Add a few initiator tests to resource timing.
Created attachment 279559 [details] Patch
This adds some tests that make sure initiator info doesn't get overridden when later requests try to store RT info for the same resource. Currently implementation relies on the fact that HashMap avoids adding entries for keys that are already in the map. Is that part of the semantics of HashMap and can be relied on, or is it necessary to add a check before attempts to add initiator info for a resource?
HashMap::add does nothing if the key is already in the map, even if you're trying to add it with a different value. HashMap::set will overwrite an existing value for the given key. This behavior won't change.
(In reply to comment #3) > HashMap::add does nothing if the key is already in the map, even if you're > trying to add it with a different value. HashMap::set will overwrite an > existing value for the given key. This behavior won't change. cool, didn't know that. In that case, no need to add further checks
Comment on attachment 279559 [details] Patch Thanks for reviewing! :)
Comment on attachment 279559 [details] Patch Clearing flags on attachment: 279559 Committed r201323: <http://trac.webkit.org/changeset/201323>
All reviewed patches have been landed. Closing bug.
Reverted r201323 for reason: One of the LayoutTests added with this change fails on all platforms. Rolling out until ResourceTiming flakiness fix is landed Committed r201343: <http://trac.webkit.org/changeset/201343>