Commit message (Collapse) | Author | Age | Files | Lines | |
---|---|---|---|---|---|
* | Allow repo in more Atom instances. | Zac Medico | 2010-10-08 | 1 | -4/+8 |
| | |||||
* | Add debug output for backtracking limit. | Zac Medico | 2010-10-08 | 1 | -0/+5 |
| | |||||
* | Allow repo atoms more often in InternalPackageSet | Zac Medico | 2010-10-07 | 1 | -3/+5 |
| | |||||
* | Allow repo since parents can be sets or args. | Zac Medico | 2010-10-06 | 1 | -2/+3 |
| | |||||
* | Fix a typo in a comment.v2.2_rc90 | Zac Medico | 2010-10-06 | 1 | -1/+1 |
| | |||||
* | Optimize slot conflict backtracking for #337178. | Zac Medico | 2010-10-06 | 1 | -0/+13 |
| | | | | | | | | This borrows logic from commit fc33f34e3d060b57e03204fb427b01971f8dafff which fixes bug #337178 in the 2.1.9 branch. With this change, BacktrackingTestCase.testBacktrackingGoodVersionFirst() is solved with one less backtracking step. | ||||
* | Add fallback backtrack_data, in case logic fails. | Zac Medico | 2010-10-06 | 1 | -6/+20 |
| | | | | | | This adds some safety in the depgraph._add_pkg() slot conflict code, in case the new logic fails somehow (it shouldn't, but given all the complexities, who knows?). Also, add some more debug output. | ||||
* | Backtracker reverses backtrack_data now. | Zac Medico | 2010-10-06 | 1 | -4/+3 |
| | | | | | | | Since commit 9ff5e9731142d389373ea6ebc949919c3b637110, Backtracker._feedback_slot_conflict() adds nodes in the order of backtrack_data and this order is then reversed when the nodes are popped from _unexplored_nodes. | ||||
* | Remove unused variable. | Zac Medico | 2010-10-05 | 1 | -1/+0 |
| | |||||
* | Fix _iter_match_pkgs installed multislot breakage | Zac Medico | 2010-10-05 | 1 | -7/+4 |
| | |||||
* | Use Atom.without_use in _show_unsatisfied_dep(). | Zac Medico | 2010-10-05 | 1 | -1/+2 |
| | |||||
* | Use match-all-cpv-only in _show_unsatisfied_dep. | Zac Medico | 2010-10-05 | 1 | -8/+4 |
| | |||||
* | Don't allow_wildcard in _iter_match_pkgs. | Zac Medico | 2010-10-05 | 1 | -1/+1 |
| | |||||
* | Mask packages that don't satisfy all atoms. | Sebastian Luther | 2010-10-05 | 1 | -19/+21 |
| | | | | This also fixes a logic error in resolver/backtracking/Backtracker._feedback_slot_conflict() | ||||
* | Fix _show_unsatisfied_dep to pass repo to _pkg(). | Zac Medico | 2010-10-04 | 1 | -0/+2 |
| | |||||
* | Add note about depgraph/portdbapi caching. | Zac Medico | 2010-10-04 | 1 | -0/+3 |
| | |||||
* | Optimize depgraph/portdbapi cache handling. | Zac Medico | 2010-10-04 | 1 | -1/+8 |
| | |||||
* | Skip redundant cpv_exists in _iter_match_pkgs(). | Zac Medico | 2010-10-04 | 1 | -2/+0 |
| | |||||
* | Clarify repo code in depgraph._pkg(). | Zac Medico | 2010-10-04 | 1 | -3/+6 |
| | |||||
* | Cache InternalPackageSet in _iter_match_pkgs(). | Zac Medico | 2010-10-04 | 1 | -2/+4 |
| | |||||
* | Clean up repo handling for depgraph pkg_cache. | Zac Medico | 2010-10-04 | 1 | -35/+18 |
| | |||||
* | Optimize slot conflict backtracking for #337178. | Zac Medico | 2010-10-04 | 1 | -0/+14 |
| | | | | | | | | This borrows logic from commit fc33f34e3d060b57e03204fb427b01971f8dafff which fixes bug #337178 in the 2.1.9 branch. With this change, BacktrackingTestCase.testBacktrackingGoodVersionFirst() is solved with fewer backtracking steps. | ||||
* | Bug #339606 - Fix broken 'missed update' message. | Zac Medico | 2010-10-03 | 1 | -4/+16 |
| | |||||
* | Prune circular nested sets from the digraph. | Zac Medico | 2010-10-02 | 1 | -1/+16 |
| | | | | | | | | | | | | | This prevents the following traceback which was reported in forum thread 847007: File "pym/_emerge/depgraph.py", line 3694, in _merge_order_bias mygraph.order.sort(key=cmp_sort_key(cmp_merge_preference)) File "pym/portage/util/__init__.py", line 808, in __lt__ return self._cmp_func(self._obj, other._obj) < 0 File "pym/_emerge/depgraph.py", line 3675, in cmp_merge_preference if node1.operation == 'uninstall': AttributeError: 'SetArg' object has no attribute 'operation' | ||||
