**** BEGIN LOGGING AT Sun Mar 25 03:00:05 2018 Mar 25 04:31:02 hmmm bursty Mar 25 04:41:06 zeekhuge: Yeah. But the link you posted does not finally involve using TIesr on BBB right? It says you are required to be approved as a member to look at the code. But I am able to access the repository just by registering. Mar 25 04:42:20 Hi illustris, will go take a look at the updates Mar 25 04:46:04 Also TIesr is capable of adapting to changing noise environment ( it is usually used in automotive controls) .So I thought it is suitable for the project since it is used by kids. Mar 25 04:46:08 I like the changes to timeline Mar 25 04:46:31 illustris ^^^ Mar 25 04:47:53 Thanks! Are there any other changes you would recommend? Mar 25 04:50:46 under implementation, I like a list of reference URLs for things that look useful for carrying out this project Mar 25 04:51:18 having that list can come in handy when a project is actually underway Mar 25 04:51:50 zeekhuge: Also it is TI's esr. So thought it will better to use that for rebuilding a TI 's product although pocketsphinx is easy to use. I still couldn't find any projects that use tiesr on beagle. Can u tell me if you know any? Mar 25 04:51:59 Wormo: As a part of implementation, or as a new section, say, "Useful resources"? Mar 25 04:53:35 I would tend to put them in as footnotes in the discussion of implementation, and the list after the main discussion Mar 25 04:54:34 but I often see it as a separate section "useful resources" with a brief description of each reference Mar 25 04:54:46 so either way works Mar 25 05:00:24 One issue that I haven't figured out yet is forwarding sequences like ctrl-c Mar 25 05:01:05 That was a problem even in the shell I made. a ctrl-c would kill the shell instead of just forwarding it to the child process Mar 25 05:03:37 There are some nice discussions of signal handling on stackoverflow Mar 25 05:03:59 you might be able to find a relevant one to refer to... Mar 25 05:06:02 πŸ‘ Mar 25 05:16:06 illustris: how are you starting that shell? Mar 25 05:16:29 non login shells tend to get killed with a SIGINT Mar 25 05:18:25 Isn't that because login shells like bash and zsh are designed to handle SIGINT without quitting? Mar 25 05:18:58 no Mar 25 05:19:26 an nonlogin shell is like a shell script, it expects to quit on SIGINT Mar 25 05:19:48 a login shell is interactive, it expects you to be running things as part of the session Mar 25 05:20:03 it is also related to controlling TTYs Mar 25 05:20:16 So if I make my custom prompt the default login shell for a user, ctrl-c won't kill the prompt? Mar 25 05:21:48 If I run bash from zsh, ctrl-c doesn't kill bash. So it must be handling the SIGINT itself, right? Mar 25 05:58:23 jyothi[m]: I"m not really into speech synthesis, but I notice you left out the link to your "hello" pull request Mar 25 06:02:07 Pac23: sorry I didn't notice your request for comments, I was afk at the time (asleep actually) Mar 25 06:03:38 you can submit a pdf more than once btw, so you can make sure to have an eligible version posted and then improve it as time allows Mar 25 06:10:48 Wormo: Mar 25 06:11:11 Can I carry forward this idea for the next GSoC proposal? Mar 25 06:11:26 By then I would be pretty sure on what to do Mar 25 06:11:40 GSoC2019* Mar 25 06:11:51 yes that could work, more likely to have success if you give yourself a good amount of time to prepare Mar 25 06:12:35 I would be submitting this proposal nonetheless because it's a first time experience for me Mar 25 06:12:55 Any other suggestions about the proposal and what I can change in it? Mar 25 06:13:21 I really like this project so even if it's not for the proposal I would like to know what I can really do with it Mar 25 06:13:29 no, not prompt Mar 25 06:13:35 login shell has nothing to do with prompt Mar 25 06:16:42 yatharthmathur[m: start thinking about some of these tasks in your timeline in more detail, pretend it's your