IRC Log Viewer » #firebreath » 2010-12-16

IRC Nick Time (GMT-7) Message
Tony_ 01:12 hi I have a problem with mac version, I tried to include some objective C source to my project, it compiled fine, except a for loop error.But when I try to link to a third party static library, I am getting error of unresolved symbols
any idea.. I have added the static library to the project by dragging it into xcode ..
stringa 01:12 anyone here?
Tonku 05:12 hi
Tonku 06:12 how to link external static library in mac project ?any idea..?
neilg_ 07:12 Tonku: Still stuck?
Tonku 07:12 :(
I am trying to link a to a C function from my own source put in mac folder..this function is calling from screencastpluginapi.cpp file
I am, getting linker error in xcode.I am illiterate about cmake. developing first time for mac , me spoiled by the VisualStudio :)
neilg any idea ..
RJanicek 07:12 Is there a way to organize files in the projects directory into sub directories?
taxilian 08:12 Yes
RJanicek: Look at how projectDef.cmake includes files in the platform dir
It is included from the parent dir
RJanicek 08:12 Ok thank you I will look at that.
taxilian 08:12 Tonku: Look at the linking libraries page on the wiki
Tonku 08:12 ok thanks I will lokk into that
taxilian 09:12 morning all
anyone here still having problems getting WiX to work?
FB_GitHubBot 09:12 FireBreath: master Richard Bateman * c6f5dce (1 files in 1 dirs): Fixed typo in wix.cmake that in some cases prevents WiX from working correctly -
FireBreath: master Richard Bateman * 712eff8 (4 files in 1 dirs): Updated JSAPISecureProxy to support weak_ptr proxying -
taxilian 09:12 hmm. question #2: anyone here using vs express?
leo 09:12 hi
taxilian 09:12 howdy
leo 09:12 here testing firebreath
i have a doubt
taxilian 09:12 a doubt? impossible! ;-)
leo 09:12 XD
taxilian 09:12 you probably mean a concern, however
leo 09:12 maybe
taxilian 09:12 tell us about it, and I'll let you know ]
leo 09:12 i wanna draw stuff on the plugin
but when i access to the width and heigth
allways obtain 0
trying to change the window with a opengl window
taxilian 09:12 you are drawing on Mac?
leo 09:12 nop
im trying something like this
not exactly
but when i call to GetWindow it allways returns 0 on widht and height
taxilian 09:12 what OS are you on, then? windows?
leo 09:12 linux
at getWindowPosition i mean
FB::PluginWindowX11 *pwnd = dynamic_cast<FB::PluginWindowX11*>(pw);
taxilian 09:12 that's a really important thing for us to know =] to be honest, none of us really know linux drawing very well, so I'm not 100% sure if the linux stuff is set up right
leo 09:12 oh
taxilian 09:12 however, I think Lta might know something about it...
leo 09:12 seems Lta is kinda afk
taxilian 09:12 and it looks like Lta is also elthariel? (I'm looking at old chat logs :-P)
leo 09:12 dual personality?
since im testing this because i wanna make the plugin multiplatform
and i guess PluginWindowX11 works
taxilian 09:12 it should be possible to do… I just don't know how to do it. he said something about changing PluginWindowX11.h so that getWindow reads: getWindow() { return GDK_WINDOW_XID(gtk_widget_get_window(m_canvas)); }
leo 09:12 oh
taxilian 09:12 and adding the include gdk/gdkx.h
so you could try that
tell me if it works =]
Lta 09:12 taxilian, im also elthariel
leo 09:12 ok
Lta 09:12 let met read the log
leo, hi
leo 09:12 hi Lta
taxilian 09:12 Lta: yeah, I hadn't put that together until just now; was reviewing my notes on our conversation about X11 drawing and saw your email address =]
Lta 09:12 it depends on the office from which im connected
taxilian 09:12 I'd actually forgotten about that; we should probably put that in trunk
Lta 09:12 leo, and taxilian. To be honest, i consider X11 drawing as the worst way to draw stuff, so generally, all that matters to me is to grab a valid xid, give it to another framework (qt or gtk) and let it handle all the nasty stuff
taxilian 09:12 lol. that's fine… as long as we make sure it can provide a valid xid =]
Lta 09:12 taxilian, the issue we talked about last time about getWindow has been solved in 1.3.0 afaik
taxilian 09:12 really? I don't remember doing that… :-P
maybe kalev took care of it.
