IRC Log Viewer » #firebreath » 2011-10-19

IRC Nick Time (GMT-7) Message
Guest47632 05:10 HI everyone
linearray 05:10 hi
Guest47632 05:10 i want to know some quation about plugins
linearray 05:10 what's up?
Guest47632 05:10 can you answer me plz
i want to use plugins to connect serial port in ubuntu
linearray 05:10 ok
Guest47632 05:10 it's plugins can do that?
linearray 05:10 yes
a plugin is just a program
Guest47632 05:10 is firefox need authenticate that plugins?
linearray 05:10 please rephrase that question.
Guest47632 05:10 I want my web can write/read COM1 by plugins , the OS is ubuntu , it's has any security problem?
sorry i am taiwanese ,type slowly
linearray 05:10 I don't know if you can write to the serial port as an unprivileged user... never used it
but my guess is No
Guest47632 05:10 you say my idea is impossible
linearray 05:10 no, it's very well possible
you just *may* have to be privileged, e.g. using sudo
try it
Guest47632 05:10 you say "never used it" what you mean?
linearray 05:10 I have no experience with serial ports
Guest47632 05:10 mmmm
i think any programe want to control the device of computer is very dangous
so the browese will stop it or the program need authenticate ,like activeX for IE
mkoch 05:10 Guest47632: it's always better trying not to directly access the devices like the serial port in your case, but to use some driver like you are doing it when you are printing (the browser doesn't send the data on the parallel or whatever port but the driver does it)
disadvantage: you have to develop/install a kind of a driver on each machine
Guest47632 05:10 now i can W/R the serial port in the computer , i know activeX can do that in IE web,
mkoch 05:10 advantages: much safer, you don't need special privileges, and the plugin itself can remain platform independent if the driver provides a god api
Guest47632 05:10 and why the activeX can?
mkoch 05:10 linearray: is the only known source on the web explaining how to display graphics in a plugin?
Guest47632: you can do that with ubuntu too. it's just not the best way to directly access the serial port from a browser
linearray 05:10 hehe, I don't know if it's the only one on the web, but the only other thing I know of are the two examples plugins
mkoch 05:10 okay... :)
linearray 05:10 !wiki drawing
FireBreathBot 05:10 8 results found. Note: Results limited to 8
"Event models and drawing models":
"Mac Plugins":
"Tips and Tricks":
"FireBreath 1.5.0RC1 Released!":
"Helpful Links":
"Building on Linux":
"Frequently Asked Questions":
"FireBreath 1.6.0RC1 Released!":
linearray 05:10 hm, nothing here
mkoch 05:10 is it the browser's or the developer's choice that which drawing (and event) model is used in OS X?
linearray 05:10 both ;)
I have yet to do drawing, but from I've read CA and ICA is almost the same
mkoch 05:10 so if you really want to be crossplatform, you need to develop your plugin for all the 3 different models?
linearray 05:10 sort of, yeah
Guest47632 05:10 excuse me, i still not know how to do about control serial port
linearray 05:10 if by cross-platform you mean including very old browsers
just look at the table. if you can live with ff4+, then cocoa/(I)CA is enough
Guest47632 05:10 linearray you talk to me or mkoch ?
linearray 05:10 to mkoch
mkoch 05:10 linearray: that's gonna be the first step then
Charlie 05:10 Hi, Can I create files on a users filesystem using Firebreath ?
linearray 05:10 yes
Charlie 05:10 Will all browsers allow this ? Can I also write to a specific directory ?
linearray 05:10 I think all browsers allow it in some sense... some may sandbox you to the Downloads directory. I haven't tried writing somewhere else yet.
But I know that Chrome prevents itself from writing anywhere else, so it's conceivable they would also do this to plugins.
Charlie 05:10 oh ok ... Thanks a lot for your help linearray !!