assigment for this week... do you have ideas on how to approach it? What things would you need to start researching to make sure you are picking a reasonable implementation? Mar 25 06:19:49 Concrete ideas on approach are lacking, so it's hard to comment on more until you start coming up with that more detailed approach Mar 25 06:21:30 ds2: Wormo: I think I'm clear on signal handling now. Simply adding a SIGINT handler that does nothing will be enough. Mar 25 06:21:35 Got any resources on how to do tab completion? Mar 25 06:23:15 Wormo: Thanks. Added the link. Mar 25 06:25:05 illustris: try readline Mar 25 06:25:34 Yeah, I was just looking into that. Found a good example on github. Mar 25 06:41:23 Wormo: Added references. Anything else? Mar 25 06:41:52 Wormo: Yeah I have been reading up on all of that since the idea came to me. Mar 25 06:42:33 I have come up with a couple ideas on how to approach the implementation Mar 25 06:43:16 yatharthmathur[m: more of that research will need to find its way into the proposal, and then there could be some iterations with potential mentors commenting on it Mar 25 06:44:12 illustris: I should get review of Pac23 updates first, so maybe tomorrow Mar 25 06:44:28 πŸ‘ Mar 25 06:44:48 @wormo btw i contacted the guy who worked on it last year and figured out what needs to be changed,transitioning the project to me shouldn't be much of a problem i guess Mar 25 06:44:52 Wormo: Mar 25 06:45:11 I have kind of figured out the image tracking part Mar 25 06:45:15 So I'll work that into my proposal Mar 25 06:45:29 Motion detection Mar 25 06:45:32 Pac23: do you have more notes on what you discussed with him? Mar 25 06:45:44 yes i do Mar 25 06:45:48 yatharthmathur[m: yes that's important to cover Mar 25 06:46:32 i had mentioned those things in the updated proposal Mar 25 06:47:15 the tx to rx conversion and back is the only thing that needs some work Mar 25 06:48:25 Pac23: sorry I meant "more" meaning aside from the new comments under Requirements Analysis Mar 25 06:49:56 yeah i do but they aren't much formal,its was more about prudaq accesing,cutting down on total result time and reducing buffer Mar 25 06:50:18 aside from that nothing else, Mar 25 06:50:32 we discussed what needs to be fixed and that isn't much as of now Mar 25 06:50:46 except for the tx to rx and back thing that is something new Mar 25 06:51:54 implementing that would mean designing a hardware circuit from scratch Mar 25 06:52:26 TI has this chip that can be integrated into it Mar 25 06:53:18 Good call not to redesign the hardware circuit, but where do you discuss "tx to rx conversion and back" in the proposal Mar 25 06:53:49 Wormo: I'll be updating my proposal as I keep researching the methods of approach Mar 25 06:54:09 Thanks a lot for your help Mar 25 06:54:14 you're welcome Mar 25 06:55:16 i did't discuss it,i have mentioned it in a single line that's it considering the designing part and the time it requires,it will simply through the original goal off track Mar 25 06:55:52 gimme 5 mins i will make a change add it in the future development goals section Mar 25 06:55:57 Sure Mar 25 06:56:05 thanks :) Mar 25 06:56:47 Oh yeah, do you have any links to your automatic water dispenser project? Mar 25 06:57:22 (Pac23) Mar 25 07:00:06 @wormo no,truth to be told i only worked on the hardware part of it,code was ripped of the internet Mar 25 07:00:35 this is was done in freshman year Mar 25 07:01:01 The cryptotracker code is on my github though Mar 25 07:04:07 Hello Mar 25 07:04:22 I am pursuing engineering. Currently in first year and learning C language which is the only one I know so I wanted to know whether I can participate in your organisation in gsco? [12:30] there's a separate channel for gsoc Mar 25 07:06:09 This is the one dude Mar 25 07:25:45 Pac23 ok np Mar 25 07:34:28 @wormo also my devops compi prize,i will upload a image asap Mar 25 07:34:39 i have made the changes sorry for the delay