**** BEGIN LOGGING AT Mon Jul 07 02:59:59 2014
Jul 07 03:29:49 <karki> alexanderhiam : I came up with my own mapping. but whenever arr[var] is involved, it poses a problem!
Jul 07 04:51:11 <karki> alexanderhiam : I have 8 bits for the instruction space. I'll use only 7 bits for all normal instructions (32 bits), ones that involve 64 bits will have the 8th bit high.
Jul 07 04:58:33 <alexanderhiam> karki: yeah, I think that's the way to go
Jul 07 05:31:47 <vmayoral> Greetings
Jul 07 05:32:10 <vvu> morning
Jul 07 05:32:57 <jkridner> gm all
Jul 07 05:32:59 <ds2> *yawn*
Jul 07 05:33:08 <jkridner> who all is around today?
Jul 07 05:33:12 <rseethamraju> gm
Jul 07 05:33:25 <jkridner> hi ds2. wasn't expecting you at this time so much. :-)
Jul 07 05:33:50 <ds2> jkridner: took a few weeks to figure out the right day ;)
Jul 07 05:33:57 <alexanderhiam> ahoy
Jul 07 05:34:03 <jkridner> howdy alexanderhiam
Jul 07 05:34:15 <ds2> this time, wednesday would be ideal ;)
Jul 07 05:34:49 <jkridner> rseethamraju: you still around?
Jul 07 05:34:54 <rseethamraju> Yes
Jul 07 05:35:05 <jkridner> k... wondering who else is.
Jul 07 05:35:12 <jkridner> vvu, vmayoral: ?
Jul 07 05:35:17 <vmayoral> I am around yeap
Jul 07 05:35:18 * vvu is here but in the shadows
Jul 07 05:35:29 <jkridner> karki: ?
Jul 07 05:35:29 <rseethamraju> Its 11am for me so no prb
Jul 07 05:35:42 <alexanderhiam> karki was here not too long ago
Jul 07 05:35:44 <karki> jkridner : hey!
Jul 07 05:35:48 <jkridner> howdy!
Jul 07 05:36:18 <alexanderhiam> jkridner, karki: did we move the time for the botspeak hangout?
Jul 07 05:36:19 * ds2 looks high and low for the illusive master of flight computing...
Jul 07 05:36:36 <karki> jkridner : you can use pru speak from your bash shell now :)
Jul 07 05:36:53 <jkridner> karki: very nice. anxious to see it.
Jul 07 05:38:33 <karki> jkridner : I'll be upgrading the compiler, for proper byte code generation (inc. variables ) and syntax error handling. as ds2 said long back, I'll be using lex/yacc
Jul 07 05:38:42 <vmayoral> tridge: good morning!
Jul 07 05:39:35 <vmayoral> karki: what functionalities do u have so far?
Jul 07 05:39:41 <jkridner> karki: interesting. those are usually pretty good for making fairly optimal tokenizers.
Jul 07 05:39:56 <tridge> vmayoral: (hi, just lurking, on phone at same time)
Jul 07 05:40:15 <jkridner> karki: I think they often use a common library that'll be interesting to be linked into the kernel.
Jul 07 05:40:57 <ds2> kernel as in Linux kernel?
Jul 07 05:41:10 * jkridner might be too tired to try to organize discussion
Jul 07 05:41:28 <jkridner> ds2: yeah, alexanderhiam and I are wanting to see the tokenizer placed in the kernel.
Jul 07 05:41:34 <tridge> vmayoral: see my email on changes in beaglepilot/master
Jul 07 05:42:03 <ds2> check the licenses to make sure they are mutually compatible
Jul 07 05:42:07 <tridge> https://groups.google.com/forum/#!topic/drones-discuss/u3AJYprDjt4
Jul 07 05:42:14 <karki> vmayoral : as of now I'm working on language completion. (that is conditionals, arrays, variables, better parsing etc.) once that is satisfactory, I'll add PWM, ADC and digital input. (Well thats the plan until the 11th week of GSoC)
Jul 07 05:42:23 <ds2> esp. if you are using the GPL'ed versions (bison, etc)
Jul 07 05:42:47 <karki> vmayoral : 10th week
Jul 07 05:43:15 <jkridner> karki: do you have the translation matrix complete/published yet?
