BaselineCompiler: Compile iteration opcodes

RESOLVED FIXED

Status

()

defect
RESOLVED FIXED
6 years ago
6 years ago

People

(Reporter: bhackett, Assigned: bhackett)

Tracking

Other Branch
x86
macOS
Points:
---

Firefox Tracking Flags

(Not tracked)

Details

Attachments

(1 attachment)

(Assignee)

Description

6 years ago
Iteration opcodes --- JSOP_ITER, JSOP_ITERNEXT, JSOP_ENDITER, JSOP_MOREITER --- are compiled by JM but not by Baseline.  Initially these can all just be stubbed, though caches for at least ITERNEXT and MOREITER will probably be needed before long.
(Assignee)

Updated

6 years ago
Assignee: general → bhackett1024
(Assignee)

Comment 1

6 years ago
Posted patch patchSplinter Review
Attachment #711479 - Flags: review?(kvijayan)
Attachment #711479 - Flags: review?(kvijayan) → review+
(Assignee)

Comment 2

6 years ago
https://hg.mozilla.org/projects/ionmonkey/rev/92621f184e4d
Status: NEW → RESOLVED
Last Resolved: 6 years ago
Resolution: --- → FIXED
(In reply to Brian Hackett (:bhackett) from comment #0)
> Iteration opcodes --- JSOP_ITER, JSOP_ITERNEXT, JSOP_ENDITER, JSOP_MOREITER
> --- are compiled by JM but not by Baseline.  Initially these can all just be
> stubbed, though caches for at least ITERNEXT and MOREITER will probably be
> needed before long.

Good prediction.  We're 10x slower on string-fasta with BL+Ion compared to JM+Ion and this seems to be responsible for a significant chunk of that.
You need to log in before you can comment on or make changes to this bug.