IRC Log Viewer » #firebreath » 2012-07-04

IRC Nick Time (GMT-7) Message
dougma 05:07 style.display = "none" | "inline" toggle doesn't re-trigger it
will try moving it in the dom...
will try style.visibility too :)
"hidden" | "visible" toggle on style.visibility doesn't do it either....
i know firefox does much more drastic things when toggling them
but seems ie9 doesn't do much at all!
dougma 05:07 yep, moving in the dom does it.
dougma 07:07 i've added a guard to prevent setReady being multiply invoked
but now the problem i see is that after moving in the dom it throws a bad_weak_ptr when trying to fire an event.
the event is scheduled on main thread ok, then exception thrown from IDispatchAPI:getIDispatch
dougma 07:07 of course that problem may pre-date my fiddling, rolling-back to test that.
dougma 07:07 indeed it does pre-date my fb-177 work, i'll open a new ticket. :(
argh... typing xcode shortcuts at visualstudio!
ch 07:07 haha :)
dougma 08:07 i suppose the problem is that event handler is attached... m_host is cached (albeit weakly) in FB::ActiveX::IDispatchAPI, plugin's host changes when plugin moved in the dom... event handler has the old host still.
this looks suspiciously like fb-181 actually.
FireBreathBot 08:07 Commit add8b37 on FB-177 by dougma: "FB-177 guard against multiple setReady calls"
FireBreathBot 09:07 Commit eb2ff80 on FB-181 by dougma: "FB-181 stop the crash"
ominds 13:07 hello guys, is this safe for handling writing of UTF8?
http://pastebin.mozilla.org/1690998
vim says file is all good, utf-8 encoding and has BOM
I wonder if it's portable to when I compile on windows
dougma 18:07 taxilian: oops... now i understand why FireBreathBot noticed my commits. Using Github for Windows I didn't realise I was branching and pushing to the FireBreath/FireBreath repo rather than dougma/FireBreath. I will be more careful in future!
taxilian 19:07 I was wondering about that =]
dougma 23:07 pretty sure fb-181 is not just from a memory leak
move object in the dom, fire event and you get that call-stack
taxilian 23:07 hmmm
interesting
we need to track it down; your hack does prevent a crash, but it hides a deeper problem
I'll have to look at it tomorrow, I guess
wait
move object in the DOM?
yeah, that would probably do it
dougma 23:07 i know.. but not crashing is a good thing i think. :)
taxilian 23:07 particularly in conjunction with the other InPlaceActivate stuff we're looking at
it's all related
true
=]
I'm for bed now
dougma 23:07 well i was testing inplaceactivate when i discovered the same crash as fb-181
taxilian 23:07 I just uploaded files to a read only filesystem on my arduino device, so I feel like I got something done today
dougma 23:07 but they are independent bugs. :)
taxilian 23:07 they are, but not unrelated
dougma 23:07 i just made a boost ios framework :)
taxilian 23:07 hehe
useful
dougma 23:07 maybe this time it will link happily too
taxilian 23:07 I have a bad habit of simply including firebreath-boost in my projects and using the header-only stuff
good luck
and good night (day =])