**** BEGIN LOGGING AT Wed Oct 10 03:00:00 2018 Oct 10 08:44:57 hello there Oct 10 08:45:09 I need to ask regarding GSOC Oct 10 08:46:45 a_tolphas: thank you for your interest, but those knowledgeable on gsoc usually hangout at #beagle-gsoc Oct 10 08:47:26 a_tolphas: and be prepared to wait quite some time for an answer, might as well try the mailing list. Oct 10 08:48:11 Thanks Oct 10 11:30:45 Hey, I have a question; how do i disable i2c!? Oct 10 11:30:53 I am using uboot overlays.. Oct 10 11:31:45 i need to free P9_20 which is used for I2C2_SDA Oct 10 11:44:34 so do i need to change cape-universal!? Oct 10 11:59:45 hexastorm: you're making your own overlay for a cape or something? Oct 10 12:02:47 hexastorm: if yes, then your overlay should overlay the cape-universals' node for this pin so that you can use it. Oct 10 12:03:48 thinkfat: I am making my own overlay.. Oct 10 12:04:31 hexastorm: then I'd say your overlay needs to disable the existing cape-universal node for this pin Oct 10 12:04:40 thinkfat: I had an error before.. but will try to simply overlay it.. Oct 10 12:05:23 maybe adding status="disabled" would be enough. I've had to do this recently but I don't recall Oct 10 12:13:45 pin 94 (PIN94) 44e10978 00000007 pinctrl-single Oct 10 12:13:49 seems to be fixed.. Oct 10 12:15:37 strange that i still see two ic2 devices via ls -l /dev/i2c* Oct 10 12:15:49 guess one of them ain't working :-) Oct 10 13:53:57 Hello, I am Uzair Mushtaq, a developer, Electronics Engineer and currently Vice President EE-Wing at IEEE Student Branch of COMSATS University Islamabad, Pakistan. We are going to organize training sessions on SBCs for students by the start of next month (November). It will be our pleasure to receive "Beagle Bone Black", so that we will be able to promote, use and test your product. Oct 10 13:54:39 Let me know if you can send us a free sample Oct 10 14:04:15 Uzair: feel free to contact the beagleboard foundation (as listed on beagleboard.org) and ask for a sample. but given the generally low price tag, better just buy one. Oct 10 14:08:01 any link? Oct 10 14:08:49 Uzair: beagleboard.org Oct 10 14:20:56 with the python gpio library I cannot set my pin to output mode Oct 10 14:21:26 there are a few questions which come to mind.. why do I need to configure my pinmode if I have already done this in my overlay Oct 10 14:22:03 and then ofcourse why doesn't it work.. I am using a very recent image Oct 10 14:24:20 i'm glad there's so much interest in touch screens on the bbb group Oct 10 14:24:47 i think i'm gonna finally fix uxmux tonight, i was supposed to do it this weekend then BOOM sick Oct 10 14:25:47 i've been toying with the idea of building libuv into uxmux but i'm not sure how sold I am on libuv's cooperative tasking structure Oct 10 14:26:13 everyone uses it but theoretically you could accidently write a blocking function that breaks it, right? Oct 10 14:29:15 hello to everyone. I've some problems with my BeagleBoard xm Rev. C and hope to find some help here. If I plug a 5V/3A DC-Supply to the board, nothing happens. The power-LED D5 is glowing, the other LEDs are off. The ethernet-LEDs are off. I connected a USB-Serial-Adapter with 115200 8n1 to the BB and while PowerUp I receive a single 0x00. Pressing 'reset' causes no serialTx. Powering-Up with no µSD produces some serial garbage. Oct 10 14:29:20 I tried all Baudrates, no readable Data. After that I tried flashing different debian or ubuntu-images to the shipped µSD: 1. Formatting FAT32, 2. Flashing with 'Win32DiskImager' or 'Etcher' - same results. Any suggestions, what i can try? Oct 10 14:30:06 ayjay_t: eww, libuv? why would you want to use that? Oct 10 14:32:11 zmatt: because it's clearly hitting the sweet spot for a lot of people Oct 10 14:32:28 ayjay_t: libuv is the sort of dependency I'd want to remove if I found it in a project I want to use Oct 10 14:32:34 do you think node.js is just popular because it's javascript or do you think it's because of the event-based cooperative tasking? Oct 10 14:32:49 i'm also thinking about just building in aio or looking at something like goroutines for C Oct 10 14:33:18 i want people to leverage the UXMux like a terminal- so that they use it to control linux processes, and i'll build a shim over processing stdin and stdout as well as some domain sockets Oct 10 14:33:27 but people will likely stil want to build async-io directly into their applications Oct 10 14:34:29 I generally use sd-event (libsystemd) whenever I need an event loop Oct 10 14:34:55 (or just bare poll()/epoll for really simplistic applications) Oct 10 14:37:08 honestly the really sexy thing about libuv's cooperative tasking structure is that it's _possible_ to go to sleep Oct 10 14:37:22 and i know thats a bit out of reach on a linux device at this point, nor do I know what sleep modes the sitara has Oct 10 14:37:34 I have no idea what you mean by that Oct 10 14:38:21 or even by "libuv's cooperative tasking structure" ... did they add cooperative tasking at some point? Oct 10 14:38:46 I don't see it mentioned in the feature list Oct 10 14:38:47 libuv basic 1) executes timers 2) executes callbacks 3) polls IO 4) repeats Oct 10 14:38:55 yes that's called an event loop Oct 10 14:38:56 so i think of those callbacks as cooperative tasks Oct 10 14:39:04 because its analogous to a protothead loop or something like that Oct 10 14:39:05 that's not what cooperative tasks are Oct 10 14:39:27 it's just an event loop library, like sd-event or many others Oct 10 14:39:58 except it has an awkward API because it tries to make linux behave like windows Oct 10 14:40:23 it's obvious to me that the even loop library is effectively a cooperative tasking structure Oct 10 14:40:26 event* Oct 10 14:40:36 i understand that it can be used otherwise Oct 10 14:41:40 .... except it isn't. cooperative tasking and an event loop are orthogonal things Oct 10 14:41:55 although they have synergy Oct 10 14:42:34 but an event loop doesn't provide cooperative tasking, and cooperative tasking doesn't require an event loop Oct 10 14:42:34 again, in no uncertain terms, the model i'm using to develop a cooperative tasking structure is built off an event loop Oct 10 14:43:02 it is such a tight fit, conceptually, there is absolutely no reason to consider it differently in this context Oct 10 14:43:15 the event driven tasking enables low power modes much better than other structures Oct 10 14:43:55 I'm honestly not sure if you're saying something that's trivial, nonsense, or a mix of the two Oct 10 14:44:24 you might want to consider a third option Oct 10 14:46:57 anyway, building out this system is going to be fun, if not a bit much. i'd really like to get to a point where i can use sitara's low power states, but i'm not really sure how well the kernel supports those Oct 10 14:47:28 event driven code tends to be more efficient than blocking code due to avoiding unnecessary context-switching, but has no effect on system power management. building cooperative tasking on top of an event loop may be nice for coding convenience, but still increase overhead again due to context switching (although typically more lightweight than context-switching of preemptive threads) and the memory ... Oct 10 14:47:34 ...