Lta 09:12 i'll check, we are working on other part of the projet curerntly so maybe i didn't checked after 1.3.0
kalev 09:12 nope, I don't think I did
taxilian 09:12 doesn't look like it
kalev 09:12 hmpf, I think GdkNativeWindow is a typedef for xid
let me look that up.
Lta 09:12 kalev, it is
taxilian 09:12 getFactoryInstance()->createPluginWindowX11(WindowContextX11(getGdkWindow(window->window)));
that doesn't look like the right call...
Lta 09:12 imho
getBrowserWindow should return the GtkPlug, and getWindow should return the canvas
taxilian 09:12 that sounds right to me as well
I think most of this I wrote, though, and I don't know much about it
leo 09:12 tested getWindow() { return GDK_WINDOW_XID(gtk_widget_get_window(m_canvas)); } but still didnt work
taxilian 09:12 but from our previous conversation, I agree
Lta 09:12 any drawing you perform on the GtkPlug is useless
taxilian 09:12 the problem is that I can't do the change myself because I have no real way to test it
I can test that it builds, but...
Lta 09:12 leo, this was working for me
kalev 09:12 I added the getBrowserWindow() call and the idea was to use that to get NPNVnetscapeWindow, not the GtkPlug
Lta 09:12 sorry, this is what i meant
taxilian 09:12 what is the difference?
Lta 09:12 gtkplug is the son of NPN...
taxilian 09:12 hmm. which is more useful? can you get one from the other? should we add a second function so that you can get both?
Lta 09:12 i guess gtk_plug_get_socket_window() does that ?
oups .. no
taxilian 09:12 well, if you folks could discuss it and someone produce a patch (or a pull request on github), it would be *really* good to get this a little more stable
Lta 09:12 kalev, is right i think, getBrowserWindow should return NPNVnetscapeWindow
the gtkPlug is not that usefull, and can be obtained from the GtkDrawingArea
taxilian 09:12 sounds good to me
Lta 09:12 leo, sorry i still didn't had a look are your problem
i'll read you link right now
leo 10:12 well
my problem is that im trying to set a opengl viewer inside the plugin window
Lta 10:12 i've read :)
leo 10:12 at the initialization i try to get the plugin window properties and get 0 as width /height that make me thing that something i was doing wrong
im initializing the viewer at onWindowAttached
Lta 10:12 actually i think the window only gets it's height and width when it is 'realized'
when it is actually mapped on screen. So, at initialization time, these values are set to 0 because this is better than random values :)
the question is: is the setWindowPosition called somewhere
leo 10:12 not at my code
at least
ill check it on fb
Lta 10:12 actually it is called here ./src/NpapiPlugin/X11/NpapiPluginX11.cpp:85
but this is the parent ("socket") position and size
so the size should be relevant for the plug, but don't know for the position
anyway, these informations are not available at initialisation time because the plugin can't predict how many space the browser will give it
i guess you should be able to force it, but this is not a good idea
leo 10:12 but when u create the plugin the width and the size of it are 2 of the parameters u give to the plugin
<object id="plugin" type="application/x-fb" width="300" height="300"> <param name="onload" value="pluginLoaded" /> </object><br />
and that corresponds with the canvas that u have later corresponding to the plugin
Lta 10:12 yes, but this information is not know 'a priori' by the plugin
and the browser can change it i guess
where do you call the getWindowPosition method ?
leo 10:12 onWindowAttached
i guessed that i have that data ready at that point
taxilian 10:12 that's just when the browser gives us info about it
there is no guarantee that it's all set up
Lta 10:12 leo, i guess the data should be ready at this point :)
taxilian 10:12 Lta: any chance you could take a look at the current PluginWindowX11.h and .cpp, fix whatever is needed, and send me the altered files? or a patch, whichever is easier for you...
Lta 10:12 i'll send you at least the patch for "return GDK_WINDOW_XID(gtk_widget_get_window(m_canvas))". i'm sorry i didn't doi it in the first place
taxilian 10:12 eh, you're busy
I understand
I'll put it in 1.4; it's enough of a change in how things work that I don't want to put it in a stable branch
Lta 10:12 i think i might have figured out what's wrong leo
the AttachEvent is fired _before_ the call to pluginWin->setWindowPosition
leo, can you try something for me ?
leo 10:12 yup
let me know what it is
Lta 10:12 mm ... i doesn't have a clone, i only have 1.2.2 and 1.3.0 archive
taxilian 10:12 Lta: are you using github?
