**** BEGIN LOGGING AT Thu Oct 11 02:59:59 2018 Oct 11 03:59:56 Hi, does anyone know the nature of the /dev/mmcblk1boot0 and /dev/mmcblk1boot1 devices, or know of relevant documentation? Why do they exist and how are they used? Oct 11 07:14:12 Has anyone got GPIO chip selects for SPI working? They all seem to be setup properly in the pin mux but nothing is coming up on my scope when I look at the pins. CS0 is working fine. Oct 11 09:06:01 ive only used spi with cs0 Oct 11 09:06:18 are you using spidev Oct 11 09:35:50 I've used gpio chipselects without a problem Oct 11 09:37:07 * zmatt knows the nature of the mmcblk*boot* devices, but regrettably robjb did not choose to stick around to receive an answer Oct 11 09:42:20 huh so sd-event w/ timerfd and signalfd Oct 11 09:42:38 there was an argument on hackernews (a long time ago) abotu whether or not a pure userspace implementation wasn't ultimately more efficient Oct 11 09:50:42 pure userspace implementation of? Oct 11 09:54:50 note that sd-event only uses one timerfd per (sd_event *e, clockid_t clock) pair, not a timerfd per timer added using sd_event_add_time() Oct 11 09:58:05 so it maintains a list sorted by timer expiry... Oct 11 10:01:22 yes, with some more details to account for timer slack Oct 11 10:02:11 (and it's not a sorted list but a priority-queue) Oct 11 10:06:35 it really does some clever things with timer slack though, e.g. when asked to schedule a really sloppy timers (more than 250ms of slack) it will try to schedule them in a way that makes all sloppy timers on the system that expire roughly at the same time will expire exactly at the same time, causing the corresponding wakeups to be coalesced hence saving power Oct 11 10:07:14 while also making sure this doesn't synchronize timer expirations across different machines, which could result in storms of network traffic (when timeouts expire) Oct 11 10:08:37 I don't think I've seen any other eventloop implementation that does this, or even supports timer slack at all Oct 11 10:21:27 that is pretty interesting Oct 11 10:21:52 i meant that libuv was an event loop, timer (probably a rbtree scheduler) in userspace Oct 11 10:22:11 but i mean, it wasn't my post and i'm just getting into this, so... Oct 11 10:22:19 any eventloop I know of does that, including sd-event Oct 11 10:22:58 its an old post and i wont try to reiterate anyone's argument Oct 11 10:24:13 like I said, sd-event only uses one timerfd per clock source (clockid_t) per event loop, which is the minimum feasible Oct 11 10:27:23 its from 2011 but this was the argument where i started on this journey: https://news.ycombinator.com/item?id=3116608 Oct 11 10:28:44 zmatt: yes, power consumption-wise that makes a lot of sense Oct 11 10:29:26 zmatt: to coalesce timer events if precision is not required so that you can have longer idle times (potentially) Oct 11 10:33:02 ayjay_t: both sides are correct, and sd-event uses the best of both worlds: it uses timerfds instead of the timeout parameter of epoll_wait to be able to support high accuracy when required (or low accuracy when possible) and different clock sources, while it manages the (potentially large number of) timers per clock source (per event loop, if you use multiple) purely in userspace Oct 11 10:33:44 thinkfat_: yep, and for many purposes you don't really need very precise timers Oct 11 10:33:52 especially things like timeouts Oct 11 10:53:04 you're probably using sd-even with C zmatt? Oct 11 10:53:07 sd-event* Oct 11 10:53:20 C++ Oct 11 10:53:33 single threaded or? Oct 11 10:53:56 preferably Oct 11 10:53:58 yeah Oct 11 10:54:08 well it'll be a fun experiment Oct 11 10:54:08 some things do require threads Oct 11 10:54:33 but typically you won't have an event loop anywhere but on the main thread Oct 11 10:55:34 yeah Oct 11 10:56:38 of course you *can* if you want to, sd-event supports multiple threads, each with their own event loop (sd_event_default()), or even multiple event loops per thread (sd_event_new()) if you have some bizarre need for it Oct 11 11:03:11 but you can't meddle with another thread's sd-event stuff... I think the only safe calls are a few information queries like sd_event_source_get_event() and sd_event_get_tid(), and even then you'll need some way to ensure the object remains alive without being able to hold an explicit ref to it Oct 11 11:07:10 hello, did anyone managed to get working with bone-debian-9.5-iot-armhf-2018-10-07-4gb ? Oct 11 11:07:18 hello, did anyone managed to get CAN working with bone-debian-9.5-iot-armhf-2018-10-07-4gb ? Oct 11 11:08:10 are you using some kind of CAPE or simply connected a CAN transceiver? Oct 11 11:09:02 just a CAN transceiver Oct 11 11:09:52 it worked with debian 8 Oct 11 11:09:56 then you'll need to manually configure the pins using config-pin (e.g. config-pin P9.24 can && config-pin P9.26 can) before attempting to bring the interface up Oct 11 11:11:10 Do I need to add CAN dtbo to uboot_overlays? Oct 11 11:11:10 or, alternatively, manually request one of the CAN overlays in /boot/uEnv.txt Oct 11 11:11:38 so it's either-or: either you configure an overlay, or you use config-pin Oct 11 12:07:57 regarding the CAN issue, we tried with both overlays and config-pin, we are able to bring the can0 interface up but there is no activity on the output pins... Oct 11 12:22:05 Hello, I encountered a problem when start PWM in different frequencies on channels, would you help to resolve it? thanks! Oct 11 12:44:19 Masood: the two channels of each ehrpwm peripheral cannot be configured to different frequencies Oct 11 12:44:28 other than that there are no restrictions Oct 11 12:46:21 zmatt when i change same frequency at all pwm pins then that error also errors Oct 11 12:47:10 zmatt when i change same frequency at all pwm pins then that error also occurs Oct 11 12:48:19 there's no need to repeat yourself Oct 11 12:49:42 you need to disable both channels of an ehrpwm peripheral before you can change its frequency Oct 11 12:49:50 waiting for