Jul 07 05:44:14 <karki> jkridner : almost done. I had an unexpected problem with instructions with "Arr[var]" operands. But me and alexanderhiam fixed that.
Jul 07 05:45:00 <jkridner> ds2: do you really need a grammar engine (bison/yacc) or is a simple lexical tokenizer (lex/flex) enough to make it work? the grammar is so simple.
Jul 07 05:45:46 <alexanderhiam> rseethamraju: none of those gst-launch commands worked for the c270
Jul 07 05:45:48 <ds2> me?
Jul 07 05:45:52 <vmayoral> tridge: thanks for the changes. Tree updated
Jul 07 05:45:52 <ds2> depends
Jul 07 05:45:55 <jkridner> rseethamraju, vvu: how are things going?
Jul 07 05:46:07 <ds2> it is more of a design thing
Jul 07 05:46:17 <karki> jkridner : the grammar is simple, but for error detection, handling variable (context), labels (again context) I'd say having yacc would make it much cleaner.
Jul 07 05:46:24 <vvu> jkridner: we have a bit of an issue with the frontend for the framebuffer stuff
Jul 07 05:46:35 <vvu> when we start the x session and direct it to displayLink
Jul 07 05:46:42 <jkridner> karki: k, I can see that.
Jul 07 05:46:46 <vvu> praveen posted on their IRC channel
Jul 07 05:47:15 <vvu> but beside that after configuring the phone we get a greeen screen which means stuff in the kernel driver are ok
Jul 07 05:47:23 <vvu> now we just need to interface X session with the kernel driver
Jul 07 05:47:31 <alexanderhiam> rseethamraju: I now have a few of the spi temp sensors, though I haven't had a chance to try them out yet
Jul 07 05:47:51 <rseethamraju> alexanderhiam:ok so opencv uses lots of different cap stuff based on which camera
Jul 07 05:48:19 <jkridner> rseethamraju: is the cap stuff needed or just optional?
Jul 07 05:48:37 <rseethamraju> I'm in clg now so after class I'll first try if my hardware is skiing with the arduino library
Jul 07 05:48:37 <rseethamraju> Working sorry
Jul 07 05:48:55 <alexanderhiam> rseethamraju: ok, have you figured out if we can do streaming over tcp with opencv?
Jul 07 05:49:42 * jkridner looks over https://gist.github.com/alexanderhiam/a238fe435b3f8b0e3bfd
Jul 07 05:50:13 <alexanderhiam> it would be a bummer to have to do all the camera detection from scratch if there's a lot to it, would be nice to use opencv for everything
Jul 07 05:50:21 <vvu> git
Jul 07 05:50:38 <vvu> oh again no coffee tells the difference
Jul 07 05:50:58 <rseethamraju> Not really but I found a Program that takes a  stream and pipes it to a gstreaner pipeline
Jul 07 05:51:17 <rseethamraju> I havent tried it yet
Jul 07 05:51:27 <alexanderhiam> think that will add a lot of delay?
Jul 07 05:51:40 <rseethamraju> Maybe
Jul 07 05:52:19 * jkridner falls asleep at the keyboard
Jul 07 05:52:34 * alexanderhiam does the same
Jul 07 05:52:39 <rseethamraju> But theres also derek molloy's program that streames from opencv  to udp
Jul 07 05:52:48 * karki wakes everyone up!
Jul 07 05:53:06 <jkridner> karki: k. what's the challenge for which you've awoken us?
Jul 07 05:53:14 <rseethamraju> Sorry about the spelling. Im in college and on my phone in class
Jul 07 05:53:29 <alexanderhiam> rseethamraju: if udp is much easier we could go with that, it's just not supported by html5 so the web interface side is harder
Jul 07 05:53:42 <alexanderhiam> rseethamraju: no problem
Jul 07 05:54:20 <rseethamraju> Theres probably a tcp version to do the same
Jul 07 05:54:28 <rseethamraju> Does it have to be in python
Jul 07 05:54:31 <rseethamraju> ?
