IRC Log Viewer » #firebreath » 2012-12-18

IRC Nick Time (GMT-7) Message
KevinXi 06:12 hi, is anybody around?
i want to know if fb is capable to act as a web installer that write register, or launch a local process? thanks!
taxilian 08:12 ConfusedDev: don't copy the build folder to the other computer
ConfusedDev on second glance looks like you dont have the xcode cmd line tools installed
jshanab 08:12 taxilian. I discovered there is a command line selector tool for the command line utilities and after that recent sdk upgrade, it must be run. the command line is a different environment than xcode gui
taxilian 08:12 yep
ConfusedDev 09:12 I just checked in Xcode and it says the command line tools are installed. Whats the command line selector tool?
jshanab 09:12 xcode-select. You will run it with a list command then sudo xcode-select -switch {your sdk dir}
KevinXi 09:12 Sorry, I am new here..
jshanab 09:12 KevinXi. Welcome and good morning
KevinXi_ 09:12 Oops... my Chrome was crashed just now...
Thanks, actually it is 11:57pm here...:)
I have an newbie question, is there any way to distribute FB plugin via webbrowser..
I mean download and install the plugin automatically while loading the page...
Are you still around?
jshanab 09:12 busy. I will get back or someone else will answer
KevinXi_ 09:12 Okay, take your time. Will take a bath now.. then off to bed..
Nice day..
ConfusedDev 09:12 So I ran sudo xcode-select -switch /Applications/ and got again. Still says Xcode is broken
jshanab 09:12 I am coming in late. You did open xcode an dinstall command line tools. AND restart the terminal (or source the profile)
ConfusedDev 09:12 Yessir.
jshanab 09:12 what version of xcode are you using BTW
taxilian 09:12 KevinXi_: there are a few options, but none of them universal or really good
the only universal way is to use an installer, which you need to download and run
ConfusedDev: what do you get when you type "which gcc" in a terminal?
ConfusedDev 09:12 When I type in which gcc I get /usr/bin/gcc
and Xcode is version 4.5.1
jshanab 09:12 taxilian. what about gcc -v which shows it is configured by llvm etc
ConfusedDev 09:12 Results from typing in gcc -v:
taxilian 09:12 interesting. that's the same as mine
I dunno; cmake is saying that it's broken, whcih means it tried to use it to compile a test program and it didn't work
you could try doing your own simple test program and see what you find
ConfusedDev 09:12 Would reinstalling CMake help?
taxilian 09:12 anything is possible =] what version of cmake is it?
ConfusedDev 09:12 2.8-10
I got "Failed create symlink installation may be incomplete: /usr/bin/ccmake when trying to install cmake after deleting the old one
jshanab 09:12 Yeah I had that issue to, it is a known one on the cmake site. I'd have to google to refresh my memory
ConfusedDev 09:12 I feel like I just screwed things up even farther. Haha
taxilian 09:12 I usually fix that by deleting the cmake symlinks first
jshanab 09:12 That sounds like the trick!
I remember that
taxilian 09:12 it's a bit of a pain; gotta remove /usr/bin/... cmake*, ccmake*, cpack, ctest, .. maybe others? I forget
ConfusedDev 09:12 cd usr
reichi 09:12 no such directory
ConfusedDev 09:12 yea... instead of waiting for an IT guy to be available, how do I get to usr?
jshanab 09:12 cd /usr
ConfusedDev 09:12 oh... haha
so there is no ccmake, cpack, ctest, ect. in bin?
jshanab 09:12 In Linux, it is all relative ;-)
reichi 09:12 try locate cpack
as in
locate cpack
usually modern linux distributions build mlocate databases regularly
ConfusedDev 09:12 oh, cd /usr/bin brought me to a different place than cd /usr and then cd /bin
reichi 09:12 yes
cd /usr/bin
is the same as
cd /usr
cd bin
cd /bin
changes to the "bin" folder in the root
pathes with / in the beginning are always absolute
jshanab 09:12 Sorry, that is the joke. It is all relative to / "root" no c: d: etc
ConfusedDev 09:12 yes! No error!
ConfusedDev 10:12 So now my plugin works in Safari and Firefox but not Chrome?!
jshanab 10:12 What does your plugin do and what in chrome fails? loading?
