Closed Bug 477052 Opened 13 years ago Closed 13 years ago

Make some nanojit classes initialize their members

Categories

(Core :: JavaScript Engine, defect)

Product:
Core
Component:
JavaScript Engine
Version:
Other Branch
Type:
defect
Priority:
Not set
Severity:
normal

Tracking

()

Status:
RESOLVED FIXED

People

(Reporter: jorendorff, Assigned: jorendorff)

Details

(Keywords: fixed1.9.1)

Attachments

(1 file)

v1
3.71 KB, patch
gal
: review+
edwsmith
: review+
Details | Diff | Splinter Review 
We go through the silliness of allocating our nanojit objects with an avmplus::GC, but the only reason for that is that LirBuffer's constructor (and maybe others) doesn't initialize its members.  All the GC does is zero them.

Low priority but I find this embarrassing.
Second this, IIRC Adobe wanted it to be optional so we'd have to #ifdef it.
Attached patch v1 — — Splinter Review 
The "#ifndef __MMgc__" here is because of comment 1.  I didn't bother in the case of LirBuffer or Fragmento, as I think they are long-lived and so the performance of the constructor doesn't matter.
Assignee: general → jorendorff
Status: NEW → ASSIGNED
Attachment #361564 - Flags: review?(gal)
Attachment #361564 - Flags: review?(edwsmith)
Attachment #361564 - Flags: review?(edwsmith) → review+
Comment on attachment 361564 [details] [diff] [review]
v1

you're right that it's both embarrasing and low priority.  in terms of code size there are plenty other places we can reduce before size of initializers really matters.  (cough) macros.  no #ifdef necessary.
Attachment #361564 - Flags: review?(gal) → review+
Pushed to the Tracemonkey branch.

http://hg.mozilla.org/tracemonkey/rev/780189ed095c
Status: ASSIGNED → RESOLVED
Closed: 13 years ago
Resolution: --- → FIXED
Flags: in-testsuite-
Flags: in-litmus-
You need to log in before you can comment on or make changes to this bug.