(mostly stack) required for a task context Oct 10 14:48:41 i agree with all that, but the other benefit of using a purely event driven loop instead of a spinning/polling loop, ie making sure that all threads will yield/block/sleep if they can, is that you can go to sleep Oct 10 14:49:10 wait, you're comparing it to spinning/polling?!? who on earth does that? Oct 10 14:49:32 the point is to provide a framework that encourages people _not_ to do that Oct 10 14:49:44 there's all kinds of skill ranges Oct 10 14:49:52 and i want lots of people to get benefits out of this Oct 10 14:49:55 like they do node.js Oct 10 14:50:36 javascript programmers tend to rely too heavily on new functions/keywords to create elegant code instead of well thought out program structure Oct 10 14:50:57 which i imagine is because often they don't have a CS/CSE background Oct 10 14:51:04 but that's not a feature unique to event loops... the same goes for blocking code (+ threading) Oct 10 14:51:10 which? Oct 10 14:51:22 yeah i mean, it's true, which is also why i'm looking at go's concurrency model Oct 10 14:51:31 allowing the system to go idle Oct 10 14:51:56 basically you have to do really weird things to end up spinning Oct 10 14:52:02 threads though have context switches, which act as insurance against bad programming to an extent but also incur a cost in context switches Oct 10 14:52:14 ? Oct 10 14:52:16 yeah but arduino programmers are basically taught to spin Oct 10 14:53:08 and also this is being integrated into the larger linux OS so i'm not sure how that'll throw a wrench in my plans for the ultimate low-power programming framework for embedded linux devices lol Oct 10 14:53:29 ayjay_t: you're late to the party I think. Oct 10 14:53:34 that seems like an unclimable party Oct 10 14:53:49 CRAP you got me Oct 10 14:53:53 inclimable mountain Oct 10 14:53:56 :-) Oct 10 14:54:03 i'm not really sure what you mean but Oct 10 14:54:12 but there are some people intereted in what i'm doing Oct 10 14:54:17 a gpio library could definitely be designed to discourage that Oct 10 14:54:33 I suspect the end result would be that ex-arduino-programmers will simply not use such a library :P Oct 10 14:54:50 yeah well that's the marketing risk of any new sol'n Oct 10 14:54:51 will people like it Oct 10 14:54:56 i mean, node.js took off Oct 10 14:54:58 golang took off Oct 10 14:55:01 ayjay_t: because you describe something that is trivial, and well known, and to which hundreds of solutions already exist Oct 10 14:55:25 thinkfat: that's not the main problem i'm solving Oct 10 14:55:25 ^ Oct 10 14:55:38 thinkfat: i'm building a lightweight gui framework Oct 10 14:55:51 i'm just trying to decide WHICH framework i want to build it on Oct 10 14:56:01 I fiddle a tiny bit with go. so far it irritated me a bit somehow Oct 10 14:56:07 *fiddled Oct 10 14:56:45 yeah it's kind of annoying. it's just basically C with a reverse variable decleration syntax, can't cast int's to pointers, it has some dynamic memory libraries built in (hash and slice), uhh goroutines, and "receivers". Oct 10 14:56:57 and thats basically all the differences Oct 10 14:57:33 being forced to comment out imports if you temporarily commented out the last use of that import Oct 10 14:57:39 >:( Oct 10 14:57:50 hmm i haven't run into that issue Oct 10 14:57:54 thinkfat do you have anything else to add? Oct 10 14:58:14 like perhaps what your favorite solution is? Oct 10 14:58:29 ayjay_t: if you really want to go for low-power, make it use wakelocks to work well with autosleep Oct 10 14:58:34 ayjay_t: define lightweight and why you think your solution is the "ultimate" one Oct 10 14:58:50 thinkfat: i'm just going to build it and launch it and let people try it Oct 10 14:58:53 that would probably be relatively novel Oct 10 14:59:18 zmatt: yes, but any existing event-loop based model could be easily extended to serve that purpose Oct 10 14:59:52 so i've been looking at libuv, also aio, and go's concurrency model, as well as all the usually suspects- epoll, libev Oct 10 15:00:03 ayjay_t: ok, then the other way 'round - what problem does it solve that has not yet been solved Oct 10 15:00:04 thinkfat: I doubt you'd add it easily to libuv Oct 10 15:00:28 but maybe i'm wrong Oct 10 15:00:52 thinkfat: why u gonna invest? Oct 