IRC Log Viewer » #firebreath » 2011-03-04

IRC Nick Time (GMT-7) Message
FB_GitHubBot 00:03 FireBreath: master Richard Bateman * 9902d2f (3 files in 1 dirs): Fixed issue #157 - onWindowAttached called excessively on X11
FireBreath: master Richard Bateman * 37c5660 (7 files in 4 dirs): Fixed line endings
FireBreath: master commits 8efb630...37c5660 -
FireBreath: firebreath-1.4 Richard Bateman * 47f0de6 (3 files in 1 dirs): Fixed issue #157 - onWindowAttached called excessively on X11 -
kalev 00:03 taxilian: not me, I just have a simple plugin without any drawing
taxilian 00:03 kalev! long time no see!
kalev 00:03 :)
taxilian 00:03 you know a bit about gtk, though, don't you?
kalev 00:03 sure, what's the problem?
taxilian 00:03 I'm trying to figure out how to initialize an opengl context
the GLX functions seem to need a Display
how do you get a Display from a GtkNativeWindow?
kalev 00:03 not sure, but this page should list all the relevant functions:
taxilian 00:03 thanks!
that's helpful
have you used 1.4 at all?
not sure if you're still working with plugins much, since I haven't seen you talk in awhiel =]
kalev 00:03 yeah, I haven't worked on the plugin for a while
but I'm thinking of giving 1.4 a try soon.
taxilian 00:03 okaoy
I'm debating if I should release it now
or wait another day
I just fixed the last issue I have been aware of
kalev 00:03 if you want me to, I could give it a try today so you can release tomorrow.
taxilian 00:03 I would appreciate your input
I don't know how much this has been tested on linux
because all the linux people I know of have been quiet lately =]
there are sure a lot of people using it these days, though;
so far we've had 200 or more hits to the website per day every day this week; most days the last 2 weeks as well, other than weekends
kalev 00:03 Oh wow, that's good news
taxilian 00:03 yeah; pretty crazy. lots of new people around lately as well
though not as many as the traffic would suggest
62 responses to the survey so far
kalev 00:03 are there any new people who do code contributions too?
taxilian 00:03 I've had a few
nothing major lately; NoAntzWk implemented POST support for streams, that's about the biggest we've had in awhile
mostly the features people are looking for seem to be there, so there have mostly been bugfixes
1.4 is a pretty major update
fixes a lot of things both in code quality and stability
kalev 00:03 I guess any little bit helps and it's pretty normal that people mostly contribute bugfixes
taxilian 00:03 yeah
well well well.
taxilian 02:03 kalev: you might find this interesting:
it doesn't work yet
crashes the browser
I'm sure I'm doing something dumb still =]
(or good day as the case may be)
alex____ 03:03 hey there
Anyone experienced the following on OS 10.6.5 (Xcode4);
./ line 10: 69198 Segmentation fault cmake -G "$GEN" -DFB_PROJECTS_DIR="${PROJDIR}" ${ARCH} "[email protected]" "${FB_ROOT}"
grapkulec 04:03 hi all
quick question: is there any way of configuring cmake to make my plugin project able to work with .net assemblies?
or do I need to use COM automation and deal with all problems that come with that?
taxilian 08:03 Everyone please test the 1.4 branch extra hard today; I think I will release it on Monday
it should be ready now
maybe I'll even tag an RC3
johnd 09:03 hey there
taxilian 09:03 howdy
aren't plugins fun? with all the weird browser quirks you have to work around?
johnd 09:03 I few days ago I was querying the dynamic way to create <object> in JS... it was not working on IE
I think I figured that out, in case anyone else cares
taxilian 09:03 how is that?
Lta: have you done any opengl on X11 with plugins?
johnd 09:03 In the attach-event, the size was set as 0,0. Then I get a resize event alter... I guess my rendering was not happy with a 0x0 window
taxilian 09:03 johnd: yeah, that's not uncommon, actually
what platform?
wait, duh
misread =]
windows, of course
what kind of rendering
johnd 09:03 it's actually OpenGL but through the Ogre engine... I haven't yet corrected it to test, but I'm guess that is the issue. I'll verify now
taxilian 09:03 quite possible; I've seen something like that before, actually
wonder if that might be what is hitting me on the linux opengl example I was working on last night
all I know for sure is it crashes the browser =]
johnd 09:03 I also have been looking a bit at SWFObject library to see if we can make a generic plugin library from it. Seems feasible, the benefit is it has workarounds for loads of corner cases and browser quirks
taxilian 09:03 agreed; that would be a good thing
that's what I eventually intended the fbinstaller.js to be
just haven't had time or enough reason to work on it
I actually added some more stuff to it
if you'd like to test it
we could even slip it into 1.4 if you can verify that it works =]
I don't have anythign right now that I can test it with
johnd 09:03 can give a look, I'm playing with it at the moment/this weekend
taxilian 09:03
I only put it in master for now
because it's not tested
johnd 09:03 btw, yes on FF the dynamic <object> gets width set immediately. but not on IE (8). I guess I'll defer rendering until I have a valid window... I don't handle resizes at all so far
taxilian 09:03 if you wait for ResizedEvent and then check the dimentions you should be okay
I wouldn't assume that FF will always do it either :-/
FB_GitHubBot 09:03 FireBreath: master Richard Bateman * d8ac980 (6 files in 2 dirs): Updated npapi headers to latest -
johnd 09:03 What on earth is FB_GitHubBot?
