IRC Log Viewer » #firebreath » 2012-08-21

IRC Nick Time (GMT-7) Message
JuanDaugherty 09:08 i don't think they've ever done it without user OK, maybe they've crossed that rubicon
taxilian 09:08 yeah, I dunno; I was able to repro this issue on 3 different computers, now I can't… with the same install file
emicastro 12:08 do you know any about ppapi?
taxilian 12:08 emicastro: I've looked into it, but not for awhile; it looked like it might be possible to make a ppapi host for firebreath, but I decided it wasn't worth the work because nothing but chrome supports it
emicastro 12:08 yeah, the same for me...
there are a new thinking way... Browsers makers (or Microsoft) wants to remove plugins from the browsers
taxilian 12:08 actually all browser manufacturers would love to get rid of plugins
emicastro 12:08 on Win8... if you download Chrome lastest version... it install two version of it... the metro and desktop app
taxilian 12:08 at least Chrome and Mozilla do; I haven't heard anything specific from Apple
we might have to come up with some creative solutions for metro; I have some ideas, but I'm not sure if they are feasible or not yet
I haven't had time to tinker with windows 8 yet
I probably should get another computer to put windows 8 on and develop with, but I haven't
emicastro 12:08 I have been programming some stuff on win8... and my plugin is a part of it... and now, I have discovered this kind of stuff....
the metro version of the browsers...
for example
taxilian 12:08 which don't allow plugins
at least, IE doesn't; dunno about Chrome
emicastro 12:08 two o three month ago, only IE had metro version
now, all browser (at least Chrome) have it...
exactly, they don't allow plugins
NPAPI and ActiveX aren't allow
taxilian 12:08 the question is, what type of actual applications can you create to run "metro style" that can be run directly from metro?
reichi 12:08 i wonder how that should ever work out
taxilian 12:08 I mean not web apps, but that we could do with c++ or c#
reichi 12:08 npapi plugins are almost everywhere all accross the web
taxilian 12:08 yeah; it's going to be a problem for us, because our software (which is used by thousands of teachers around the US) requires an NPAPI plugin to function
emicastro 12:08 yeah... I mean on microsft world... they don't like plugins... they want to migrate everything to HTML5
taxilian 12:08 so my thought is that it might be possible to create an application with an embedded browser that basically loads a given web page and makes your plugin available inside it
reichi 12:08 i like the basic idea behind that
(getting rid of npapi)
but i just don't see how that could work
taxilian 12:08 yeah; plugins are a huge security concern
but they are also invaluable
emicastro 12:08 removing private plugins to handle videos (flash player) for example, this is a lie cause they has made an agreement with Adobe to keep Adope plugin working on metro :)
reichi 12:08 there's just too much things you cannot solve without npapi
at least currently
emicastro 12:08 yes
your approuch is good... but this web page must be remotly..
if your app run on metro environment
taxilian 12:08 emicastro: what we do currently for iOS is we us a UIWebView and load our app, then we use hooks so that we can call into the Obj C code and say "put the 'plugin' in this rect"
seems like it might be possible to make a fairly generic shell for firebreath plugins that would do the same thing
the nice thing is you could even limit it to your own website
emicastro 12:08 you cann't loopback from metro apps... you cann't access to localhost from a metro app...
taxilian 12:08 it wouldn't solve all problems, but it would help from some things
emicastro 12:08 yeah that right
taxilian 12:08 interesting
but it seems like that might be the best solution for us, at least; it would probably work well for many others as well
emicastro 12:08 yes, but Metro environment will be a headache for us in a feature... Metro Apps opening external links launching a browsers that allow plugins only on the desktop version... but you cann't tell on which version (metro or desktop) that link must be opened...
taxilian 12:08 yeah
emicastro 12:08 new paradigm may be...