10 15:01:22 i mean i don't really care if it does, tbh Oct 10 15:01:26 because i really like it Oct 10 15:01:36 ayjay_t: well, that's of course reason enough ;) Oct 10 15:01:55 the basic idea is that my gui framework can start and stop processes and read stdin/stdout and also pipes Oct 10 15:01:57 ayjay_t: I do a lot of stuff that has no apparent merit, just because I really like it Oct 10 15:02:03 but i want to build aio and tasking into the gui itself Oct 10 15:02:20 because a) i want to let people not have to context switch from process to process for IO Oct 10 15:02:28 and b) i have a render loop to deal with Oct 10 15:02:49 so i'm thinking libuv since that is pretty popular, but i bet a lot of it has to do with using javascript Oct 10 15:03:58 ayjay_t: libuv uses thread pools and stuff, it is not efficient Oct 10 15:04:06 yeah for disk io Oct 10 15:04:13 also for socket io Oct 10 15:04:15 iirc Oct 10 15:04:19 not for network, id on't think Oct 10 15:04:30 but you're right, and basically they said "we're leaving it this way because otherwise cross platform is a pain" Oct 10 15:04:40 but this sol'n obviously doesn't need to be cross platform Oct 10 15:05:01 but libuv has _such_ a large active community, that's a huge benefit to anything built on it Oct 10 15:05:01 hence my suggestion of sd-event Oct 10 15:05:01 Hello Oct 10 15:05:27 arduino also has a large active community, therefore surely doing things the polling style is better ;) Oct 10 15:05:28 sd-event hmm Oct 10 15:05:41 i'm making compromises, zmatt Oct 10 15:05:46 hi Gaurav_, ask your question :-) Oct 10 15:06:10 do you think i'll be able to replace PID 1 w/ my framework tho/ Oct 10 15:06:16 because that would be super cool Oct 10 15:06:16 ayjay_t: hey it's your pet project, your choices Oct 10 15:06:47 ayjay_t: replacing pid 1 is simple. init=/bin/yourthing Oct 10 15:06:59 thats true, but now i don't have a network and go knows what else Oct 10 15:07:00 the question is rather what you miss after that. Oct 10 15:07:03 yeah Oct 10 15:07:27 maybe it would be better just to have systemd somehow give my framework highest priority Oct 10 15:07:38 /shrug Oct 10 15:07:55 (not a systemd expert) Oct 10 15:07:57 (clearly) Oct 10 15:08:15 systemd has rich options for configuring prioritization and resource limits Oct 10 15:08:47 yeah i don't think that systemd is the source of problems in such cases Oct 10 15:09:30 well either way its going to be excited. i'm gonna try to stream dev tonight, if i can stay awake... and maybe have made some decent progress by this weekend Oct 10 15:09:39 exciting* anyway, ill be around, thanks for the input guys Oct 10 15:09:42 ayjay_t: for sd-event see /usr/include/systemd/sd-event.h or "man sd-event" ( https://www.freedesktop.org/software/systemd/man/sd-event.html ) Oct 10 15:10:13 although i need to mention after that conversation that sometimes the tone in these conversations can be pretty intimidating Oct 10 15:10:38 (thanks zmatt) Oct 10 15:12:19 I didn't mean to appear intimidating. I apologize Oct 10 15:12:25 actually no, thinkfat, i really want to know what framework you'd suggest Oct 10 15:13:24 oh boy, the last time I used a "framework" I hacked something up just using select() coupled with a timer list Oct 10 15:14:05 and apart from that I've been mainly using Qt for GUI Oct 10 15:14:16 but that doesn't really count as "lightweight" any more Oct 10 15:15:08 I use either sd-event or some really minimal epoll-wrapper I wrote that we still use in a few services, but I kinda want to phase out the latter in favor of sd-event for uniformity's sake Oct 10 15:15:47 the stuff I hacked up mainly used select and pipes for event notifications Oct 10 15:16:01 so that threads could report to the mainloop Oct 10 15:16:15 you might actually like my solution then Oct 10 15:16:25 ill take a look at sd-event