Lta 10:12 ok, just wait for a few minutes. I'll fork the project and send you the version to test
i am :)
taxilian 10:12 cool =]
Lta 10:12 leo, you'll be able to try. i didn't test it because im so lazy and ill today :-/
leo 10:12 ok
Lta 10:12 if it doesn't fix it i'll have a deeper look
leo 10:12 where it is
Lta 10:12 so you can find my fork here
afk for a few minutes, pancake times :)
leo 10:12 :)
taxilian 10:12 leo: let me know if that seems to work for you
i'll pull it into the trunk
Lta 10:12 ok the code doesn't event build
sorry leo, just wait a few more minutes
taxilian 10:12 lol
Lta 10:12 i'm sorry. i'm sick and start to loose my mind a little bit
taxilian 10:12 Lta: thanks for doing this — hope you feel better soon
Lta 10:12 i hope so, cause i'll start driving lessons next week :)
leo 10:12 gl
on driving lessons
Lta 10:12 taxilian, off topic, on which timezone d'ya live ?
taxilian 10:12 GMT-0700
I'm in Utah, USA
Lta 10:12 ok :D. i'm in france (iirc gmt +1 depending on the season)
taxilian 10:12 cool. I don't think I'm aware of any other users in France (though I'm sure there are some)
I'd give a lot to know how many people are actually using FireBreath
I'm tempted to add a little ping to the cmake prep script :-P but some people probably wouldn't appreciate that
leo 10:12 gmt+1 here too
Lta 10:12 taxilian, actually i wouldn't have appreciated it
the more i crash firefox, the longer it takes to start
taxilian 10:12 yeah… and long term it wouldn't help all that much. it would just be nice to know how many people are using it =]
you could disable the mozilla crash reporter
that would probably help
Lta 10:12 leo, taxilian. i tested the getWindowPosition, and it works on my system
you could now try leo
can i disable that ^&*$#) menu ?
taxilian 10:12 menu?
the crash reporter? yes
Lta 10:12 the crash reporter
taxilian 10:12 in fact, it makes debugging a lot easier
Lta 10:12 sounds fun :)
taxilian 10:12 find Application.ini (depends on your system where it is). there is a setting in there
not sure where to find it on linux
mac and windows are easyish
Lta 10:12 i'll go eat my pancakes :)
taxilian 10:12 prolly in ~/.mozilla
bon apetite
leo 11:12 i wasnt using the git and im downloading the stuff now
u changed it on master branch?
taxilian 11:12 he did, on his fork
stringa 11:12 does firebreath just download the whole game to a temp directory? So when the user is not on the internet, they could browse to the directory and just run the executable?
Lta 11:12 leo, git clone git://
stringa, firebreath is a library intented to developper about browser plugin
stringa 11:12 play my DirectX game in the browser
Lta 11:12 meaning it only runs inside a browser (except for activex i guess, which should be embeddable in other application)
stringa 11:12 I'm curious about protecting my product...
taxilian 11:12 stringa: firebreath doesn't do any of that; it just makes it easy for you to write a browser plugin
you'll have to implement the details
and you can do it more or less however you want
Lta 11:12 you have to implement a security model by yourself, this is not the goal of firebreath
taxilian 11:12 some people have implemented games as you suggest, to where the game exe is just downloaded
stringa 11:12 I can't have that
taxilian 11:12 others implement the game inside the DLL
that's harder, because you live in the browser process
others use a .exe but make it communicate wtih the plugin process so that the plugin process has to be running for the game to work
leo 11:12 Lta: it worked
Lta 11:12 leo, nice :)
taxilian 11:12 some even have the plugin process give the game process the window to draw in, so that it is all inside the browser
stringa 11:12 thanks for the info
taxilian 11:12 it's just a windows application — you have to take care of the details
Lta 11:12 did you also succeed in drawing on it
leo 11:12 not yet
just tested that i can get width and height
is kinda late here
i think ill draw it tomorrow and let u know
Tonku 11:12 hi taxillian..
