Last modified: 2014-02-18 16:19:45 UTC
Flow tests run on CloudBees/Sauce Labs occasionally fail with "failed insert" editing an existing post. Overlapping test sessions may attempt to edit the same topic or post on the Talk:Flow_QA board, which leads to an edit conflict. It would be better if the test edited a known topic or post rather than the first one on the Flow board. Previous feature tests create topics and posts, so (commence handwaving...) the edit_existing test could take an optional parameter naming a particular text string or item ID to edit The alternative is for the test to handle edit conflict, but it's a rare occurrence so the code would be untested until it breaks. A related minor enhancement: to aid debugging issues it would be nice if edit strings like "Post edited0.2921675098405858" included the BROWSER_LABEL.
Change 104511 had a related patch set uploaded by Cmcmahon: add BROWSER_LABEL to edit string per Bug 59011 https://gerrit.wikimedia.org/r/104511
Change 104531 had a related patch set uploaded by Cmcmahon: different topics per browser to prevent edit conflict Bug 59011 https://gerrit.wikimedia.org/r/104531
Change 104511 merged by Cmcmahon: add BROWSER_LABEL to edit string per Bug 59011 https://gerrit.wikimedia.org/r/104511
Change 104531 merged by jenkins-bot: different topics per browser to prevent edit conflict Bug 59011 https://gerrit.wikimedia.org/r/104531
Tests fail if BROWSER_LABEL not defined. see my comment in Gerrit change #104511.
Change 104783 had a related patch set uploaded by Cmcmahon: Revert "add BROWSER_LABEL to edit string per Bug 59011" https://gerrit.wikimedia.org/r/104783
Change 104785 had a related patch set uploaded by Spage: Add some comments https://gerrit.wikimedia.org/r/104785
Change 104785 merged by jenkins-bot: Add some comments https://gerrit.wikimedia.org/r/104785
Change 104783 merged by Cmcmahon: Revert "add BROWSER_LABEL to edit string per Bug 59011" https://gerrit.wikimedia.org/r/104783
S / Chris: All four patches were merged a while ago - is there more work left to do here (if yes: please reset the bug report status to NEW or ASSIGNED), or can you close this ticket as RESOLVED FIXED?