Guest47632 05:10 linearray : can you tell me more about NPAPI plugins control hard device ,can I complete it by use FireVreath?
linearray 06:10 unfortunately I have no idea
Guest47632 06:10 thx your help ,
linearray 06:10 but if you hang around more knowledgable people will arrive!
Guest47632 06:10 i will post my question in Forums
linearray 06:10 ok, that works as well
Guest47632 06:10 i just want know it possible or not ,somebody say no somebody say yes,
Charlie 06:10 Guest47632 : We are also looking to control devices using Firebreath... I think its possible..
Guest47632 06:10 O
what are you want to control?
Charlie 06:10 Telephony device through USB
Guest47632 06:10 i find some soft use activeX plugins to call activeX in Windows,
Charlie 06:10 Its possible using ActiveX, dont know about NPAPI
Guest47632 06:10 so you sucess?
i know activeX can
Charlie 06:10 We're still trying to come up with a prototype....
Guest47632 06:10 my Boss say activeX can do it, and NPAPI can too
but he dont know what is NPAPI
he only know activeX==plugins
mkoch 06:10 :)
linearray 06:10 Charlie: I just checked, the sandbox is not applied to plugins.
so you can write anywhere you want
Guest47632 06:10 ??
you say the sandbox dont check the plugins security
in firefox?
linearray 06:10 Chrome
Guest47632 06:10 so we can control the device via web? how dangous
linearray 06:10 no, it has nothing to do with that
Charlie: oh, and use boost::filesystem. It will reduce the headache.
Charlie 06:10 Great !!! Thanks linearray !!
mkoch 06:10 re
mkoch 07:10 the manual says that there is no guarantee that the browser won't detach and reattach the window anytime during the runtime
what happens exactly in such a case?
is the pluginwindow object destroyed and the browser will give the plugin a new one then?
dougma 07:10 i interpret that as onWindowAttached/onWindowDetached can be called at anytime
mkoch 07:10 yes, that's clear but what should a plugin do when it happens anytime other then loading the plugin the 1st time or leaving the page (closing the plugin)?
dougma 07:10 so you track the FB::PluginWindow * it gives you and be aware that it can be null i suppose
that's what i do. :)
mkoch 07:10 but I guess you do it in your redraw routine then, right?
dougma 07:10 when i want to be redrawn: if (m_window) m_window->InvalidateWindow()
mkoch 07:10 that's what I meant
dougma 07:10 ok then
mkoch 07:10 but then is there anything in those two functions for you?
dougma 07:10 yes, if onWindowDetached is called, i null my m_window
mkoch 07:10 ah okay :)
dougma 07:10 remarkable it all seems to work
well, drawing works
mkoch 07:10 which platform are you using btw?
dougma 07:10 windows and mac
mkoch 07:10 and are you by any chance working on some open source project? :)
dougma 07:10 just firebreath. :)
the plugin isn't opensource
mkoch 07:10 okay, then I won't ask for the code then :)
cyberaxes 08:10 anyone here?
i've been using SpicIE for plugin development and am thinking of moving to firebreath
i'm trying to solve a problem that i'm not sure is really solvable….can two instantiations of IE/Firefox plugins share information?
without having to IPC
linearray 09:10 hmm
unless you count files
or databases :)
cyberaxes 09:10 that's what i was afraid of, are there any restrictions on what you can access? like memory mapped io, files etc.? some security settings?
does firebreath offer any tools to help with sharing data?
