[Webkit-unassigned] [Bug 58122] chromium rebaseline script did the wrong thing with snowleopard expectations

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=58122


Dirk Pranke <dpranke at chromium.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |enne at google.com,
                   |                            |koz at chromium.org,
                   |                            |mihaip at chromium.org




--- Comment #1 from Dirk Pranke <dpranke at chromium.org>  2011-04-08 09:27:17 PST ---
Hm. So the chromium-mac directory actually contained baselines that would pass on Leopard but fail on SL? That certainly seems plausible, at least when introducing new o/s versions.

Unfortunately, fixing this so that the script is smart enough to realize that a baseline should move *down* the search path rather than up (or be deleted) requires the code to have a more holistic point of view than it currently does (since it basically only does a port at a time).

I wonder if there's a way to change the code to update/move/delete files in two passes, or if it's just best to leave this as an edge case and hope the new tool koz is working on will deal with this.

It seems like the "workaround" would be to just re-run the tool after the bots had cycled again, at which point SL would pass but Leopard would fail, and the tool could rebaseline the Leopard bots to work properly.

Also adding Adrienne to this, as she might've seen a similar problem a couple days ago.

-- 
Configure bugmail: https://bugs.webkit.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.