taxilian 10:03 it's the github bot that pops in and lets us know anytime someone pushes something to the repo
how does everyone feel about the state of 1.4? should I release it now?
nitrogenycs 10:03 I got a little problem upgrading. Before I got the latest firebreath from git I was on some 1.3ish git. I followed much of the things here: . Also removed my custom boost and let firebreath get boost via git.
I upgraded the projectDef.cmake and the CMakeLists according to what I found in the example plugin
Now I get errors like "boost/shared_ptr.hpp": No such file or directory when compiling my files
it doesn't happen when compiling the firebreath "core" stuff, just with my own project files
obviously it's missing some include paths
I wonder how to get those into cmake and also wonder if they shouldn't be already there assuming my cmake files are setup properly
also fails on FactoryBase.h, PluginEvent.h and JSObject.h etc
johnd 11:03 I had a boost problem too, I deleted the boost folder and and re-prepped and it worked.
might not be the same thing
nitrogenycs 11:03 hmm, it's not just boost I think, it's all kinds of include pathes. My own project doesn't get the include paths it should. Dunno if it's related to the fact that I have an out-of-tree project
nitrogenycs 11:03 btw, is anybody using a tool like SWIG to interface their c++ code with javascript?
taxilian 11:03 nitrogenycs: are you on master or firebreath-1.4 branch?
nitrogenycs 11:03 master
taxilian 11:03 nitrogenycs: not sure what SWIG is
then you're on 1.5, not 1.4
and there are additional breaking changes
including a new way to do include paths for your project
it's on the wiki, hang on
ahh. no, I haven't heard of anyone using SWIG
might be interesting, though
nitrogenycs 11:03 ahh found it:
taxilian 11:03 yep
nitrogenycs 11:03 I didn't see the subpage
taxilian 11:03 I should add a link from the version history page
nitrogenycs 11:03 right, I already removed common.cmake from my files :)
taxilian 11:03 there were a lot of weird cmake things that I have wanted to fix for a long time
this really really really simplifies the cmake project
and makes it possible to use firebreath as a subproject
doubletap 11:03 yo!
taxilian 11:03 welcome
doubletap 11:03 my boys are tentative to try out firebreath. i am here to watch and see what i can discover to get them to try it out.
taxilian 11:03 what are you thinking to do with it?
doubletap 11:03 everything we already do :)
wallet, pswd manager, etc
taxilian 11:03 you do currently with a plugin?
nitrogenycs 11:03 duh, I had all the changes already done because I copied from the example plugin, but the overlook the single line with the include dependencies....
taxilian 11:03 nitrogenycs: lol =]
did you see that we have POST support in browserstreams now, btw?
nitrogenycs 11:03 yes, I read on the list and even looked through the changesets
doubletap 11:03 taxilian: I do all the client side stuff. There are others doing the npapi stuff.
taxilian 11:03 doubletap: okay; just making sure you're already using npapi or plugins or something. so the question is whether FireBreath would be better than maintaining your own browser plugin stuff?
doubletap 11:03 well, we are changing things up with more stuff moving to the client side but having all the variations is a pain across the platforms.
so, it is an ROI evaluation.
nitrogenycs 11:03 taxilian: the post stuff looks good, I might take a look at it quite soon
doubletap 11:03 what is the migration like?
taxilian 11:03 doubletap: migration depends a lot on how heavy your abstraction is
johnd 11:03 is there any kind of roadmap for 1.5, 1.6, etc or is it dictated by what you feel like doing?
doubletap 11:03 i think we are pretty close to the floor for each platform.
taxilian 11:03 doubletap: then it probably wouldn't take long. it all depends on how much you're relying on things that are specifically NPAPI. I suggest looking at the video tutorial; you can literally have a simple plugin up and running easily in under an hour
if you do a lot of drawing, that can be trickier sometimes to port
but not terrible
the APIs are far, far easier in FireBreath than raw NPAPI (probably FB's greatest strength), but it may take some time to port the existing code to use normal C++ types instead of NPVariants
nitrogenycs 11:03 taxilian: swig is a nice tool, I used it to wrap our c++ engine to python. Unfortunately there's no working javascript module out there already. It would be cool if there was one to generate firebreath jsapi stuff automatically for a given c++ header file. However, I don't have the time to do it right now. Instead using some IPC between my plugin and the python wrapper. The python wrapper can inspect itself and create some javascript on the f
from that. Then I can call my engine from javascript through python. That's not performant in any way, but it's ok for a start.