Sudi 10:10 hello everyone
taxilian 10:10 hello, Sudi
busy night
for those who have been trying to answer questions, a few points
in IE on Vista/win7 w/ UAC enabled you are restricted where you can write; any other browsers you are not
the reason for the IE restriction is that your process is running in low integrity mode
on linux, accessing a serial port will work the same as from any other application; if non-elevated applications can normally access a serial port then a browser plugin can too
as a general rule of thumb, if you can do it in an activex plugin you can do it in NPAPI, at least as relates to permissions
taxilian 12:10 nirvdrum: gotta tell you, it may disappoint you to hear it, but I really like the iPhone, regardless of the existence or absence of android. In fact, the only real compelling thing I can think of that android has that I might consider is that it isn't Apple
(yes, I do have a bad tendency to respond to tweets in completely unrelated mediums :-P)
nirvdrum 12:10 taxilian: I've had an iPhone for 3 years. I just switched to Android.
taxilian 12:10 what convinced you to switch?
nirvdrum 12:10 You can be happy with either. But your happiness doesn't have to come at the expense of shitting on others.
taxilian 12:10 hehe. agreed
nirvdrum 12:10 I've yet to see anyone praise iPhone 4S. I've just seen them crap on Android 4.0.
taxilian 12:10 really? I really like the 4s
it's definitely the most reliable phone I've used; granted, I'm coming from the 3gs, so I can't compare it to the iphone4
nirvdrum 12:10 A few different things. I really wanted a physical keyboard. I find I don't ever actually reply to emails on my iPhone because it's just too frustrating to type on. And with the keyboard, I can now reasonably SSH into machines. I prefer the open ecosystem, even if the apps aren't as nice. And I really wanted to switch to T-Mobile.
taxilian 12:10 I can't comment on android 4.0 because I haven't used it; I do have a tablet running 3.2, and I have to say that it's terrible compared to my iPad, but I think the phones are a lot better from what I've seen
all reasonable reasons
nirvdrum 12:10 Up until the day before the 4S came out, I was planning on getting a 4S. But AT&T was trying to screw me out of another $30 - $40 / month.
And the CDMA networks don't work where I live.
taxilian 12:10 why would your price go up?
nirvdrum 12:10 They did away with limited SMS plans. Your options are now pay per message or pay for an unlimited plan.
taxilian 12:10 I was really interested in the open ecosystem until I looked through the android app market with my tablet and realized that because it's so open, 80% of the apps I found were crap, 20% I woudln't want my kids to see, and at least 30% looked to be pretty obviously malware… =]
seriously? they don't even have the $10/mo for 1000 anymore?
nirvdrum 12:10 And I was going to get my wife the free 3GS, but they only would allow me to choose a 700 min. / month phone plan (we're on the 550 and have plenty), so that was going to be an extra $10 / month.
taxilian 12:10 (I'm on a family plan and it's cheaper to pay the $30 for unlimited for all of us)
nirvdrum 12:10 I think you can be grandfathered in, but she's on the original iPhone plan, so she would have had to get a new plan. And we're on a family plan, so it would have cascaded.
Between the two of us, we maybe hit 70 SMS a month.
So maybe the metered billing would have been fine, but it's a tad absurd, too.
taxilian 12:10 that stinks. I'm with AT&T primarily because anything else would be too expensive; Verizon works *way* better here, but it would cost a lot more than I'm paying
nirvdrum 12:10 So, I finally made the switch. It's taken some adjusting, but overall it's more my style.
It's only been a few days, but I find I'm getting a lot more utility out of my phone because I can quickly type on it.
And having options and nifty integration points are cool. E.g., Google Voice is just my voicemail provider now.
taxilian 12:10 fair enough. I've been considering getting an android phone as a second phone just so I know how to use it and can learn to program it; if I can find a way to pay just for data and then have prepaid voice I might do it
nirvdrum 12:10 I can't comment on the tablets. I can't see a need for either of them.
taxilian 12:10 I use tablets mostly for ebooks
kylehuff 12:10 I switched from iPhone to android mostly on principal. overall, I am pleased with the switch, but that is not to say that the iPhone isn't a technological wonder. I just hate they way they implement it
nirvdrum 12:10 Yeah. I can't stand reading that way. I need e-ink or actual paper.
taxilian 12:10 and I have a really extensive ebook library, both recreational and technical reading
hehe. haven't had a problem with my ipad, but I've been reading on PDAs for many many years
nirvdrum 12:10 kylehuff: I was happy with an iPhone. I would've gotten it. But like I said, I really wanted a physical keyboard.
taxilian 12:10 what things does your android phone have that you like better than the iphone?
nirvdrum 12:10 But I'm pretty sure a large number of people dumping on Android have never actually used it.
