**** BEGIN LOGGING AT Mon Jul 21 03:00:00 2014
Jul 21 04:24:50 <jkridner|m> not going to make it another hour tonight
Jul 21 04:45:28 <vvu> morning
Jul 21 05:26:17 <tridge> good afternoon
Jul 21 05:26:46 <tridge> jkridner|m: did you manage to stay awake? :-)
Jul 21 05:27:43 <tridge> i guess thats a no :)
Jul 21 05:28:38 <Abhishek_> Hi all
Jul 21 05:34:48 * alexanderhiam is not going to be conscious much longer, you stuck on anything rseethamraju?
Jul 21 05:35:05 <rseethamraju> not really
Jul 21 05:35:13 <rseethamraju> still trying the camera
Jul 21 05:35:26 <rseethamraju> but accelerometer is going well
Jul 21 05:35:46 <rseethamraju> I should be mostly done by tomorrow’s meeting
Jul 21 05:35:57 <alexanderhiam> oh awesome
Jul 21 05:36:05 <alexanderhiam> I forgot about that
Jul 21 05:37:39 <rseethamraju> I had only camera and accelerometer this weekend
Jul 21 05:37:51 <rseethamraju> did I forget anything?
Jul 21 05:38:15 <alexanderhiam> I don't think so
Jul 21 05:38:59 <rseethamraju> camera audio isn’t working
Jul 21 05:39:11 <alexanderhiam> naturally
Jul 21 05:39:46 <alexanderhiam> I'm fine with not having audio for a first version
Jul 21 05:39:51 * mranostay pokes in
Jul 21 05:39:52 <rseethamraju> ok cool
Jul 21 05:40:27 <rseethamraju> so I tried sending a audiotestsrc to tcpserversink
Jul 21 05:40:36 <rseethamraju> that doesn’t work
Jul 21 05:41:22 <rseethamraju> I used the <audio> tag in html
Jul 21 05:41:50 <alexanderhiam> that didn't work at all?
Jul 21 05:42:51 <rseethamraju> no
Jul 21 05:43:40 <rseethamraju> I hit play and nothing happens
Jul 21 05:44:10 <alexanderhiam> does the pipeline work? did you try with a filesink?
Jul 21 05:44:22 <rseethamraju> yes
Jul 21 05:44:25 <alexanderhiam> does gstreamer use alsa as a source?
Jul 21 05:44:26 <rseethamraju> that does
Jul 21 05:44:40 <rseethamraju> I mean with audio test source
Jul 21 05:45:02 <rseethamraju> so its sending random audio
Jul 21 05:45:10 <alexanderhiam> ah
Jul 21 05:45:43 <rseethamraju> either way in our case <video> should play both right so this won’t matter, right?
Jul 21 05:46:38 <alexanderhiam> you mean both video and audio?
Jul 21 05:46:49 <rseethamraju> yes
Jul 21 05:46:59 <alexanderhiam> it should
Jul 21 05:47:13 <rseethamraju> and all the pipelines I’ve seen used alsasrc for audio
Jul 21 05:47:58 <alexanderhiam> so you source them separate then mix them together before the tcpsink?
Jul 21 05:49:07 <rseethamraju> yes, I think so
Jul 21 05:50:29 <rseethamraju> what’s alsa btw?
Jul 21 05:50:49 <rseethamraju> oh got it
Jul 21 05:51:53 <rseethamraju> so the camera program without the audio works without putting it in an object and functions
Jul 21 05:53:03 <alexanderhiam> so the next step is getting it wrapped up in a Webcam object that can stream, record and save snapshots
Jul 21 05:54:20 <alexanderhiam> have you figured out yet if you're going to share the same source or have 3 totally separate pipelines?
Jul 21 05:55:14 <rseethamraju> I was tring to do element_factory_make in the __init__ but since snapshots are taken by opencv maybe I can do them completetely in the methods
Jul 21 05:55:23 <rseethamraju> 3 for what?
