**** BEGIN LOGGING AT Wed May 23 03:00:04 2018
**** BEGIN LOGGING AT Wed May 23 08:28:11 2018
May 23 15:57:22 <jkridner[m]> @vaishnav98_ great patch. applied.
May 23 15:57:32 <jkridner[m]> probably should make a 0.7.0 release soon to start getting some users.
May 23 15:57:46 <jkridner[m]> maybe do one later this week? thoughts?
May 23 15:58:02 <jkridner[m]> via the -beta tags maybe?
May 23 16:06:02 <vaishnav98_> jkridner[m]: Thanks :) , I support making a --beta release soon , so that we could get more feedback  about the changes.
May 23 16:13:03 <ds2> ...
May 23 16:20:03 <cwicks[m]> Hi GSOC friends
May 23 16:20:14 <cwicks[m]> jkridner: I will miss most of today's meeting - I'm going to stop by NI Week
May 23 16:26:27 <vaishnav98_> Hi all
May 23 16:28:54 <anirban1998[m]> Hello everyone!
May 23 16:29:55 <muneeb17[m]> Hello all
May 23 16:30:01 <ravikp7> Hi all
May 23 16:30:19 <abhishek_[m]> Hi all
May 23 16:32:58 <welsh[m]> hi
May 23 16:33:04 <abhishek_[m]> All students are present.
May 23 16:33:52 <anirban1998[m]> Hi welsh abhishek_
May 23 16:39:35 <jkridner> sorry I'm late.
May 23 16:39:37 <jkridner> who all is here?
May 23 16:40:04 <zeekhuge> jkridner  : Will be here in 10 minutes.
May 23 16:40:16 <abhishek_[m]> Students: Are everyone on track? Is anyone stuck with anything that we can help with?
May 23 16:40:22 <jkridner> vaishnav98_, ravikp7, muneeb17[m], anirban1998[m]: who is here?
May 23 16:40:39 <vaishnav98_> jkridner: I am here.
May 23 16:40:40 <ravikp7> I'm here
May 23 16:40:51 <muneeb17[m]> jkridner:  I'm here
May 23 16:40:53 <anirban1998[m]> jkridner: Me
May 23 16:40:55 <abhishek_[m]> jkridner: Everyone are
May 23 16:41:22 * jkridner remembers he needs to give some feedback to some folks that still didn't feel they got good feedback on not being accepted and will do that via direct-message now if anyone would like.
May 23 16:42:07 <jkridner> sweet.
May 23 16:42:35 <jkridner> shall we go around the room, maybe in alphabetical order?
May 23 16:42:49 <jkridner> anirban1998[m]: got any blocking issues?
May 23 16:43:15 <jkridner> anirban1998[m]: did you have a status report since 5/21?
May 23 16:43:23 <jkridner> nm, that was Monday! :-D
May 23 16:43:59 <jkridner> anirban1998[m]: still having headset problems?
May 23 16:44:21 <anirban1998[m]> jkridner: yeah
May 23 16:44:27 <anirban1998[m]> I think it is a power issue
May 23 16:44:44 <ds2> anirban1998[m]: how are you powering things?
May 23 16:44:47 <jkridner> got a powered hub?
May 23 16:45:09 <jkridner> what model? does it show up in a supported list of Linux devices?
May 23 16:45:43 <anirban1998[m]> jkridner: Will have to borrow or buy a USB power hub from somewhere...it is not getting enough power from the PC USB port
May 23 16:46:29 <welsh[m]> The audio adapter is:  https://www.amazon.com/Plugable-Headphone-Microphone-Aluminum-Compatibility/dp/B00NMXY2MO/
May 23 16:47:29 <jkridner> ah, that guy. works for sure. Guess power *could* be an issue. What about the headset itself?
May 23 16:48:39 <anirban1998[m]> The headset is fine...working on my PC...but not on pocketbeagle due to the power problem I feel
May 23 16:49:48 <jkridner> anirban1998[m]: is this blocking or just inconvenient?
May 23 16:51:45 <jkridner> anirban1998[m]: I think we need to pay you back for some duties. Would some kind of cash reimbursement help with getting a powered hub?
May 23 16:51:55 * jkridner isn't sure what could happen fast.