Free SSH client that can do public key auth.
taxilian 13:10 that is probably true; this is why I don't dump on android out of general principle… except the tablet version, which the nicest thing I can say is "it's pretty nice if you can't afford an ipad" =] it's young, though, and I'm sure it'll improve
nirvdrum 13:10 I'm aware there are SSH apps for iPhone, but I don't need a "cheerful SSH app".
taxilian 13:10 heh. I use iSSH, and it's worked really well. I think I did pay $5 or so for it a long time ago, though
nirvdrum 13:10 Plus I couldn't imagine trying to type commands to run on my server with the touchscreen, autocorrect keyboard.
linearray 13:10 issh is bearable
taxilian 13:10 ConnectBot does work pretty well, though
linearray 13:10 but no match for connectbot :)
taxilian 13:10 at least on my tablet
kylehuff 13:10 my big gripe was with app and carrier portability. I dont' want to purchase apps that I can only use on one line of devices put out by the same manufacturer who shits all over the people who wrote the codebase they are using.
nirvdrum 13:10 The Google Voice stuff is really nice.
The camera app I have is much nicer, but I think it's specific to my model of phone.
The calendar stuff works a lot better.
I have an SNES emulator, which is just a lot of fun :-)
Made more fun by the physical keyboard.
kylehuff 13:10 yes, but do you have Contra? lol.. best game ever...
taxilian 13:10 kylehuff: conversely, I am not really excited about the idea of developing for android because there are dozens of devices that all have different hardware characteristics
nirvdrum 13:10 kylehuff: I could get an NES emulator too.
taxilian 13:10 lol. I have been surprised that I haven't really missed my physical keyboard all that much (from when I had a winMo phone)
nirvdrum 13:10 Keep in mind, I had a 3GS. I skipped the 4 entirely.
I wanted the flexibility afforded by not being on contract.
Oh, and TMobile has this wifi calling app for Android that I love.
Basically it's a built-in microcell. So I can get phone coverage all throughout my house through my wifi.
taxilian: There seems to have been a lot of disappointment from diehards with the 4S announcement. And then it quickly turned to "well, it's still better than Android"
Hardware-wise, I think the Android phones have mostly caught up and in some cases started to surpass.
I'm also looking forward to trying out a Scala project on this thing. And not having to pay anyone $100 for the privilege to do so.
kylehuff 13:10 taxilian: yes, from a developer point of view, I agree there... that would probably be a pain. however, I like the flexibility to be able to not give a certain company money in order to use my apps if they behave in a manner that violates my moral or spiritual principals.
nirvdrum 13:10 And then the one last thing I love is having a microSD slot. I wanted 32 GB. So I was going to have to shell out an extra $100 on the 4S to get that. With this phone I can just plop a $30 32 GB microSD card in.
The 4S really should have gone with 32 GB base if they were going to add an 8 megapixel camera and 1080p video.
But again, I wholly get why people like the iPhone. I just know people that haven't used anything but iPhone and iPad for 5 years. So they haven't used Android at all. And yet they crap on it.
It's kinda like Mac users that still crap on Windows, when they haven't touched it since Win 98.
taxilian 13:10 one big reason I like the iPhone is that I just don't have to mess with it; everyone that I know with an Android phone ends up flashing custom firmwares, etc, which sounds like a lot of fun but I wasted way too much time on that. The iPhone has surprisingly little bloatware
nirvdrum 13:10 I'm surprised how much I'm excited by NFC. I didn't really care about it until I realized it could be a way to get rid of all my customer loyalty cards and pay for access on the public transit system.
taxilian 13:10 hehe. I crap on Windows, but I use it every day :-P I just find Mac to be much more powerful
around here NFC would probably not help me much (if at all)
nirvdrum 13:10 I'm about to wipe my Mac and just run Ubuntu. This 10.7.2 update has been terrible to me.
taxilian 13:10 really? what problems are you having?
kylehuff 13:10 I crap on windows and mac, but moreso windows... hehe
nirvdrum 13:10 It's esoteric, but I have two different keyboard layouts. It keeps randomly choosing them on me.