Jul 21 05:55:48 <alexanderhiam> oh right, forgot you were using opencv for that
Jul 21 05:57:20 <alexanderhiam> the advantage of creating the pipelines right in the startStreaming() and startRecording() methods is that you could pass in some configuration, like width and height, etc.
Jul 21 05:58:51 <rseethamraju> yes
Jul 21 05:59:06 <rseethamraju> oh you want me to do a startRecording() too
Jul 21 05:59:35 <rseethamraju> so they give the name of the file and it records?
Jul 21 05:59:42 <alexanderhiam> that's what I was thinking
Jul 21 06:00:18 <rseethamraju> so in that case, wouldn’t it be better if the file was not an .ogg file?
Jul 21 06:01:56 <alexanderhiam> you thinking mp4 instead?
Jul 21 06:02:55 <rseethamraju> yes
Jul 21 06:03:10 <rseethamraju> I have a pipeline for mjpeg
Jul 21 06:03:26 <rseethamraju> but that worked for 290
Jul 21 06:03:28 <alexanderhiam> whichever's easier
Jul 21 06:03:40 <rseethamraju> anyway filesink is easier to achieve
Jul 21 06:04:05 <rseethamraju> and its easier for me to write the pipelines seperately so startStreaming() and startRecording() would work
Jul 21 06:04:18 <alexanderhiam> ok cool
Jul 21 06:04:35 <rseethamraju> the other issue is ctrl-c doesn’t stop the python program
Jul 21 06:05:31 <rseethamraju> it would be nice if you call a method called stop() the pipeline goes to a stop state
Jul 21 06:05:43 <rseethamraju> I still have that left to do
Jul 21 06:07:54 <alexanderhiam> we could probably figure that part out on Tuesday
Jul 21 06:13:44 <rseethamraju> ok
Jul 21 06:13:55 <rseethamraju> that’s it though
Jul 21 09:40:54 <praveendath92> vvu: ping
Jul 21 09:44:41 <vvu> pong
Jul 21 09:44:46 <vvu> praveendath92:
Jul 21 09:48:21 <praveendath92> I got a the raw from over USB.
Jul 21 09:48:34 <vvu> jackpot!
Jul 21 09:48:36 <praveendath92> I rendered it on phone.
Jul 21 09:48:44 <vvu> what was hacky there?
Jul 21 09:49:15 <praveendath92> The initial green screen was rendered correctly but I'm getting total black screen with xserver start screen.
Jul 21 09:49:34 <praveendath92> I removed all the loops and added my USB bulk code.
Jul 21 11:31:00 <Abhishek_> praveendath92: Good to know :)
Jul 21 13:25:29 <jkridner> karki: you around?
Jul 21 13:25:53 <jkridner> karki_: you around?
Jul 21 13:25:58 <jkridner> we started a g+ hangout.
Jul 21 13:26:56 <vvu> praveendath92: the last git commit is an actual usb data? or just paste raw from /dev/dl?
Jul 21 13:27:10 <Abhishek_> hi jkridner
Jul 21 13:27:23 <praveendath92> vvu: Actual usb data.
Jul 21 13:27:36 <praveendath92> A bit tricky though.
Jul 21 13:27:42 <vvu> why so ?
Jul 21 13:28:17 <praveendath92> I'm reading all the data to a file.
Jul 21 13:28:18 <praveendath92> Once I see that the transfer is complete. I rename the file to bard.raw
Jul 21 13:28:28 <vvu> mhmh
Jul 21 13:28:52 <vvu> can't you include some of those CMDs
Jul 21 13:28:56 <vvu> to find out where to put the line ?
Jul 21 14:51:29 <jkridner> hi Abhishek_
Jul 21 15:02:56 <praveendath92> av500: ping
Jul 21 15:05:49 <karki_> alexanderhiam : the lbl inst won't work for forward jumps
Jul 21 15:06:43 <alexanderhiam> it can, the interpreter just has to have a way to resolve instruction nnumbers to absolute instruction addresses
Jul 21 15:09:02 <av500> praveendath92: pong
Jul 21 15:09:35 <praveendath92> You know what a page in the framebuffer contains?