Jul 07 05:55:15 <jkridner> alexanderhiam, rseethamraju: for streaming, mjpg-streamer seems to be rather popular as used by they OpenROV guys. https://github.com/OpenROV/openrov-software
Jul 07 05:55:20 <alexanderhiam> rseethamraju: well we can wrap C or C++ code with extensions of course, just gotta make sure the licensing works
Jul 07 05:55:33 * jkridner checks out what google says; http://stackoverflow.com/questions/20205358/how-do-i-stream-an-opencv-video-to-an-html-webpage
Jul 07 05:55:51 <rseethamraju> Wait did you try on chrome?
Jul 07 05:56:23 <alexanderhiam> yeah
Jul 07 05:56:26 <alexanderhiam> and firefox
Jul 07 05:56:42 <rseethamraju> Maybe the issue is the .ogg
Jul 07 05:56:55 <rseethamraju> We havent tried mpeg
Jul 07 05:57:34 <alexanderhiam> one of the stackoverflow answers jkridner pointed to used ffmpeg, which I think is how Derek Molloy did the streaming as well, right?
Jul 07 05:57:50 <rseethamraju> Yes
Jul 07 05:57:57 <alexanderhiam> but it was working for both of us with the c920 as ogg
Jul 07 05:58:01 <rseethamraju> That a shell command
Jul 07 05:58:04 <rseethamraju> Can we use that?
Jul 07 05:58:11 <rseethamraju> Yes
Jul 07 05:58:26 <alexanderhiam> yeah that's fine, we can run it as a subprocess
Jul 07 05:58:32 <tridge> vmayoral: (off the phone now)
Jul 07 05:58:46 <tridge> vmayoral: I have the GPS working, but I don't yet have the telemetry port working
Jul 07 05:58:47 <rseethamraju> I just got another uvc camera that I'll try today
Jul 07 05:58:54 <alexanderhiam> let's look into the open-rov stuff too
Jul 07 05:58:58 <tridge> vmayoral: have you tried telemetry via serial on your PXF?
Jul 07 05:59:11 <alexanderhiam> thanks jkridner
Jul 07 05:59:31 <tridge> jkridner: a bit of an update on BeaglePilot. Victor, Anuj and Sid have now all receiverd their PXF capes
Jul 07 05:59:43 <rseethamraju> Ok then that would work coz derek molloys prg uses opencv and then a bunch of shell commands
Jul 07 05:59:51 <tridge> jkridner: we have them booting correctly, and the MPU9250 driver Victor wrote is working
Jul 07 06:00:20 <rseethamraju> This is so weird Im on my phone under the bench
Jul 07 06:00:30 <tridge> jkridner: so we have baro(ms5611), IMU(MPU9250), compass(HMC5883), PWM(via PRU) and GPS(ublox) all working.
Jul 07 06:00:48 <tridge> jkridner: just missing RC input as core driver now
Jul 07 06:00:49 <rseethamraju> And the teacher kinda figured that Im texting
Jul 07 06:00:56 <ds2> MPU9250 and HMC5883 are kernel drivers?
Jul 07 06:00:58 <alexanderhiam> rseethamraju: mjpeg-streamer looks pretty promising (just looking at the description): http://sourceforge.net/projects/mjpg-streamer/
Jul 07 06:01:06 <tridge> ds2: no, user space for now
Jul 07 06:01:10 <jkridner> howdy tridge. hope things are ok.
Jul 07 06:01:25 <ds2> isn't that reinventing the wheel?
Jul 07 06:01:30 <tridge> ds2: using kernel SPI layer (spi-dev.h)
Jul 07 06:01:31 <alexanderhiam> rseethamraju: well don't get in trouble ;) sounds like you have plenty to work on for now
Jul 07 06:01:43 <tridge> ds2: doing a kernel driver would have been re-inventing the wheel actually :-)
Jul 07 06:01:51 <ds2> there are drivers for them in the kernel
Jul 07 06:01:53 <tridge> ds2: as ardupilot already has a bunch of drivers
Jul 07 06:02:07 <ds2> the MPU9250 should work fine iwth the 6050 driver
Jul 07 06:02:09 <tridge> ds2: a autopilot needs some fairly specialised behaviour from these drivers
Jul 07 06:02:13 <ds2> 6000
Jul 07 06:02:19 <tridge> i'd be surprised if the kernel driver would actually do what we need
Jul 07 06:02:28 <tridge> but I haven't tried them
Jul 07 06:02:37 <ds2> and spidev is a total complete absolute waste of time
Jul 07 06:02:55 <vvu> if [ "$MSYSTEM" == "MINGW32" ]; then
Jul 07 06:02:55 <vvu>   MINGW_PACKAGE_PREFIX="mingw-w64-i686"
Jul 07 06:02:55 <vvu>   MAKEPKG_CONFIG="/etc/makepkg_mingw32.conf"
Jul 07 06:02:55 <vvu> else
Jul 07 06:02:55 <vvu>   MINGW_PACKAGE_PREFIX="mingw-w64-x86_64"
Jul 07 06:02:56 <tridge> oh? cpu overhead you mean?