I lost my iPhone backup when upgrading to iOS 5 because it switched the layout when prompting me for a password.
I never managed to recover it.
taxilian 13:10 really? I use 3 different layouts and I've never had that problem
are you sure you don't have a hotkey set funny somewhere?
nirvdrum 13:10 I use QWERTY and Dvorak.
taxilian 13:10 I use Qwerty, Dvorak, and Russian :-P
but my Dvorak is really rusty now… my Russian ruined it
I should get it back up to speed
nirvdrum 13:10 I had disabled it. But this happens even just when a dialog pops up.
taxilian 13:10 really? you might want to just try reinstalling
nirvdrum 13:10 Like I'll be using Firefox it'll be in QWERTY and then a basic auth dialog will appear and it switches to Dvorak.
taxilian 13:10 because I have 2 layouts set on all of my computers and I've never had that issue
on any version
nirvdrum 13:10 It used to do this in 10.6, but only with Chrome oddly.
It went away in 10.7.
And now it's back and worse than ever in 10.7.2.
taxilian 13:10 so I'd almost bet you have some weird setting somewhere that you aren't aware of
nirvdrum 13:10 Maybe Apple does. I don't. All I did was run their crappy software update.
taxilian 13:10 have you tried calling apple support?
(yes, I realize that shouldn't need to be done)
nirvdrum 13:10 Nay.
taxilian 13:10 ironically I've had problems like that far more on windows… never had an issue like that on Mac
really bizarre
nirvdrum 13:10 Anyway, if you dislike something based on your experience.
kylehuff 13:10 the keyboard layout switcher for gnome is kind of crappy, but mostly because it doesn't let you assign an arbitrary hot-key/combination. you have to use pre-defined crap that is common to other utilities.
nirvdrum 13:10 But at least expose yourself to it first.
taxilian 13:10 hehe. agreed. to be fair, though, there are just as many "anti-iphone" people out there who have never used one as there are "anti-android"
kylehuff 13:10 fortunately for me though, it switches the layout to RTL, so it is obvious when it switches due to accidental key combination
taxilian 13:10 I agree that it's annoying, though
Windows isn't my system of choice, but I use it and there are times when it is the best choice… it annoys me anytime someone tries to claim that one system or another is "the only possible choice"
nirvdrum 13:10 taxilian: I don't see them on Twitter or HackerNews :-P
At the end of the day, they all suck in their own special ways.
taxilian 13:10 really? 'cause I see them all the time. I'll copy one for you the next time I see it :-P lately most of them have been talking about how terrible the iphone4s is =]
so have any of you tried 1.6RC2?
I'm thinking of releasing 1.6.0 final today or tomorrow
nirvdrum 13:10 I haven't. I'm somewhere post RC1.
linearray 13:10 taxilian: do you have any ideas what may cause some systems to default to clang instead of gcc?
I'd love to see the fix for that in 1.6 :)
I'm on 10.6, maybe that's the problem?
taxilian 13:10 linearray: I really don't know; ideally, we should just remove main
but I'd feel more comfortable if I knew for sure how far back you have to go before that breaks something
linearray 13:10 well, I did the tests for firefox
so I guess we still need IE. anything else?
taxilian 13:10 IE?
this is mac only
linearray 13:10 oh
main() is mac only?
taxilian 13:10 yes
linearray 13:10 even better! so, what's left to do?