Jul 21 15:10:04 <praveendath92> I means a page = line ?
Jul 21 15:10:05 <praveendath92> or something else?
Jul 21 15:10:18 <av500> no
Jul 21 15:10:23 <av500> a page is 4K
Jul 21 15:10:27 * vvu never heard work page except mmu page
Jul 21 15:10:29 <av500> a framebuffer can have any size
Jul 21 15:10:35 <vvu> s/work/word
Jul 21 15:10:46 <praveendath92> Oh.
Jul 21 15:10:57 <av500> a 320x240 framebuffer would waste a lot of space if every page was a line only
Jul 21 15:11:06 <av500> praveendath92: but I dont understand
Jul 21 15:11:20 <av500> from the kernel point of view, the framebuffer is just a piece of memory
Jul 21 15:11:26 <praveendath92> I was able to get the 1st frame right by simply making a buffer of each line placed side by side.
Jul 21 15:11:27 <av500> where do pages come into play?
Jul 21 15:12:01 <praveendath92> When there is an update in frame, it is checking for page faults and sending only those frames that are damaged.
Jul 21 15:12:13 <praveendath92> I'm talking about from the second frame.
Jul 21 15:13:30 <praveendath92> I tried putting all the pages into a buffer but it is reading a complete black image.
Jul 21 15:13:39 <praveendath92> :(
Jul 21 15:15:38 <av500> ah
Jul 21 15:15:45 <av500> clever
Jul 21 15:15:53 <av500> but who is "it"?
Jul 21 15:16:00 <av500> the udlfb driver?
Jul 21 15:16:39 <praveendath92> Yes.
Jul 21 15:17:59 <av500> but again, that should not be your concern
Jul 21 15:18:02 <av500> you have 2 buffers
Jul 21 15:18:07 <av500> one local, one remote
Jul 21 15:18:11 <av500> same size etc...
Jul 21 15:18:19 <av500> the driver tells you what bytes to transfer, you do that
Jul 21 15:18:26 <av500> then you redraw on android side
Jul 21 15:18:29 <praveendath92_> Yes.
Jul 21 15:18:49 <av500> so?
Jul 21 15:19:04 <praveendath92_> When starting the xserver, all the data changes so, i assume everything is sent.
Jul 21 15:19:30 <praveendath92_> I mean all pages => all lines.
Jul 21 15:19:49 <av500> ok
Jul 21 15:19:51 <praveendath92_> But I'm getting a complete black image with that
Jul 21 15:20:17 <praveendath92_> PS: The 1st frame worked right. A full green screen.
Jul 21 15:20:35 <av500> is it really all black?
Jul 21 15:20:39 <av500> looked at the bytes
Jul 21 15:20:50 <av500> or maybe you render it wrong?
Jul 21 15:20:54 <av500> RGB vs RGBA or so?
Jul 21 15:20:58 <praveendath92_> I will check.
Jul 21 15:21:08 <vvu> praveendath92_: hexdump is your best friend
Jul 21 15:21:11 <praveendath92_> The driver is written to handle RGB565
Jul 21 15:21:34 <praveendath92_> Yes. vvu, i will try with the hex dump.
Jul 21 15:21:39 <av500> is the framebuffer 565?
Jul 21 15:21:46 <av500> on the linux side?
Jul 21 15:21:51 <praveendath92_> Just tried w/o using page fault mechanism.
Jul 21 15:21:52 <praveendath92_> Yes
Jul 21 15:22:19 <praveendath92_> I got a page fault error :/
Jul 21 15:22:36 <praveendath92_> I will check the hex dump
Jul 21 15:22:37 <av500> pastebin the code pls
Jul 21 15:22:42 <av500> and the error
Jul 21 15:23:25 <praveendath92_> av500: code is on git.
