Closed
Bug 1005442
Opened 11 years ago
Closed 11 years ago
memory reporter diff gets confused by null principals
Categories
(Toolkit :: about:memory, defect)
Toolkit
about:memory
Tracking
()
RESOLVED
FIXED
mozilla32
People
(Reporter: bholley, Assigned: n.nethercote)
Details
(Whiteboard: [MemShrink:P2])
Attachments
(1 file)
|
5.42 KB,
patch
|
mccr8
:
review+
|
Details | Diff | Splinter Review |
(Not sure which component this belongs to - njn, can you reclassify?)
I was getting very confused over a diff between two memory reports, because it showed this:
│ │ ├──3.61 MB (25.62%) ++ (226 tiny)
│ │ └──1.17 MB (08.31%) ++ compartment(moz-nullprincipal:{09969844-1311-494d-8b90-6a50783f523b}, XPConnect Compilation Compartment)
And I was very surprised that the Compilation Compartment grew by so much. As it turned out, when I expanded "tiny", I found:
-1.08 MB (-7.70%) ++ compartment(moz-nullprincipal:{72419579-0050-4249-9dcb-e4ab09a7f156}, XPConnect Compilation Compartment)
Which makes a lot more sense. But this was very confusing. Given that null principals are generated with a random UUID which varies between runs, maybe we should avoid including those in the memory report, at least in the case where the compartment also has a name?
| Reporter | ||
Updated•11 years ago
|
Flags: needinfo?(n.nethercote)
| Assignee | ||
Comment 1•11 years ago
|
||
The diff tool currently anonymizes certain random things such as addresses and PIDs. Would doing the same for UUIDs fix this problem?
Flags: needinfo?(n.nethercote)
Comment 2•11 years ago
|
||
Also, I think tiny shouldn't include large negative values.
Component: General → about:memory
Product: Core → Toolkit
| Reporter | ||
Comment 3•11 years ago
|
||
(In reply to Nicholas Nethercote [:njn] from comment #1)
> The diff tool currently anonymizes certain random things such as addresses
> and PIDs. Would doing the same for UUIDs fix this problem?
Yeah. Though would that cause us to lose information about addons, which use UUIDs that are not randomized at runtime?
(In reply to Andrew McCreight [:mccr8] from comment #2)
> Also, I think tiny shouldn't include large negative values.
I agree.
Comment 4•11 years ago
|
||
I filed bug 1005489 for that.
Updated•11 years ago
|
OS: Mac OS X → All
Hardware: x86 → All
Whiteboard: [MemShrink]
| Assignee | ||
Updated•11 years ago
|
Whiteboard: [MemShrink] → [MemShrink:P2]
| Assignee | ||
Comment 5•11 years ago
|
||
This patch strips the UUIDs just from null principals, thus avoiding add-on
UUIDs.
Attachment #8422146 -
Flags: review?(continuation)
| Assignee | ||
Updated•11 years ago
|
Assignee: nobody → n.nethercote
Status: NEW → ASSIGNED
Comment 6•11 years ago
|
||
Comment on attachment 8422146 [details] [diff] [review]
Strip moz-nullprincipal UUIDs when diffing memory reports
Review of attachment 8422146 [details] [diff] [review]:
-----------------------------------------------------------------
woot
Attachment #8422146 -
Flags: review?(continuation) → review+
| Assignee | ||
Comment 7•11 years ago
|
||
Well that was fast! :)
https://hg.mozilla.org/integration/mozilla-inbound/rev/83b7b82834f0
Comment 8•11 years ago
|
||
Status: ASSIGNED → RESOLVED
Closed: 11 years ago
Resolution: --- → FIXED
Target Milestone: --- → mozilla32
You need to log in
before you can comment on or make changes to this bug.
Description
•