerifier so that it doesn't try to rename the file ifZac Medico2010-09-251-1/+3
| | | | it doesn't exist.
* Make MetaDataXML.format_maintainer_string() check a values to makeZac Medico2010-09-251-4/+6
| | | | sure they are non-empty and not just whitespace.
* Implement PORTAGE_BINHOST ssh:// protocol support by spawning sshZac Medico2010-09-251-1/+18
| | | | | | | | | | and using it to stream the Packages file on stdout. This assumes that you have ssh-agent running, for authentication. For now, you won't be able to use this unless you define your own FETCHCOMMAND_SSH and RESUMECOMMAND_SSH values. It's trivial to do this if you create a helper script that calls rsync after transforming the URI into the host:/path format that rsync uses.
* In EbuildPhase, don't display Repository if the value is not available.Zac Medico2010-09-251-1/+2
|
* Move all imports in bintree.py to the top.Zac Medico2010-09-251-22/+14
|
* Make BinpkgFetcher use AsynchronousLock.Zac Medico2010-09-251-3/+7
|
* Fix catpkgsplit() to validate the category part. This fixesZac Medico2010-09-251-1/+3
| | | | vardbapi.cp_all() to reject invalid categories.
* Bug #338588 - Allow quickpkg to create packages even though theZac Medico2010-09-251-3/+10
| | | | current user doesn't have permission to lock /var/db/pkg.
* Bug #337465 - Enable EbuildIpcDaemon on Darwin and FreeBSD sinceZac Medico2010-09-251-8/+0
| | | | | | it seems fixed now. This reverts commit 0906fee12ffe0ebd07e8952316f03927962af29b.
* Make EbuildIpcDaemon._input_handler() handle EAGAIN from os.read()Zac Medico2010-09-231-1/+7
| | | | for FreeBSD (bug 337465, comment #35).
* Merge branch 'master' of git+ssh://git.overlays.gentoo.org/proj/portageFabian Groffen2010-09-231-2/+2
|\
| * Add binpkg-logs to the list of supported features.Arfrever Frehtes Taifersar Arahesis2010-09-231-2/+2
| |
* | always use eroot with CONFIG_MEMORY_FILEFabian Groffen2010-09-231-2/+2
|/
* Bug #337465 - Make EbuildIpcDaemon._input_handler() use os.read()Zac Medico2010-09-231-3/+11
| | | | | since array.fromfile() and file.read() are both known to erroneously return an empty string from this non-blocking fifo stream on FreeBSD.
* Add support for FEATURES=binpkg-logs which is similar to unmerge-logs,Zac Medico2010-09-223-1/+14
| | | | but applies to all binary package merges.
* Fix a typo in a comment.Zac Medico2010-09-211-1/+1
|
* In EbuildIpc._receive_reply(), close the input_file when finished.v2.2_rc86Zac Medico2010-09-211-0/+2
|
* In EbuildIpc._wait(), use waitpid() instead of wait() since we reallyZac Medico2010-09-211-1/+1
| | | | only need to wait on one process.
* Bug #337465 - Make ebuild-ipc use an array in order to force aZac Medico2010-09-211-3/+16
| | | | single atomic read of a whole pickle.
* Bug #337465 - Disable EbuildIpcDaemon on Darwin and FreeBSD sinceZac Medico2010-09-211-0/+8
| | | | | | it reportedly triggers a tight loop in emerge. This reverts commit 6085103bdd7623d941e3c6ae5905e66658c49fd7.
* Make ebuild-ipc use select() for timeouts, instead of alarm signals.Zac Medico2010-09-211-29/+44
|
* Bug #338267 - Fix broken PORTAGE_BUILDDIR assertion in Binpkg whenZac Medico2010-09-211-1/+3
| | | | PORTAGE_TMPDIR is a symlink.
* Bug #338190 - Handle more exceptions inside getbinpkg.dir_get_metadata().Zac Medico2010-09-211-1/+11
| | | | | | The ftplib.all_errors attribute is a useful collection of exceptions to handle here. Thanks to Christian Kauhaus <kc@gocept.com> for the suggestion.
* Fix ConfigTestCase.testLicenseManager() breakage from from previousZac Medico2010-09-211-1/+1
| | | | commit.
* Bug #300455 - Store LicenseManager._license_groups values as frozensetZac Medico2010-09-211-0/+3
| | | | in order to avoid false warnings about circular group references.
* reposyntax: Add support in _sets/base.pySebastian Luther2010-09-201-12/+17
|
* Fix default() and default_src_install() definitions to account forZac Medico2010-09-201-0/+10
| | | | the new EAPI 4 default src_install implementation.
* Make ebuild-ipc use a normal read() call instead of array.fromfile()Zac Medico2010-09-201-13/+3
| | | | since that should work fine for blocking IO.
* Remove obsolete comment about non-blocking fromZac Medico2010-09-201-4/+0
| | | | EbuildIpcDaemon._send_reply() since ebuild-ipc now uses blocking IO.
* Make ebuild-ipc kill the child IO process before it exits due to timeout.Zac Medico2010-09-201-1/+8
|
* Fix ebuild-ipc to correctly handle non-zero return codes fromZac Medico2010-09-201-4/+8
| | | | has_version ipc calls.
* Rewrite ebuild-ipc timeout handling to use forks.Zac Medico2010-09-201-62/+85
| | | | | | | Use forks so that the child process can handle blocking IO un-interrupted, while the parent handles all timeout considerations. This helps to avoid possible race conditions from interference between timeouts and blocking IO operations.
* Remove obsolete docstring from _RepoDisplay.__str__().Zac Medico2010-09-201-4/+0
|
* Make ebuild-ipc show an additional 'read failed' message if the bufferZac Medico2010-09-201-1/+8
| | | | happens to be empty after the read loop.