Jul 21 15:23:50 <praveendath92_> Since it is better to use pagefault checking, i will try the hex dump
Jul 21 15:39:12 <av500> what do you change to disable this checking?
Jul 21 15:40:02 <praveendath92_> modprobe udlfb dl_defio=0
Jul 21 15:40:34 <av500> ok, disable that for now
Jul 21 15:40:38 <av500> it should work without
Jul 21 15:40:52 <praveendath92_> fb_defio*
Jul 21 15:41:01 <av500> yes
Jul 21 15:41:03 <av500> I see that
Jul 21 15:43:52 <praveendath92_> I just tried. Looks like it is still sending pages.
Jul 21 15:46:18 <praveendath92_> Hexdump has some whites too.
Jul 21 15:46:40 <praveendath92_> I should check the image reader in java.
Jul 21 15:53:17 <av500> ok
Jul 21 16:05:36 <praveendath922> vvu: ping
Jul 21 16:05:52 <praveendath922> vvu: http://imagebin.ca/v/1U67gzqdxPyk
Jul 21 16:06:49 <praveendath922> Can you confirm if that page has an image. I'm getting a broken image page everytime I try. Just want to make sure it's not because of my net.
Jul 21 16:08:51 <karki_> my friend is interested in OS development but not hardware! any tips where he can start off?
Jul 21 16:13:29 <av500> try SW maybe? :)
Jul 21 16:13:43 <vvu> praveendath922: i have no clue what to check
Jul 21 16:13:50 <av500> there is a ton of books on OS design
Jul 21 16:13:52 <vvu> you are puting data wrong in the buffer i think so
Jul 21 16:14:18 <av500> praveendath922: no image
Jul 21 16:14:23 <vvu> karki_: the dinosaur book
Jul 21 16:15:41 <Abhishek_> jkridner: Just to let you know that there are no more non-technical/urgent blockers
Jul 21 16:18:47 <praveendath922> I should the udlfb to verify the data I'm sending.
Jul 21 16:19:11 <vvu> praveendath922: do a hexdump of a single frame from udlfb and compare it to the stuff u have on android
Jul 21 16:19:16 <vvu> check what is wrong there
Jul 21 16:19:16 <praveendath922> I'm guessing that some data is missing might be the problem here.
Jul 21 16:19:43 <praveendath922> vvu: Just tried that.
Jul 21 16:20:09 <av500> praveendath922: you can use imgur
Jul 21 16:20:16 <praveendath922> The 1st line itself differs.
Jul 21 16:20:20 <praveendath922> :(
Jul 21 16:21:09 <karki_> vvu : he really wants to build a OS, in ARM assembly while handling interrupts etc. I don't know how he will go about it if he doesn't really like hardware!
Jul 21 16:21:24 <karki_> vvu : on an actual board
Jul 21 16:21:58 <vvu> karki_: this is way out of my league if so :)
Jul 21 16:22:17 <praveendath922> av500: Ah. That's what I was looking for. The site that can read raw images. I got a wrong one there before.
Jul 21 16:22:29 <karki_> vvu : I would normally get started by building some simple io control - something he doesn't want to do :p
Jul 21 16:22:48 <praveendath922> karki_ that sounds like a cool task.
Jul 21 16:23:02 <karki_> yeah I know
Jul 21 16:23:03 <praveendath922> Wondering how difficult that would be.
Jul 21 16:23:09 <karki_> he is in his 1st year
Jul 21 16:23:14 <karki_> thats the problem
Jul 21 16:23:16 <praveendath922> Interrupts is fine I suppose.
Jul 21 16:23:41 <karki_> yep, but not going into h/w will give him a lot of trouble later!
Jul 21 16:23:45 <Abhishek_> karki_: building an OS for an ARM core is not that hard, but there are so many OSes for ARM out there that he might want to use them - or if he really has a requirement that the other OSes don't have, or wants to learn it in the process, then it should be good enough
Jul 21 16:23:46 <praveendath922> Ah. The beginner enthu. That's nice actually.