May 23 16:52:04 <jkridner> for time's sake, do we need to keep going around the room?
May 23 16:54:16 <jkridner> hendersa, ds2, welsh[m], others: comments/questions before moving to muneeb17[m]? anirban1998[m] can continue to answer in the background and we can all work together to fix the headset issue.
May 23 16:55:16 <ds2> other then the 'sphinx accuracy may still be a stumbling block but we will get to it once the HW works better'
May 23 16:55:50 <jkridner> muneeb17[m]: got any blocking issues?
May 23 16:56:03 <anirban1998[m]> jkridner: No problem...I will buy the powered hub as soon as possible...you can pay back my duties anytime
May 23 16:56:32 <muneeb17[m]> I'm on track. Working with the swig bindings now. They're working fine for python and perl. Yet to try with other languages. No blocking issues here.
May 23 16:56:44 <ds2> anirban1998[m]: for just powering, do you have a recent phone charger?
May 23 16:57:13 <jkridner> just so I understand, the swig if for the UIO API or something else?
May 23 16:57:50 <jkridner> maybe mmap?
May 23 16:58:00 <jkridner> why is __init__.py empty?
May 23 16:58:19 <muneeb17[m]> jkridner:  its just a tool to generate languages bindings
May 23 16:58:32 <jkridner> yes, but what libraries are you binding?
May 23 16:58:36 <anirban1998[m]> ds2: Yeah I have.
May 23 16:59:00 <ds2> anirban1998[m]: do you have the FTDI cable too?
May 23 16:59:22 <muneeb17[m]> jkridner:  the C functions for remoteproc/rpmsg
May 23 16:59:51 <muneeb17[m]> so that i dont have to write for each language separately
May 23 16:59:53 * jkridner didn't know remoteproc had userspace C functions.
May 23 16:59:57 <anirban1998[m]> ds2 I have it.
May 23 17:00:25 <muneeb17[m]> jkridner:  I'm writing functions using the sysfs interfaces
May 23 17:00:39 <jkridner> odd.
May 23 17:00:59 <jkridner> if it is sysfs interfaces, why wouldn't you just natively connect each language?
May 23 17:01:14 <jkridner> zeekhuge: you guys aligned on the approach?
May 23 17:02:38 <jkridner> abhishek_[m]: ?
May 23 17:02:45 <zeekhuge> jkridner:  The remoteproc and RPMsg interface is based on sysfs. So yeah. we are alinged on it.
May 23 17:02:57 <zeekhuge> abhishek_: ^ you think so . right ?
May 23 17:03:45 <jkridner> even writing a single-language (C?) API and then using swig to interface python/perl?
May 23 17:03:48 <zeekhuge> muneeb17: You need to make commits and push them on daily basis.
May 23 17:04:07 <jkridner> does the kernel not expose any ioctl or other interface for RPMsg/remoteproc?
May 23 17:04:38 * jkridner has found communication with vaishnav98_ to be primarily based on commits and pull requests.  has been very efficient.
May 23 17:04:43 <abhishek_[m]> jkridner's question is whether generating bindings for other langs should be done through swig or bindings be generated individually for each lang
May 23 17:04:46 <zeekhuge> muneeb17:  Also, bonus marks for OOP based code. Like creating Classes and their methods and that.
May 23 17:05:47 <jkridner> abhishek_[m]: exactly. if you are just poking sysfs file entries, doesn't seem like making a single language API is that desirable.
May 23 17:05:49 <abhishek_[m]> jkridner: No ioctl's. Just read and write so far
May 23 17:05:49 <muneeb17[m]> zeekhuge:  Yeah, i'll modify the code for OOP
May 23 17:06:13 <zeekhuge> jkridner:  We already know that the project will be about writting interface code for sysfs. Although I wasnt aware of tool like `swig` until soon.
May 23 17:06:41 <abhishek_[m]> ZeekHuge: Had the same thought on OOP approach. An object for each PRU which has methods like start/stop and stuff
May 23 17:07:15 <jkridner> muneeb17[m]: so, nothing blocking?
May 23 17:07:26 <jkridner> abhishek_[m], zeekhuge: sounds like you guys are moving forward without barriers?