answer Oct 11 12:50:07 people in here are often busy, have patience Oct 11 12:50:13 k Oct 11 12:58:08 Hi ! Can anyone tell me how to know when the flash is finished? I read the 4 LEDs should stay on? but they beat the heart for 10 min so I think it's done? can someone confirm? Oct 11 12:58:32 bruce1495: sounds like you're not flashing at all but have simply booted a system Oct 11 12:58:38 indeed Oct 11 12:58:52 you should see "cylon lights" for for the time it flashes Oct 11 13:00:32 by cyclon you mean led 1 then led2 then led3 ...? I saw this but I left about 30 min Oct 11 13:00:58 that's weird Oct 11 13:01:22 how do you flash the emmc then ? i followed these steps Oct 11 13:01:23 http://172.22.1.158/bone101/Support/BoneScript/updates/ Oct 11 13:01:50 oops these https://elinux.org/Beagleboard:BeagleBoneBlack_Debian#Flashing_eMMC Oct 11 13:01:51 cylon lights means: 0 - 1 - 2 - 3 - 2 - 1 - 0 - 1 - 2 - 3 - 2 - 1 - etc Oct 11 13:02:19 yes i ve seen this for a while before I left :) Oct 11 13:02:21 yes you uncomment the appropriate line in /boot/uEnv.txt on the sd card and then boot from that card, that is the correct procedure Oct 11 13:02:33 what you're describing sounds like it somehow got power-cycled Oct 11 13:02:41 possibly after flashing had completed Oct 11 13:02:44 dunno Oct 11 13:03:07 (and not merely rebooted, I mean power-cycled, as in not something that software can do) Oct 11 13:10:49 oh i simply wrote reboot in command line :s Oct 11 13:10:52 will try again then Oct 11 13:10:54 thanks ;) Oct 11 13:11:05 ??? Oct 11 13:11:20 what? Oct 11 13:11:45 I don't understand what you're saying or why you think you need to try again Oct 11 13:12:02 you need to find out how the board got power-cycled in your 30 minutes of absence Oct 11 13:13:47 my english is bad i dont understand everything :s Oct 11 13:14:12 how do I know if this worked ? Oct 11 13:14:44 actually I don't really have any explanation of how it can have booted into a system after you observed it flashing Oct 11 13:15:41 I dont think the card is booted because I cant log in at all with putty Oct 11 13:16:07 when you say "they beat the heart" you mean a heartbeat pattern on led0 only? Oct 11 13:16:28 and other leds off? or random intermittent activity? Oct 11 13:16:40 the 4 Leds are heartbeating Oct 11 13:17:02 that's a pattern I've never heard of. don't you simply mean all four leds flashing? Oct 11 13:17:33 a heartbeat is on off on off (pause) on off on off (pause) etc Oct 11 13:17:56 ahah I broke the Beaglebone :p Oct 11 13:18:04 yes i m sure its a hearbeat on all 4 leds Oct 11 13:18:17 o.O Oct 11 13:18:29 maybe that's some failure condition I'm unfamiliar with Oct 11 13:19:31 :s ok then should i unplug the card and try to boot again ? Oct 11 13:19:36 oh they do flash in a heartbeat on flasher failure Oct 11 13:19:45 I thought they were simply flashing Oct 11 13:19:53 I guess I misunderstood (or it changed, dunno) Oct 11 13:20:24 hmm, I don't think I've ever *seen* a flasher failure myself, and other people always referred to it as "flashing" rather than "heartbeat" Oct 11 13:20:35 anyway, it's a flasher failure Oct 11 13:20:49 are you sure you're not trying to flash a 4GB image onto an older beaglebone that only has 2GB of eMMC ? Oct 11 13:22:54 mmh dont know someone gave me this BBB without docs i ll try to find out Oct 11 13:23:23 revision may be listed on a sticker on the side of the board Oct 11 13:24:02 or you can look up the part number of the eMMC chip itself Oct 11 13:25:05 i see revB3 on the side Oct 11 13:25:22 B3 is not a revision that exists Oct 11 13:25:25 rev B does Oct 11 13:25:28 but not B3 Oct 11 13:28:06 https://beagleboard.org/static/ti/product_detail_black_lg.jpg the large chip in the topright corner is the eMMC. if yours has the same logo as this photo does (micron), check the code on the second line (starting with "JW") Oct 11 13:32:47 ok i found it and its 2Gb only :s Oct 11 13:33:12 its JW 896 Oct 11 13:34:10 2GB Oct 11 13:34:11 yes Oct 11 13:35:11 ok so I need another BBB or a smaller image to do this :) Oct 11 13:35:17 thanks for your time zmatt ;) Oct 11 13:38:32 yep. older iot images used to fit, but they've grown too obese Oct 11 13:38:37 console image is still fine of course Oct 11 13:39:17 but it is deliberate minimal, you'll need to install whatever stuff you need Oct 11 13:49:05 another question ^^' Oct 11 13:50:05 now i removed the # in uEnv.txt ... it starts flashing emmc each time when I boot the card Oct 11 13:50:31 correct, so be sure to remove the card after flashing Oct 11 13:53:04 is there a way to put that # back in uEnv.txt file ? without booting the card ? Oct 11 13:53:21 because i cant do anything to the card now :s Oct 11 13:53:45 easiest is to just wipe or reflash the card (presumably you still have the .img file you downloaded) Oct 11 13:54:02 or you can mount it on any linux system (including the beaglebone after it has booted) and modify the file on the card Oct 11 15:00:42 When first time I run the dimLED.py, it does work. Butt, if I re-run the code by changing frequency at PWM pins its show that error “RuntimeError: Problem with a sysfs file“. But when I again restart the beaglebone and change the frequency at PWM pins its work first time and afterward it shows the same error. Oct 11 15:01:00 waiting for comments and solution Oct 11 15:20:23 14:49 < zmatt> you need to disable both channels of an ehrpwm peripheral before you can change its frequency Oct 11 15:21:42 how can i disable both channels of an ehrpwm peripheral? Oct 11 15:21:49 set them to 0 Oct 11 15:23:10 by using which command? Oct 11 15:23:33 please give me one example Oct 11 15:27:56 ? same as the normal way you change pwm channels Oct 11 15:28:03 set the duty cycle to 0, I would guess Oct 11 15:28:42 I don't know what library you're using (other than being something in python apparently) or what its syntax is Oct 11 15:29:07 but you've already been able to set pwm channels successfully so... Oct 11 15:33:03 hey zmatt... why didn't