Closed Bug 462103 Opened 16 years ago Closed 16 years ago

TM: We don't trace some variants of string + other type

Categories

(Core :: JavaScript Engine, defect, P1)

defect

Tracking

()

RESOLVED FIXED
mozilla1.9.1b2

People

(Reporter: gal, Assigned: brendan)

Details

Attachments

(1 file, 2 obsolete files)

Attached patch patch (obsolete) — Splinter Review
      No description provided.
Attachment #345241 - Flags: review?(brendan)
Assignee: general → gal
Status: NEW → ASSIGNED
OS: Mac OS X → All
Priority: -- → P1
Hardware: PC → All
Target Milestone: --- → mozilla1.9.1b2
I can land this if it's good -- let me know if I missed any nits. Interdiff to see what changed.

/be
Assignee: gal → brendan
Attachment #345244 - Flags: review+
Attachment #345244 - Flags: review?(gal)
Attachment #345244 - Flags: review?(gal) → review+
Comment on attachment 345244 [details] [diff] [review]
reorder booleanAtoms and typeAtoms, pick nits and perf buglets

Relying on booleanAtoms overflowing into typeAtoms is going to come back and bite you one day. And I will be there and say I told you so ;)
(In reply to comment #2)
> (From update of attachment 345244 [details] [diff] [review])
> Relying on booleanAtoms overflowing into typeAtoms is going to come back and
> bite you one day. And I will be there and say I told you so ;)

Only if you remove the static asserts already there, and ones I'm adding in the next patch -- which happen to cover your habit of hardcoding 2 for JSVAL_TO_BOOLEAN(JSVAL_VOID) :-P. Igor's static asserts save the day here.

/be
Attached patch patch to commitSplinter Review
Attachment #345241 - Attachment is obsolete: true
Attachment #345244 - Attachment is obsolete: true
Attachment #345246 - Flags: review+
Attachment #345241 - Flags: review?(brendan)
Fixed in tm:

http://hg.mozilla.org/tracemonkey/rev/fb35f0d7e0c0

Hope this makes the m-c train.

/be
Comment on attachment 345246 [details] [diff] [review]
patch to commit

>+++ b/js/src/jstracer.cpp
>@@ -1934,17 +1934,17 @@ TraceRecorder::checkType(jsval& v, uint8
>-        /* We sink i2f conversions into the side exit, but at the loop edge we have to make
>+        /* We sink i2f conversions into the side exit, but at the loop edge we have to
>            sure we promote back to double if at loop entry we want a double. */
Was this accidental?
Yes, sorry about that -- I am restoring the "make" in another patch.

/be
This is in m-c now.

/be
Status: ASSIGNED → RESOLVED
Closed: 16 years ago
Resolution: --- → FIXED
My patch. I should have closed it. I have no good way to track these though, so I take only partial blame.
added to js1_8_1/trace/trace-test.js in http://hg.mozilla.org/mozilla-central/rev/8155aba03537
Flags: in-testsuite+
Flags: in-litmus-
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Created:
Updated:
Size: