Filip noticed that m_doubleConstantsMap should not use a double as a key, because it means two unfortunate things: - It will probably break for zero. - It will think that -0 is the same as +0 under some circumstances, size -0==+0 even though they are distinct values (for example 1/-0 != 1/+0). Will fix.
Created attachment 233260 [details] the patch. This patch has passed the JSC tests and the layout tests with a debug build on x86_64.
Are there ever cases where different NaNs would need to map to the same value?
Comment on attachment 233260 [details] the patch. r=me
(In reply to comment #2) > Are there ever cases where different NaNs would need to map to the same value? In user code, there are only ever two NaNs (positive unsignaling and negative unsignaling), and they do not need (or want) to map to the same value.
Comment on attachment 233260 [details] the patch. Clearing flags on attachment: 233260 Committed r170082: <http://trac.webkit.org/changeset/170082>
All reviewed patches have been landed. Closing bug.
Re-opened since this is blocked by bug 134006
Created attachment 233309 [details] patch 2: fixed build problem, and properly tested on 32-bit build.
Comment on attachment 233309 [details] patch 2: fixed build problem, and properly tested on 32-bit build. r=me
Thanks. Landed in r170109: <http://trac.webkit.org/r170109>.