a fully patched OS X 10.5 installation might be helpful
taxilian 13:10 10.4 would be nice to check
and 10.5
I have a 10.5 ppc that I can try, I suppose
and/or do some searches
someone may have documented it
linearray 14:10 hmm, finding someone with 10.4 can be a challenge... it's really really old
hm, somehow it never occurred to me that FB might support anything but i386/x86-64
taxilian 14:10 actually, as far as I know I'm the only one supporting ppc :-P
ironic, really
but that is who our clients are
so new information; just verified that main was only used in FireFox
you tested back to 3.0?
linearray 14:10 yes!
good news
taxilian 14:10 alright. out it comes
the decision, then, is do I really want to make a change like that between the last RC and the release
my thought is yes; what do you all think?
linearray 14:10 I'm not impartial, so I will remain silent.
taxilian 14:10 I already counted your vote =] looking around, though, I don't know if anyone else in here cares
let me check with the one person who might (my boss =])
linearray 14:10 regarding the people who asked questions today
good to know about IE + UAC
chrome actually has the sandbox for plugins in place (at least on mac), but it's default off.
and regarding the other guy with the webcam: I could barely understand what he wanted to know, but there's an interesting question there: how *do* you run code that requires elevated privileges in a plugin on *ix?
I can think of hacks like putting it in a file and running/setuid'ing it with 'gksudo'
taxilian 14:10 basically, I think you are correct; you'd need to install some sort of a broker process with the suid sticky bit set like sudo that can launch your process for you
or us gksudo or something similar
however you do it, you'll have to launch an elevated process and have that do the work
same for windows/IE w/ protected mode
linearray 14:10 ah ok
taxilian 14:10 decision is in; we remove main
and then I verify that it works on 10.5 ppc
linearray 14:10 nice
taxilian 14:10 and then we party
have you already done it locally? can you submit a patch?
linearray 14:10 hm, I didn't remove any code
just the function from the export list
taxilian 14:10 ahh
well, should be simple enough, then
linearray 14:10 btw.
taxilian 14:10 yeah; fortunately that won't affect us
it's all tech that we have never supported
linearray 14:10 isn't that exactly what this is about?
main() and CFM.
taxilian 14:10 not most of it; most of it predates even what we are using
main() didn't just apply to CFM
linearray 14:10 alright
linearray 15:10 should these '#pragma GCC visibility push(default)' statements remain?
taxilian 15:10 hmm. I honestly don't remember =]
linearray 15:10 *sigh
I created a new branch, I did the changes, committed, pushed and tried a pull request
it still wants to send 5 commits instead of 1
ironically the "files changed" is correct
taxilian 15:10 git remote add trunk git://
git fetch trunk
git rebase trunk/master
after it rebases you should be good
you'll have to git push -f origin master
linearray 15:10 hehe, that reduced it to two commits
tell you what, here's the commit
I'm super tired and will fix everything tomorrow!
FireBreathBot 15:10 871f6fc by Max Amanshauser: Removed unnecessary main(.) to allow building with clang.
taxilian 15:10 awesome; I'll pull that in and do some testing
cristcost 15:10 hello everybody!
taxilian 15:10 hello
cristcost 15:10 anyone can tell me the difference between a FB::PluginWindowWin and FB::PluginWindowlessWin? I understand why there are 2 implementation, but when do Firebreath use one or the other?
taxilian 15:10 FireBreath uses the first when it's a windowed plugin, the second when it's a windowless
!wiki isWindowless
FireBreathBot 15:10 3 results found. Note: Results limited to 8
"class FB PluginCore isWindowless":
"FireBreath 1.4 Beta 4 released!":
"PluginCore base class":
cristcost 15:10 thank you! I take a look at the links!
so it's me who have to choose if creating a windowless or windowed plugin... and which one is better? Do they are both supported in both NPAPI and IE?
taxilian 16:10 yes, they are both supported by both. Which is better? it depends on your needs
windowless interacts with the page better, but you don't get an hWND
cristcost 16:10 and on Mac and Linux, the two concepts does not apply?