Jul 07 06:02:57 <vvu>   MAKEPKG_CONFIG="/etc/makepkg_mingw64.conf"
Jul 07 06:03:00 <vvu> fi
Jul 07 06:03:17 <rseethamraju> alexanderhiam: ya I'll try that today. Mjpeg streamer not with the c920
Jul 07 06:03:18 <ds2> it would be really nice if we trying to help get a single set of drivers for those senssors
Jul 07 06:03:41 <ds2> I see more versions of MPU6050/MPU6000/MPU6500/ITG3200/etc drivers then I have fingers for
Jul 07 06:03:55 <tridge> ds2: one thing to remember is ardupilot is a portable autopilot, it runs on a wide variety of boards, only a small number of boards run linux
Jul 07 06:04:00 <rseethamraju> alexanderhiam what about the temp sensor?
Jul 07 06:04:12 <tridge> ds2: drivers inside ardupilot means when we fix issues we fix them for all boards
Jul 07 06:04:31 <ds2> tridge: *nod* but IMO, that is a very narrow view... for Linux, there is value in trying to unify all the drivers out there
Jul 07 06:04:39 <ds2> otherwise, there is no value in running Linux
Jul 07 06:04:45 <alexanderhiam> rseethamraju: I'll test it here soon and let you know if I get wierd values
Jul 07 06:04:54 <tridge> ds2: there is still a lot of value in running Linux I think
Jul 07 06:05:14 <rseethamraju> Ya and ill try with the arduino first
Jul 07 06:05:25 <ds2> what I am getting at is... this is the same as writing non unstreamable code for an opensource project
Jul 07 06:05:34 <tridge> i would like to see a comparison of running the kernel driver as well, but certainly the easiest way to start is with the existing ardupilot drivers
Jul 07 06:05:40 <rseethamraju> alexanderhiam: oh and what about eqep?
Jul 07 06:05:50 <rseethamraju> I sent u a mail
Jul 07 06:06:02 <ds2> but since I have said I've dropped out of that project...I'll hush now.
Jul 07 06:06:05 <alexanderhiam> rseethamraju: saw it, I haven't had a chance to think about it yet
Jul 07 06:07:23 <tridge> ds2: I believe Jimmy Johnson is working on the MPU6000 kernel driver
Jul 07 06:07:39 <tridge> ds2: but we seem to have a MPU6000 hardware fault on the PXF cape
Jul 07 06:07:49 <tridge> ds2: so we've switched to MPU9250
Jul 07 06:07:51 <rseethamraju> Ok tomorrow then
Jul 07 06:08:06 <tridge> (the cape has 3 IMUs, a MPU6000, a MPU9250 and a LSM9DSO)
Jul 07 06:08:24 <alexanderhiam> rseethamraju: yeah, sorry, I'm running out of steam here
Jul 07 06:08:42 <rseethamraju> Its ok good night!
Jul 07 06:09:16 <alexanderhiam> jkridner: I don't see an invite for the hangout tomorrow, is it still on?
Jul 07 06:09:31 <jkridner> with Chris?
Jul 07 06:09:37 <alexanderhiam> yeah
Jul 07 06:09:46 <jkridner> I don't know the time we would have it.
Jul 07 06:09:58 <jkridner> never got finalized on time.
Jul 07 06:09:59 <alexanderhiam> did we talk about moving it to 9?
Jul 07 06:11:18 <jkridner> sounds right... I'll setup an event for that time.
Jul 07 06:16:07 <rseethamraju> So is the meeting done?