* | Use _pkg_visibility_check in _loadResumeCommand. | Zac Medico | 2010-10-01 | 1 | -1/+1 |
| | |||||
* | Enable depgraph._load_vdb() with --buildpkgonly. | Zac Medico | 2010-10-01 | 1 | -2/+1 |
| | | | | | | The --buildpkgonly option does not negate the need for _load_vdb(), at least not with the --resume option since otherwise the select_pkg_from_graph() method fails to select installed packages. | ||||
* | Bug #339316 - Fix --resume to handle repo masks. | Zac Medico | 2010-10-01 | 1 | -3/+17 |
| | |||||
* | Tweak depgraph._show_unsatisfied_dep() display for AtomArg instances. | Zac Medico | 2010-09-29 | 1 | -2/+11 |
| | |||||
* | Add basic repo atom support for emerge --resume. | Zac Medico | 2010-09-28 | 1 | -10/+20 |
| | |||||
* | Fix breakage in depgraph._serialize_tasks() blocker handling due | Zac Medico | 2010-09-27 | 1 | -5/+9 |
| | | | | to addition of repo to Package._hash_key. | ||||
* | Fix depgraph._pkg() to not raise PackageNotFound when myrepo is None | Zac Medico | 2010-09-27 | 1 | -1/+3 |
| | | | | and a package is found, so --resume works. | ||||
* | Omit repo from mtimedb resume mergelist entries since the format | Zac Medico | 2010-09-27 | 1 | -4/+8 |
| | | | | is intended to be frozen. | ||||
* | reposyntax: Fix backtrace caused by cache changes | Sebastian Luther | 2010-09-26 | 1 | -3/+3 |
| | |||||
* | reposyntax: Handle the "give me a Package for an ebuild from any repo" case ↵ | Sebastian Luther | 2010-09-26 | 1 | -7/+23 |
| | | | | better | ||||
* | reposyntax: Print cpv from all repos for "masked by" messages | Sebastian Luther | 2010-09-26 | 1 | -50/+64 |
| | |||||
* | Fix breakage from the last commit | Sebastian Luther | 2010-09-26 | 1 | -1/+1 |
| | |||||
* | reposyntax: Next try at fixing the package cache | Sebastian Luther | 2010-09-26 | 1 | -1/+15 |
| | |||||
* | reposyntax: Get the Package cache to work again | Sebastian Luther | 2010-09-26 | 1 | -1/+1 |
| | |||||
* | Fix breakage with old style virtuals from the last commit | Sebastian Luther | 2010-09-26 | 1 | -10/+10 |
| | |||||
* | reposyntax: Fix ignored slot spec if the same cpv is in several repos with ↵ | Sebastian Luther | 2010-09-26 | 1 | -0/+3 |
| | | | | different SLOT | ||||
* | reposyntax: Add support all over the place | Sebastian Luther | 2010-09-26 | 1 | -43/+48 |
| | |||||
* | backtracking: Take all branches in case of slot collisionsv2.2_rc88 | Sebastian Luther | 2010-09-26 | 1 | -98/+83 |
| | |||||
* | Revert one more change from bug #337178: parent_atoms is supposed | Zac Medico | 2010-09-26 | 1 | -1/+1 |
| | | | | to refer to the new pkg, not the one to_be_masked. | ||||
* | Revert slot conflict behavior backtracking behavior changes from | Zac Medico | 2010-09-26 | 1 | -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 by | Zac Medico | 2010-09-26 | 1 | -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(). | ||||
* | Tests: add resolver/test_output: Make sure we don't backtrace in the merge ↵ | Sebastian Luther | 2010-09-20 | 1 | -16/+12 |
| | | | | list printer | ||||
* | Move mergelist printing into resolver/output.py | Sebastian Luther | 2010-09-20 | 1 | -799/+4 |
| | |||||
* | Always mask the higher version when backtracking due to a slot conflict | Sebastian Luther | 2010-09-19 | 1 | -4/+10 |
| | |||||
* | Fix depgraph._complete_graph() to preserve initial_arg_list when it | Zac Medico | 2010-09-19 | 1 | -4/+6 |
| | | | | updates args with _set_args(). | ||||
* | Fix incorrect depgraph_sets reference in depgraph._load_favorites().v2.2_rc84 | Zac Medico | 2010-09-18 | 1 | -1/+1 |
| |