commit bb24de7ac7c82d9d2c85e143c0d1655b168c5990 Author: Dan Callaghan Date: Wed Mar 26 14:16:45 2014 +1000 Automatic commit of release 0.16.1 commit 5ad04ba33d4f83c5922434f008bfd6a1f439fde6 Author: Dan Callaghan Date: Wed Mar 26 14:15:37 2014 +1000 add bz1080685 to 0.16.1 release notes commit 7ff6422cf254acbced057de6eb0e1d96ef2df417 Author: Dan Callaghan Date: Wed Mar 26 09:45:30 2014 +1000 fix handling of release_action in save_power controller Since commit da6966ce the power config form validator is actually applied, therefore release_action is now already a valid ReleaseAction enum value. Calling ReleaseAction.from_string() on it again results in a ValueError. Bug: 1080685 Change-Id: Iefda403ac457aa2321d5f9ac37a93f28eb97aec8 commit 81913ed60e95abf1b6e6bae8d3d6bad496be70bd Author: Dan Callaghan Date: Tue Mar 25 17:01:13 2014 +1000 release notes for 0.16.1 Change-Id: Ia9101fd7a883d3407b313d85d384c114cb17a04f commit e23d2ea3991179f90eb86ba2f0e565ccd0952018 Author: Dan Callaghan Date: Mon Mar 24 13:54:23 2014 +1000 split 'reboot' commands into 'off' and 'on' on the server side This way the quiescent period will correctly be applied to both the 'off' and 'on' commands. Currently there is no delay between 'off' and 'on' when rebooting. Bug: 1079603 Change-Id: I907d6f7b6546124769d2b267482345685e779fee commit e0cc7afa9dd5a22fdacb4534c2d84a651ca133e1 Author: Dan Callaghan Date: Mon Mar 24 16:53:23 2014 +1000 obey quiescent period in consecutive power commands We need to wait for the quiescent period after waiting for all predecessor commands, otherwise the quiescent period delay will happen concurrently. Also record the last command timestamp only when the command has finished, rather than before it is started, since it may take some time to complete. Bug: 1079816 Change-Id: Id4bbf79ae739a669b54ffde0101329f5174424e1 commit 4ae5b753dedf20c2def95bb034a75ca51e8d6865 Author: Dan Callaghan Date: Mon Mar 24 16:49:49 2014 +1000 insert commands into the top of the queue instead of appending System.command_queue is supposed to be in descending order (most recent command at the top). Normally the insertion position makes no difference because we will be querying System.command_queue back in a separate transaction, which means it will always end up ordered correctly. This change is only significant for tests which enqueue commands, expecting them to end up in the right position in the list. Change-Id: I69ee2c34535fffd662ea40c00a83b7b785da09f1 commit 3e18efcedd3aa822fc6a40bff8caf637bbdc19c1 Author: Dan Callaghan Date: Mon Mar 24 15:47:38 2014 +1000 tests: actually test quiescent period is obeyed The previous version of the test was catching AssertionError, but TestCase.fail() raises AssertionError so it would always pass. We can also skip the log inspection in this case. Change-Id: I644d0dba57db3723087a28f22df55c17c2158412 commit f28e30b42cb6e1fe283071c6c1e0e6a40eb7f838 Author: Dan Callaghan Date: Mon Mar 24 02:59:23 2014 +0000 Revert "change aarch64 pxedir to keep admins happy." This reverts commit 537ceb0f47636041f7873813100abb7b084e510a. Change-Id: I96e6c34ce0d192cd3208ea5f9ecff672c32e0c7b commit c71cd22812e8bf22bd49eab2f6254d64036a409e Author: Dan Callaghan Date: Wed Mar 19 09:39:19 2014 +1000 define a 'remote' for group typeahead instead of relying on prefetch If the set of groups in Beaker has changed since it was prefetched, new groups would never appear in the suggestions even when they were typed. Now the group typeahead will always fetch fresh suggestions. Bug: 1076322 Change-Id: I36122bfd6dde2b50394d4962de43bdf5e2bea11f commit 6246bbb8bbcd534422bbb230ecfc1c5bdbb42c67 Author: Dan Callaghan Date: Wed Mar 19 09:34:08 2014 +1000 access policy editor: "Add" button for explicitly adding new rows The current behaviour is to add a new row as soon as a typeahead completion is made. That's problematic because typeahead.js doesn't have an official way to indicate if a match has been made, and the current hack does not work when the typed value matches a suggestion that was not in the prefetch. It is also subject to a race if the user blurs the input before the typeahead suggestions have been displayed. With this patch the typeahead is treated purely as a suggestion (which is the way it is intended to be used), it's up to the user to enter or select the right value and then click Add or press Enter to add the row to the policy. If they enter a non-existent user or group we show an error when they attempt to save the policy. Change-Id: I4b5284a36d2a6d63748e89494abf854ee86b8dde commit acdd130802fb11df61a4caee9b399657d49fe352 Author: Dan Callaghan Date: Mon Mar 24 08:50:15 2014 +1000 rename OSMajor.osversion and OSMajor.osminor properties to OSMajor.osversions Up until 0.16 we had both OSMajor.osversion and OSMajor.osminor meaning the same thing (list of OSVersion). During the conversion to SQLAlchemy declarative in 0.16, OSMajor.osversion was accidentally dropped. This patch renames them both to OSMajor.osversions, which more accurately reflects what the property really is. Bug: 1078941 Change-Id: I90497858c2487ad578fec4673dae8e7b15871a93 commit 49810df70932ee6b302ba0c976812611b8c7236f Author: Matt Jia Date: Fri Mar 21 12:43:27 2014 +1000 Fix duplicate packages causing a server fault when submitting a new job Bug: 883887 Change-Id: Ie3679c9bcea0014f273131cc94b09fd342e20137 commit 537ceb0f47636041f7873813100abb7b084e510a Author: Bill Peck Date: Thu Mar 20 11:53:56 2014 -0400 change aarch64 pxedir to keep admins happy. commit 7eb31bbbc92a30f6e637269ccd84c3aa9fd8ecf8 Author: Dan Callaghan Date: Thu Mar 20 16:01:56 2014 +1000 tests: attempting to fix timing-sensitivty in test_job_delete Sometimes the job deletion tests fail because they go to the recipe page, expecting the recipe to be deleted, but it's not yet. This appears to be because they are clicking through before the deletion has completed. I'm hoping that this patch fixes it by correcting the XPath used to look for the deleted table row. Change-Id: I33b7ddd188d572f30cb168f166c5437992201df2 commit 0d8708a9eb5a469dd599b78467bd6245caf94839 Author: Dan Callaghan Date: Thu Mar 20 14:15:27 2014 +1000 set power_quiescent_period to 5 when upgrading to 0.16 We are using a Python default rather than a SQL default for this column, but that means we need to explicitly set the column to its default value for all existing rows when upgrading. Bug: 1078620 Change-Id: I3f07b49229c8ab6f864308cc673e2729bc9677fa commit 0d7f1c7cf63c3720c338e92ec9e9c1326ffbea27 Author: Raymond Mancy Date: Fri Mar 14 16:24:24 2014 +1000 added missing TimeoutProxyHTTP and urlparse to bkr.common.xmlrpc Change-Id: I598858eb87015236b7f51a4746faa7f340de1b58 commit 2cea44372e0aa0f5a94add97bb490d9c4d7bb07c Author: Dan Callaghan Date: Wed Mar 19 11:50:43 2014 +1000 docs: recommend running bkr watchdogs-extend during upgrades Change-Id: Idfa58f655ad8002ede4f3455bd25642777bebe0c commit 12dfaadae3db4e070c86dd53124f654481e11c57 Author: Alexander Todorov Date: Wed Mar 5 11:51:35 2014 +0200 add new workflow-installer-test command Bug: 1026730 Change-Id: I81bd8fb2c72c54e51b251a14ade90bbc5966310d commit b4d47a0c22524a637111a69e6b45515399e272dc Author: Amit Saha Date: Fri Mar 14 13:43:29 2014 +1000 Don't check if beah harness repo exists when alternative harness is used Bug: 1005865 Change-Id: I004c37ec04c9c0fe5614e720198f3acc94fb0913 commit 250f6a8c6c251312e4a1cf8ae883b77b14f5c1d7 Author: Dan Callaghan Date: Mon Mar 17 09:34:35 2014 +1000 tests: don't assert errno for invalid createrepo command Normally this should be ENOENT but for some reason in Brew the invalid command is failing with EACCES instead. Change-Id: I9d18e3bb8f01ba5500de4e526b9192790954d352 commit a92b4b578224b7fc614b42dab65f14c114792623 Author: Dan Callaghan Date: Mon Mar 17 07:57:33 2014 +1000 client build needs to create /etc/beaker because the RPM owns it now Change-Id: Ic9f8d2c833b139e2d0ecaba21905d5b940a7855e commit beea49bd475c0f075eeab5ae8d6cbdf36687df92 Author: Dan Callaghan Date: Mon Mar 17 07:54:02 2014 +1000 nose is needed at build time, as of commit ab1c06fd Change-Id: Ibd00a6ed964c7c8025bae58dbfc152c8f37e62a8 commit 8324973d6b14f8d9a8395ea3fbab34feb36f26e0 Author: Dan Callaghan Date: Mon Mar 17 07:46:28 2014 +1000 CherryPy is not required for client tests Change-Id: I1f09de8370d75072253524d1b526fa7600e47d00 commit b2aadbea30308ab37f2f74ac6b8dd9cd8f76c337 Author: Dan Callaghan Date: Thu Mar 13 10:34:04 2014 +1000 base package needs to own the bkr directory Currently no package owns it. Change-Id: I0fe46b8501a825518c1da7e1e6b6d1ca5031bfdb commit 2616a905df2ad0f73d68479de897992fb74696da Author: Dan Callaghan Date: Tue Mar 11 15:20:45 2014 +1000 report nice error for missing krb credential cache Bug: 1074832 Change-Id: I7504a74bf692409e8d2c15ac9ff1a97c28eef054