111333 - leaks up 40 bytes after landing for bug #110850

Status: RESOLVED FIXED

(SeaMonkey :: General, defect)

Description

[(not reading, please use [e-mail] instead)]

leaks up 40k after landing for bug #xxxxx

we're investigating.

Comment 1

[(not reading, please use [e-mail] instead)]

Attachment: bloat log, before

Comment 2

[(not reading, please use [e-mail] instead)]

Attachment: bloat log, after

Comment 3

[David :Bienvenu]

Either the history ab isn't getting closed (perhaps because shutdown
notifications dont' work reliably anymore), or I screwed up the leak stats for
mork objects

Comment 4

[(not reading, please use [e-mail] instead)]

one thing that shows up is that we're leaking 40 morkObjects, but those appear 
to be only 40 bytes per instance.

Comment 5

[(not reading, please use [e-mail] instead)]

40K, no.  40B yes.  I'm a moron, who can't read a leak log if his life depended 
on it.

keeping this open, as we appear to be leaking a morkObject.

this may not be a new leak, it's just that david's recent fixes now make it so 
we report those leaks.

Comment 6

[David :Bienvenu]

believe me, we leaked a lot more than that before. Purify doesn't show any
morkObject leaks, so I'm a little unsure who to trust. It might be a global
object...

Comment 7

[Alec Flett]

I don't think the bloat tests even start mail, so this might be a global history
leak.

Comment 8

[David :Bienvenu]

yes, it's global history, and it's almost certainly the global factory. Anyway,
I'm looking at it, but it's not a high priority since actual leaks have gone way
down.

Comment 9

[Alec Flett]

Attachment: fix global history leak

how exciting. turns out it was pretty darn easy. we were just leaking an
nsIMdbFactory

can I get sr=bienvenu, r=sspitzer?

Comment 10

[Alec Flett]

assigning to ME since I fixed it :)

Comment 11

[Alec Flett]

hmm.. now I'm getting an assertion on shutdown with this leak. I'm not sure if
its really me or not though.

Comment 12

[David :Bienvenu]

Famous last words ... not sure if it's really me... I'll try the patch and see
if I see assertions.

Comment 13

[Alec Flett]

actually, see bug 112058 for a more complete patch that cleans up lots of
non-mork leaks.

Comment 14

[David :Bienvenu]

Hey, Alec, I'm not seeing assertions - what are the assertions? Assuming they're
not mork assertions, r=bienvenu on the patch.

Comment 15

[David :Bienvenu]

Comment on attachment 59227 [details] [diff] [review]
fix global history leak

r=bienvenu,assuming assertions aren't new.

Comment 16

[David :Bienvenu]

OK, I see the assertion if I just start up with the browser and shutdown. I'll
attach a patch for that assertion.

Comment 17

[David :Bienvenu]

Attachment: fix for assertions.

Comment 18

[David :Bienvenu]

Alec, the assertion is caused by your change, but the attached patch should fix
it. Do you want to try it and check it in with your patch?

Comment 19

[(not reading, please use [e-mail] instead)]

Comment on attachment 59266 [details] [diff] [review]
fix for assertions.

sr=sspitzer

Comment 20

[(not reading, please use [e-mail] instead)]

Comment on attachment 59227 [details] [diff] [review]
fix global history leak

sr=sspitzer

Comment 21

[Alec Flett]

Comment on attachment 59266 [details] [diff] [review]
fix for assertions.

awesome!
r=alecf and I'm checking both in.

Comment 22

[Alec Flett]

thanks guys, fix is in.