you like the Slack bridge again? you said you don't like it, right? Oct 11 15:33:32 every bridge that's a single user in chat here is annoying Oct 11 15:33:56 but the bridge prints a username, yes? Oct 11 15:34:12 riot = good, b/c they all have unique users? Oct 11 15:34:34 * jkridner wonders if Slack has a better IRC bridge. Oct 11 15:34:54 if it creates a user for each user then there's no problem, that can't be distinguished for people using irc clients Oct 11 15:34:57 * jkridner loves IRC, but newbs, well, you know. Oct 11 15:35:07 k Oct 11 15:35:26 @jkridner yeah newbs Oct 11 15:35:39 newbs use the web client :P Oct 11 15:36:07 @jkridner newbs are lost Oct 11 15:36:43 I wonder if https://blog.irccloud.com/slack-integration/ would work Oct 11 15:37:40 @zmatt Please Oct 11 15:37:43 hmmm... maybe that isn't supporting freenode. Oct 11 15:38:30 rd: ? Oct 11 15:38:35 not with that newfangled sasl auth anyway I'd say Oct 11 15:39:22 zmatt: rd is pushing to make barriers lower for newbs.... we've been chatting on Slack. Oct 11 15:39:47 one of the challenges he's trying to address is the ability to share his screen to get live answers. Oct 11 15:40:08 no irc gateway is going to help with that Oct 11 15:40:10 there's nothing better to help a new than to point him to what he needs to learn... Oct 11 15:40:10 not that many of us who love IRC would engage in such a process often... Oct 11 15:40:30 as it hurts our ability to really multitask. Oct 11 15:40:38 yup Oct 11 15:40:43 #zmatt I am a manufacturing engineer but I have been having a great deal of difficulty with the books and other documentation available not matching my screen. Oct 11 15:40:44 thinkfat_: sure, but sometimes you gotta walk them off the ledge. Oct 11 15:40:45 realtime, shudder.... Oct 11 15:41:00 I don't see slack-irc integration going to help with that in any way Oct 11 15:41:22 zmatt: I try to avoid creating forks in the community where I can... Oct 11 15:41:33 and "hey, can you look at my screen and help me out" is getting in the sort of terrain I'd expect from commercial support, not community support Oct 11 15:41:42 if there is a push to add another central place for support, I try to figure out how I can bridge the existing community. Oct 11 15:41:42 * thinkfat_ nods Oct 11 15:42:20 zmatt, thinkfat_: I agree, but I think there are some "spare timers" who might engage on such an endeavor every now and again. Oct 11 15:42:38 jkridner|web: I suspect that such a bridge is more likely to piss off or alienate the existing community than integrate them, since if you want to use slack for enhanced functionality then you're also making the bridged irc users "second-tier" Oct 11 15:42:42 within that community Oct 11 15:42:55 one of those community impedance mis-match sort of issues, but I always want to explore new on-boarding possiblities. Oct 11 15:43:06 e.g. "can you look at my screen", "go away, I'm using irc" Oct 11 15:43:10 it's not going to help Oct 11 15:43:12 :-) Oct 11 15:43:28 It might not be a place for developers. It might work if each successive person down the path was allowed and encouraged to leave a trail of bread crumbs. Oct 11 15:43:37 guess I'll look at slack as a newb gateway until they are ready for IRC. Oct 11 15:43:50 what is slack, anyway? Oct 11 15:43:53 seeding Slack with some experienced users then becomes the challenge. Oct 11 15:43:55 good place to hang out? Oct 11 15:44:06 thinkfat_: another random proprietary chat app. Oct 11 15:44:15 we get to use it "free" as a non-profit. Oct 11 15:44:15 oh, alright Oct 11 15:44:34 "free" as in "we own your bits, so we don't charge you $$$", not freedom. Oct 11 15:44:55 but, it has lots of "features" Oct 11 15:44:58 "except we charge you $$$ if you want " Oct 11 15:45:01 like screen sharing. Oct 11 15:45:10 e.g. access to your history before a certain time Oct 11 15:45:20 zmatt: indeed. but the zero-$$$ tier for non-profits is pretty good. Oct 11 15:45:36 voice calling and drawing on a community or shared screen. Oct 11 15:45:45 full history access is one of the features a non-profit gets without $$$. Oct 11 15:45:54 but, crazy walled garden crap. Oct 11 15:45:58 :) Oct 11 15:46:26 for those that don't know how important freedom is, however, it might be good to use to help them learn. Oct 11 15:46:43 thing is, the limitations of irc match pretty well with the sort of support I'm willing to give. I would not engage in voice calling or screen sharing regardless Oct 11 15:46:58 * jkridner|web shows his colors as a raging pragmatist. Oct 11 15:47:04 but other people may perhaps Oct 11 15:47:32 Those of us on the other end of the spectrum. Lost and desperate. Oct 11 15:47:41 I told rd that I didn't make the bridge because 'zmatt didn't like it and he's a monster in support on #beagle" Oct 11 15:47:53 True Oct 11 15:48:02 * jkridner|web does not like pissing off crazy-valuable community contributors. Oct 11 15:48:29 fun fact: the c++ slack is very large and incredibly good. like, "want to talk to the commitee members? sure, here they are." Oct 11 15:48:42 guess we are mostly wrapped on this discussion for now, but I still have to figure out how to seed good devs that would look at someone's screen. Oct 11 15:48:45 if a bridge bot like that were here and active, I'd either put it on /ignore or simply leave Oct 11 15:48:56 and i just tend to treat it like another irc reaplcement. Oct 11 15:49:02 I value freedom to the point that it is difficult to hold down a job. Oct 11 15:49:04 LetoThe2nd: thanks for the feedback. Oct 11 15:49:25 LetoThe2nd: maybe you want to join beagleboard.slack.com? :-) Oct 11 15:50:02 jkridner|web: i actually did watch it already for some time. if i can get it into my screen habits, sure. but no promises given. Oct 11 15:50:25 there is a #general that I used to bridge here before killing the bridge. Oct 11 15:50:32 another reason I wouldn't be a fan of slack is that it would be yet another browser tab. I have no problem being in a dozen IRC channels, but I