Chrome is most like firefox in therms of Firebreath plugins
ConfusedDev 10:12 The plugin doesn't load and after like 20 seconds a banner pops up saying the plugin crashed
Glanzmann 10:12 I would like to write a google chrome extension that tunnels an rdp connection over websocket, by listening on a local socket, connect to the destination server over rdp on tcp port 3389 using the browsers websocket support (including the ability to use proxy.pac and proxy servers). And than spawn a local programm (rdesktop/xfreerdp on Linux or mstsc on Windows to the local port).
jshanab 10:12 Chrome on windows or mac? firefox and safari are 64 bit and chrome cannot do 64 bit plugins (to my knowledge) you will need to build for both i386 and x86_64 making any dependent libs also fat
(that was a Mac statement)
Glanzmann 10:12 Is it possible to listen on a tcp socket using google chrome without using npapi?
This looks like it:
kylehuff 10:12 taxilian: I pulled down firebreath-v1.7, and I don't have any (new) issues on Linux
ConfusedDev 10:12 Running on Mac. The plugin actually crashed on Firefox and Safari too after a minute (or at least it says it crashed)
taxilian 10:12 Glanzmann there are websockets
it isn't a raw TCP socket, but it's usable for many things
kylehuff: awesome, thanks!
particularly since I released 1.7 yesterday… :-P
kylehuff 10:12 yeah, I saw that. unfortunately my leave was delayed
taxilian 10:12 no worries
Glanzmann 10:12 taxilian: I know that. I basically want to build a port forwarder. Chrome listens on localhost and forwards the port over websocket to a http server ip and port.
But at the end of the day I want a tcp socket.
Tunneled using websockets.
browser listens and connects using websocket to a remote ip adress and port over a websocket capabable webserver.
kylehuff 10:12 general question - what is the best way to expose my plugin's methods so developers can easily access those methods in an intelligent javascript console like in webkit browser?
taxilian 10:12 kylehuff: introspection works as long as you aren't accessing it through the DOM element
so if you were to make a method that returned shared_from_this() you could see the methods and properties on it with introspection
actually it works through the dom element too sometimes, but I'm not sure why it doesn't always
jshanab 10:12 glanzmann. Are you inside the localnet with the socket? Would ZMQ do the trick?
taxilian 10:12 he left
jshanab 10:12 I am slow today
taxilian 10:12 hehe
kylehuff 10:12 ah, okay. I noticed it worked with properties, just not always with methods.
kylehuff 11:12 hmmm. It seems that at times when I reload my extension, google-chrome segfaults with "third_party/tcmalloc/chromium/src/] Memory corruption detected." since moving to firebreath-1.7
taxilian 11:12 awesome!
that's, err, a feature!
what were you on before?
kylehuff 11:12 lol... on 1.6 before, I checked out 1.7 and also boost 1.50.0
taxilian 11:12 hmm. I don't know of any reason that would occur; that's a little concerning
is this linux only or do you see it elsewhere?
kylehuff 11:12 only on linux ATM. I can't reproduce it every time. but 3 times in the last hour chrome has segfaulted on reload of my extension
taxilian 11:12 hmm
let me know if you learn anything new on it :-/
kylehuff 11:12 yeah, will do. I'm trying to figure out how to reproduce it now so I can identify the cause. I'm hoping it is unrelated to the upgrade.
taxilian 11:12 me too =]
kylehuff 12:12 off topic, is this the correct syntax for FB_OPTIMIZE? 'set(FB_OPTIMIZE "size")'
also, is it best to put that in Win/projectDef.cmake since it is only for windows, or in PluginConfig.cmake?
taxilian 12:12 I think that's correct
actually you'll have to put it in the prep script
as an argument
kylehuff 12:12 ah, okay. too easy
taxilian 12:12 jshanab has started on a nifty little python script for giving you a simple gui for the prep scripts
almost tempted to make a c++ one so there is no python dependency; have a simple script for "bootstrapping" it for your platform
hmm. in fact, if we did that we could easily pre-build the core libraries and thus dramatically decrease the compilation time
kylehuff 12:12 yeah, that'd be cool
I might have to spend some time on the XPI and CRX buildscripts. They appear to be specific to windows only. Also, the CRX buildscript doesn't account for compilation on a box without the chrome application binary, so the build+sign would fail.
jshanab 12:12 tui :-) aka Text User Interface. C++ would be ok, but would have to be built for each platform. Isn't python already needed. Only Windows users would really need to install it i think
I kinda would like to see the elimination of the prep script. Something where you could use the cmake gui. This would also may allow direct cmake usage in kdevelop4 instead of prepmake and import
ConfusedDev 12:12 Hey guys! Thanks so much for helping me get my plugin compiled! One last thing is the fact that it doesnt want to work on Chrome...
taxilian 12:12 you need to build it 32 bit as well as 64 bit
ConfusedDev 12:12 I've been building it 32-bit all this time. Chrome is 64 bit?