May 23 17:07:28 <zeekhuge> *I mean, came to know about the `swig` tool only after the project started
May 23 17:07:41 <anirban1998[m]> welsh ds2 hendersa Have you reviewed my code? Is there anything up to now on which I can improve in my code?
May 23 17:08:10 <muneeb17[m]> jkridner:  no issues, as such
May 23 17:08:26 <muneeb17[m]> abhishek_:  sounds good. I'll work on that.
May 23 17:09:15 <jkridner> ravikp7: any blockers? close to providing that pull request to etcher?
May 23 17:09:31 <zeekhuge> jkridner: Yeah. looks like no barriers for now.
May 23 17:09:34 <jkridner> ravikp7: got any users? got a built mac image I can just install without building?
May 23 17:09:43 <jkridner> not that mac will work. :-)
May 23 17:09:48 <ravikp7> not any blocking issue, but spl usb boot still doesn't work on osx and windows
May 23 17:09:48 <hendersa> anirban1998[m]: I have not had an opportunity to review your code yet. Perhaps one of the others can offer a recommendation. My primary concern is getting your hardware up and running. :)
May 23 17:09:49 <jkridner> so, Linux I mean. :-)
May 23 17:10:13 <jkridner> ravikp7: do you have a methodology for data-logging all USB transactions?
May 23 17:10:41 <jkridner> ravikp7: I think there's something with SPL now providing both ECM and RNDIS interfaces.
May 23 17:10:45 <ravikp7> jkridner: https://github.com/ravikp7/etcher/tree/beagleboot , I think I've fixed all bugs, it's ready for pull request
May 23 17:10:46 <abhishek_[m]> Wireshark?
May 23 17:11:01 <jkridner> ravikp7: but for Linux only, no?
May 23 17:11:14 <jkridner> ravikp7: does it "smartly" not impact other OSes?
May 23 17:11:28 <ravikp7> jkridner: I can't on mac, I think I'll give it a shot on Windows
May 23 17:11:47 <welsh[m]> anirban1998:  Unfortunately, I haven't had a chance to review the code either.  hoping to take a look this weekend.
May 23 17:11:57 <abhishek_[m]> Hmm, ECM interface? Haven't seen that before.
May 23 17:12:41 <ds2> anirban1998[m]: I am looking mainly at the speech stuff
May 23 17:12:44 <anirban1998[m]> welsh: hendersa ds2 I am aiming to add more robust features to my speech recognition model...right now...am successful in launching games...using only specific keywords...want to improve on that...any suggestions?
May 23 17:13:13 <ravikp7> jkridner: couldn't understand it
May 23 17:13:35 <zeekhuge> muneeb17: Can you create a "wip" (Work In Progress) branch and upload a sample of the code/bindings created by `Swig` for the mentors to review it  ?
May 23 17:14:09 <ds2> anirban1998[m]: are you just working on this yourself? it may be useful if you get other people to try it (maybe later) to get a better picture of hte speech stuff
May 23 17:14:19 <muneeb17[m]> heres the repo https://github.com/MuneebMohammed/PRUSS-Bindings
May 23 17:14:23 <anirban1998[m]> anujdeshpande: It would be very useful if you provide some audio files...for generating accuracy statistics...
May 23 17:15:10 <jkridner|pd> ravikp7: Wireshark, as abhishek_[m] mentions, is what I was thinking as the cheapest approach. Might also be possible to put another BeagleBone in the loop to sniff, but don't know if the pass-through drivers really work anymore.
May 23 17:15:55 <muneeb17[m]> zeekhuge:  as you suggested, i'll merge pypruss with it and other bindings as well.
May 23 17:17:17 <jkridner> ravikp7: I think you likely need some help with the u-boot stuff. Your query might get some response on the u-boot list if posed properly.
May 23 17:17:33 <anirban1998[m]> ds2 Didn't get u..
May 23 17:17:56 <ravikp7> jkridner: I'll setup cross compile on Windows this week, and will use wireshark to get the usb transactions info
May 23 17:18:30 <jkridner> Does Wireshark run on Windows?