would really not want to be in a dozen slacks Oct 11 15:50:48 there is also a #support and that is probably where I want to start sending newbs if I can get some critical mass. Oct 11 15:51:06 perhaps from https://beagleboard.org/chat where a lot of them show up here. Oct 11 15:51:17 and where I'm logged in from this screen. Oct 11 15:51:33 zmatt: yeah, thats really a pain point. but like the other pragmatist here put it, for certain slacks the gain outweighs the downsides, at least for me. Oct 11 15:51:47 * jkridner|web just wishes people other than zmatt would learn to ping me here rather than direct messaging me where I'll *never* answer. Oct 11 15:52:09 jkridner|web: hum, if you want me to i can ping you all day. but what for? Oct 11 15:52:14 LetoThe2nd: for me that is currently only one, and that's the one for the company I work for :P Oct 11 15:52:21 LetoThe2nd: we had a huge problem with GSoCers using Google Hangouts.... Oct 11 15:52:32 we were losing all the good community that IRC offers.... Oct 11 15:52:51 jkridner|web: is it that time of the year again? Oct 11 15:52:55 we *required* them to go back and use IRC and kill the G+ walled garden nonsense. Oct 11 15:53:10 LetoThe2nd: naw, we already wrapped up GSoC for the Summer. Oct 11 15:53:28 anyway, I'd look at Slack as a bridge... Oct 11 15:53:32 or maybe a filter.... Oct 11 15:53:40 filter is probably a better word.... Oct 11 15:53:46 well see. Oct 11 15:53:48 jkridner|web: https://beagleboard.org/chat could really use a few quick points about how to behave on irc in a way that maximizes chance of getting help (and not a link to a long article) Oct 11 15:53:51 make sure they know enough not to waste people's time here too much. Oct 11 15:53:59 training ground... Oct 11 15:54:01 Yeah. I saw the Google Hangouts as well. There are a number ways to get information but I have no idea where to begin. Oct 11 15:54:08 e.g. don't send private messages, just ask your question (with sufficient detail), and BE PATIENT Oct 11 15:54:11 wait so what is the best channel for n00bs? Oct 11 15:54:18 is it really slack? gitter? discord? Oct 11 15:54:19 at least force them to read http://www.catb.org/~esr/faqs/smart-questions.html Oct 11 15:54:28 jkridner|web: tl;dr Oct 11 15:54:38 and not all of it is really relevant Oct 11 15:54:52 rd: well, this is the very best place for live chat info, but you have to put together your expectations and questions right.... Oct 11 15:55:01 and real-time support is the wrong expectation. Oct 11 15:55:04 That article scared me from saying anything when I went to this chat room and it looked like a bunch of higher level conversation. Oct 11 15:55:07 like, it's a useful doc to read, but it's not useful to link to as a way to explain how to behave on irc Oct 11 15:55:24 rd screenshare solution is probably tmux? if you're talking about console Oct 11 15:55:36 ayjay_t: for a real newb, we are trying to see if we can put something together... Oct 11 15:55:41 rd: there's never a problem with asking a question at any time. just be prepared to pay attention if multiple conversations are going on at the same time Oct 11 15:55:46 this channel is the best place for support for semi-experts. Oct 11 15:56:04 that is, people that know some basics and just need to share in the tribal knowledge. Oct 11 15:56:08 rd: but that's happening even now, hence I tend to explicitly address people in such cases to avoid ambiguity Oct 11 15:56:30 jkridner|web: heh, that's a good description Oct 11 15:56:51 Part of the problem is that as a newb I do not know the language to create well defined descriptions of my problems. Oct 11 15:56:59 rd, zmatt: indeed, adding who you are speaking to before you start typing more stuff is very helpful to keep a thread going. Oct 11 15:57:38 I mean, noobs can get help too... knowledge is not required at all, but like jkridner said you need the right expectations and ask the right sort of questions Oct 11 15:57:54 zmatt, rd: interesting feedback on the doc. didn't know it was scaring people away Oct 11 15:58:11 jkridner|web: I didn't know it was scaring anyone away, I simply assumed noone read it Oct 11 15:58:14 my goal was to give people some rules that helped them be productive in here. Oct 11 15:58:33 I can show you a picture and make it clear that I do not know what I am talking about and you can instantly point me in the right direction. Oct 11 15:58:36 #1 issue always seems to be people dropping "stupid" questions and then closing the window in about 120 seconds. Oct 11 15:58:41 a few really short points, inline on the bbb.io/chat page itself, is all I'd expect people to read Oct 11 15:58:45 i feel like the discourse forum is #1 right now for support? Oct 11 15:58:57 jkridner|web: or just say "hello" and leave Oct 11 15:59:04 it is fine to ask a "stupid" question, as long as you keep trying, following up and taking the feedback to heart. Oct 11 15:59:08 expecting it to work like 1-on-1 helpdesk or something Oct 11 15:59:09 (and having thick skin) Oct 11 15:59:25 ayjay_t: there's a "discourse forum"? Oct 11 15:59:31 its a brand of forum Oct 11 15:59:45 discourse.org Oct 11 15:59:48 ayjay_t: but people are using it to support Beagle users? Oct 11 15:59:57 no sorry Oct 11 16:00:25 for what its worth, i'm the admin of the beaglebone black facebook group and we get a decent amount of international action there Oct 11 16:00:34 rd: sometimes people will click on links to images, but it will reduce the response-probability Oct 11 16:00:35 i inherited it Oct 11 16:00:47 there's a couple... Oct 11 16:00:49 Is there a BeagleBone Blue group? Oct 11 16:01:08 https://www.facebook.com/groups/127802247417133/ Oct 11 16:01:23 thats the beaglebone black group i moderate Oct 11 16:01:43 @ayjay_t Thanks Oct 11 16:01:46 rd: it also depends on context, e.g. "wtf is happening with my lcd output" could easily be accompanied by a link to an image, or possibly even a video Oct 11 16:02:31 yeah, jkridner|web you should join it, i'll make you an admin