taxilian 12:12 no, chrome is 32 bit
other browsers are all 64
so the most common problem is to accidentally build it 64 only
ConfusedDev 12:12 Ok, so they should be working with Chrome since they are compiled for 32?
taxilian 12:12 if it is built with 32 bit then it should be working with chrome
kylehuff 12:12 (chromium and google-chrome on linux is by default 64bit on a 64bit box)
taxilian 12:12 if that isn't it, we'll need a lot more info before we can even guess
ConfusedDev 12:12 Changed the architectures to 32-bit Intel and ran it on Chrome and it still said the plugin has crashed
the architecture was i386 x86_64 before I changed it by the way
keith__ 13:12 Saw taxillian's call for help. Anything I, a newbie, can do to help?
A newbie to firebreath, anyway. :)
taxilian 13:12 I'm sure there is lots you can do
what are your skills / interests?
keith__ 13:12 I'm a web developer. Currently using Ruby/Rails. Java before that. Know jQuery. Know unix admin. I set up AWS VPC stuff.
taxilian 13:12 any php exerience?
keith__ 13:12 I can fumble my way around. Everything starts to look like Perl to me after a while.
taxilian 13:12 hehe. so we have kinda an ongoing issue that I haven't had time to work on
there is a log viewer php app that someone made a year or so ago that we were using here
but it is really lacking in a lot of important functionality
keith__ 13:12 I build start-ups professionally, but that doesn't seem relevant here. :)
taxilian 13:12 if you were interested, that is probably one of the higher priority sysadmin/web development issues we have right now
basically I have logs of everything that occurs in this irc channel
keith__ 13:12 I'd be happy to try and help. How would I get started?
taxilian 13:12 but we need them viewable and searchable on the web
let me show you the existing project
I don't really care how we do it; if someone wants to extend the current one I can get it back up (changed servers recently)
keith__ 13:12 Ok.
taxilian 13:12
do you want a tarball of the log files?
keith__ 13:12 That's be great. What's the order of magnitude for log file sizes? Gigs?
I've used (Sphinx) and (Lucene) for search in the past.
taxilian 13:12 I think it'd be awsome to have a real text search tool in it
keith__ 13:12 Do you have any specific search capability requirements? Wildcard, stemming, etc.
taxilian 13:12 16M total =]
oh, that includes the tar archive...
keith__ 13:12 And the production env is some unix variant running apache and mod_php?
taxilian 13:12 the production env is the box that runs off of; it's running centos 6.2 (or 6.3, I forget) and can optionally have apache and mod_php if that's useful
currently it does not
it has nginx and is running the atlassian products in seperate tomcat containers
I dont' care in the slightest what language the tool ends up in =]
basically first priority would be to simply be able to view all logs by date and time
second priority to search in the logs and get some of the logs with context
ultimately it would be really cool to be able to save snippets of conversation as favorites by name, etc, so that important conversations could be linked to
but I'll take what we can get =]
keith__ 13:12 ok. I'll poke around the irc-logviewer project and try to figure things out.
taxilian 13:12 I have looked at several other options for irc log viewing; I haven't found any that I both consider to be useful and can make work
that would be awesome
keith__ 13:12 What's an easy way to get the 16M of logs?
taxilian 13:12 honestly this could easily become its own open source project; lots of projects could benefit
there they are
keith__ 13:12 Cool. I have them downloaded. I'll look at everything this evening, when I'm not supposed to be working.
Is irc the best way to keep in touch?
taxilian 13:12 hehe. awesome. other easy ways to help include doing screencasts, writing blog posts, tweeting about the project, updating the documentation, releasing sample plugins, etc
and when you are feeling more comfortable with the project I'd be happy to help you get started with adding a feature or fixing a bug =]
keith__ 13:12 *nod* Right now I'm re-writing the firebreath plugin Hopster uses. (day job)
taxilian 13:12 Hopster?
keith__ 13:12 I'll start with the IRC log parsing. It seems accessible and something I can do in the evening.
taxilian 13:12 that would really be a huge help. I have people asking about that 2-3 times a week
and I just don't have time to get to it
keith__ 13:12 and recent startup who is one of the many companies using firebreath.
er, a recent startup
I appreciate firebreath and I'll try to give a little love back.
taxilian 13:12 cool. you should add a link on the "in the wild" page on the wiki =]
keith__ 13:12 I'm on linked in if you'd like a more formal introduction.
And, where I'll go fork irc-logviewer.
