IRC Log Viewer » #firebreath » 2011-01-17

IRC Nick Time (GMT-7) Message
jerome_ 00:01 hello all, i have some question.
if you read this, please e-mail to me
i need to develop use firebreath
before, we used ATL + NPAPI custom framework
and used wxwidgets.
so, we compile firebreath, and wxWidgets code in there
but wxThread ( internally called creatthread(win32) ) is terminate when i create that
probalby, the problem is SEH
i don't understand why. because before framework ( ATL + NPAPI custom framework ) is works fine.
anyways, when i call directly the createthread on firebreath that is worksfine.
but wxThread has problem. if you know why or has problem in firebreath please reply to me
my email is [email protected]
men 02:01 hi
iaincollins 08:01 taxilian: thanks again for ... very straightfoward
iaincollins 08:01 any reason (other than an unrelated bug in my code, say) why I wouldn't always see a call back triggered?
In FB 1.3.2 don't always get an event back, maybe 1 time out of 10. Could be bogus error on my part (investigating), just curious if there are any issues to be aware of.
taxilian 08:01 iaincollins: Dont remember the timing, but there was an issue that could have caused that... Thout it was fixed in the 1.3.2 release, though
I would try it with 1.4... The breaking changes are pretty minor
iaincollins 08:01 kk thanks
neilg_ 08:01 Hey all. I'm trying not to lurk in here so much this week. :)
taxilian 09:01 Note that there is currently a crash on Chrome... I should have it fixed sometime etoday
neilg_ 09:01 Though at least I've been helping on the mailing list! lol
taxilian 09:01 As soon as I figure out what is causing it :)
Lol. Thanks for that neilg_ :)
So why are you trying to kick the FireBreath IRC habit?
neilg_ 09:01 What, of lurking? :)
taxilian 09:01 Actually I meant thanks for being helpful on the maiilng l ist
Or did yiu mean youre going to try to be more active instead of lurking?
neilg_ 09:01 I'm going to try and be more active. :)
taxilian 09:01 Sounds good to me :)
neilg_ 09:01 I've pretty much been ignoring the channel but since I have a fair amount of knowledge in the area now...
I'm just better at mailing lists because it's in my face! lol
taxilian 09:01 Hehe. I was happy to see so many responses to that entleman on the list who ask how active the community is
neilg_ 09:01 Yes, the irony wasn't missed by me! :) It was nice to see so many people reply. I feel bad because you're always doing so much to help people out so...
taxilian 09:01 Many have suggested a forum instead of just the google group, but I dont know if I would keep up with it
I have been getting a bit overwhelmed, So I appreciate it
neilg_ 09:01 A forum with a mail gateway might be a good idea for people like me that tend to leave things open and forget about them... But I always check my email!
taxilian 09:01 Yeah... Maybe Ill try to find something like that. A hybrid
Not real familiar with the forum software out there
neilg_ 09:01 The only one I have a lot of experience with isn't free - which is vBulletin. And I'm not sure whether that has an email gateway or not though it seems likely. But I'm only familiar with the concept because of libsdl which has been using one for a few months now
taxilian 09:01 Hmm. Something to think about, I guess
i need to set up Jira sometime soon
The google code issues are just getting annoying
Need everything more integrated
neilg_ 09:01 Sure, I can see why that would be useful. launchpad maybe?
taxilian 09:01 Be back in a bit
neilg_ 09:01 Have fun!
taxilian 09:01 I think the confluence/jira combination will work pretty well
just need to find time to set it all up
taxilian 09:01 lol
iaincollins 09:01 Oh did I see people talking about passing messages between plug-ins and other Windows applications? (here or on the mailing list)
taxilian 09:01 probably
iaincollins 09:01 would be interested if anyone is doing that, have plenty of "Plan B's" in case there are issues there in some browsers, but if there are any tips would be good to know.
taxilian 09:01 what do you need to accomplish?
iaincollins 09:01 I am thinknig about having an option from a Windows Service (with admin rights) to be able to let a plug-in known an event has occured
taxilian 09:01 ahh. well, the way I'd probably start is to look into boost IPC
iaincollins 09:01 oh boost has an IPC thing? cool
taxilian 09:01 I've never actually used to, however. I *have* used named pipes for interprocess communication, and I know that works
but if boost IPC works with a low integrity process that's probably a lot more powerful
iaincollins 09:01 yeah, have used that with C# (calling unmanaged code)
but yeah, thanks, can imagine if boost has it it's likely a pretty good implementation
neilg_ 09:01 I've actually had bad experiences with boost's IPC on Windows - it just doesn't work very well
And hasn't been maintained on Windows for several versions
taxilian 09:01 really? that's too bad
neilg_ 09:01 :(
taxilian 09:01 what version of boost was that?
neilg_ 09:01 1.44
taxilian 09:01 hmm. don't remember any notes about it in 1.45
that's really too bad
neilg_ 09:01 They've made minor changes in 1.45 but nobody seems to be supporting it well for Windows
iaincollins 09:01 neilg_: oh :( ..hmm thanks for the heads up
neilg_ 10:01 In any case IPC doesn't seem to work well on low-integrity processes in any case. A named pipe will work though. We actually got around it using sockets on localhost which is allowed.