May 23 17:18:34 <abhishek_[m]> ravikp7: Cc me in the email to uboot list
May 23 17:18:36 <ravikp7> jkridner: sure, will post there with usb logs
May 23 17:19:08 <ravikp7> I think so
May 23 17:19:09 <ds2> anirban1998[m]: while sphinx is suppose to be speaker independent, diferent speakers can have different accuracy
May 23 17:19:15 <abhishek_[m]> jkridner: It can, but not sure how it will capture usb tea
May 23 17:19:16 <zeekhuge> abhishek_:  (Wormo might be late here) : do we agree to use `swig` instead of actual code ?
May 23 17:19:23 <abhishek_[m]> *traffic
May 23 17:19:41 <welsh[m]> Wireshark has a windows version:  https://www.wireshark.org/download.html
May 23 17:19:48 <abhishek_[m]> ZeekHuge: Wormo might not make it today
May 23 17:19:50 <jkridner> zeekhuge: do you not consider 'swig' actual code? I mean, you still need something for a swig to call.
May 23 17:20:26 <jkridner> abhishek_[m]: yeah, I'd be concerned about that.
May 23 17:20:30 <jkridner> it does work on Windows.
May 23 17:20:55 <welsh[m]> Wireshark can capture RNDIS over USB pretty easily.  However, to do straight USB, there are a couple other things you have to add.
May 23 17:21:03 <jkridner> er, Mac
May 23 17:21:12 <ravikp7> jkridner, abhishek_: if you can, please review and try the etcher fork with beagleboot, so that I can then send them pull request
May 23 17:21:22 <jkridner> welsh[m]: can you give pointers?
May 23 17:21:33 <abhishek_[m]> Let me see if I can try over the weekend.
May 23 17:21:36 <jkridner> ravikp7: do you have a pre-built Linux image?
May 23 17:21:47 <jkridner> I don't want to deal with building right now.
May 23 17:21:51 <zeekhuge> jkridner:  I am not sure the code quality it will produce. I think code-quality should be good and should be understandable enough for someone to hack/modify/learn from it
May 23 17:22:21 <anirban1998[m]> ds2 Yeah...true...but will that change in accuracy not affect the end results? And even if it would...I feel that we would need to add a feature to ask to re-enter the words
May 23 17:22:21 <zeekhuge> jkridner: And I have not used that tool. Would you suggest using it ?
May 23 17:22:24 <welsh[m]> https://wiki.wireshark.org/CaptureSetup/USB
May 23 17:22:27 <welsh[m]> There is a windows section.
May 23 17:22:49 <ravikp7> I'll try to give you pre built installers to try
May 23 17:22:59 <abhishek_[m]> jkridner: The way SWIG will work is one writes the code in C, and SWIG generates the python and other language stuff from that automatically
May 23 17:23:13 <jkridner> abhishek_[m]: yes, I know.
May 23 17:23:57 <jkridner> abhishek_[m]: you'd still generally write both C and Python around the SWIG.
May 23 17:24:01 <ds2> anirban1998[m]: there are 2 things - 1 is just knowing how useful it is and 2 - how to work around
May 23 17:24:36 <jkridner> before we completely run out of time...
May 23 17:24:40 <jkridner> vaishnav98_: any blockers?
May 23 17:25:16 <zeekhuge> jkridner: abhishek_ : I would suggest that we use `swig` to generate basic code and then edit it to make it more readable etc. ?
May 23 17:25:46 <jkridner> vaishnav98_: how do you feel you are on pace for some of the stretch goals?
May 23 17:26:40 <abhishek_[m]> ZeekHuge: I'd prefer an automated approach, manually editing generated code seems like a repetition of effort
May 23 17:27:24 <anirban1998[m]> ds2 Are you talking about the accuracy detection?