Thanks again. ttyl
taxilian 13:12 cool. I'll add you =]
I'm… well, Taxilian. =] a.k.a. Richard
keith__ 13:12 Heh. You're well known to anyone who uses firebreath. Positively famous. :)
taxilian 13:12 hehe. a fact which would boggle the mind of anyone who knows me that doesn't use firebreath, believe me.
I appreciate you responding to the request for help; others have responded, you are the first to actually come out and say "what can I do?" instead of "yeah, I'll try to help more"
keith__ 13:12 Oh. Well. That makes me a bit sad.
taxilian 13:12 well, that's not totally fair as there are some in this room who have been helping in small ways as they can, but the number of people actively trying to help vs how many use the project is really disproportionate
you and me both, my friend
hopefully things are moving to being a bit better, though. I have many, many ideas and dreams for FireBreath but they are all severely limited by time. I wish that I could have worded that email differently and said "I don't have time to do bug fixes, so I'm going to concentrate on building new features" and let others ensure stability, but that day is still a long ways off if ever
keith__ 13:12 Next up, actually doing something. :)
I don't know enough about FB to contribute directly at this point. Maybe in a few weeks. I'll do my best to help around the edges for now.
taxilian 13:12 I appreciate whatever you can do to help
I do recommend you update to 1.7 if you haven't; that's a good place to start
ConfusedDev 13:12 So what are some things I can check that would affect my Chrome compatibility?
taxilian 13:12 ConfusedDev: Honestly I have no idea. the only common issue I know of is the 32 vs 64 bit thing
I'd look at your Console
see if there are any unusual messages
jshanab 13:12 Chrome was the last browser I got my plugin to work on. For me it was because I had a window built up around the hwnd passed to me and event handling is a bit different in chrome. Symptom was resize or scroll crashed chrome
That was on windows though
I will say that chrome has a much more likely chance of a more asnychronous javascript than FF. (on windows this is true but add that IE has NO chance of async javascript calls)
taxilian 14:12 I'm not sure what you mean by IE has no change of async js calls?
jshanab 14:12 I discoverd that the javascript engine in IE is a lot slower. Turns out that it is really "single minded" for lack of a better phrase. Very intense blocking when compared with FF or Chrome. (obviously i do not have access to source code, jsut an observation in the debugger)
taxilian 14:12 ahh
jshanab 14:12 The point I was making was that the two areas that I have seen a differnce in chrome vs firefox on the event handling and the ability for it to expose synchronization issues with a much faster and more multithreaded javascript engine (part advertisement, part subjective interpolation :-) )
taxilian 14:12 makes sense
timing of callbacks across different browsers is never something you can safely rely on
to be the same all the time
ConfusedDev 15:12 so I tried the code from the last svn tag we made for the plugin and it compiles and has the same issues with chrome as the new plugin. I am doing something wrong in the compiling process. I set everything to 32-bit which threw up 68 errors so I went back and just set the overall FireBreath project to 32. Any ideas for what else I can check?
taxilian 15:12 if you really want it to just be 32 bit you should pass that into the prep script command line
(search the website for prep scripts, instructions are there)
ConfusedDev 15:12 so I just run ./ -D CMAKE_OSX_ARCHITECTURES="i386" ?
taxilian 15:12 yeah
ConfusedDev 15:12 What would I do with the old build folder? Just delete it with its contents and let prepmac make a new one?
taxilian 15:12 yeah, though usually you don't need to fully recreate the build folder I probably would for a change like that
ConfusedDev 15:12 oh man.. and now I have 24 of these types of errors: Undefined symbols for architecture i386: "_IOHIDDeviceClose", referenced from:
taxilian 15:12 sounds like you're missing some libs
ConfusedDev 15:12 Hahaha, I'm an idiot. Thanks
The output file was half the size and crashed on all browsers. Hmmm....
jshanab 16:12 Here I think you are running into my issue. I don't yet have all my dependent libs build for 32 bit so I am restricting to 64 bit until I get the feature set working. At least on APPLE we have the fat binary concept!
Do you have libraries that your developers built for iOS that you include and link?
ConfusedDev 16:12 Dont have any libraries that were built. Documents just say that I should include IOKIT.framework
taxilian 16:12 and did you?
('cause that's what I would recommend as well)
ConfusedDev 16:12 yea, I did except for that one oops I had earlier, its been included.
kylehuff 16:12 I am thinking my segfault issue is possibly related a recent change in google-chrome; it has segfaulted on reload of the extension (and plugin) while the plugin was already defunct, so it might not be related to the plugin at all.
I will have to add an unpacked extension without a plugin and see if it still occurs when I reload it.