Oct 10 15:16:42 i'm realizing that one of the reasons people abuse libuv and it's callback based loop is because they can nest all these lambda functions in javascript Oct 10 15:17:09 thats not really the way to do things in C so i'm wondering how that'll effect the solution Oct 10 15:17:51 event-based code in C does require more manual labour obviously Oct 10 15:18:23 i mean, considering that nested labmda call backs are basically the cancer of javascript, i don't think the restriction is a bad thing Oct 10 15:18:32 having said that, sweeping explicit state under the carpet by e.g. using javascript's async/await may seem nice at first, but it has definitely bitten me in the ass too (e.g. if you want to be able to abort operations in progress) Oct 10 15:18:42 but that feature in javascript is basically waht allows beginners to free-write their programs Oct 10 15:19:24 well the deeply nested callback mess has hopefully mostly died out in favor of async/await Oct 10 15:19:51 ultimately my solution in javascript was to build a state manager and a function that uses next-tick to communicate callback results back to the calling module Oct 10 15:19:55 instead of promises or w/e else Oct 10 15:20:14 o.O Oct 10 15:20:35 its not bad Oct 10 15:21:05 or you building node.js on elinux or in a server system? Oct 10 15:21:09 like a cloud system i mean Oct 10 15:21:32 we use nodejs both on the beaglebone and on servers Oct 10 15:22:05 not a fan, but I'll write js code if I'm forced at gunpoint ;) Oct 10 15:22:49 makes sense, alright! Oct 10 15:22:54 e.g. I wrote a dbus library because I was unhappy with the existing options (I tried to clean up an existing one but I gave up) Oct 10 15:23:19 yeah dbus Oct 10 15:23:24 one of those systemd things. Oct 10 15:23:26 logs and comms and stuff Oct 10 15:23:31 pub sub i think? Oct 10 15:23:35 dbus predates systemd Oct 10 15:23:44 o rly Oct 10 15:24:36 it's structured IPC, i.e. remote procedure calls and messages (including broadcasts) Oct 10 15:25:16 ohhh it includes rpc eh Oct 10 15:25:23 something else i should look into Oct 10 15:35:01 it's also slow as molasses Oct 10 15:35:03 ;) Oct 10 15:40:08 btw do you pronounce it "cyoot" or "cutie" Oct 10 15:47:00 cutie Oct 10 15:51:51 upgraded to new image fixed issues Oct 10 15:52:04 still have one question... if i run a sudo apt update & upgrade on the latest images Oct 10 15:52:07 i get a failure Oct 10 15:52:18 something about not being authenticated Oct 10 15:52:36 like the repo are not properly signed Oct 10 15:53:47 ayjay_t: for example, this is what poettering's silly simple calculator dbus service (augmented with a property) looks like using my own dbus library in nodejs: https://pastebin.com/SE6ceRNM Oct 10 15:54:26 ayjay_t: and a client using my work-in-progress C++ wrapper of sd-bus: https://pastebin.com/FS8uMAMK Oct 10 16:06:04 ayjay_t: and playing a bit with it from the commandline: https://pastebin.com/K5seF5sH Oct 10 16:08:29 it's nice for gluing different services written in different languages running as different users when performance isn't critical Oct 10 16:08:57 different users eh? Oct 10 16:09:11 yeah i mean, it remains to be seen how much the performance cost will really impact us Oct 10 16:09:25 it's not like we're fielding 10000 req/sec Oct 10 16:09:43 exactly, it's fine for lots of purposes Oct 10 16:09:44 i'm mostly concerned with grabbing data from an ADC and writing it to the touch screen Oct 10 16:10:15 thats why the basic architecture for uxmux is gong to be like, "you just build the gui here, and then write seperate scripts/programs for your other functional/business logic" Oct 10 16:10:45 but i still think in program aio is important to consider long-run Oct 10 16:10:51 a framework to help the n00bs Oct 10 16:10:55 and yeah, dbus has config files for defining a security policy about