Lta 11:12 it is :)
leo 11:12 thank you very much
so dark
hate being at U this late
Tonku 11:12 is this the way to set additional libs to the xcode project..?
set (LIBRARY_PATH ${FireBreath_SOURCE_DIR}/path/to/library) target_link_libraries(${PLUGIN_NAME} debug "${LIBRARY_PATH }/debug/library.lib") target_link_libraries(${PLUGIN_NAME} optimized "${SANDSTONE_DIR}/release/library.lib")
taxilian 11:12 hehe.
Tonku: more or less, yes
obviously make sure the library paths are all correct
and remember that has to happen in the cmake files after the library is added (the target that you're adding libraries to)
Tonku 11:12 Pls excuse me I am novice ..I have edited my cmake file in the Mac folder like this and I am getting wired error on linking :(
set (LIBRARY_PATH ${FireBreath_SOURCE_DIR}/projects/AppsharePresenter/mac) target_link_libraries(${PLUGIN_NAME} debug "${LIBRARY_PATH }/libturbojpeg.a")
I get this error..Command /Developer/usr/bin/g++-4.2 failed with exit code 1
I tried to drag and drpop this lib into my xcode project and it failed to link :(
taxilian 11:12 okay, but what is the message associated with that?
don't modify the xcode project directly
Tonku 11:12 i686-apple-darwin10-g++-4.2.1: /libturbojpeg.a: No such file or directory
I put this lib in the Mac folder
taxilian 11:12 probably because you have a random extra space in your ${LIBRARY_PATH }
it should be ${LIBRARY_PATH}
Tonku 11:12 ok let me correct it and try again
Lta 11:12 taxilian, i'm testing the getWindow() changes and i'll send you the pull request
regarding getWindowPosition(). I works on my system, it worked for leo, i hope this is enough testing
taxilian 11:12 absolutely
Lta 11:12 taxilian, the getWindow() stuff is working
taxilian 11:12 awesome
I'll pull it now
Lta 11:12 taxilian, i never sent pull request
taxilian 11:12 that's okay, I know where it is
Lta 11:12 so don't hesitate to check
another stuff, i'm sorry i never fixed my emacs configuration, there is some whitespace changes :-/
FB_GitHubBot 11:12 FireBreath: master Julien 'Lta' BALLET * 0410344 (2 files in 1 dirs): PluginWindowX11, getWindow now return m_canvas XID (a GtkDrawingArea), insted of browser window
FireBreath: master Julien 'Lta' BALLET * 35857e4 (1 files in 1 dirs): NpapiPluginX11, window's position and clipping wasn't set when firing AttachEvent
FireBreath: master Julien 'Lta' BALLET * de14339 (1 files in 1 dirs): Little typo, replaced gtk/gtkx.h by gdk/gdkx.h
FireBreath: master commits 712eff8...de14339 -
Lta 11:12 wow ! never seen the github bot as well.
taxilian 11:12 useful, no? =]
Lta 11:12 it's kind of a tornado
it is
FB_GitHubBot 11:12 FireBreath: master Richard Bateman * 445d124 (75 files in 22 dirs): Fixed whitespace errors (tab->space) -
Tonku 11:12 Hi Taxillian It linked :)
thanks alot
Lta 11:12 taxilian, ty for accepting my pull request :). This getwindow stuff is finally a resolved problem
taxilian 11:12 Lta: the pull request is done and closed
FB_GitHubBot 11:12 FireBreath: cmake_patch Richard Bateman * 5f8918d (2 files in 2 dirs): Made changes to work with patched cmake. Hopefully this patch will be in cmake 2.8.4 -
taxilian 11:12 and thank you!
Lta 11:12 you are welcome, this is nice not to have to maintain small patchsets everywhere
i'll leave you, it's nearly 8 o'clock here
taxilian 11:12 yeah; seems everyone has an easier time doign things like this with github
Lta 11:12 i use it for most of my projects now, and when i don't, i try to have an synchronized github mirror
Tonku 11:12 hi Taxillian me again..