taxilian 11:03 johnd: not really; currently I don't have much planned for 1.5 other than what is in it. right now I'm implementing things as they become important to me or requested a lot
johnd: I think right now we almost need examples more than we need features on FB itself; bugfixes are mostly the order of business at the moment.
nitrogenycs: that would be cool; did you know that you can actually register a method on another class? it doesnt' have to be the JSAPI class
so some people have created a subclass that knows nothing about FireBreath and registered methods on that class; created small wrappers where needed
johnd 11:03 as a noob I'd probably agree with that, examples and docs can go a long way to attract a user base
doubletap 11:03 thank you, taxilian.
taxilian 11:03 doubletap: if you have questions, just let us know. FireBreath is actually pretty solid at this point; bugs are usually fixed pretty quickly, and anything missing can be added
the architecture is pretty well designed, in my (completely unbiased ;-)) opinion
doubletap 11:03 great. i would love to get our team moved. i am chatting with my boss about it.
taxilian 11:03 it is being used (so far just in development) by many companies; Echostar Advanced Technologies (subcompany of Echostar) and Facebook among them
doubletap 11:03 taxilian: any thoughts on the possibility of the mobile plugins like in FF or some next gen Safari?
taxilian 11:03 johnd: I started an opengl example project last night; not sure how fast I'll get it running, but feel free to contribute your own examples as well =]
nitrogenycs 11:03 taxilian: no, I didn't. Right now I have only a handful of api methods anyway. Mainly one to send an event from javascript to the process I launch with the plugin and one api method to go the other way (send event from python process to javascript). That way I can do lots of stuff without changing the plugin api
taxilian 11:03 doubletap: if it supports NPAPI it'd probably be doable and relatively straightforward; if it is a completely different API it is probably doable but harder
nitrogenycs: just thought you'd find it interesting =]
doubletap 11:03 cool
taxilian 11:03 doubletap: seriously, though, try the tutorial; FireBreath itself is the best marketing I can show you
doubletap 11:03 i wont try it. too much to do. but, my boss is going through your classes now :) I am just an advocate.
taxilian 11:03 it is almost ridiculously easy to get simple things to work; complex things are mostly complex simply by themselves, and there are of course things you just have to know about living in a browser; but your NPAPI guys will know those things
tell him to pop in if he has any questions
doubletap 11:03 will do
nitrogenycs 11:03 taxilian: sure I do, it's cool :)
irc having really severe hickups today, got no messages for 3 minutes, then the last 20 at once
taxilian 11:03 nitrogenycs: the funny thing is I had never even thought of that =] it wasn't designed to work that way intentionally, it just does
nitrogenycs 12:03 taxilian: did the config.h go away?
taxilian 12:03 no, but it moved. global/config.h
that's one of the 1.4 breaking changes
nitrogenycs 12:03 the global folder is supposed to be inside my project directory?
taxilian 12:03 global is a generated folder
it's in build/projects/<plugin>/gen/global
nitrogenycs 12:03 ahh right, there was a generated folder too
I need to get up to speed again, haven't been to fb for a couple of weeks/months
taxilian 12:03 heh. no problem
nitrogenycs 12:03 suggestion: include those files into the "Generated" folder in the visual studio solution
taxilian 12:03 not a bad idea
file an issue so I don't forget =]
or submit a patch, whichever you prefer =]
nitrogenycs 12:03 I won't touch cmake if I don't have to :)
taxilian 12:03 lol
it's not that bad, just takes some getting used to
nitrogenycs 12:03 woo, my plugin compiles again with master
:) thanks taxilian
taxilian 12:03 =] no problem. let me know how it works
nitrogenycs 12:03 yes, runtime testing pending
taxilian 12:03 I have a new coding philosophy: "Always code as if the guy who ends up maintaining your code will be a violent psychopath who knows where you live."
nitrogenycs 12:03 kudos to whoever did the firebreath_sign_plugin macro
taxilian 12:03 that was me =] thanks
I'm proud of that one =]
nitrogenycs 12:03 :) I was just trying to figure out how add_custom_command in cmake works when I found it
taxilian 12:03 lol
nitrogenycs 12:03 is there a way to build my project via cmake from the command line (without opening the visual studio solution first)?
taxilian 13:03 cmake —build <build dir>
req cmake 2.8 or later
nitrogenycs 13:03 taxilian: ahh thanks. I also found another way to start visual studio in the background and make it build the app.
taxilian 13:03 yeah, you can call devenv.exe from the command line and tell it to build as well
but the cmake way is cross-platform
nitrogenycs 13:03 instead of running devenv.exe just run
yes, I am going to pursue the cmake
taxilian 13:03 (it does devenv behind the scenes)
nitrogenycs 13:03 thanks taxilian, that worked wonderfully
Lta 13:03 taxilian: no, haven't done any opengl with firebreath. Never used opengl btw