Comment #3 on issue 3040 by [email protected]: Truncated division
breaks on ia32 for certain inputs
http://code.google.com/p/v8/issues/detail?id=3040
The bug itself was really architecture-independent, only the symptom
depends on the platform. If the platform under consideration has the right
semantics (ARM? MIPS? Don't know...), you can still remove the overflow
check, but you have to do it on the Lithium->machine code level, not as a
Hydrogen "optimization". Note that this optimization was not part of the
fix.
The previous unit test actually didn't test what it was supposed to test,
because the division by -1 was lowered to a negation, which wouldn't
trigger a FP exception. Lesson learned: Look at the actual machine code of
your test...
--
You received this message because this project is configured to send all
issue notifications to this address.
You may adjust your notification preferences at:
https://code.google.com/hosting/settings
--
--
v8-dev mailing list
[email protected]
http://groups.google.com/group/v8-dev
---
You received this message because you are subscribed to the Google Groups "v8-dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
For more options, visit https://groups.google.com/groups/opt_out.