who can register which names on the system bus and who can make which calls Oct 10 16:10:57 sd-event or libuv, maybe goroutines who knows Oct 10 16:11:01 i wonder what rust does Oct 10 16:12:29 but anyway MORE OPEN TABS TO READ Oct 10 16:12:40 my educational todo list gets longer and longer the more time I spend on IRC Oct 10 16:12:48 i guess that's better than spending time on twitter where i just get angry Oct 10 16:12:58 lol Oct 10 16:15:22 hahaha.... https://www.youtube.com/watch?v=xCoH2AORcEQ Oct 10 17:01:28 hi Oct 10 17:01:57 i am a newbie Oct 10 17:02:07 can someone help me Oct 10 17:02:32 only if you explain your problem or ask your question Oct 10 17:02:59 ohk Oct 10 17:03:18 :) Oct 10 17:03:18 i wanted to apply in gsoc 2019 in beagle Oct 10 17:03:45 but i have no prior experience with beaglebone black Oct 10 17:04:05 the only hardware in gsoc is beagle Oct 10 17:04:43 i have basic knowledge of arduino Oct 10 17:04:44 I'll just quote LetoThe2nd here: Oct 10 17:04:47 "thank you for your interest, but those knowledgeable on gsoc usually hangout at #beagle-gsoc" Oct 10 17:04:53 "and be prepared to wait quite some time for an answer, might as well try the mailing list." Oct 10 17:05:23 ohk Oct 10 17:05:26 I don't know anything about gsoc myself, so Oct 10 17:05:38 n where exactly is beagle-gsoc link Oct 10 17:05:50 you can just do /join #beagle-gsoc Oct 10 17:06:40 or use https://webchat.freenode.net/?channels=beagle-gsoc Oct 10 17:07:08 see also https://beagleboard.org/gsoc for more resources Oct 10 17:07:31 (I found that page as the top google hit, so you could have found it too) Oct 10 17:11:32 i actually did Oct 10 17:11:40 besides gsoc Oct 10 17:11:56 can u tell how to get started with beaglebone Oct 10 17:12:30 http://lmgtfy.com/?q=beaglebone+getting+started Oct 10 17:14:09 i was expecting it to be more specify Oct 10 17:14:27 your question isn't specific either Oct 10 17:14:30 like resources like ebooks for videos Oct 10 17:14:36 or* Oct 10 17:15:50 no idea, sorry Oct 10 17:16:28 ohk how u started Oct 10 17:16:38 learning beagle Oct 10 17:17:44 I'm not sure how to describe that. I'm rather good at picking things up, and I already had experience with linux, embedded systems in general, and another TI SoC Oct 10 17:18:21 "learning beagle" is rather vague anyway. for the most part it's just an embedded linux system like many others Oct 10 17:18:47 there are hardware-specific details of course, but for that you really need to ask a more concrete and specific question Oct 10 17:19:27 ohk that clears things Oct 10 17:20:02 so can u tell me what is needed in terms of software for beagle Oct 10 17:20:18 linux is core i guess Oct 10 17:20:54 but what specifically in linux i need to learn in order to get going on beaglebone Oct 10 17:20:54 \ Oct 10 17:21:09 I don't know how to answer that, sorry Oct 10 17:21:16 i know some commands on terminal Oct 10 17:21:31 i know bit bash scripting Oct 10 17:21:47 is that helpful Oct 10 17:21:48 ? Oct 10 17:21:57 so is there something else Oct 10 17:22:11 curiosity and a search engine Oct 10 17:22:23 and a willingness to dive in and try things Oct 10 17:22:26 as all youtube videos are on these things only Oct 10 17:22:43 yep, what vagrantc just said Oct 10 17:22:59 i tired that Oct 10 17:23:05 that what ended me here Oct 10 17:23:14 so irony Oct 10 17:24:16 as i was going through project the prerequisite was Oct 10 17:24:23 linux kernel programming Oct 10 17:24:33 ?? Oct 10 17:24:40 what project? Oct 10 17:24:45 i when i searched google for corresponding courses Oct 10 17:24:49 i got nothing Oct 10 17:25:04 gsoc beagle projects Oct 10 17:25:18 not important to the matter Oct 10 17:25:26 quite important to the matter Oct 10 17:25:36 but what is linux kernel programming/developmnt Oct 10 17:25:52 ohk if u say Oct 10 17:26:10 linux kernel programming is