I am trying to link a to a C function from my own source put in mac folder..this function is calling from screencastpluginapi.cpp file
but I am getting linker error in xcode..any clue..
taxilian 11:12 Tonku: I'm not sure I understand; if the function you need is in your code, and it is being built in the project and you still have link errors, it's likely an issue with your .h file not matching the .cpp
Tonku 11:12 initially I was trying to link a C function on a .m file from the cpp file (screencastpluginapi.cpp equivalent), I called this c function with extern..I got linker error
then I made a .h and .c files for the purpose and included this .h file in the .cpp file..
taxilian 11:12 is the .m file showing up in the xcode project?
and being compiled?
Tonku 11:12 yes this m file is being compiled, and the lib i was trying to link was for this .m file
Actually I have manually added these files to the xcode project , after placing these source file in mac folder
taxilian 11:12 then you probably have a problem in your header file
don't do that
add the file by editing the projectDef.cmake file
Tonku 11:12 pls advice me the correct way
ok any wiki page for adding source files ..?
taxilian 11:12 probably not; just look at where it is adding Mac/*.h, Mac/*.cpp, and add a Mac/*.m
Tonku 11:12 I did that already
taxilian 11:12 and if you'd like you're welcome to add a wiki page when you are done =]
did you run again?
Tonku 11:12 yes ..
I will give you the edited code
taxilian 11:12 then you shouldn't have to add the file by hand...
pastebin it, don't paste it here
Tonku 12:12 pastebin..?I am so illiterate infront of you :(
how to do patebin..?
taxilian 12:12 is a good one is the original (AFAIK)
Tonku 12:12 ok thank you
here is it..
now shoud I manually remove the manually added fiels from xcode project..?
taxilian 12:12 no, every time you run it should make you reload the project
that liks like it should be fine
Tonku 12:12 before running the shoud I delete the build directory in firebreath folder..?
taxilian 12:12 you shouldn't need to
but if xcode doesn't ask you to reload the project, you'll want to do so yourself
Tonku 12:12 is there any problem if i delete the build folder before running .will the script regenerate everything..?
taxilian 12:12 it will
Tonku 12:12 let me try now ..
kylehuff 12:12 is building on linux from trunk supposed to be broken ATM?
taxilian 12:12 is not supposed to be
but we did just push some changes
let me check
running a test build on both lin32 and lin64
eh, and mac and windows sjust for fun
both build fine for me
kylehuff: there were some changes that could have potentially been breaking to the way the drawing stuff works on x11
kylehuff 12:12 I see
taxilian 12:12 hmm. are you using FB_GUI_DISABLED, by chance?
kylehuff 12:12 yes
taxilian 12:12 hmm. that could be why; most likely Lta didn't take that into account when he changed things this morning
let me look
kylehuff 12:12 ah, okay.. makes sense..
taxilian 12:12 kylehuff: I'll have the fix up in a minute
and thank you for developing against trunk so that we find issues like this quickly =]
kylehuff 12:12 ok, taxilian; I will pull+test as soon as I see the commit
FB_GitHubBot 12:12 FireBreath: master Richard Bateman * 5fc28d3 (1 files in 1 dirs): Fixed FB_GUI_DISABLED flag to work again -
taxilian 12:12 I really like the fb_githubbot =]
kylehuff 12:12 it is handy
(building now)
can you specify the release from the CLI on linux? i.e. MinSizeRel, etc?
taxilian 12:12 yes
you have to pass it into cmake...
hang on, I'll find it… =]
kylehuff 12:12 ok, I can look for it
not a big deal yet
kalev 12:12 I think it was -DCMAKE_BUILD_TYPE=Release
taxilian 13:12 that sounds right
kylehuff 13:12 cool, thanks kalev
speaking of flags
why do I still see 'setenv ARCHS "i386 x86_64"' in the PhaseScriptExecution after adding the -DCMAKE_OSX_ARCHITECTURES=i386 and using the CLI command 'xcodebuild ARCH="i386"'?
on OSX
kalev 13:12 did you try removing the build directory and creating it from scratch with that flag?