Oct 11 16:02:39 on the other hand logs or other text output should be shared using pastebin.com or a similar site, and taking a screenshot of it and sharing that will annoy people Oct 11 16:03:21 I can't select and copy-paste an error message that appears in a screenshot (to search for it) Oct 11 16:03:21 rd: I'm not aware of a Blue-specific support channel. Oct 11 16:03:38 rd: I can say that James Strawson is on our Slack #dev channel. Oct 11 16:04:09 and he's probably the best expert, along with Eric Welsh and, after a few other people, me. Oct 11 16:05:03 What I am interested is being able to accommodate the "WTF" question by automatically routing them to a most likely answer. Perhaps not even chatting with the person but filling in holes in an automated response. There are times when people are at the emotional point where they feel like failures because they can not figure out a problem for themselves. Not the best emotional state for formulating a coherent question. Oct 11 16:05:14 zmatt: yeah, I think the complexities of GUI support are easily lost on people just getting started with Linux.... but I hope to convert them to the "Light Side" nontheless. Oct 11 16:05:42 rd: one of the ideas you've been championing is classification of failure modes for newbs. Oct 11 16:06:34 rd: if we can create some documentation that can help a newb really describe the point in which they encountered a challenge, it can really help the community help them get un-stuck. Oct 11 16:06:47 honestly, better documentation is really needed Oct 11 16:06:55 where are you fielding this hard-to-decipher questions? Oct 11 16:06:56 zmatt: +1,000 Oct 11 16:06:59 is it all on slack? whats the link? Oct 11 16:07:24 rd: and that's where I really want to focus your guidance to me and other devs. Oct 11 16:07:40 rd: I want to improve the documentation incrementally, every day. Oct 11 16:08:24 basically the current state is Oct 11 16:08:27 rd: if you can show up here or on Slack every day and ask for a small fix to the docs, that'll go a LONG way in 3-4 months. Oct 11 16:08:27 21:54:58 < fasdgghwegh> zmatt: do you know of any good free online resources that retells what you just showed me? Oct 11 16:08:31 21:55:17 < zmatt> there might be an irc log of this channel online ;) Oct 11 16:08:36 hahaha Oct 11 16:08:42 no but seriously where are these n00bs Oct 11 16:08:47 I have a degree in manufacturing engineering with minors in math and business with 20+ years of experience. I make documentation for a living. I am not able to bridge the gap despite numerous books and several years of intermittent effort. Oct 11 16:09:02 * jkridner|web knows Blue needs an SRM like https://github.com/beagleboard/beaglebone-black/wiki/System-Reference-Manual Oct 11 16:10:10 that SRM is my go-to for BeagleBone Black... Oct 11 16:10:27 i'm sure i'm qualified to help some people answer questions, i'm active on three different BBB channels i've taught like 4 interns how to use the BBB Oct 11 16:10:31 rd: I can imagine that experience in those fields is not going to be sufficient to bridge the gap Oct 11 16:10:34 i mean i'm not as qualified as zmatt or other people Oct 11 16:10:37 and similarly https://github.com/beagleboard/pocketbeagle/wiki/System-Reference-Manual for PocketBeagle Oct 11 16:10:56 rd: I mean, it's quite a leap from there to embedded linux development Oct 11 16:11:18 ayjay_t: but, you and rd are both qualified to help me understand where I need to focus energy on fixing documentation for newbs. Oct 11 16:11:33 but i need to _meet_ more newbs to answer that question Oct 11 16:11:39 i want to get a sense for those "common failure modes too" Oct 11 16:11:43 and i want to know what channels they find most popular Oct 11 16:12:05 maybe we can do better to improve community awareness of already popular channels like IRC and google groups Oct 11 16:12:07 on-ramping people to the Linux Foundation training feels to be a salient goal. (or bootlin, etc.) Oct 11 16:13:06 ayjay_t, rd: one thought I've had and haven't followed-up on is publishing FAQs on many areas within the https://beagleboard.org/discuss forum. Oct 11 16:13:12 @zmatt I am not asking to make any giant leaps. (I greatly appreciate your time by the way) but I am suggesting that we do a better job of only answering one question one time and then putting our efforts into leaving behind a better trail of bread crumbs. Oct 11 16:13:35 my thought was to edit https://beagleboard.org/faq and other FAQ sites to point to the pinned FAQ in the discussion Google group. Oct 11 16:14:21 well i use IRC and i like it because it makes me feel like a "real hacker" :cry: but it also turned out to be one of my top 3 sources of education Oct 11 16:14:24 rd: the main problem I think is that there's not really good getting-started documentation that's accurate, uptodate, and helpful for people who are genuinely just *getting started* Oct 11 16:14:30 seems like I haven't followed-up on that idea since 2013: http://beagleboard.org/discuss?place=msg%2Fbeagleboard%2Fu28ytaoNenU%2FoudgXICbR58J Oct 11 16:14:45 rd: you don't want people asking "how do I get started?" on irc, it's not useful Oct 11 16:14:55 @zmatt oh hell yeah! Oct 11 16:15:09 zmatt: please also report any fixes you'd like to see on https://beagleboard.org/getting-started Oct 11 16:15:12 okay but that being said, how do you "diagnose" that question? i mean you need to fire back the "what do you want to make?" Oct 11 16:15:23 it is my thought that that page is the go-to for newbs. Oct 11 16:15:24 i feel like without some idea of what the user is into, its hard to know how to make suggestions Oct 11 16:15:38 links out to various FAQs seems like the thing to do? Oct 11 16:16:22 rd, zmatt, ayjay_t: the overall documentation is crowd-sourced, but please help set my priorities on what can be fixed on https://beagleboard.org ! Oct 11 16:16:30 @jkridner Much of this points to what I have suggested about routing people based on them logging into the beagleboard.org