IRC Log Viewer » #firebreath » 2011-07-21

IRC Nick Time (GMT-7) Message
tarang 00:07 hi
dougma 00:07 hello
tarang 00:07 plz suggest me how to access data from sqlite database in this plugins
because i am create a project by firbreath and i want to access data from sqlite db
dougma 01:07 what's the problem?
it's just the same as accessing any sqlite database
tarang 01:07 plz tell me how i start this
dougma 01:07 read the sqlite documentation?
tarang 01:07 have any link for that type of sample
dougma 01:07 no
tarang 01:07 k
taxilian 02:07 I gave him a google search that would be adequate to find all sorts of examples… I guess he doesn't like to read
Mac OS Lion has some definite oddities
someone-noone 02:07 how to make plugin window full-screened on mac?
dougma 02:07 just lioned up my laptop too
someone-noone 02:07 To make full-screen: should i create a new NSWindow as here
AFAIK, plugins on mac are windowless. So it's impossible to use that technique
am i right?
arriu 11:07 I am getting a weakHost.expired() assert fail when I close my plugin. I found this related page: it says to use a BrowserHostWeakPtr. How do I do this?
I can't even locate the required header for BrowserHostWeakPtr
taxilian 11:07 arriu: you need to learn how boost::shared_ptr and boost::weak_ptr work
BrowserHostWeakPtr is a typedef for boost::weak_ptr<BrowserHost>
similarly, BrowserHostPtr is a typedef for boost::shared_ptr<BrowserHost>
… where is FireBreathBot?
tha'ts weird
linearray 11:07 is he hosted with amackera?
taxilian 11:07 no, I host him
there we go
arriu: this is probably the best article I'm finding right off on shared_ptr / weak_ptr:
very very very useful things
arriu 11:07 oh thanks for looking :)
taxilian 11:07 no problem
linearray 11:07 but you wrote one yourself :)
taxilian 11:07 along with being a jerk I try to be nice once in a while, just to keep people off balance
did I really?
linearray 11:07 I believe so
taxilian 11:07 huh
what'd'ya know
I did
arriu: that one is okay as well
though is it firebreath specific
jshanab_wcw 12:07 I am pulling my hair out again, really stuck. I havea plugin on FB1.5 that displays video form cameras. I can display 14 cameras, select 1 and view it. However, when I go back to display the 14 cameras it might flash them up there for 1/2 a sec(like old window buffer) but then it doesnt' show anything, when i referesh it crashes the page. Or if i open a second browser to the same multi... view. The funny part...
Is that the backtrace is meaningless, shutting down the browser in ntUserSetParent. Just wondering if that is a clue
"Unhandled exception at 0x76602d79 (user32.dll) in firefox.exe: 0xC000041D: An unhandled exception was encountered during a user callback."
taxilian 12:07 is your plugin loaded at all?
jshanab_wcw 12:07 Maybe something unloads them, that would explain the flash I get of about 1/2 the cameras before it updates them all to white.
arriu 12:07 When you hit the refresh button, a new plugin is initialized then the old one is shut down. Maybe you have some global variables that are overwritten?
jshanab_wcw 12:07 OK. I am gonna set some breakpoints and check that. Up until now i thought is was the SDL lib that I am changing too that was my issue
taxilian 12:07 brb
arriu 12:07 What happens when you open the plugin in more than one tab?
jshanab_wcw 12:07 I have not tried that!
Works fine, well until i go to a camera and them come back
arriu 13:07 I still cannot figure out how to make the weakHost.expired() assert go away :(
taxilian 13:07 well, what that means is that you have a shared_ptr somewhere that isn't getting released
it's a memory leak
most often that's because of a circular reference, but it can be tricky to track down
arriu 13:07 all I am doing is creating a class to log messages using the host, like so:
taxilian 13:07 I'll be back a bit later; I just got called by a company who needs someone to fix a PBX
odd the things I end up doing :-P
that class is almost certainly not your problem
because you're using the weak ptr correctly, from what I can see
look elsewhere
good luck!
arriu 13:07 thanks
arriu 14:07 the problem is with htmlLog(), when it is there I get the error, when it is not all is fine.
jshanab_wcw 14:07 I had to rip out all the SDL code, I just could not get it stable or even predictable. So i am back to DirectX on windows and I have a nuisense bug. If I scroll the browser window it goes white if i scroll medium speed or if it gets clipped. I am responding to Redraws, Is this just a browser artifact and I need to live withit, or am I failing to do something
arriu 15:07 don't use directx, opengl runs much better in the browser (speaking form experience)
jshanab_wcw 15:07 I like OpenGL and I use it elsewher, but how do you avoid the performance penalty of YUV to RGBA conversion and 4 bytes/pixel across the bus without a custom shader?
arriu 15:07 why can't you use a shader?
jshanab_wcw 15:07 I could not get the SOB to work. :-(
It is code that goes on the card
arriu 15:07 so you had problems with the shader or with the plugin?
or both haha
jshanab_wcw 15:07 The shader is loaded by the plugin onto the card. that depends on the card model. SFML has a bug when talking to ati cards and it kept segfaulting. After a few days i had to set it aside for later.
Ziggy 16:07 hey
anyone there?
I'm creating a new plugin
using the guide online
im at the step where it says "Plugin has no UI [false]"
does this mean that it just does not have it own ui independent of the extension im attatching it to?
linearray 16:07 not exactly sure where you see that, but I think it means whether the plugin has its own window
yes, that's what it means
taxilian 16:07 Ziggy: it means do you plan to do any actual drawing with the plugin
linearray 16:07 and I meant drawing, yes :)
taxilian 16:07 as opposed to just calling javascript APIs and drawing with html
Ziggy 16:07 O ok thanks all!
it is in cmd
when im going through this:
its not in the list of 'series of prompts'
taxilian 16:07 yeah; it's newish
Ziggy 16:07 may be good for someone more knowledgeable to add it in wiki
taxilian 16:07 go ahead and add it and I'll edit it if need be
the problem is that everyone waits for "someone more knowledgeable" to document things, so nothing gets documented
because I can't do it all
Ziggy 16:07 lol ok
taxilian 16:07 I monitor changes; if you put something in that isn't correct I'll fix it
Ziggy 16:07 Im just a total novice thats why i suggested but ill add it
taxilian 16:07 thanks
Ziggy 17:07 Just added it though I couldn't figure out how to make the background of the cell grey
taxilian 18:07 Ziggy: thanks; I updated the wording slightly and made it a heading column
as a hint, it's often easier to see how things work on the wiki if you change to the source view instead of tee wysiwyg
Ziggy 18:07 oic Thanks! the guide was super helpful and I just finished building my first plugin!
taxilian 18:07 awesome!
thank you for being one of those who reads the instructions and figures things out instead of asking us to walk them through things that are already written down
ahh, good catch. some of those documents haven't been updated in awhile
Ziggy 18:07 Thanks!
Skype 21:07 I am throwing a FB::script_error("useful msg"); IE displays this fine in try/catch but Firefox just says "Error calling method on NPObject" is this a known issue with FF?
taxilian 21:07 unfortunately, yes
ff 4 and later
and all recent versions of chrome and safari as well
awesome, huh?
Skype 21:07 f$#@, then I think the best way for me to report errors would be to event, right?
taxilian 21:07 if you want them consistent, probably so
another way would be to store the message in your JSAPI object and then call getLastError()
(which you owuld have to write)
actually, that's really not a bad idea for a long-term solution; I should add that to JSAPI
quick, add a jira ticket so I don't forget ;-)
or implement it for me and submit a patch, whichever you'd prefer
Skype 21:07 I am going the event way but how will getLastError work across threads?
How about dev still throws FB::script_error, but FireBreath could be more clever, if a OnFbError or similar event is registered then it throws and calls the event?
taxilian 21:07 it would work fine, because *all javascript calls happen on the same thread*
no need for that
here is how it would work:
whenever a javascript call is made
and a FB::script_error exception is thrown
the message would be stored in a variable and then the exception would be thrown to javascript
now, on most browsers these days that won't give you the message... you'll just get the "Error calling method on NPObject" crap
so you just call plugin.getLastError() and it returns the message
that would work fine *unless* the object is invalid and that's the reason for the error
Skype 21:07 and JS dev would try/catch(err) { alert(plugin.getLastError();) } nice
taxilian 21:07 right
the flow would be the same
it could potentially be a special method on the browser object rather than on the JSAPI object
that has the disadvantage of making it impossible to override it, but the advantage of being able to work with any error
Skype 21:07 I can live with that, am used to Windows GetLastError for the last couple of days:-)
taxilian 21:07 hehe
it'll be a bit before I'll have time to get to it; I have quite a laundry list to get to right now, I'm afraid
I've been swamped
but at least add a jira ticket or I'll forget
Skype 21:07 no worries, btw do you know why the FF and family decided to go with not reporting nicely, security?
ok I'll add a Jira ticket for that