Jul 21 16:24:01 <karki_> :)
Jul 21 16:24:21 <karki_> Abhishek_ : he wants to learn, thats it.
Jul 21 16:24:40 <praveendath922> Abhishek_ has a good point.
Jul 21 16:24:44 <Abhishek_> karki_: he could start with protothreads
Jul 21 16:24:55 <av500> praveendath922: why raw?
Jul 21 16:24:59 <av500> convert it
Jul 21 16:25:05 <av500> I doubt they can read raw
Jul 21 16:25:07 <karki_> building a decent OS is not easy, unless you have something to work with before hand
Jul 21 16:25:33 <praveendath922> karki_ Making one from scratch could be overwhelming.
Jul 21 16:25:37 <karki_> Abhishek_ : think mmu, io, scheduler, semaphores etc
Jul 21 16:25:41 <Abhishek_> karki_: The real issue with building an OS is portability - take a look at the Linux kernel, for example :)
Jul 21 16:26:01 <karki_> how will protothreads help?
Jul 21 16:26:06 <praveendath922> av500: The last time you sent me a raw image. What was that site?
Jul 21 16:26:19 <av500> forgot
Jul 21 16:26:27 <praveendath922> That's the raw data I got on Android device.
Jul 21 16:26:31 <praveendath922> For 2nd frame.
Jul 21 16:26:34 <av500> so?
Jul 21 16:26:39 <av500> *look* at it
Jul 21 16:26:48 <av500> you dont need imagebin to do that
Jul 21 16:26:51 <av500> you have eyes, no?
Jul 21 16:26:58 <av500> also, you can look at the hey data
Jul 21 16:27:04 <av500> hex
Jul 21 16:27:06 <av500> and compare
Jul 21 16:27:10 <av500> you have all the info
Jul 21 16:27:25 <praveendath922> My imageviewer isn't working on RGB565 images.
Jul 21 16:27:30 <av500> so?
Jul 21 16:27:40 <av500> you are not the first one to have that problem
Jul 21 16:27:43 <av500> I am sure
Jul 21 16:27:59 <av500> then look at the hex
Jul 21 16:27:59 <praveendath922> Yes. I looked at the data. There are differences.
Jul 21 16:28:03 <av500> what do you see?
Jul 21 16:28:14 <av500> you have an xterm in the top tright colors
Jul 21 16:28:17 <av500> corner
Jul 21 16:28:22 <av500> the rest is black
Jul 21 16:28:27 <av500> you should able to see a patterns
Jul 21 16:28:31 <av500> its only a few colors
Jul 21 16:28:33 <av500> etc..
Jul 21 16:28:48 <praveendath922> A lot of 0s and fs. So, it can't be a completely black image for sure.
Jul 21 16:29:01 <av500> right
Jul 21 16:29:02 <Abhishek_> How could I use git send-email over a SOCKS proxy? I was able to set up pull & push but this is a little painful for me.
Jul 21 16:29:08 <praveendath922> Yes. yes.
Jul 21 16:29:15 <av500> so, now find a SW to convert 565 to something you can look at
Jul 21 16:29:18 <av500> I need to leave
Jul 21 16:29:19 <av500> bbl
Jul 21 16:31:02 <praveendath922> imagemagik.
Jul 21 16:31:03 <praveendath922> I will do the conversion and see.
Jul 21 16:32:23 <Abhishek_> praveendath922: You have a raw image file right now?
Jul 21 16:33:10 <praveendath922> Yes
Jul 21 16:33:29 <praveendath922> Many of them.
