Last modified: 2014-07-01 10:18:50 UTC
The siteinfo API uses an empty string to indicate presence of an attribute. The "protorel" property was being set to 'true' instead. Change it to '' to be consistent.
Change 140132 had a related patch set uploaded by Cscott: The siteinfo API uses '' as an attribute value, not `true`. https://gerrit.wikimedia.org/r/140132
Change 140130 had a related patch set uploaded by Cscott: The siteinfo API uses '' as a value to indicate true. https://gerrit.wikimedia.org/r/140130
Change 140132 merged by jenkins-bot: The siteinfo API uses '' as an attribute value, not `true`. https://gerrit.wikimedia.org/r/140132
Not fixed until the corresponding Parsoid patch is merged ( https://gerrit.wikimedia.org/r/140130 ).
Change 140130 merged by jenkins-bot: The siteinfo API uses '' as a value to indicate true. https://gerrit.wikimedia.org/r/140130
(In reply to C. Scott Ananian from comment #4) > Not fixed until the corresponding Parsoid patch is merged This bug is filed against the API, which is fixed, and says nothing about Parsoid. There's no reason that Parsoid not being fixed should hold an API bug open. The issue in Parsoid should probably have its own bug depending on this one. Although the point is moot now since I see your Parsoid patch has been +2ed.
Change 142444 had a related patch set uploaded by Cscott: Fix parserTest setup: siteinfo API returns '' to indicate true. https://gerrit.wikimedia.org/r/142444
Change 142444 merged by jenkins-bot: Fix parserTest setup: siteinfo API returns '' to indicate true. https://gerrit.wikimedia.org/r/142444
Sorry, Brad, when I file a bug, and it is assigned to me, and the issue applies to both Parsoid and core I usually use the same bug # for both components and try to arrange that the patches be merged together. I don't like for the bug to disappear from my bugzilla radar until that has happened. I probably should have changed the component of the bug over to Parsoid after the core patch landed. Or opened a new bug, but that actually obscures the connection between the parsoid and core changes.
(In reply to C. Scott Ananian from comment #9) > I probably should have changed the component of the bug over to Parsoid > after the core patch landed. That would've worked for me, although it would mean that someone trying to do statistics on API bugs wouldn't find it.
(In reply to C. Scott Ananian from comment #9) > Or opened a new bug, but that actually > obscures the connection between the parsoid and core changes. That's what originally the "Blocks" and "Depends on" fields were for in Bugzilla (but the [ab]use in this Bugzilla is pretty broad).
