IRC Log Viewer » #firebreath » 2011-05-05

IRC Nick Time (GMT-7) Message
FireBreathBot 03:05 JIRA issue http://jira.firebreath.org/browse/FIREBREATH-50 issue created by sylvain.chevalier
jshanab_wcw 06:05 taxilian. I was reading some of the discussion on the new fbgen, will the command line one remain? i would hate to be stopped on createing a new project becuase me 3G card is out again. (think SVN vs GIT :-) )
FireBreathBot 06:05 JIRA issue http://jira.firebreath.org/browse/FIREBREATH-11 issue commented by gmurphy "Just to chime in with my own experience, we ran into the same issue with IE9, the meta tag which ..."
taxilian 08:05 jshanab_wcw: Yes, the old one will remain, but I likely wont maintain it.. Someone else is welcome to
FireBreathBot 08:05 JIRA issue http://jira.firebreath.org/browse/FIREBREATH-50 issue commented by richard "Please show the c++ code you are using to fire the event: this is most likely something yu are do..."
JIRA issue http://jira.firebreath.org/browse/FIREBREATH-11 issue commented by richard "You shouldnt need the workaround in fb 1.4.3 and 1.5.* as this issue has been fixed"
amackera 09:05 I never thought I'd say this, but java applets are actually kind of cool
taxilian 09:05 hehe. all technologies can be cool if used correctly
amackera 09:05 Good point, the designers of any technology must have thought it would be cool / useful
nobody intentionally sets out to make a bad piece of software :P
taxilian 09:05 hehe =] the problem you usually have is that people get excited about the latest "silver bullet" software/technology and try to use it for everything
including things that it just doesn't do well
FireBreathBot 10:05 JIRA issue http://jira.firebreath.org/browse/FIREBREATH-50 issue commented by cstevens "Each element of a variant list passed to FireEvent() is passed as a unique named argument to the ..."
taxilian 10:05 !wiki events
FireBreathBot 10:05 8 results found. Note: Results limited to 8
"Firing Events from JSAPI": http://goo.gl/9LGTk
"file PluginEventSource.cpp": http://goo.gl/xVPob
"file MacEventCarbon.h": http://goo.gl/cP4Dw
"file PluginEventSink.h": http://goo.gl/wTeri
"file MacEventCocoa.h": http://goo.gl/Bn8B7
"file PluginEventSource.h": http://goo.gl/rBxHY
"class FB PluginEventSink shared_ptr": http://goo.gl/Ojh6t
"class FB PluginEventSource shared_ptr": http://goo.gl/F6bwM
JIRA issue http://jira.firebreath.org/browse/FIREBREATH-50 issue commented by richard "Well said, Clay. Even better is to use the new typed events, like this:
JIRA issue http://jira.firebreath.org/browse/FIREBREATH-50 issue commented by richard "incidently, documentation for firing events in this manner are here: http://www.firebreath.org/di..."
taxilian 13:05 stuartmorgan: so aparently the CALayer retention problem with Safari gets even better… if Safari is running in 64 bit mode, things work according to spec. In 32 bit mode, they work the way Safari originally worked… and thus we get a memory leak in 64 bit safari
our workaround for now is to set it to only do the retain if the currently running process is Safari (which only happens in 32 bit mode on a 32 bit plugin, since 64 bit safari is out-of-process)
stuartmorgan 13:05 awesome
Hm, but you'll crash in other WebKit browsers, won't you?
e.g., iCab or OmniWeb
taxilian 13:05 possible
I'm open to other ideas =] but since the version is the same and our plugin is 32 bit (in this case) I don't know another way to detect it
stuartmorgan 13:05 Hmm... you could do the inverse, and check to see if you are running in the plugin host bundle
That's not future-proof
But at least the penalty for being wrong is a leak instead of a crash...
taxilian 13:05 hmm. something to think about, definitely. supporting lesser-known browsers isn't top of my priority list, but at the same time I'd prefer to support them if it's not going to require a great deal of effort
stuartmorgan 13:05 Oh, but if you reverse the logic you'd have to special-case all the browsers that are right, like Firefox and Chrome
So that would be ugly
taxilian 13:05 :-/
stuartmorgan 13:05 Make it crash in all non-compliant cases and then release a plugin popular enough Apple can't ignore the problem, and fixes WebKit? ;)
taxilian 13:05 lol
stuartmorgan 13:05 The only other thing I can think of is something really ugly with a global map, timers, and inspection of retain counts (which you are never actually supposed to do)
And that's probably worse than anything else
taxilian 13:05 lol. yeah, probably so
I think I'll leave it with this fix until someone complains, at least
FireBreathBot 20:05 JIRA issue http://jira.firebreath.org/browse/FIREBREATH-50 issue commented by sylvain.chevalier "Thank you Richard and Clay, I misunderstood the way to use it and did not realize that elements o..."