site. Oct 11 16:16:49 well i don't know exactly what to do, it seems like a lot of work Oct 11 16:16:57 I mean, if https://beagleboard.org isn't a helpful resource for newbs, that's a problem Oct 11 16:16:58 but if n00bs are flocking to a certain channel, i'll happily sit on it and try to help people Oct 11 16:17:08 jkridner|web: I don't have specific fixes... the getting-started seems fine for getting existing linux developers started on the beaglebone Oct 11 16:17:26 "logging into" is an overloaded term. I hope you just mean visiting. Oct 11 16:17:35 When you do formal training, you check off a box with where you are at in the process. A flow chart would guide you to the information you need and away from information that would cause you confusion. Oct 11 16:17:41 except I guess you should very strongly advise either reflashing eMMC or erasing eMMC Oct 11 16:17:57 since otherwise an old u-boot on eMMC combined with booting from a new system on sd card creates tons of problems Oct 11 16:18:06 zmatt: indeed, which is why "step 0" is there now, which is making a microSD to boot from... Oct 11 16:18:09 haha rd that's a fun idea, gamify it! Oct 11 16:18:21 but the dang old u-boot issue has really made our lives hard. :-( Oct 11 16:18:42 jkridner|web: booting from microSD is a problem, not a solution. I'd rather you provided flashers Oct 11 16:18:53 holding the "S3/BOOT/SD" button when applying power is really tough to teach to absolute nebs. Oct 11 16:19:01 I am still lost as to, "is my driver working" That stops me cold in my tracks on page 14 of 200+ in the "Beaglebone Robotics Projects" book. Oct 11 16:19:07 jkridner|web: it's also too inconvenient Oct 11 16:19:25 zmatt: what about PocketBeagle users? just automatically don't flash if you find that's what you are booting? Oct 11 16:19:48 we used to provide flashers by default, but the fear was that newbs would lose their work on their eMMC.... Oct 11 16:19:53 * ayjay_t never had a problem with the hold-button-to-flash thing Oct 11 16:20:00 maybe we should just let them lose it? Oct 11 16:20:01 jkridner|web: was that a frequently-reported problem? Oct 11 16:20:31 zmatt: probably just an irrational fear. can't say I remember many reports, despite seeing it in action once or twice. Oct 11 16:20:40 I frequently get people here who have trouble making a flasher, or who have problems fixable by reflashing Oct 11 16:21:04 zmatt: if I re-think it, I can believe that the people who experienced losing their data weren't incredibly upset by it and simply "learned a lesson" Oct 11 16:21:08 accidental reflashing is not something I've frequently heard of as a problem Oct 11 16:21:10 yeah it's frustrating when you are having trouble booting Oct 11 16:21:23 but honestly a lot of the fun is overcoming difficult problems and mastering the skills Oct 11 16:21:36 and its helpful to have people who care about the community at least there to encourage people Oct 11 16:21:43 zmatt: at one point I proposed a u-boot change on the list involving holding the boot button to initiate flashing... Oct 11 16:21:57 good times Oct 11 16:22:00 * ayjay_t gets back to work Oct 11 16:22:08 some people even mistakenly document that the button is somehow involved in initiating flashing as of now. Oct 11 16:22:35 jkridner|web: really a big part of the confusion is that u-boot on eMMC will load linux from SD card Oct 11 16:22:41 * jkridner|web refers to https://www.youtube.com/watch?v=k04Lv2HXWcM Oct 11 16:22:43 The difficulty is not the problem. The lack of/not mated correctly to the correct information/solution. Oct 11 16:22:50 u-boot should only load linux from the same device it was located on Oct 11 16:22:55 that would solve a ton of confusion Oct 11 16:23:55 i.e. if you want to boot from sd card, either use the S2 button (if your desire to boot from SD card is temporary) or wipe eMMC (if you want to permanently boot from sd card) Oct 11 16:23:55 zmatt: yeah, I didn't like that rcn-ee did that at the time, but the cat is out of the bag there. Oct 11 16:24:35 * jkridner|web didn't mind that the button had to be held down to boot from microSD, but didn't balk at the right time. Oct 11 16:25:02 rd: docs that are crisp, clean, clear, and consise are hard to come by, yep. Oct 11 16:25:13 it's also inconvenient that bootrom doesn't provide a boot order that places eMMC after SD Oct 11 16:25:41 zmatt: to confuse matters more, AM57x does. Oct 11 16:25:50 @jkridner "logging into" is an overloaded term. Not entirely sure what you mean but I was referring to a user experience like when I set up my google router. I scanned the QR code and the app knew how to get me to the correct information and not present me with anything that does not apply. How ever that is accomplished..... Oct 11 16:26:54 well, https://beagleboard.org/getting-started is indeed generic, but I can create short-cut URLs specific to certain boards. Oct 11 16:27:00 jkridner|web: speaking of the am57x, why are there packages specific to the am57x installed on the images specific to the beaglebone? :P Oct 11 16:27:20 I don't want to maintain an entirely different version of that page for every board from a practical maintenance viewpoint. Oct 11 16:27:56 zmatt: for the same reason there is only one https://beagleboard.org/getting-started.... the overhead of customization. Oct 11 16:28:18 deleting the stuff automatically could probably be done... that was done for Green customization at one point. Oct 11 16:28:44 I doubt you have to "delete" anything, just don't install it. since this is done by a script I don't really see the problem Oct 11 16:29:04 and I think I can do the same easily on https://beagleboard.org/getting-started .... I can use a pull-down to set a cookie and I can use a cookie to hide any non-relevant information. Oct 11 16:29:32 zmatt: fair enough as different images exist today. guess rcn-ee would need to answer. Oct 11 16:29:55 When I talk about "logging into the website" I am wanting to register the board I have so I can