RFC: Backport fix for Sun Bug 6539464 (Math.log() produces inconsistent results)

[Andrew Haley]

Andrew John Hughes wrote:
> 2009/5/12 Andrew Haley <aph at redhat.com>:
>> Omair Majid wrote:
>>> Hi,
>>>
>>> There is an issue in HS14 on x86_64 where calls to Math.log() can return
>>> different values for the same input value. This patch backports the fix
>>> [1] from OpenJDK7 (b56) to IcedTea6
>>>
>>> Sun's bug entry:
>>> http://bugs.sun.com/view_bug.do?bug_id=6539464
>>> Red Hat's bugzilla entry:
>>> https://bugzilla.redhat.com/show_bug.cgi?id=500077
>>>
>>> ChangeLog:
>>> 2009-05-11  Omair Majid  <omajid at redhat.com>
>>>
>>>     * patches/hotspot/default/icedtea-6539464.patch: New file. Backport
>>>       fix for Sun Bug 6539464 from OpenJDK7
>>>     * Makefile.am (ICEDTEA_PATCHES): Apply the above.
>>>     * HACKING: Document the above.
>> Well caught!  Please push this upstream straight away.  I imagine it'll be
>> approved immediately.

> Except there's still no upstream hs14 to push to...

Well, yeah: tell us something we _don't_ know!  No reason not to submit
it upstream for approval.

Andrew.