a rather advanced topic, and if you're not already comfortable with it you should maybe not pick a project that has it as prerequisite Oct 10 17:26:11 matt_murdock: that is a very broad topic, but sounds like it might not be the best fit for you at the moment Oct 10 17:26:33 ohk Oct 10 17:26:56 so if i aim to do this in next 10months where do i start Oct 10 17:27:05 or someone part of it Oct 10 17:27:21 you might need a few years if you're just starting with bash scripting Oct 10 17:27:38 you're really asking questions that are better asked in #beagle-gsoc or on the gsoc mailing list Oct 10 17:27:46 they exist for a reason Oct 10 17:27:49 ohk Oct 10 17:27:53 i get it Oct 10 17:28:14 but how i wanted to do these things Oct 10 17:28:33 n no one is replying ther Oct 10 17:29:02 you might need to ask more refined questions about specific issues Oct 10 17:29:38 how can i if i dont know anything about it Oct 10 17:29:47 matt_murdock: not getting a reply in the appropriate forum to ask means you either need more patience or need to ask better questions, it doesn't mean you should ask the same question in a wrong place Oct 10 17:30:16 i am waiting Oct 10 17:30:30 it open in other tab Oct 10 17:30:32 you're asking for an amount of handholding that very few people will be willing to provide Oct 10 17:31:19 u guys are pro i get it Oct 10 17:31:40 but I don't know how gsoc generally works, which is exactly why you should ask for guidance in the gsoc channel and not here Oct 10 17:31:55 you can ask technical questions here Oct 10 17:32:05 ohk Oct 10 17:32:14 but not something as vague as "how do I get started" Oct 10 17:32:16 my 1st time Oct 10 17:32:18 my bad Oct 10 17:32:22 it's ok Oct 10 17:32:26 matt_murdock: basically, you've asked something equivalent to "how do i do math?" Oct 10 17:32:47 try asking that in #math ... I'd be interested to know the answers you get ;) Oct 10 17:32:51 yeah i get your point Oct 10 17:32:54 true that :) Oct 10 17:33:16 i will also be displeased with such question Oct 10 17:33:50 i there any other link for gsoc where i can try Oct 10 17:34:01 rather than siting ducks on this one Oct 10 17:34:22 https://beagleboard.org/gsoc is really the best place to get started Oct 10 17:34:28 the links it provides Oct 10 17:35:20 i try finding the mentors names through this but i couldn't find a way of contacting them Oct 10 17:35:33 i tried linkedin Oct 10 17:35:38 mails Oct 10 17:35:43 but i got nothing Oct 10 17:35:59 so someone suggested to go to forums Oct 10 17:36:16 i can these ppl names on side Oct 10 17:36:28 but how to actually contact them Oct 10 17:36:28 just ask in #beagle-gsoc and leave your client connected to get answers... people may be in a different timezone or only sporadically look at irc Oct 10 17:36:48 ohk i will try thata Oct 10 17:37:17 n really sorry for killing u ppl with such question Oct 10 17:37:22 also, isn't it a bit odd to ask how to get started with gsoc in october? I mean, gsoc 2018 is over and gsoc 2019 hasn't even remotely begun yet Oct 10 17:37:50 i have a lot to learn to finally make it Oct 10 17:38:01 so i thought to start early Oct 10 17:38:16 I suppose, but that might explain why you're having trouble getting a response Oct 10 17:38:26 it's a high risk high reward thing Oct 10 17:38:33 yeah i guess Oct 10 17:38:47 if you want to get a headstart, just get a beaglebone and start playing with it Oct 10 17:38:56 i got one Oct 10 17:38:59 yeterday Oct 10 17:39:30 dont want to fuck $50 board Oct 10 17:39:51 so thougth get some basics covered n then go Oct 10 17:40:10 it is virtually impossible to fuck the board as long as you stick to software Oct 10 17:40:23 even if you mess the system up completely you can always reflash it Oct 10 17:40:33 ohk there are ways Oct 10 17:40:50 power requirements always differs Oct 10 17:41:09 i have fucked boards