be given specific information related to that board. Older information can still exist but be hidden from the newb. Oct 11 16:30:01 yeah I should mail him... except I don't care *quite* enough to give me the energy to do so ;) Oct 11 16:30:03 * jkridner|web wants to get it down to 1 image. Oct 11 16:30:12 1 image is never going to happen Oct 11 16:30:38 zmatt: 100% sure? no foo will do the trick? Oct 11 16:30:51 I know I could never figure out the xM / BeagleBone foo. Oct 11 16:31:17 but, I thought that having multiple SPL images that could "fail" properly could do the trick. Oct 11 16:31:29 hmmmm Oct 11 16:31:42 I'll admit I'm not sure what happens if SPL returns to bootrom Oct 11 16:31:53 I would focus on getting the user to the correct image with a flow chart. The current method is all sorts of information that does not apply on the same page. Oct 11 16:31:58 I think the signatures expected by the ROM are different, but I don't know for sure. Oct 11 16:32:17 I don't think it is Oct 11 16:32:52 rd: what about the cookie/pull-down approach? don't like it? Oct 11 16:33:19 * jkridner|web needs to streamline the time spend on this on a daily basis. Oct 11 16:33:47 I would suggest a question and answer approach with the flow chart. Something to guide the user on what to choose. Oct 11 16:33:48 nice to spend some extra time to get this bootstrapped, but I really need some small actions broken down for me on a daily basis to get this moving forward. Oct 11 16:33:50 jkridner|web: configuration header is the same on x15 and bbb Oct 11 16:34:30 rd: well, the pull-down could be in response to a question. I'm fine with basic HTML, but something that needs a graphic designer would need to be in a phase 2. Oct 11 16:34:40 zmatt: :-( Oct 11 16:34:52 jkridner|web: https://pastebin.com/raw/gWtSqyq8 Oct 11 16:35:38 I see where the constraints come in. Perhaps a dumb question box with check boxes.... Oct 11 16:35:41 and the MLO itself doesn't have a signature, its header is just struct { u32 length; u32 addr; }; Oct 11 16:36:09 zmatt: why can't the same SPL run on each again? Oct 11 16:36:09 hmm, maybe the load addr used on the am335x is invalid on the x15 Oct 11 16:36:22 zmatt: using 'if' rather than '#ifdef'? Oct 11 16:36:29 jkridner|web: won't fit Oct 11 16:36:45 u-boot is too bloated Oct 11 16:36:48 even SPL Oct 11 16:37:20 well, if we limit the boot sources (only same medium) and maybe even punt on DDR config if it is too big (ie., load TPL), I think we could get there. Oct 11 16:37:47 @zmatt I am just trying to get past "driver failed" This is the point where I want to use a picture and an arrow. Oct 11 16:37:49 * jkridner|web wishes we could kick back to ROM if on the wrong processor. Oct 11 16:38:00 yeah it could be theoretically possible, but it would require a lot of work probably Oct 11 16:38:05 I don't exclude the possibility you can Oct 11 16:38:31 rd: the easiest answer to get past "driver failed" is to start by loading the latest image from https://beagleboard.org/latest-images Oct 11 16:38:36 I think bootrom invokes the MLO like a function, rather than jumping to it, but I'm not sure Oct 11 16:38:55 zmatt: so, you think the stack has a return pointer?!? Oct 11 16:39:04 lr, not stack Oct 11 16:39:26 lemme check if the parts of bootrom I've reverse-engineered include that call Oct 11 16:39:31 zmatt: whatever... compiler detail. Oct 11 16:39:33 its been a while Oct 11 16:40:10 I do not know what all of this means on the page or if it applies to my situation. "To turn these images into eMMC flasher images, edit the /boot/uEnv.txt file on the Linux partition on the microSD card and remove the '#' on the line with 'cmdline=init=/opt/scripts/tools/eMMC/init-eMMC-flasher-v3.sh'. Enabling this will cause booting the microSD card to flash the eMMC. Images are no longer provided here for this to avoid people accidenta Oct 11 16:40:17 zmatt: well, some state variables would also need to be preserved to know it should try to load the next offset. Oct 11 16:40:53 rd: we are off in a tangent for a broader user experience issue... that of picking the right image to put on the microSD card. Oct 11 16:41:25 rd: oh, sorry, answered the wrong question. Oct 11 16:41:33 @jkridner I know you are off talking about that. LOL Oct 11 16:41:52 rd: it means that we don't provide images that re-flash the eMMC by default. Oct 11 16:42:35 rd: it tried to provide the process in a terse way for modifying the image to reprogram the eMMC on-board flash in the case you choose to do so. Oct 11 16:42:44 @jkridner I am a newb and have no idea of the significance of any of those statements. That is exactly the sort of problem I am referring to. Oct 11 16:42:58 rd: it is the thing that zmatt asserted should be done by default (or at least provided as a standard image download). Oct 11 16:43:06 jkridner|web: I think I've found the call from bootrom into MLO, and it looks like it calls it as a function with 3 arguments and depending on whether its return value is zero or not proceeds into two different paths Oct 11 16:43:45 * jkridner|web just needs a prioritized action. Oct 11 16:43:57 @jkridner I kinda gathered that. Hence my desire to interject and lobby the need for information that matches the user needs. Oct 11 16:44:27 anyway, time for some shopping Oct 11 16:45:00 hmmm.... I wonder if my proposal to alter u-boot to enable flashing based on the BOOT button went to a private list. Oct 11 16:45:57 I don't think I'd be in favor of using the S2 button to flash though Oct 11 16:46:09 you want to be able to use it to force sd-boot for recovery Oct 11 16:46:22 @jkridner I went to the store and bought a 32 Gb Micro SD card so that I could "do over" and try to get past the "driver error". Oct 11 16:46:35 what is giving a "driver error" ? Oct 11 16:46:53 A Beaglebone blue and windows 10 Oct 11 16:47:22 jkridner|web: do you have analytics solution or beagleboard? so you can see what your users are doing? Oct 11 16:47:27 ugh Oct 11 16:47:30 for beagleboard.org* Oct 