taxilian 16:10 on Mac everything is windowless
on Linux nobody has bothered to figure out how windowless works
cristcost 16:10 :_)
I thank you,
I'll make some test
good night (or equivalent, here is midnight, I don't now where are you ;-)
taxilian 16:10 =]
it's about 4pm
but sleep well =]
cristcost 16:10 :-)
FireBreathBot 16:10 JIRA issue issue created by miketalbot
FireBreathBot 16:10 JIRA issue issue updated by miketalbot
taxilian 16:10 !findfile Platform_windows.cpp
FireBreathBot 16:10 Found 1 matching file(s) in the master branch. First 1 are:
JIRA issue issue commented by miketalbot "Patch attached that fixes the problem."
Commit ecc7704 on master by miket: "FIREBREATH-134: Added ATL_INCLUDE_DIRS and ATL_LIBRARY to fi..."
Commit ce6bbad on firebreath-1.6 by Richard Bateman: "Merge branch 'master' of"
Commit 83f4138 on firebreath-1.6 by Kyle L. Huff: "Converted to use SOAP interface"
Commit 718e7a3 on firebreath-1.6 by Richard Bateman: "Merge remote branch 'kylehuff/master'"
Commit 908839a on firebreath-1.6 by Kyle L. Huff: "this should fix the documentation upload"
Commit fe8e601 on firebreath-1.6 by Richard Bateman: "Merge remote branch 'kylehuff/master'"
Commit 518e5e3 on firebreath-1.6 by Kyle L. Huff: "specifically cast Page-ID as long"
Commit 6616fd7 on firebreath-1.6 by Richard Bateman: "Merge remote branch 'kylehuff/master'"
Commit b1efd3f on firebreath-1.6 by Kyle L. Huff: "specifically cast all Page-IDs as long..."
Commit 038c396 on firebreath-1.6 by Kyle L. Huff: "Use custom SOAPpy parser to deal with duplicate XML nodes"
Commit 5decd5a on firebreath-1.6 by Kyle L. Huff: "fixed usage of potentially null variable in doc script"
Commit 823e80b on firebreath-1.6 by Richard Bateman: "Merge remote branch 'kylehuff/master'"
Commit 0286c33 on firebreath-1.6 by Richard Bateman: "Fixed minor doc error"
Commit 651d14b on firebreath-1.6 by Kyle L. Huff: "Fixes for doxygen parsing"
Commit db3d565 on firebreath-1.6 by dougma: "initialise PluginWindowlessWin members"
Commit ecc7704 on firebreath-1.6 by miket: "FIREBREATH-134: Added ATL_INCLUDE_DIRS and ATL_LIBRARY to fi..."
Commit a9e1db5 on firebreath-1.6 by Kyle L. Huff: "fixed parsing in over-zealous cleaning method"
Commit 8176d65 on firebreath-1.6 by Richard Bateman: "Merge remote branch 'kylehuff/master'"
Commit ded1e86 on firebreath-1.6 by Richard Bateman: "Merge remote branch 'kylehuff/master'"
Commit ec51ba5 on firebreath-1.6 by Richard Bateman: "Merge pull request #34 from dougma/unitialised-members
JIRA issue issue resolved by richard "I'll have to revisit this; most of that file doesn't need to be there anymore, and I doubt it rea..."
Feeling 18:10 hi all
how to update plugin on IE? i mean, how to get the previous plugin vresion?