Jul 07 06:16:42 <alexanderhiam> jkridner may have fallen asleep on the gavel
Jul 07 06:16:53 * jkridner uses head as gavel
Jul 07 06:16:59 <jkridner> *slam*
Jul 07 06:17:18 <rseethamraju> Good night!
Jul 07 06:17:30 <alexanderhiam> good night/morning
Jul 07 06:17:50 <tridge> night!
Jul 07 06:32:03 <karki> is docs.google.com working when accessed from India?
Jul 07 06:45:46 <vvu> praveendath92: ping
Jul 07 06:46:29 <praveendath92> pong
Jul 07 06:46:34 <vvu> any answer on the IRC channel ?
Jul 07 06:47:17 <praveendath92> Nope.
Jul 07 06:48:50 <praveendath92> I tried a couple of things. If I remove fb from screens list, there is no error.
Jul 07 06:48:50 <vvu> :(
Jul 07 06:49:08 <vvu> try their ML
Jul 07 06:49:13 <vvu> i am a bit off hand here
Jul 07 06:49:28 <vvu> ds2: still here ?
Jul 07 06:49:43 <praveendath92> Definitely something with displaylink driver. Try to traceback.
Jul 07 06:49:43 <praveendath92> Unfortunately no kernel logs.
Jul 07 06:52:14 <praveendath92> vvu: I will.
Jul 07 06:56:59 <av500> praveendath92: hi
Jul 07 06:57:05 <av500> what is the status?
Jul 07 06:57:22 <praveendath92> av500: Hello
Jul 07 06:57:51 <praveendath92> av500: I see a complete green screen in framebuffer.
Jul 07 06:58:24 <praveendath92> When I try to start an xserver using dl driver I get a segmentation fault.
Jul 07 06:59:21 <av500> hmm
Jul 07 06:59:24 <av500> green? where?
Jul 07 06:59:36 <praveendath92> I tried posting on xorg irc with the logs but no response.
Jul 07 07:00:03 <praveendath92> The whole green screen is when the setup is successful for dl devices.
Jul 07 07:00:16 <av500> but where do you see green?
Jul 07 07:00:36 <praveendath92> When I read the current frame in the framebuffer.
Jul 07 07:00:48 <praveendath92> Using fbgrab
Jul 07 07:00:52 <av500> ah
Jul 07 07:01:38 <praveendath92> I have the logs here. https://gist.github.com/praveendath92/905160a386fb03a3ef3d
Jul 07 07:02:31 <praveendath92> attached - with android device attached (segmentation fault)
Jul 07 07:02:40 <praveendath92> not-attached - no device attached. No errors. Display starts.
Jul 07 07:03:24 <av500> so you can start x?
Jul 07 07:03:32 <av500> ah
Jul 07 07:03:39 <av500> not attached means not your fb
Jul 07 07:03:46 <praveendath92> Yes.
Jul 07 07:03:56 <praveendath92> So, that is not a helpful situation.
Jul 07 07:04:18 <av500>  /usr/lib/xorg/modules/drivers/displaylink_drv.so
Jul 07 07:04:20 <av500> er
Jul 07 07:04:24 <praveendath92> But it tells that the issue is with dl.
Jul 07 07:04:25 <av500> so a dedicated driver?
Jul 07 07:04:34 <praveendath92> Hmm.
Jul 07 07:04:53 <praveendath92> I followed this - http://nameq.wordpress.com/2011/09/21/displaylink-dl-165-running-on-linux-fedora/
Jul 07 07:04:55 <av500> did you tell it to use a displaylink driver?
Jul 07 07:05:40 <praveendath92> Yes. There are a lot dl(0) in the logs.
Jul 07 07:07:14 <praveendath92> Wondering how to analyze the stacktrace.
Jul 07 07:07:57 <av500> I am not sure your udlfb and this X driver are in sync
Jul 07 07:08:16 <av500> I thought one could run X on a generic framebuffer /dev/fb
Jul 07 07:08:53 <praveendath92> I think so too.
Jul 07 07:09:02 <praveendath92> Not in sync as in?
Jul 07 07:09:20 <praveendath92> The EDID info in the logs is read from udlfb.
Jul 07 07:17:19 <av500> X ran on fb long before there was EDID
Jul 07 07:20:35 <praveendath92> line?
Jul 07 07:21:24 <av500> line?