previously Oct 10 17:41:12 beaglebone generally only get destroyed things like exposing the board to overvoltage on its I/o when connecting external hardware Oct 10 17:41:28 yes exactly Oct 10 17:41:44 that what to my many other boards Oct 10 17:41:55 happened* Oct 10 17:42:15 poor boards Oct 10 17:42:24 may they rest in peace Oct 10 17:42:28 they were cheaper than beagle Oct 10 17:42:34 so cant risk beagle Oct 10 17:42:41 yeah they are Oct 10 17:43:01 so don't connect external hardware, at least at first Oct 10 17:43:25 i am keeping that in mind Oct 10 17:44:50 how old are u ? if u dont mid Oct 10 17:45:05 mind* Oct 10 17:45:16 35 Oct 10 17:45:26 oh my Oct 10 17:45:29 i forget how i destroyed one of my beaglebone black boards ... something to do with plugging it into the LCD7 cape either in the wrong orientation, or plugging in the power to the wrong place (instructions were very clear about using the LCD7 power and *not* the board when it's connected) Oct 10 17:46:03 but otherwise it's been pretty tolerant to abuse Oct 10 17:46:19 how many yrs of exp in linux industry Oct 10 17:46:39 i have direct supply for wall socket Oct 10 17:46:44 I can report that having a +7V supply wire accidently fall on top of the bbb is also an excellent way to kill one Oct 10 17:47:04 i have to bring it down appropriately Oct 10 17:47:37 i will keep those things in mind Oct 10 17:48:13 and we destroyed the eMMC of a few beaglebone by some badly written logging software that want bezerk and wrote about 16GB of data per hour to eMMC Oct 10 17:48:21 *that went Oct 10 17:48:42 i've managed to flash a bad bootloader onto the eMMC ... and that was a little troublesome to recover from, but eventually managed to boot from microSD after some fiddling Oct 10 17:48:47 i have no idea what all that means but sure Oct 10 17:48:59 vagrantc: hold S2 :) Oct 10 17:49:44 I'd classify that problem as "trivial to recover from", but I guess that's only true once you know how Oct 10 17:49:45 zmatt: sure ... though actually being sure it's held was required some fine motor skills and/or dumb luck Oct 10 17:50:03 yes I made a little "tool" for it Oct 10 17:50:18 to be able to press the button more comfortably Oct 10 17:50:44 also kind of fiddly, but i like boards that support that sort of thing with jumpers or switches Oct 10 17:51:21 instead of using the button you can connect P8.43 to ground (preferably via a resistor) Oct 10 19:49:26 what happens if you flash a bad bootloader to the eMMC? can't you just boot from the uSD card? Oct 10 20:23:50 yep Oct 10 20:24:46 but it does require taking eMMC out of the boot order by powering on with the little finicky S2 button held down Oct 10 20:37:43 a minor inconvenience Oct 10 20:56:59 tis but a flesh wound Oct 10 21:39:29 interesting, the "automatic nesting" feature of the pruss interrupt controller seems to be designed in a way that is impossible to make use of without introducing a race condition Oct 10 22:03:07 zmatt: Do you remember my project for a arb. function gen with the bbb and a ad5676? Oct 10 22:13:14 so does await block or not Oct 10 22:13:56 is async basically process.nextTick with a forced return process? Oct 10 22:14:06 why are all languages keywords, functions, and syntax so arbitrary Oct 10 22:14:12 i hate high level stuff Oct 10 22:54:04 ayjay_t: roughly speaking the use of the 'await' keyword turns the remainder of the code into a callback that is executed when the promise expression you're awaiting is fulfilled Oct 10 23:49:47 ayjay_t: await can be pretty nice though, and difficult to cleanly replicate using promises and callbacks, especially w.r.t. error-handling Oct 10 23:49:57 e.g. this: https://pastebin.com/kUvHydHA Oct 10 23:51:00 is basically equivalent to function() { return Promise.resolve( bar() ).then( foo ); } if we ignore all error handling Oct 10 23:52:18