Bug 164390 - REGRESSION (r208349) StringHasher::hashMemory behavior changed, causing API tests to fail
Summary: REGRESSION (r208349) StringHasher::hashMemory behavior changed, causing API t...
Status: RESOLVED FIXED
Alias: None
Product: WebKit
Classification: Unclassified
Component: Web Template Framework (show other bugs)
Version: WebKit Nightly Build
Hardware: Unspecified Unspecified
: P2 Normal
Assignee: Brady Eidson
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2016-11-03 17:10 PDT by Brady Eidson
Modified: 2016-11-04 11:05 PDT (History)
4 users (show)

See Also:


Attachments
Patch (2.78 KB, patch)
2016-11-03 17:12 PDT, Brady Eidson
no flags Details | Formatted Diff | Diff

Note You need to log in before you can comment on or make changes to this bug.
Description Brady Eidson 2016-11-03 17:10:16 PDT
REGRESSION (r208349) StringHasher::hashMemory behavior changed, causing API tests to fail

Link to the change: https://trac.webkit.org/changeset/208349

The difference is that memory hashing used to mask the top 8 bits of the hash, just like string hashing needed.
Now it no longer does.

We can either restore the masking behavior, or accept the change in behavior (as an improvement, I think) and update the test.

Note: There was a FIXME in hashMemory that called out how odd it was that we masked the top 8 bits for memory hashing when it wasn't needed like for string hashing.
Comment 1 Brady Eidson 2016-11-03 17:12:52 PDT
Created attachment 293830 [details]
Patch
Comment 2 Geoffrey Garen 2016-11-04 10:41:03 PDT
Comment on attachment 293830 [details]
Patch

Specifically, we don't see any reason that memory hashing needs to mask off the high bits of the hash code, even though string hashing does so in order to support an implementation of StringImpl that stores flags in the high bits of the hash code.
Comment 3 WebKit Commit Bot 2016-11-04 11:05:31 PDT
Comment on attachment 293830 [details]
Patch

Clearing flags on attachment: 293830

Committed r208387: <http://trac.webkit.org/changeset/208387>
Comment 4 WebKit Commit Bot 2016-11-04 11:05:35 PDT
All reviewed patches have been landed.  Closing bug.