kylehuff 13:12 no, I haven't. I will do that now
taxilian 13:12 hey, while I have so many of you here
I finally got the cmake book, and I'm (very) slowly finding time to read it
FB_GitHubBot 13:12 FireBreath: master Kalev Lember * d1d0f03 (1 files in 1 dirs): NpapiPlugin: Fixed the order of initializers in NpapiStream constructor ... -
taxilian 13:12 we are really not using cmake correctly =]
hey, thanks, kalev
I appreciate you keeping an eye on that
kalev 13:12 that was my gcc warning fix for today :)
taxilian 13:12 anyway, I'm really kinda wanting to change things so that we don't have prep scripts (except on mac until the patch makes it through) and we use cmake for the configuration
probably also change it so that you run cmake on the directory of the plugin project which includes the firebreath cmake stuff, instead of the other way around
however, that would be a pretty significant change in how everything works
which could confuse a lot of people
nitrogenycs 13:12 I wasn't even aware that wrong initialization order would give a warning
taxilian 13:12 nitrogenycs: it does on gcc
kalev 13:12 yes, sounds like a very good plan (if you remember I pointed it out that I'd like it that way around when you last reworked the prep scripts)
the prep scripts might still be handy, but running cmake inside a plugin folder would be really nice improvement
taxilian 13:12 yeah; I just didn't know enough to do it then. I've been thinking about it, though =]
well, I'd still recommend to do an out-of-source build
but yeah
it's also much less important to be able to have multiple plugin projects together now that we support multiple mime types
kylehuff 13:12 kalev: specifying that flag with the prepmac script seemed to fix the problem; thanks
kalev 13:12 kylehuff: glad I could help
taxilian 13:12 so I think I'll at least modify things so that you can run a prep script on a signle plugin project instead of a directory with multiple projects
but I'm trying to decide if it is worth breaking all the docs and the workflow to change it to use cmake the "right" way
kalev 13:12 yep, building a single plugin project instead of the directoryful is a nice first step
taxilian 13:12 that I could probably do in a few minutes
kalev 13:12 anyway, turning the build system upside down like that also gets us a step closer to having Firebreath as a proper library
taxilian 13:12 just didn't before because we didn't have the multi-mimetype thing going
true; that's still a ways off on windows in particular, however
there are a *lot* of things that have to be generated currently
and I don't see a way to change that
so we could make it a better "cmake-enabled" library, but I don't see any way we could remove cmake from the picture
kylehuff 13:12 this really doesn't matter, mostly just curious; but why does this work: API->FireEvent("onkeygenstatus", FB::variant_list_of("complete")); but not API->FireEvent("onkeygenstatus", "complete"); ?
sorry, by "work" I mean, compile..
taxilian 13:12 because FireEvent takes an array of arguments
kalev 13:12 VariantList is just a typedef of std::vector<variant>. If VariantList was a class of its own, then it could conceivably have a non-explicit constructor from std::string and the conversion would happen automagically. But as it is only a std::vector, there is no such constructor.
taxilian 13:12 a vector
FB::variant_list_of() creates a vector
kylehuff 13:12 okay, that makes sense..
thanks taxilian
taxilian 13:12 you can also use it like FB::variant_list_of(shared_ptr())("complete")(72.4)(42)
and it will create a vector with all of those items
kylehuff 13:12 cool
taxilian 13:12 I guess we could add an overload to accept just one parameter, but that sounds like a lot of work for something easy to do with different syntax =]
kylehuff 13:12 yeah, I was just curious - it makes sense the way it is
does `xcodebuild` not generate the project.plugin package?
taxilian 13:12 I would expect it does
though it's easier to use cmake —build build/
kylehuff 13:12 it generated the .so, but I don't see the .plugin anywhere
taxilian 13:12 (that's two - -)
the .so?
did you use or
kylehuff 13:12 prepmac
taxilian 13:12 there shouldn't be a .so anywhere, then...
Tonku 13:12 hi taxillian... I have a c style function on one of my .m file.How can I call that function from my Auto-generated AppsharePresenterAPI.cpp.I tried to call using extern , but I am getting unresolved symbol error..:(
taxilian 13:12 however you would normally call a c style function, I would assume
I'm not sure how that is
Tonku 13:12 Sorry I am not getting you ..
kylehuff 13:12 using xcodebuild (which I don't think would work if I had done, as it wouldn't generate the xcode project) - it put into build/projects/gauPlugin/MinSizeRel/
taxilian 13:12 Tonku: that is not a cmake or a firebreath question; that's a question on how to call a c function from a cpp file. I don't know the answer
kylehuff: weird...
