671150 - Add memory reporter for the JS PropertyTable

Status: RESOLVED FIXED

(Core :: JavaScript Engine, defect)

Description

[Nicholas Nethercote [inactive]]

Bug 563700 says that js::PropertyTable::init() and js::PropertyTable::change() together account for 4.3% in one profile of gmail.

Comment 1

[Nicholas Nethercote [inactive]]

Attachment: patch

Fairly straightforward.

Comment 2

[Nicholas Nethercote [inactive]]

I'm seeing this reduce heap-unclassified by about 2% on normal (for me) browsing.  Not as much as I'd hoped, but better than a kick in the teeth.

Comment 3

[Bill McCloskey [inactive unless it's an emergency] (:billm)]

Maybe this should go in a different bug, but it would be nice also to count the size of the hashtable holding each shape's children.

Comment 4

[Nicholas Nethercote [inactive]]

(In reply to comment #3)
> Maybe this should go in a different bug, but it would be nice also to count
> the size of the hashtable holding each shape's children.

Just to clarify (for my own benefit):  Shape::kids can point to a KidsHash (which is a typedef for a HashSet, and that's not accounted for.

Comment 5

[Brendan Eich [:brendan]]

Comment on attachment 547331 [details] [diff] [review]
patch

entriesSize vs. sizeOf name nit: how about sizeOfEntries instead of entriesSize? More self-consistent, prettier because it avoids the sS conjunction.

r=me with that.

/be

Comment 6

[Nicholas Nethercote [inactive]]

http://hg.mozilla.org/integration/mozilla-inbound/rev/2d63d7fd35a0

Comment 7

[Marco Bonardo [:mak] (Away Apr 25 - May 5)]

http://hg.mozilla.org/mozilla-central/rev/2d63d7fd35a0