Closed Bug 839918 Opened 13 years ago Closed 2 years ago

Huge CPU usage after seeing a lot of consequent data-URI images

Categories

(Core :: General, defect)

defect

Tracking

()

RESOLVED INCOMPLETE

People

(Reporter: mounir, Unassigned)

Details

I've been debugging some reftests so I was copy-pasting the results and expected results data-URI's to my Firefox instance. Those data-URI are screenshots of Firefox windows (I'm not sure of the size). After that debugging session, and closing all the tabs, my Firefox was very slow and taking more memory than it was before. Given that I would bet it's not the first time this happens, I thought it would be worth opening a bug. I could experiment more if you guys want.
If you can get an about:memory dump and a profiler log from when firefox was going slowly and using more memory than you expected, that would let us dig in here...
Mounir, can you provide us with any info here? Otherwise we should just close this.
Flags: needinfo?(mounir)
Whiteboard: [MemShrink]
I tried to reproduce this by copy-pasting a few URIs but I didn't find anything interesting but RyanVM had the same problem, I've heard. FWIW, this is a problem I had a few times. The only hint I've found is that SessionRestore is taking quite some memory to keep track of the data URIs. I'm not sure though if that could have a real impact.
Flags: needinfo?(mounir)
Yeah, I hit something like this last week. Mochitest failures print a data URI to the log with a screenshot of the builder. I looked at about:memory at the end of the day and saw ~100MB worth of them even though they were all from earlier in the morning. If I can reproduce it again, I'll post an about:memory here.
Severity: normal → S3
Status: NEW → RESOLVED
Closed: 2 years ago
Resolution: --- → INCOMPLETE
You need to log in before you can comment on or make changes to this bug.