Those failures have been started recently and seem to happen more frequently but not permanent across platforms. test2.js::testDirectUpdate_Download New wizard page has been selected - got downloading, expected finished
For better results: http://qa-masterblaster.mv.mozilla.com:8080/job/mozilla-central_update/1621/ Build from: https://ftp.mozilla.org/pub/mozilla.org/firefox/nightly/2012/05/2012-05-29-03-05-18-mozilla-central/firefox-15.0a1.en-US.mac.dmg *** Updating to branch 'default' pulling from mozmill-tests searching for changes no changes found 0 files updated, 0 files merged, 0 files removed, 0 files unresolved TEST-START | /var/folders/vq/bds4whmx4jg5n19nm5p_9xfr0000gp/T/tmpUaJnjZ.mozmill-tests/tests/update/testDirectUpdate/test1.js | setupModule TEST-PASS | /var/folders/vq/bds4whmx4jg5n19nm5p_9xfr0000gp/T/tmpUaJnjZ.mozmill-tests/tests/update/testDirectUpdate/test1.js | test1.js::setupModule TEST-START | /var/folders/vq/bds4whmx4jg5n19nm5p_9xfr0000gp/T/tmpUaJnjZ.mozmill-tests/tests/update/testDirectUpdate/test2.js | setupModule TEST-PASS | /var/folders/vq/bds4whmx4jg5n19nm5p_9xfr0000gp/T/tmpUaJnjZ.mozmill-tests/tests/update/testDirectUpdate/test2.js | test2.js::setupModule TEST-START | /var/folders/vq/bds4whmx4jg5n19nm5p_9xfr0000gp/T/tmpUaJnjZ.mozmill-tests/tests/update/testDirectUpdate/test2.js | testDirectUpdate_Download TEST-UNEXPECTED-FAIL | /var/folders/vq/bds4whmx4jg5n19nm5p_9xfr0000gp/T/tmpUaJnjZ.mozmill-tests/tests/update/testDirectUpdate/test2.js | test2.js::testDirectUpdate_Download TEST-START | /var/folders/vq/bds4whmx4jg5n19nm5p_9xfr0000gp/T/tmpUaJnjZ.mozmill-tests/tests/update/testDirectUpdate/test2.js | teardownModule TEST-PASS | /var/folders/vq/bds4whmx4jg5n19nm5p_9xfr0000gp/T/tmpUaJnjZ.mozmill-tests/tests/update/testDirectUpdate/test2.js | test2.js::teardownModule
The regression has been started on May 23rd and is most likely because of the background updates. Passed: 20120522145033 (http://mozmill-ci.blargon7.com/#/update/report/f87375a634b1a5ba746e5f763a6e3512) Failed: 20120523112744 (http://mozmill-ci.blargon7.com/#/update/report/f87375a634b1a5ba746e5f763a7386b1) Pushlog: http://hg.mozilla.org/mozilla-central/pushloghtml?fromchange=b038090f07c2&tochange=320b16daa7c0 Ehsan, which of the patches could have regressed this? As what I can see it's because of applying the patch in the old updater ui. There should have probably been a new wizard page but now we re-use the 'downloading' one. Simplest solution for us would be to wait a bit longer for the finished pane once the download has been finished.
FWIW, there are no background updates related patches in that range. If this is indeed caused by background updates, the correct solution is to increase the timeout, as I think I noted to Anthony some time ago. Otherwise, you need to bisect the range to see what is causing this failure to happen.
Sorry forgot that the updates were disabled for a couple of days last week. So the regression range is wrong. In fact it is a fallout from bug 307181, especially: http://hg.mozilla.org/mozilla-central/diff/c20d415ef1b5/toolkit/mozapps/update/content/updates.js#l1.113 Given that no new pane has been used for the 'apply' state, waiting until the current pane is not the downloading pane should be sufficient for our case. I will attach a patch in a minute which works fine for me.
Summary: Update tests fail with 'New wizard page has been selected - got downloading, expected finished' → Update tests have to obey 'Applying update...' state since bug 307181 has been landed
Fixed the alphabetical order in the declaration section and a single typo.
Comment on attachment 627978 [details] [diff] [review] Patch v1.1 Looks good, and tests are passing for me. Landed as: http://hg.mozilla.org/qa/mozmill-tests/rev/3878a9948a38 Do we need to transplant this?
Attachment #627978 - Flags: review?(dave.hunt) → review+
Thanks Dave for landing this patch! No I do not think that we want to transplant it because it would need a different patch for the other branches, and once we have to merge one more patch is necessary. So lets bake it on mozilla-central and let it flow with the usual merge process.
You need to log in before you can comment on or make changes to this bug.