I have no idea what would cause that...
kylehuff 13:12 I will use the xcode IDE to generate it and see if I get something diffent
Tonku 13:12 Ok thaxillian..I will try to findout how to do it
kalev 13:12 Tonku: in your Mac/projectDef.cmake file there's a glob which starts with this: file (GLOB PLATFORM RELATIVE
Tonku 13:12 ok...
kalev 13:12 I bet it's currently only adding .cpp files to the list; try adding *.m in there too
taxilian 13:12 kalev: he already has the file included in the project
kylehuff 13:12 I'm gonna go find food, I will let you know when I get back what the result is using the IDE build
taxilian 13:12 he just can't get it to link when he tries to call the function in his .m file, which is a c style function
Tonku 13:12 I am getting my.m files compiles fine..
kalev 13:12 ah. no idea then.
Tonku 13:12 I was trying to make a link between the firebreath module and my source
kalev 13:12 anyway, I'm a lousy mac programmer, but I got my .m and .mm files working, so if you want to take a peek:
Tonku 13:12 ok :) thanks guys for the help.. btw firebreat is really cool its great :)
If I firebreath was not there I would have to compromise to a webkit plugin , that works only in safri and chrome :)
kalev 13:12 yep, I agree: Firebreath is awesome
taxilian 13:12 I do think it would be cool to make it so that FireBreath could create webkit plugins as well, though
amackera 13:12 Oo that would be cool
How different are they from NPAPI?
taxilian 13:12 oh, more or less completely
but so is ActiveX
you'd basically just need to add some extra info when we register the JSAPI methods because it'd have to create an ObjC object dynamically to wrap the JSAPI objects
but most everything else would probably be close enough it wouldn't be a big deal
kylehuff 14:12 taxilian: that works, I can build my plugin on linux using FB_GUI_DISABLED with the latest commit
taxilian 14:12 awesome
kylehuff 14:12 I am still trying to end up with a .plugin on OSX; I rebuilt the build directory and am waiting for it to compile now
was an issue ever created regarding the hard-coded file-version in firebreathWin.rc always being 1,0,0,1? I will create one if not. (I'm on the issues page and I don't see it..)
taxilian 14:12 I think there was an issue created
but the reason I haven't worked on it
is because I'm honestly not sure how best to fix it
and I'm open to ideas
how would you expect to change that?
kylehuff 14:12 okay, I just wanted to make sure it got opened. as for fixing it; I don't know - I don't (yet) know how that file is even generated. if I can think of a way to correct it, I will fork it and report back
taxilian 14:12 oh, that's easy enough
the file is in gen_templates/
it just replaces any cmake variables and copies it to the gen directory
the problem is how to define that version string from in PluginConfig
or where to get it otherwise
you don't even have to fork; just copy the file into your plugin dir
and it'll use that one insetad
you can then change the version yourself
or come up with a better way to do it
without even changing fb files
kylehuff 14:12 I currently maintain the version in PluginConfig.cmake via set(PLUGIN_VERSION, "x.x.x")
taxilian 14:12 but if you put that into the rc file it'll break
because it wants a specific format
it would have to be x,x,x,x
kylehuff 14:12 ah, okay.. (I haven't even tried to mess with it yet..)
taxilian 14:12 you could easily enough automatically create another variable like "PLUGIN_VERSION_RC" that just replaces "." with ","
but you'd still need 4
at least AFAIK
marrone 14:12 curious to know how our company can license your plug-in?
taxilian 14:12 marrone: you don't need to license anything; FireBreath is available under the new BSD license
basically that states that you can do just about anything you want as long as you don't remove the copyright notices (I am not a lawyer, that's just my interpretation)
also, FireBreath isn't actually a plugin, just a framework for making plugins. If you tell us a bit more what you're hoping to do, we might be able to give you some better guidance
kylehuff 14:12 okay, I deleted /build and re-ran, and this time compiled it using the xcode IDE - same result - in /build/projects/gauPlugin/Release there is only
taxilian 14:12 you're using branch master?
kylehuff 14:12 taxilian: no, branch cmake_patch
marrone 14:12 very good, thank you
kylehuff 15:12 should I be using master?
taxilian 15:12 kylehuff: and you're using the patched cmake?
kylehuff 15:12 taxilian: yes
taxilian 15:12 hmm
I don't know; we may have lost something in that branch when I rebased it
kylehuff 15:12 I am behind 4 commits apparantly..