taxilian 10:01 this Chrome thing is so weird… it's not actually crashing. the plugin does something that Chrome doesn't like and chrome shuts it down
neilg_ 10:01 What Chrome thing is that?
taxilian 10:01 oh, 1.4 beta 1 doesn't work on Chrome in Windows
nothing serious, right?
iaincollins 10:01 I had Chrome doing that to my plug-in when I did something that took too long to respond
taxilian 10:01 no, this *always* does it
iaincollins 10:01 (not actually crashing, just blocking)
taxilian 10:01 kinda un-optimial
un-optimal even
iaincollins 10:01 *nod*, just noting it has special behaviour for watching plug-ins
taxilian 10:01 yeah
neilg_ 10:01 But it only does it with FB 1.4 beta 1 and is fine with other versions of FB?
taxilian 10:01 1.3.2 worked fine
actually, I think I just found it
so I just need to figure out what is causing it
hurray, it works!
iaincollins 10:01 \o/
What was it? ... or do you still not know? :)
taxilian 10:01 aparently if you make a call to the DOM too early it confuses poor chrome
iaincollins 10:01 Oh, weird, and that's *after* the plug-in is loaded? Buggy :/
taxilian 10:01 yep
FB_GitHubBot 10:01 FireBreath: firebreath-1.4 Richard Bateman * 363a35e (3 files in 2 dirs): Fixed crash on Chrome / Windows -
iaincollins 10:01 Hmm weird, I'm not seeing events raised in FB 1.3.2 my plugin-in in Google (but other browsers are fine)...
not a bug report, recall seeing similar behaviour mentioned, but not the specifics
taxilian 10:01 hmm.
try the latest 1.4 is all I can suggest
if that works, I'll see if I can figure out what the difference is
iaincollins 10:01 yeah, am not too worried as going to be moving to that :)
will let you know
Firefox, IE, Safari all being fine (I have some JS that catches the timeout so gets dealt with okay in my case anyway)
and errm it's started working now
(after failing every time, about 50 times)
taxilian 10:01 lol
iaincollins 10:01 hmm I quit the browser window and tried it again, and it's gone back to doing that... weird (all other browsers fine)
once it started working, it was working every time (tried it about 5-6 times in a row before closing and re-opening the window)
Not enough information for anything bug report, just mentioning for reference as it's definately specific and odd to Chrome (have 4 browsers up and running on the same VMware instance), though certainly not saying it's not triggered by a bug of mine.
taxilian 10:01 yeah; I appreciate the info
let me know if you find anything more
iaincollins 10:01 will do, will look at 1.4
FB_GitHubBot 11:01 FireBreath: firebreath-1.4 Richard Bateman * ba9f5f3 (2 files in 1 dirs): Added attached/detached events for testing -
FB_GitHubBot 11:01 FireBreath: master Richard Bateman * 363a35e (3 files in 2 dirs): Fixed crash on Chrome / Windows
FireBreath: master Richard Bateman * ba9f5f3 (2 files in 1 dirs): Added attached/detached events for testing
FireBreath: master Richard Bateman * 74c190b (5 files in 3 dirs): Merge branch 'firebreath-1.4'
FireBreath: master commits b62b65a...74c190b -
sabotaged|wk 13:01 hey taxilian
so i see you made some commits for the async calls / shutdown issue
should i give it a try now?
taxilian 14:01 yep
should hopefully be fixed
I think even without any leaks
though that required a little bit of convoluted logic… :-/ darn race conditions
sabotaged|wk 14:01 ok great
taxilian 14:01 let me know if it works, please =]
taxilian 15:01 sabotaged|wk: had a chance to test that yet?
sabotaged|wk 15:01 not yet, soon
meetings :/
actually i got some link errors
unresolved symbol getDefaultZone
taxilian 16:01 hmm. give me a sec to double check that
what branch are you on, sabotaged|wk?
sabotaged|wk 16:01 firebreath-1.4
taxilian 16:01 hmm. it builds for me...
even on the build server
sabotaged|wk 16:01 hmm, well i did merge that into my subversion repo, maybe something went wrong
taxilian 16:01 that's possible. do you have local changes? Just curious why you bother keeping your own copy of the repo instead of just using the public one
sabotaged|wk 16:01 well i was thinking that if i needed to make my own changes, but i havent really had to yet
also we keep some other open source projects like this too so i was following suit
taxilian 16:01 I think I have windowless plugins working on IE
makes sense, I guess
if I were you, though, I'd make my own fork on github
then use the svn read/write support if you want to use svn
sabotaged|wk 16:01 that might be a bit better, but then i wonder about the reliability of github
taxilian 16:01 I've been pretty impressed with it
up to you of course =]
just seems like that'd be easier
sabotaged|wk 16:01 oh cool, clean and rebuilt and they're gone
taxilian 16:01 nice
ahh… yeah, probably needed to rerun the prep script
or even wipe the build dir and run prep
sabotaged|wk 16:01 yeah i re-ran prep script, just needed a clean before that too
taxilian 16:01 ahh
sabotaged|wk 16:01 awesome, looks like the reference count for my object is what it should be now
taxilian 16:01 sabotaged|wk: glad to hear it =]
sabotaged|wk 16:01 thanks for fixing it