May 23 17:27:27 <muneeb17[m]> zeekhuge:  Swig works somewhat differently. It generated swig specific (.py file for example) and a wrapper to generate the module
May 23 17:27:33 <ds2> anirban1998[m]: yes
May 23 17:27:33 <vaishnav98_> jkridner: I was trying to look into the issue of installing bonescript on snappy Ubuntu core , I saw that it isn't officially supported now for BB could get an image from here http://releases.ubuntu.com/15.04/ubuntu-15.04-snappy-armhf-bbb.img.xz   tried it : but enable classic mode fails throwing errors that many packages weren't installed. I am not sure how to proceed with this
May 23 17:27:34 <abhishek_[m]> ZeekHuge: see https://github.com/MuneebMohammed/PRUSS-Bindings/blob/master/C/pruss.py
May 23 17:27:50 <muneeb17[m]> https://github.com/MuneebMohammed/PRUSS-Bindings/blob/master/C/pruss.py like this
May 23 17:28:15 <abhishek_[m]> muneeb17: Learn about using gitignore
May 23 17:28:33 <abhishek_[m]> Once again you are committing binary files
May 23 17:29:09 <muneeb17[m]> abhishek_:  Ok, will take care of that.
May 23 17:29:29 <zeekhuge> And intermediate files too ?
May 23 17:29:40 <jkridner> vaishnav98_: hmm.... I need to ping ogra.
May 23 17:29:44 <abhishek_[m]> Yep
May 23 17:29:54 <zeekhuge> also, please organize it a bit. Like `bindings` in different directory and `config` file or whaterver they are in different directories
May 23 17:31:01 <anirban1998[m]> ds2 I know...that's why automated the accuracy checker...based on it's accuracy statistics have to tune the AGC mic accordingly I feel
May 23 17:31:24 <abhishek_[m]> jkridner: Are we done with going around the room?
May 23 17:31:27 <jkridner> k, pinged ogra_ on #snappy
May 23 17:31:40 <jkridner> abhishek_[m]: just wrapping up vaishnav98_
May 23 17:31:44 <muneeb17[m]> zeekhuge:  Ok, will do that.
May 23 17:31:46 <zeekhuge> muneeb17:  I dont get the README. Why is it telling me "how to use `swig`" instead of describing the bindings ?
May 23 17:31:51 <jkridner> he's got some Snappy blockers..
May 23 17:32:19 <jkridner> vaishnav98_: can you start with the latest Ubuntu 18.04 instead? anyway to downsize it to snappy?
May 23 17:32:26 <abhishek_[m]> jkridner: Okay.
May 23 17:32:51 <vaishnav98_> jkridner: I feel that it would be possible to work on the stretch goals if everything goes smoothly. I had a look at https://github.com/Microsoft/pxt-microbit  , would it be feasible to start from it for the blocky-bonescript project?
May 23 17:33:29 <muneeb17[m]> zeekhuge:  just instructions to generate the bindings. Yeah, doesn't make sense actually. Sorry, I dont know what i was thinking.
May 23 17:33:59 <vaishnav98_> jkridner: I will try to and will inform the results.
May 23 17:34:35 <jkridner> vaishnav98_: https://rcn-ee.net/rootfs/bb.org/testing/2018-05-18/bionic-ros-iot/
May 23 17:35:07 <jkridner> vaishnav98_: isn't that Makecode instead of Blockly? are they the same?
May 23 17:35:28 <jkridner> vaishnav98_: MIT license, so looks reasonable.
May 23 17:36:09 <jkridner> hmmmm.... Electron app... looks interesting to merge with BeagleBoot and a buildroot image.
May 23 17:36:39 <jkridner> is Makecode actually Blockly?
May 23 17:37:00 <ds2> anirban1998[m]: just pointing out one voice is not enough
May 23 17:37:00 * jkridner doesn't really have a strong feel for which to use... just make something work.
May 23 17:37:14 * jkridner thinks the bonescript simulator and BBUI would be needed.
May 23 17:37:27 <vaishnav98_> jkridner: also I have some doubts regarding bringing in support for HCSR04 sensor , what would be the best way to bring pru support to bonescript, I was thinking about remoteproc sysfs, is it okay?
May 23 17:38:40 <vaishnav98_> jkridner: Makecode is based on blockly , thanks for the image.
May 23 17:38:46 <anirban1998[m]> ds2: Then How to make it applicable easily for all users?
May 23 17:39:16 <zeekhuge> muneeb17: Can you please clear the `https://github.com/MuneebMohammed/PRUSS-Bindings` repo of all intermediate and binary files, organize the source code a bit and then ping me ?