taxilian 15:12 well, try a git pull
I rebase it periodically to get the latest and so that I don't have merges everywhere
kylehuff 15:12 yeah, I thought I had done that, but apparently it failed and I must not have noticed
taxilian 15:12 you might have to pull --force
honestly, I'd forgotten anyone was on that branch
FB_GitHubBot 15:12 FireBreath: cmake_patch Richard Bateman * 201af06 (2 files in 2 dirs): Made changes to work with patched cmake. Hopefully this patch will be in cmake 2.8.4 -
taxilian 15:12 kylehuff: I just pushed changes again, so make sure you get them
I'm doing a test build to make sure it works
kylehuff 15:12 I am sooo out of memory...
taxilian 15:12 lol
I probably shouldn't tell you what is on the desk I'm sitting at...
kylehuff 15:12 shouldn't tell me because it is illegal? or I might hate you forever? (or both? lol)
taxilian 15:12 I have 2 laptops (both 8GB ram, macbook pros) 1 27" imac w/ 6GB of ram, a 23" 1080p screen, and a 37" 1080p screen...
though granted at the moment I'm only using one of the laptops and the imac
kylehuff 15:12 oooo nice..
I am on an intel 1.2GHz laptop with 2GB of ram, running Debian linux as the host with a WinXP guest and an OSX guest..
at one point I had my plugin building in all three o/s'
taxilian 15:12 lol
not bad, considering =] that's really impressive
kylehuff 15:12 okay, I am compiling again with the fresh pull and subsequent (for good measure)
taxilian 15:12 it worked for me, btw
kylehuff 15:12 cool
I actually forgot what I need to test under OSX now...
taxilian 15:12 lol
kylehuff 15:12 ohh, I remember.. I have a boost::thread that on return, was causing the plugin to crash
(only on OSX)
I think I found the problem, but I might be picking your brain about it if not.
taxilian 15:12 ok'
kylehuff 15:12 taxilian: same result; I only have a .so and no plugin bundle under /build/projects/gauPlugin/Release
taxilian 15:12 kylehuff: I can only conclude that you probably aren't using the patched version of cmake
but it doesn't sound like you really need to
so it might be better to just use master
kylehuff 15:12 I will checkout master and try again
taxilian 15:12 the only thing the patched version and cmake_patch give you is the ability to use makefiles, which you aren't using
taxilian 15:12 amackera: do you have an ETA on when you can look at IE windowless? as soon as that is in I want to lock down features on 1.4
and put it in beta
kylehuff 16:12 taxilian: master builds the .plugin as intended
taxilian 16:12 kylehuff: you're not using the patched version of cmake, then
shouldn't matter for what you're doing
amackera 16:12 taxilian: Hm... I will really try to get to it this weekend
(on Saturday)
taxilian 16:12 that would be awesome
I'll be at Facebook next week, Mon-Wed, so don't know how much I'll be around those days
kylehuff 16:12 taxilian: I don't know I how I could not be, I completely removed cmake and built/installed it from the github repo
taxilian 16:12 when I get back I'll rpboably not be around much
kylehuff: I don't know either =]
sabotaged|wk 16:12 anyone have any idea how to get the path of the firebreath plugin at runtime?
mac and windows, though i imagine it might be platform specific code
taxilian 16:12 yeah
m_filesystemPath in your plugin object
there is an example in FBTestPlugin
sabotaged|wk 16:12 oh sweet
oh the docs say only supported by windows?
ah GetModuleFileName
kylehuff 17:12 okay, so from trunk:master I should be able to build the plugin with FB_GUI_DISABLED flag set, and not have a dependency on libgtk, right?
(on linux)
taxilian 17:12 that's the theory; I'm not 100% certain it's working right now
kylehuff 17:12 I see.. I am almost 100% certain it isn't working right now.. =c )
taxilian 17:12 hehe. sorry :-/
kylehuff 17:12 it's all good
taxilian 17:12 I'll have to look at it probably the week after next
kylehuff 17:12 no rush.. I know it was working the other week, so it is probably getting hosed due to a recent change.
taxilian 17:12 well, kinda
there is a complication
there is actually an issue filed on it, if you want to go read up on what is going on
kylehuff 17:12 ahh, okay, I see now.. well, by "was working" I just mean it was not linking against gtk with the no gui flag set..
taxilian 17:12 yeah