| commit | f5abe3d780c70eccc20bd2c41b0670b750079ff7 | |
| author | Madelaine Boyd <madelaine@fb.com> | |
| Wed, 26 Oct 2016 13:01:51 +0000 (26 06:01 -0700) | ||
| committer | Hhvm Bot <hhvm-bot-bot@fb.com> | |
| Wed, 26 Oct 2016 13:06:15 +0000 (26 06:06 -0700) | ||
| tree | 8ab4e6ce909445facdf427989ea4076e372365da | treesnapshot (tar.gz zip) |
| parent | b73bd4539f4efa28d710c5bf87193145362fdaa5 | commitdiff |
Special-case $x**3
Summary:
Problem - I pop a Pow and push two MulIntOs. Diff as it stands causes a stack corruption in the updated unit test.
What is the proper way to exit from here?
Realizing now that I need to handle the possibility of (LARGE_INT**3),
where (LARGE_INT*LARGE_INT)*LARGE_INT--> the first multiplication can overflow to double. I would need to check in the second multiplication pass
whether this calls for MulIntO or MulDbl.
Reviewed By: ricklavoie
Differential Revision: D4063317
fbshipit-source-id: c6b5edbb598f9c4b3fcb52f4baaaf2e903e63eec
Summary:
Problem - I pop a Pow and push two MulIntOs. Diff as it stands causes a stack corruption in the updated unit test.
What is the proper way to exit from here?
Realizing now that I need to handle the possibility of (LARGE_INT**3),
where (LARGE_INT*LARGE_INT)*LARGE_INT--> the first multiplication can overflow to double. I would need to check in the second multiplication pass
whether this calls for MulIntO or MulDbl.
Reviewed By: ricklavoie
Differential Revision: D4063317
fbshipit-source-id: c6b5edbb598f9c4b3fcb52f4baaaf2e903e63eec