Closed Bug 969210 Opened 10 years ago Closed 10 years ago

make test-perf fail to collect mem stats

Categories

(Firefox OS Graveyard :: Gaia::PerformanceTest, defect, P1)

x86_64
Linux
defect

Tracking

(b2g-v1.3T fixed)

RESOLVED FIXED
Tracking Status
b2g-v1.3T --- fixed

People

(Reporter: hub, Assigned: hub)

References

Details

(Keywords: perf, Whiteboard: [c=automation p=1 s= u=])

Attachments

(1 file)

46 bytes, text/x-github-pull-request
julienw
: review+
Details | Review
Since very recently, collecting memory stats on buri/hamachi files.

b2g-info output starts with "Expected to read 4 fields from fscanf(/proc/2942//stat), but got 2." which is not expected. I can't figure why but we could make the meminfo.js code deal with it better.
Assignee: nobody → hub
Keywords: perf
Whiteboard: [c=automation p=1 s= u=]
Ok, this is caused by not having this revision of gonk-misc on my buri:

ac1b69cc36ce8cf64c8ce585826e7a940b349072

I think the simple solution is to actually update it on my device.
Attached file Pull request
I think that even though I need to update b2g-info on my device, we can still have an error message pop out. Make this more robust.
Attachment #8372054 - Flags: review?(felash)
Comment on attachment 8372054 [details] [review]
Pull request

I agree with the reasoning, but I think we can do a more reliable function. See github :)
Attachment #8372054 - Flags: review?(felash)
Priority: -- → P1
See Also: → 971657
Comment on attachment 8372054 [details] [review]
Pull request

Updated the pull request.
Attachment #8372054 - Flags: review?(felash)
Component: Gaia → Gaia::PerformanceTest
Comment on attachment 8372054 [details] [review]
Pull request

r=me with nits for the commit "Rewrite the parser to be more permissive and foresee more changes."
Attachment #8372054 - Flags: review?(felash) → review+
Status: NEW → RESOLVED
Closed: 10 years ago
Resolution: --- → FIXED
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Created:
Updated:
Size: