IRC Log Viewer » #firebreath » 2012-05-30

IRC Nick Time (GMT-7) Message
Dyblast 01:05 Good morning everyone
!
taxilian 08:05 Dyblast: good morning =]
Dyblast 08:05 good afternoon now :D
taxilian 08:05 not for me it aint =]
Dyblast 08:05 have you see the video in comment i have made https://github.com/firebreath/FireBreath/pull/62#issuecomment-5972841 ?
taxilian 08:05 I did, but without having had time to look through your code it doesn't tell me much
I'm behind on some projects at work and then I'm moving on Friday, so this has not been a good couple of weeks for me to help people :-/
Dyblast 08:05 haha
no problem :)
taxilian 08:05 what you probably need to do is look in the code that is used to "convert" a JSAPI object to a NPObject; what it should be doing is detecting that an NPObject wrapper for that jsapi is already created and returning the same object each time
if it isn't doing that, hopefully you can figure out why by either stepping through or adding log statements (or both)
Dyblast 08:05 I sure that retainJSAPIPtr is called multiple time for the same JSAPIPtr
+s
taxilian 08:05 that shouldn't be a problem
well, kinda
the reason it would be called multiple times is that multiple NPObject wrappers are created
but that shouldn't happen
so if multiple NPObject wrappers are being created when they shouldn't be you need to figure out *why* multiple NPObject wrappers are being created
and fix that
Dyblast 08:05 ok what you tell that my patch is not enough close to the problem?
taxilian 09:05 your patch would actually cause problems
it probably seems to fix it, but what actually will happen is that when one of the NPObjects is garbage collected by javascript all of them will stop working
so in essence it's the opposite of a memory leak -- the memory will go away before everything is done using it
this wouldn't actually cause a crash, but it would cause odd problems in javascript
Dyblast 09:05 hum ok
understood
i will investigate
Dyblast 12:05 hummm
taxilian 13:05 you don't say
Adam__ 14:05 Is there a "best practice" approach in locking down a plugin to only be called from a specific site?
taxilian 14:05 well, the simplest way is to check the page location
but that doesn't protect against XSS or DNS spoofing
Adam__ 14:05 Hmmm.... I saw this "http://stackoverflow.com/questions/7214460/how-to-write-a-secure-npapi-plugin-that-would-interact-only-with-a-specific-chro" ... how does this work?
taxilian 14:05 that's not a plugin, that's an extension with a plugin inside
in chrome extensions only it's possible to make an extension with a plugin inside that the plugin can only be accessed by the extension
Adam__ 14:05 I'm using the plugin I created within a Chrome extension...
taxilian 14:05 then you just have to set private=true in the manifest, is my understanding
but I have never made one myself
so I could be misremembering
http://code.google.com/chrome/extensions/npapi.html
public: false
Adam__ 14:05 Hmmm... It seems there would have to be something within the plugin to detect what or who is making the call.
Reading now...
Thanks
Alex_I 16:05 hi. I have subclassed the window in my NPAPI plugin, but still not receiving WM_KEYDOWN
taxilian 16:05 the plugin has to have focus or you don't get any events
Alex_I 16:05 must I response for some OnFocus(), or just have to navigate to that window ?
I mean : must the plugin respond on some OnFocus() ? or just a user have to navigate onto it ?
Dyblast 16:05 Hey ... added to this it can be cool if we can add cmake function in order to create on windows the .cab for internet explorer
taxilian 16:05 alex you probably just need to click on the window
Alex_I 16:05 because Beep() inside of PluginWinProc() works, but not for WM_KEYDOWN
taxilian 16:05 Dyblast: be my guest =]
Dyblast 16:05 https://github.com/diorcety/linphone-web/blob/master/Win/projectDef.cmake
a draft
of function :)
at the end
taxilian 16:05 make a macro out of it
and submit it as a pull request
Dyblast 16:05 ok i will (try)
Dyblast 17:05 taxilian: done
taxilian 17:05 git pull request
FireBreathBot 17:05 7 open pull requests:
diorcety: Add Exe wrapper to msi and cab generation http://tinyurl.com/bqfe5ye
diorcety: FIREBREATH-176: Multiple references to one JSAPIAuto http://tinyurl.com/czavgdz
medusade: Modified InvalidateWindow in the PluginWindowX11 class to se... http://tinyurl.com/c463duz
Geal: Add a CMake function to build a Chrome plugin http://tinyurl.com/ckfmopw
Geal: Add a CMake function to build a XPI package using XPISigner http://tinyurl.com/blhoc5s
Geal: DMG installer generation http://tinyurl.com/carck32
taxilian 17:05 wow. I need to get on some of those :-/
FireBreathBot 17:05 wroberts: Unsolicited npp newstream http://tinyurl.com/crz3qxw
taxilian 17:05 if anyone wants to review one that isn't theirs it would help a lot
I didn't realize I was so far behind
Dyblast: awesome. if nobody else has time I'll try to look at it next week or the week after. I'm really sorry I'm so far behind on accepting pull requests :-(