IE catch the plugin instance until terminate browser when i called the property or method.
taxilian 18:10 Feeling: I'm not sure I understand your question
once you have installed the new plugin (registered a new file) it should thereafter load that file instead of the old one
dougma 19:10 Feeling: if you're using .cab files it's easy: you append the version to the codebase attribute of the object tag
Feeling 19:10 but, should terminate to browser if load old one
dougma 19:10 not required...
only if the dll is locked
Feeling 19:10 i think i will use msi or nsis
dougma 19:10 but you can't do anything about that.
actually, i'm not sure i understand you either sorry
what do you mean "terminate if load old one"?
taxilian 19:10 Feeling: don't overwrite the old file
put a new file in
and flag the old one to be deleted
dougma 19:10 good idea. :)
Feeling 19:10 i already do that but, plugins instance is old dll
dougma 19:10 put the version in the filename or install path
taxilian 19:10 once you have registered the new dll, any new instances you create will use hte new one
if you remove the object tag and add it again you'll get the new one
dougma 19:10 taxilian: do you deliver your plugins via msi?
taxilian 19:10 yes, though I have recently started packaging the MSI inside a .exe file
dougma 19:10 not msi wrapped in cab?
taxilian 19:10 because windows roaming profiles loses the MSI file
so you can't uninstall
I hate cab files
don't use 'em
don't do any browser specific installers
too much trouble, too many problems in the past
dougma 19:10 ok.
does your install webpage loop until the new object is available?
i've noticed chrome reloads all webpages having plugins when you register a new plugin
Feeling 19:10 how about use this? -> <object codebase="#version=some version" onerror="ReInstallorUpdate()"
taxilian 19:10 dougma: my install loop is in the firebreath codebase
dougma 19:10 feeling, but i'm going msi now. :)
taxilian 19:10 !findfile .*\.js
FireBreathBot 19:10 Found 1 matching file(s) in the master branch. First 1 are:
dougma 19:10 how did I miss that! nice :)
Feeling 19:10 so, how to delete old dll if i depoly new version(different name) of dll?
dougma 19:10 msi upgrade should take care of that for you right?
i need to do more testing on this. :/
Feeling 19:10 hum.. when reboot computer? or browser?
dougma 19:10 not sure what will happen if the dll is locked on upgrade
dll = old dll
Feeling 19:10 get the version
call the property or method
dougma 19:10 is that a question?
taxilian 19:10 MSI upgrade will take care of it
if the DLL is locked it'll mark it to be deleted when you restart the computer
Feeling 19:10 no, dll is locked when you will call property or method
i see
so will i upgrade dll and remove the object tag and add it again? right?
dougma 19:10 something like that, check the code taxilian mentioned
Feeling 19:10 ok. thanks very much.
staylor 20:10 is there a cross platform library commonly used, like sdl, for doing 2d graphics with firebreath?
taxilian 20:10 not really; some have had some success with sdl, but it's a little weird on Mac at best
staylor 20:10 so better just using the native apis then eh, alright thanks
taxilian 20:10 probably, unfortunately
staylor 20:10 wonder if using the browser's canvas from a plugin would be viable
taxilian 20:10 not real great, no
there isn't a good method for sending binary data to it
staylor 20:10 I see
so I see that firebreath has support for keyboard/mouse events, does that mean that the input handling can be cross platform event through I implement the drawing per-platform? the same view can have platform specific drawing code but cross-platform input code?
taxilian 20:10 that is the theory, yes
staylor 20:10 but does it work? say for gdi using bitblt to draw to the window I get from getHWND() I could use events from firebreath instead of using SetCapture(hwnd)?
taxilian 20:10 you might have to setcapture to get the events
I have a coregraphics bitblt somewhere
with a decent abstraction you can keep the platform specific code pretty clean
staylor 20:10 just trying to figure out if I will write platform specific input handling or use firebreath api, I'm just targeting windows for now but the less code I have to write for linux/mac the better
taxilian 20:10 not usually too hard to refactor to be cross platform
if you use firebreath's event abstraction then that code should work the same
gotta run for now
taxilian 23:10 I love it when things work the first time without needing to fix anything
FireBreathBot 23:10 Commit e95018d on master by Richard Bateman: "Added support for multiple WiX installers (per-machine and p..."
Commit e95018d on firebreath-1.6 by Richard Bateman: "Added support for multiple WiX installers (per-machine and p..."