Jul 07 07:21:26 <av500> not line
Jul 07 07:21:37 <av500> I mean, once upon a time there was no EDID
Jul 07 07:21:40 <av500> :)
Jul 07 07:21:54 <praveendath92> Oh. That :)
Jul 07 07:21:55 <praveendath92> https://gist.github.com/praveendath92/905160a386fb03a3ef3d#file-not-attached-L248
Jul 07 07:21:56 <av500> so please remove that dl driver
Jul 07 07:22:38 <praveendath92> My initial attempts were to start X on fb w/o dl driver.
Jul 07 07:22:45 <av500> yes
Jul 07 07:22:53 <praveendath92> You know how to do that. Starting X on a fb ?
Jul 07 07:24:08 <praveendath92> BTW, that link was pointing the log where error occurs is when dl driver is unloaded when device is not attached.
Jul 07 07:26:39 <av500> Loading sub module "fbdevhw"
Jul 07 07:26:50 <av500> Loading /usr/lib/xorg/modules//libfb.so
Jul 07 07:26:54 <av500> do you have that one?
Jul 07 07:27:59 <praveendath92> Nope.
Jul 07 07:28:35 <praveendath92> A lib used by dl driver I think.
Jul 07 07:33:25 <av500> https://packages.debian.org/wheezy/armhf/xserver-xorg-core/filelist
Jul 07 07:33:34 <av500> should be in core
Jul 07 07:34:09 <av500> I have /usr/lib64/xorg/modules/drivers/fbdev_drv.so
Jul 07 07:36:34 <praveendath92> Oh.
Jul 07 07:37:12 <praveendath92> What is it for?
Jul 07 07:39:53 <praveendath92> This looks promising. http://www.tldp.org/HOWTO/Framebuffer-HOWTO/x1010.html
Jul 07 07:56:42 <av500> praveendath92: its a fb driver for x11
Jul 07 07:56:46 <av500> try to use this one
Jul 07 08:00:23 <Abhishek__> Hi everyone
Jul 07 08:01:01 <Abhishek__> Sorry I couldn't make it to the meeting
Jul 07 08:09:20 * karki notes the extra __ 
Jul 07 09:15:41 <Abhishek__> hi panto
Jul 07 09:17:29 <panto> hi Abhishek__
Jul 07 12:40:37 <karki> jkridner : don't know if I can make it to the meeting. pathetic internet connection. and I mean *pathetic* :(
Jul 07 12:40:43 <karki> I'll try though
Jul 07 13:15:31 <jkridner> karki: no one coming? panto, alexanderhiam?
Jul 07 13:15:36 <jkridner> no Chris either. :(
Jul 07 13:15:59 <panto> jkridner, I got a conflict...
Jul 07 13:16:04 <karki> jkridner : bad net connection. goes off every 5 min
Jul 07 13:16:10 <panto> can we reschedule for two hours from now?
Jul 07 13:16:16 <karki> have to keep fixing it.
Jul 07 13:16:31 <alexanderhiam> oops, lost track of time!
Jul 07 13:16:41 <alexanderhiam> I can do now or two hours rom now
Jul 07 13:16:45 <alexanderhiam> from*
Jul 07 13:27:24 <alexanderhiam> karki: can you push your latest stuff?
Jul 07 19:27:35 <mranostay> quiet quiet in here
Jul 07 19:41:30 <ds2> hmm
Jul 07 23:02:46 <VoltVisionFrench> hello DiegoTc & jkridner !!
Jul 07 23:02:52 <jkridner> hi VoltVisionFrench....
Jul 07 23:02:54 <DiegoTc> ho!
Jul 07 23:02:56 <jkridner> trying to find the event.
Jul 07 23:03:01 <DiegoTc> Hi VoltVisionFrench & jkridner
Jul 08 00:37:55 <Abhishek_> morning
Jul 08 02:33:18 <karki> alexanderhiam , jkridner : Done with the encoding stuff (yesterday). figured out what to put where. only thing I have to do is give each command an opcode. :D bad news is that all the data is in my notebook, I have to type it out onto a doc :( and google docs is banned!!
Jul 08 02:39:22 <karki> worst case was that I required 48 bits - for commands where  both operands are of type : Arr[var]
**** ENDING LOGGING AT Tue Jul 08 02:59:58 2014