May 23 17:39:45 <ds2> anirban1998[m]: that is the challenge. once things get far enough, getting at least some idea of how it will behave w/other voices (and accents) would be useful/helpful
May 23 17:39:50 <muneeb17[m]> zeekhuge:  yeah, just a min
May 23 17:39:51 <ds2> anirban1998[m]: it won
May 23 17:39:55 <ds2> t be prefect
May 23 17:40:26 <ds2> anirban1998[m]: this is partially why they have moved to ML for alot of the speech stuff thesesday but that is (AFAIK) too heavy for an embedded device
May 23 17:40:27 <zeekhuge> muneeb17:  btw, use `gitignore` to clear the repo.
May 23 17:40:52 <jkridner> vaishnav98_: ogra_ has daily snappy builds at http://people.canonical.com/~ogra/snappy/all-snaps/daily-stable/current/
May 23 17:40:55 <vaishnav98_> jkridner: sure , I was proposing that only for the blocky-bonescript part , will start working on BBUI as soon as the boards arrive, got confirmation that it cleared customs.
May 23 17:40:58 <jkridner> use #snappy for support.
May 23 17:41:08 <jkridner> I just got that from him.
May 23 17:41:15 * abhishek_[m] wonders if Mozilla DeepSpeech can run on the PB
May 23 17:41:22 <jkridner> I'd also think you'd want to try Robert's Bionic builds.
May 23 17:41:28 <jkridner> Bionic == Ubuntu 18.04.
May 23 17:41:35 <ds2> abhishek_[m]: wanna port it? :D
May 23 17:42:13 <abhishek_[m]> ds2: Not yet.
May 23 17:43:19 <vaishnav98_> jkridner: sure i will try both the sources, and use the #snappy channel for support
May 23 17:44:31 <anirban1998[m]> ds2 I will surely work on it...yeah it won't be perfect...but need to keep improving on it...will nltk libraries also prove heavy for embedded?
May 23 17:44:54 <jkridner> sorry, very distract4ed by #snappy channel
May 23 17:45:04 <jkridner> I need to run too.
May 23 17:45:10 <jkridner> all wrapped up?
May 23 17:45:22 <jkridner> sorry we started late and ran long...
May 23 17:45:30 <jkridner> I'll try to manage our time a bit better next time.
May 23 17:45:52 <ds2> anirban1998[m]: maybe... but I would avoid using that, at least initially. I suspect going that way may be outside of the GSoC timeframe
May 23 17:46:33 <vaishnav98_> jkridner: it would be really helpful if you could suggest something about the pru firmware loading support issue.
May 23 17:46:57 <anirban1998[m]> ds2: Yeah
May 23 17:47:38 <zeekhuge> I'll be leaving too. muneeb17 : ping me when its done.
May 23 17:47:38 <abhishek_[m]> vaishnav98_: What is your issue with PRU fw loading?
May 23 17:48:35 <abhishek_[m]> vaishnav98_: Put your issue here, or send me an email and let me see if I can take a look.
May 23 17:49:00 <abhishek_[m]> PRU related
May 23 17:50:37 <vaishnav98_> abhishek_[m]: I was planning to integrate pru support in bonescript possibly through sysfs manipulation, I just wanted to know if that would work as a possible solution , or if there are any better methods , thanks for looking into the matter :)
May 23 17:52:29 <abhishek_[m]> Co-ordinate with muneeb17 and see if you both can work together there with regards to generating bindings for nodejs
May 23 17:54:52 <vaishnav98_> abhishek_[m]: sure, will work with him on it.
May 23 18:05:11 * anujdeshpande[m] missed the meeting courtesy of some other meeting
May 23 18:05:22 <anujdeshpande[m]> anirban1998: thanks for reminding! Had almost forgotten about it
May 23 18:13:58 <muneeb17[m]> zeekhuge:  done
May 23 18:46:34 <anujdeshpande[m]> anirban1998:  you around?
**** BEGIN LOGGING AT Wed May 23 19:15:04 2018
May 23 19:45:32 <anujdeshpande[m]> anirban1998: I have sent you a mail with the samples - https://groups.google.com/d/msg/beagleboard-gsoc/2GdTc9WMsA8/vRZCBo9HCwAJ
**** ENDING LOGGING AT Thu May 24 03:00:03 2018