taxilian 12:08 either way it's going to be a bit of a nightmare
emicastro 12:08 i don't know
reichi 12:08 thank god i am on linux only
it seems like THE blessing
taxilian 12:08 hehe. as long as you don't need to release a binary distro… =]
reichi 12:08 well, it's totally under our control
from the drivers to the kernel
up to wha actually lands in the firmware
taxilian 12:08 yeah; in that case you're fine =]
people keep asking me why GradeCam doesn't have a linux version; they dont' understand how few teachers use linux nor how difficult it is to put out a binary distribution of a plugin to work on all flavors of linux
just not cost-effective
reichi 12:08 well
you could extend that to "how many people except developers and admins use linux"
and still what have an ridiculously small number
emicastro 12:08 reichi: customers will decide where the plugin has to run, not you (the developer) :) so.... we have to fight with this kind of things like run behind a beta OS like Win 8 for example
taxilian 12:08 yeah. fortunately for me windows 8 changes the game so drastically that most schools won't adopt it for at least another year, but I'm going to have to start worrying about it soon.
emicastro 12:08 dodging problems that they don't know that exist :)
taxilian 12:08 the really sad thing is how many of our customers are still on IE7
IE8 is sorta understandable because so many are on XP
but IE7?
at least we were able to drop IE6 support
emicastro 12:08 yeah! this is one of the most amazing things on web development too...
office computers that are plugged it to intranets running some intern software....
they still use IE7
taxilian 13:08 heh. many of those still use IE6
reichi 13:08 emicastro: i develop for an embedded platform, so in my case, that decicsion is fully ours ;)
taxilian 13:08 heh. braggart ;-)
reichi 13:08 grudger
emicastro 13:08 you're lucky :)
reichi 13:08 but well, fun aside
i don't know if it's a good decision
to ban npapi from metro browsers
taxilian 13:08 I can understand the reasoning, but it shows a frightening lack of understanding of how things work
reichi 13:08 i could imagine that at least the 3rd party browser may take a step back there
taxilian 13:08 even more frustrating is that they have reportedly made an exception for flash
reichi 13:08 yeah
sure npapi should be gotten rid of
taxilian 13:08 but only if you're providing a viable alternative
reichi 13:08 but before that there has to be some way to offer somehow similar functionlitay
let it be javascript or whatsoever
taxilian 13:08 even javascript wouldn't work for what we need; our image analysis wouldn't be performant enough in javascript, and even if it were so we would definitely *not* want the code that is our "secret sauce" to be easily downloaded by anyone with a halfway decent javascript code formatter
reichi 13:08 that is very true
it's the same for us
taxilian 13:08 but seriously the only thing that makes our stuff viable is an insane amount of compiler optimization that simply wouldn't happen in javascript
reichi 13:08 how stable is the 1.7 branch at the moment?
taxilian 13:08 pretty stable
I'm using it in production
I mainly just haven't taken the time to release it
reichi 13:08 i guess i'll upgrade than
taxilian 13:08 it's ready, AFAIK
reichi 13:08 and merge my changes into the 1.7. branch
(which hopefully doesn't turn into a merging nightmare)
taxilian 13:08 hehe
well, not too much has seriously changed in 1.7; there are some definite fixes you'll want
we're now using PCH on both windows and mac, but I don't know how to use them with makefiles linux-style, so I don't have them on for that platform. FB::variant has been updated
FireBreathBot 14:08 Commit d92c402 on master by Byron Clark: "Move these messages to LogLevel_Trace.
Commit 3636a14 on master by Richard Bateman: "Merge pull request #78 from byronclark/getvalue-less-verbose..."
Binbo 20:08 taxilian: hi, I am really not a good programmer and I really stucked on calling mypluginAPI method from myplugin class as I have asked you last time. Really need an example snippet to understand it =(
taxilian 20:08 sorry, you'll have to refresh my memory on your question
Binbo 20:08 taxilian : I just use default class given. let say my plugin named "myplugin", how to invoke mypluginAPI's method from myplugin class?
taxilian 20:08 FB::ptr_cast<mypluginAPI>(getRootJSAPI())->someMethod()
and I recommend you spend some time studying C++; also learn about boost::shared_ptr
that will help you understand what is going on
Binbo 20:08 definitely, my c++ fundamental is totally garbage since I seldom practice it
taxilian 20:08 FB::ptr_cast is just an alias for boost::dynamic_pointer_cast, FYI
Binbo 20:08 thanks!
working hard to master it =)