Jul 21 16:33:42 <praveendath922> :)
Jul 21 17:54:55 <mranostay> Abhishek_: google it. i'm sure someone has done it :)
Jul 21 17:55:12 <mranostay> inside the borg the stupid proxy is the bane of my existance :)
Jul 21 17:55:42 <mranostay> also fun when the proxy address is different between campuses :)
Jul 21 17:59:51 <Abhishek_> we have about 5-6 different proxies, but all are the same in terms of blocking traffic and almost all the ports
Jul 21 18:50:31 <Abhishek_> mranostay: figured it out . "usewithtor git send-email blah"
Jul 21 18:52:16 <Abhishek_> that was it
Jul 21 18:53:06 <vvu> rambooting has been defeated!
Jul 21 19:53:30 <karki_> alexanderhiam, jkridner : what should the PRU do if it finds an out of bound array access?
Jul 21 20:01:23 <alexanderhiam> karki_: it should probably abort if it's runnign a script. Ideally it would report the error, which could be as simple as writing a single byte error code to a reserved location in the shared memory so the userspace program could look there to find out what went wrong
Jul 21 20:01:43 <karki_> hm..
Jul 21 20:01:59 <karki_> when would the userspace prog check the ret val?
Jul 21 20:02:22 <karki_> how would it be informed of a crashed script?
Jul 21 20:02:33 <karki_> thats what I meant
Jul 21 20:09:13 <karki_> alexanderhiam?
Jul 21 20:10:34 <alexanderhiam> hmm, you could have an "I'm alive" flag that the interpreter sets every time through its loop, then clear the flag every once in a while from the userspace program
Jul 21 20:10:54 <alexanderhiam> then if the flag remains clear for > some time you know the program is stuck
Jul 21 20:11:21 <alexanderhiam> or just clear the error flag every time a script starts and check it every once in a while from the userspace program
Jul 21 20:11:41 <karki_> ok :)
Jul 21 20:12:28 <alexanderhiam> you could use a threading.Timer instance in the Python program to check it every second
Jul 21 20:12:38 <karki_> yeah
Jul 21 20:13:05 <karki_> I was also thinking some other ways, but we'll keep it simple for now!
Jul 21 20:13:32 <karki_> actually there is a sysfs called pru_speak_state
Jul 21 20:14:01 <karki_> that prints 1 if you cat it while script is running.
Jul 21 20:14:07 <alexanderhiam> is it guaranteed to report correctly though?
Jul 21 20:14:10 <karki_> yes
Jul 21 20:14:19 <karki_> 100%
Jul 21 20:14:38 <karki_> unless the interpreter itself crashes......
Jul 21 20:14:50 <karki_> in which case we will have a kernel log instead
Jul 21 20:15:03 <alexanderhiam> right, that's the benefit of the "I'm alive" flag
Jul 21 20:15:20 <karki_> I added it long back for easy debugging
Jul 21 20:15:22 <karki_> :)
Jul 21 20:15:44 <karki_> anyway, almost 2:00AM here. I'll code for a while and sleep :p
Jul 21 20:16:25 <Abhishek_> I'm going to stay up for some more time today and push the web interface harder
Jul 21 20:20:52 <karki_> jason will be helping me out with my web interface ;)
Jul 21 20:21:32 <Abhishek_> nice :) . Do you have a mockup?
Jul 21 20:22:33 * Abhishek_ would call for help once he's ready with a basic interface
Jul 21 20:23:54 <karki_> "help" => jason will be doing it for me ;)
Jul 21 20:24:12 <karki_> atleast the frontend
Jul 21 20:24:25 <karki_> I'll be doing the webserver part
Jul 21 20:24:41 <Abhishek_> in twisted?
Jul 21 20:24:52 <karki_> nah
Jul 21 20:25:12 <karki_> no frameworks
Jul 21 20:25:37 <karki_> simple HTTP server that comes inbuilt into python
Jul 21 20:26:14 <karki_> maybe later move to socket.io
Jul 21 20:39:13 <Abhishek_> hmm, yes, so you will be able to concentrate better on the JIT part\
**** ENDING LOGGING AT Tue Jul 22 03:00:01 2014