[Webkit-unassigned] [Bug 153438] LayoutTest imported/w3c/indexeddb/keyorder-private.html flaky on mac-wk1

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=153438

--- Comment #9 from Brady Eidson <beidson at apple.com> ---
The plot thickens...

The MemoryObjectStore keeps both a HashMap of keys-to-values, and a std::map of ordered keys.

Every time we insert a record, we record the key-value-mapping in the HashMap, and insert the key into the std::map for ordered-ness.

The bug is happening because the HashMap *thinks* it has a record for this particular IDBKey. But the std::map does *NOT* have a record for the key.

Lots of printf-style debugging confirms that we've never tried to put this key in either the HashMap or std::map before, nor have we tried to put anything with its hash in those maps before...

Did this test discover a bug in our HashMap?

-- 
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.webkit.org/pipermail/webkit-unassigned/attachments/20160204/2a7e1d14/attachment-0001.html>