Mar 25 07:34:59 Ok I'm still here for a bit, I'll take a look Mar 25 07:36:15 Yeah that makes sense now Mar 25 07:36:22 under future goals Mar 25 07:41:18 also added a cloud service section Mar 25 07:42:26 using azure would be fine right ? or there would be some issues given microsofts history and lawsuits Mar 25 07:46:13 Wormo: Mar 25 07:46:13 https://elinux.org/BeagleBoard/GSoC/%22beagle%22bone Mar 25 07:46:27 Can you take a look at it again and tell me if this is any better? Mar 25 07:46:47 this is only being suggested because if a server is setup on the beagleboard it would require a dynamic dns service to resolve to the ip Mar 25 07:46:48 I have included a separate section for approach and I have tried to fit things into a timeline Mar 25 07:46:55 Does it feel any more realistic? Mar 25 07:47:01 since cellular ips change and a static one would be required Mar 25 07:49:49 Pac23: take a look at https://www.noip.com Mar 25 07:50:10 It's free, and they have a linux client for updating when IP changes Mar 25 07:51:19 thankyou @illustris i will look into this Mar 25 07:52:11 24.95 for the whole year is nice Mar 25 07:53:32 That's for the paid domains. If you're fine with .no-ip.org, you can get it for free Mar 25 07:55:30 With that, cellular connectivity won't be a problem for your project Mar 25 07:56:03 saw it,yup Mar 25 07:57:26 so basically its a again a dyanmic ip routing service Mar 25 07:59:39 doing it on the cloud offer bit more control imo rather than hosting a server on a remote node itself deployed somewhere, Mar 25 08:00:33 and i have free acess to azure through msdn so for gsoc this should be enough Mar 25 08:15:45 wormo what is you opinion on dumpin data into a file and sending the file over mail instead ? Mar 25 08:21:24 "free azure" probably isn't relevant to most bbb users, so that's not worth spending time on during gsoc Mar 25 08:22:56 regarding email, I'm not following -- why would users prefer email over a simple web ui? Mar 25 08:24:22 the dyndns thing seems relatively convenient Mar 25 08:24:35 I've used such a service before Mar 25 08:25:04 some people just buy domains for their personal websites Mar 25 08:25:17 so normal dns is feasible too Mar 25 08:37:12 (there are some really cheap domains available, and let's encrypt provides free ssl certs) Mar 25 08:41:42 I knew somebody using static dns with home ppoe, his ISP changed ips so rarely he just updated it manually when it broke, since the home stuff was for fun nothing he depended on Mar 25 08:47:43 the graph generation would be at the client end anyways so why not implement a simple python script that send the raw data via email and then use that data to generate graphs or whatever in the ui,from a cost perspective this seemed nice Mar 25 08:47:47 Pac23 it would be nice to see this turn into the kind of thing that a high-school science teacher might pick as a class project, might just have a beagle available on local LAN for classroom access Mar 25 08:48:14 i do agree with you on the dynamic dns service Mar 25 08:48:18 that can be done too Mar 25 08:49:04 yes i have considered that too in my proposal,both remote and over lan Mar 25 08:49:46 basically the station collects only the data,all the processing happens on the client side Mar 25 08:51:50 I'm just saying that simple protocols for data access are fine, getting good documentation and easily reproducible setup seems more important for outside adoption Mar 25 08:52:31 in some projects cloud access is central, but not here Mar 25 08:56:40 kind of sounds like you're thinking a higher level research project, where gaps in data would be more serious, so maybe we just have different concepts of who might be attracted to using this thing Mar 25 08:58:05 yes,thats why the client end packaged gui,simple installation input the data files,generate data however you want and the email setup would be to simply assigning the mail adress to and from using argparse during the initial station setup Mar 25 08:59:40 what i will do is i will add both methods(email,dns service) and later if i my proposal gets accepted this can be brainstormed during the community interaction week. Mar 25 09:01:44 ok Mar 25 09:11:06 Wormo: can you check my proposal once more, there are a few more edits I have done... Mar 25 09:11:13 https://elinux.org/BeagleBoard/GSoC/%22beagle%22bone Mar 25 09:22:23 Hey my elinux account is not approved yet it seems how long does it take? Mar 25 09:22:25 yatharthmathur[m: I'm running out of time for the night, but did you notice the pull request at https://elinux.org/BeagleBoard/GSoC/Ideas#General_requirements that also supposed to be part of applying Mar 25 09:22:55 Sidharth: when did you apply? Mar 25 09:23:07 12 hours ago Mar 25 09:24:22 I don't know, but you can mention it again if it's still not active 24 hrs from when you requested Mar 25 09:26:21 @wormo i did'nt notice your last reply before my reply, ah yes i was thinking about a research level thing,i get what you are saying i will make the changes accordingly Mar 25 09:31:37 it's good to have a realistic target audience and keep them solidly in mind Mar 25 09:31:51 afk for the night... Mar 25 09:34:11 bye,gn Mar 25 09:34:43 Thanks for the help wormo gn Mar 25 09:39:56 Are we supposed to print the elinux page as PDF and upload to summerofcode? Mar 25 09:41:59 yes the final draft should be a pdf Mar 25 09:42:05 @illustris Mar 25 09:42:39 Yeah, but are we supposed to print the elinux page as a PDF? Mine came out looking terrible Mar 25 09:45:16 by print do you mean like a physical print ? and then a scanned pdf or a direct print to pdf ? Mar 25 09:59:35 obviously not πŸ˜… Mar 25 10:12:34 :P Mar 25 13:29:03 @jkrinder i had a suggestion regarding the wiki ,rpi has a list of supported hardware peripehrals beagle bone should have one too Mar 25 13:30:02 this could be documentation project someone can pick maybe during nextyears gsoc or something Mar 25 13:30:40 because what i have realised it finding supported hardware means digging deep into the specific hardwares docs etc hit or a miss thing Mar 25 13:30:51 a centralized place with links would be nice Mar 25 14:13:10 Why is it that hard ? If a hardware module has the interface that BBB supports, then it the hardware module is supported. Mar 25 14:22:12 yes for us it isn't but for newbies a place that lists it would be nice Mar 25 14:23:55 usb peripherals Mar 25 14:24:24 good evening zeekhuge Mar 25 14:25:07 I tried to fix my proposal Mar 25 14:25:17 can you please check if it's any good Mar 25 14:25:24 https://elinux.org/BeagleBoard/GSoC/%22beagle%22bone Mar 25 14:34:25 zeekhuge: Mar 25 14:34:26 Can you suggest me some improvements Mar 25 14:34:27 https://www.elinux.org/BeagleBoard/GSoC/ModernSpeak-Spell Mar 25 15:29:32 jm_donthineni: "Implementing the voice recognition system" what does this mean ? Mar 25 15:30:17 A complete week in just one sentence ? That's not a "good" plan. Mar 25 15:34:01 Also, a small suggestion, divide the timeline in terms of "week 1,2,3...". That makes it a bit easy to understand. Mar 25 15:34:57 Okay. I will make the section detailed. Mar 25 15:39:06 jm_donthineni: have you tried installing the libraries on emulator or something ? Mar 25 15:40:50 yatharth.mathur: the project looks tooooo ambitious. 12 weeks is not a lot of time. Mar 25 15:40:51 Your proposal talks about andy-app, food dispenser, cloud integration and so many things . its just a lot. Mar 25 15:41:40 No. I haven't tried. Mar 25 15:53:41 jm_donthineni: try it. Also, you need to decide on exactly which library you are going to use. Mar 25 15:53:42 I am stating this again and again. The proposal needs to convince the mentors that you can pull of the project, we do not know any of the students personally. And 1-2 lines like "implementing this and that" will not be able to build much confidence. So include answers of "how will you implement 'this'", " what will you use to implement "that" etc. Mar 25 15:54:13 And that too, for each task and maybe even for each subtask. Mar 25 15:54:28 As much as you know about the project. Mar 25 15:54:37 Make it 10 pages long if you want. But include the details. Mar 25 15:55:33 Although, preferably not more than 5 pages :P Mar 25 15:57:20 Okay. Will change it accordingly. Thanks Mar 25 16:03:34 zeekhuge: I need some help in choosing the speech recognition library.I just found "please bear in mind that TIesr is "demo" quality and was not intended to be used in a production environment. It was purely to demonstrate the voice triggering capabilities of the C5000 DSPs." Mar 25 16:04:06 The above statement in tiesr forums Mar 25 16:04:42 Do u think pocketsphinx is more appropriate for the project Mar 25 16:10:43 zeekhuge: Mar 25 16:11:05 Any mentor online zeekhuge nerdboy Mar 25 16:14:54 Please review my proposal and recommend me mentor for it please Mar 25 16:14:59 https://elinux.org/GSoC_Proposal_Stereo_Vision Mar 25 16:18:19 Please anyone recommend me changes Mar 25 16:20:12 I was not able to contribute because my exams are going on but i will show my potential and work from Tuesday please furnish me with recommendations in my proposal Mar 25 16:21:09 is there a mentor interested in stereo vision? Mar 25 16:27:54 I just need mentorship in packaging Mar 25 16:29:00 Rishu: well, you need a mentor to get the project accepted .. So ... Mar 25 16:29:11 Yes sir Mar 25 16:29:44 zeekhuge: will you be my mentor sir Mar 25 16:29:50 And I actually kinda like the idea, but I am not sure if I'll be able to help on with the project. Mar 25 16:30:37 Sir i will do the project with minimum help i have done some work on this topic Mar 25 16:32:21 I am really interested to do work on electronics and information technology combined topic and this could be one of great place for me Mar 25 16:32:44 Just i need is getting my proposal accepted Mar 25 16:36:01 zeekhuge: can you help me with it sir Mar 25 16:40:10 I would actually prefer to be a co-mentor of the project. Mar 25 16:41:24 zeekhuge: sir can you suggest me any mentor Mar 25 16:43:08 I need to mention him/her before final submission Mar 25 16:43:15 Prepare a kickass proposal, and I think you will find mentors. Mar 25 16:43:54 Suggest me changes or any changes for my proposal Mar 25 16:44:10 Improvement Mar 25 16:44:12 Rishu: you mean you need to mention it in the proposal ? Mar 25 16:45:05 Its the basic proposal layout to mention mentors in proposal Mar 25 16:47:38 zeekhuge: can I mention you as my co mentor sir Mar 25 16:48:02 Naah ! Thats "Possible mentors". Mar 25 16:48:13 Sure. Mar 25 16:48:57 And its not "required". But ofc its better to mention. Mar 25 16:51:43 zeekhuge: thanks sir Mar 25 18:20:12 jkridner: Here is a draft I have done: https://elinux.org/GSoC_2018_Proposal:_Update_PRU_Python_loader--Bill_Morrisson. Please your suggestions will be very appreciated. Mar 25 18:20:46 I also plead with other mentors who can look at it and give some suggestions. Mar 25 19:01:07 does the final proposal have a specific format like the elinux draft had ? Mar 25 20:38:10 spiritus: Did we have conversation about this project ? or was it with muneeb17 ? Mar 25 20:40:12 Rishu: It would be better if you give a more detailed description Mar 25 20:40:21 maybe in "Description" section itself ? Mar 25 20:40:41 tell us about the drivers you are taking about. Mar 25 20:41:02 and what do you mean by "base model" Mar 25 20:41:57 also, how would you caliberate the setup etc etc. Mar 25 20:43:37 zeekhuge: Should I include detail in my proposal of which I'm not sure how would it play out Mar 25 20:45:49 Sidharth: Definitely, and then ask mentors to go through it. Mentors will correct you then. Mar 25 20:46:31 Oh so nice!! I was trying to tie those loose ends, but kept getting more uncertainity!! Mar 25 20:46:31 Also, its little about "if you know exactly what to do", its more about "if you can find out what to do or not". Mar 25 21:01:02 Although its always better to more of ""you know exactly what to do"". That natural. Mar 25 21:01:07 *thats Mar 25 21:05:29 zeekhuge: I had it with jkridner Mar 25 21:06:41 Cool ! but you realize that the project is actually very simple for 12 months long time period. right ? Mar 25 21:08:40 spiritus: Infact, let me ask you this : Mar 25 21:08:57 how will you implement rproc_boot(...) / rproc_shutdown(...) ? Mar 25 21:17:46 Well, you do not. Mar 25 21:17:47 Those are kernel level APIs Mar 25 21:18:05 and you just place the compiled PRU firmware at /lib/firmware/ and modprobe the echo "bind" in one of the sysfs Mar 25 21:18:17 and all these APIs are internally invoked to boot and load the PRUs Mar 25 21:30:09 zeekhuge... seems it is β€˜start’ now, no need to bind/unbind. Mar 25 21:31:17 zeekhuge... thanks for being active today. Mar 25 21:34:21 Wormo: thanks for being around today too. Mar 25 21:36:06 jkridner: zeekhuge suggested we add DMA to PyPRUSS too. What do you think? Mar 25 21:36:23 Sounds like a good idea. Mar 25 21:36:58 ravikp7: I had something I wanted to add to beagle-boot.... a TCP/IP proxy.... reviewing the proposal again now. Mar 25 21:37:21 all: make sure you have entries in both summerofcode.withgoogle.com and elinux.org. Mar 25 21:39:23 zeekhuge: jkridner: I went through the pru dma from last year... thinking of how to integrate it to make an API as components there on kernel as well as firmware side Mar 25 21:40:17 although each example provides a sysfs interface.. I want to make it more generalised Mar 25 21:42:34 do we have a good set of examples in the schedule now? I like the idea of having existing examples enumerated to move to new images and python loader. Mar 25 21:43:58 yeah, the existing pypruss examples and also the ones like in BeagleScope/pru-support-package Mar 25 21:44:43 Though, I didn't make the changes, I'll do that with pru dma Mar 25 21:49:03 jkridner: your suggestions on how to include dma? Mar 25 21:50:04 ravikp7: You should look at https://beagleboard.org/updater for some info on parsing https://beagleboard.org/latest-images Mar 25 21:52:24 muneeb17: I'd think the Python library should provide some way to allocate memory and either transfer to or from the PRU, along with synchronization events at the completion of the DMA. There should be a PRU-side library that includes initiating the DMA transfers. Mar 25 21:52:39 and testing for completion events. Mar 25 21:52:52 that should have been part of the previous project for the most part. Mar 25 21:54:53 I'd be a big fan of getting a printf implementation that used RPMsg and to replace the version in the standard library so that printf "just works". I'd also look to limit the arguments to 4 (always 32-bit) and have the host-side code always do the formatting. Parsing the strings in the PRU ELF should enable strings to be passed by reference only. Mar 25 21:57:09 * jkridner[m] wonders if there are details to http://processors.wiki.ti.com/index.php/Printf_support_in_compiler Mar 25 21:57:15 and http://processors.wiki.ti.com/index.php/Tips_for_using_printf#Reducing_printf_code_size_with_--printf_support Mar 25 21:57:58 I think it is the LOG_printf one that is cool: http://processors.wiki.ti.com/index.php/DSP_BIOS_FAQ#Q:_How_can_I_convert_printf.28.29_statements_to_LOG_printf.28.29.3F Mar 25 21:58:23 probably getting carried away now. ;-) Mar 25 21:58:52 but, I think it is helpful to think about how people will debug their PRU code from PyPRUSS. Mar 25 22:00:51 jkridner: Yeah. I'll look into the info you shared. Also, about the existing DDR/PRU-mem functions, should we use /dev/mem + mmap()? Mar 25 22:05:04 zeekhuge: I thought of them as being calls to be made in pypruss to execute them. Mar 25 22:05:20 I did not have a clear vision. Mar 25 22:50:05 muneeb17: I'd think so. **** ENDING LOGGING AT Mon Mar 26 03:00:02 2018