11 16:47:59 mixpanel, hotspot, google analytics Oct 11 16:48:11 facebook pixel Oct 11 16:48:42 grrr... proposal is indeed on a private group. Here was my thought: https://pastebin.com/QSSGaumR Oct 11 16:48:47 rd: that doesn't really answer anything. you don't need to install any driver these days, it should just work plug&play Oct 11 16:49:09 @zmatt I have about 14 pages of screen captures and arrows that I showed @jkridner in a screen share on Slack. (you can also do small webinars) Oct 11 16:50:40 poor jkridner Oct 11 16:50:47 ;) Oct 11 16:51:23 @zmatt I am sorry, I know that does not answer anything. That is part of the hesitation of asking for help. It does not make any sense. This is the reason why I am lobbying for an error code/serialization that helps define how far along a person has made it successfully. Oct 11 16:52:10 rd: are you booting from eMMC or sd card? Oct 11 16:52:46 rd: I did some webinars at https://beagleboard.org/show/e14-webinar but could do something better for newbs. Oct 11 16:52:49 I have a micro SD card to my knowledge. I would have to Google what an eMMC card is. Oct 11 16:52:57 eMMC is the on-board storage Oct 11 16:53:03 we used to have a nice getting started video for Black. Oct 11 16:53:09 it is out-of-date now. Oct 11 16:53:27 pretty much everything is out-of-date Oct 11 16:53:29 I also did a webinar on getting started with BeagleBone Blue for Autodesk. Oct 11 16:53:31 @jkridner Yeah I tried following those. I am stuck with the "Driver error" Oct 11 16:54:00 rd: for BeagleBone Blue, you should start by using the access point and not plugging via USB for the best experience. Oct 11 16:54:15 also, you should start by downloading and using the latest image from https://beagleboard.org/latest-images Oct 11 16:55:25 rd: "Driver error" without further details also sounds like a Windows problem to me rather than a beaglebone problem, which is part of what makes it hard to give useful feedback Oct 11 16:55:58 @zmatt Yeah, now I recall. At the moment, I am following the Dr. Richard Grimmett book BeagleBone Robotics Projects. Because the Beagleboard.org wiki page for BeagleBone Blue is blank. I am wanting to build industrial robots and automation. Oct 11 16:57:12 @zmatt I do not want to burn "social capital" by asking questions that are irksome or tiresome to more advanced users. Oct 11 16:58:55 @zmatt On the non text based chat room I can show screen captures of what is going on. It allows me to communicate without knowing what I am doing. Oct 11 16:59:34 there are plenty of ways to share images, e.g. upload them to imgur.com or google photos and give a link here Oct 11 17:00:18 although whenver possible, errors should be copy-pasted as text (directly into IRC if single-line or into a paste service like pastebin.com if multi-line) Oct 11 17:00:35 rd: fyi, I replaced the "BeagleBone Blue" link on https://beagleboard.org/getting-started so that it goes to a non-blank wiki.... not that following that link to get started was ever what I intended. Oct 11 17:00:36 @jkridner "using the access point" ? I will have to go read up on what you are referring to. I thought I was supposed to plug in a USB cable. Oct 11 17:01:23 rd: in general with BeagleBone, yes, but Blue has an even-easier way in my opinion. Oct 11 17:01:25 wireless versions of the beaglebone (i.e. bbbw and bbblue) act as a wifi access point to which you can connect. I'm not sure why jkridner thinks this yields a better experience though Oct 11 17:01:45 zmatt: you don't think it is easier? Oct 11 17:02:04 zmatt: I just think people are used to connecting to access points already. Oct 11 17:02:06 jkridner|web: does it yield a portal where you can easily connect the beaglebone to your wifi network? Oct 11 17:02:26 zmatt: whereas there are perhaps-never-used drivers involved with using the USB. Oct 11 17:02:43 otherwise it's basically useless, since if your computer has wifi then you're probably using that to connect to the internet Oct 11 17:02:46 zmatt: nope, just serves Cloud9 IDE on https://192.168.8.1:3000 Oct 11 17:02:52 i.e. useless Oct 11 17:03:05 usb doesn't require driver installation, hasn't required it in a long time Oct 11 17:03:09 zmatt: we installed wifi-dog at one point to provide a captive portal, but rcn-ee didn't feel comfortable supporting it. Oct 11 17:03:34 usb should work plug&play on all platforms nowadays Oct 11 17:04:31 rd: have you previously done some driver installation related to the beaglebone specifically? Oct 11 17:04:44 since if so I'm wondering if that's actually causing problems Oct 11 17:04:54 but who knows with Windows Oct 11 17:05:15 there would never have been any problems if microsoft bothered implementing the USB standard :) Oct 11 17:05:24 @jkridner The page https://beagleboard.org/latest-images is full of things that I do not know the significance of choosing one over the other. When I click on a BeagleBone Blue link it takes me to the page for the Beaglebone Blue but not directly to the software I need to download. Oct 11 17:05:53 rd: first download link, Debian 9.5 2018-10-07 4GB SD IoT Oct 11 17:06:08 yeah I don't know why the product names are links there Oct 11 17:06:42 jkridner|web: also, maybe remove the pocketbeagle and blue from the list of supported devices for the LXQT image? Oct 11 17:06:51 since it's completely nonsensical to install that on those devices Oct 11 17:07:12 @zmatt I have tried to download and use beaglebone products for several years with some success with the Black and cloud nine. I am stalled with BeagleBone Blue. Oct 11 17:07:52 that's unfortunate Oct 11 17:08:28 @jkridner are you still sharing my screen? Oct 11 17:09:12 I regret to say, the driver-related problems are entirely Microsoft's fault for not properly implementing USB Oct 11 17:09:39 although the latest images have worked around that successfully Oct 11 17:09:56 rd: I can look. Oct 11 17:09:59 (or so people report anyway, I don't use Windows myself so I can't confirm or deny it) Oct 11 17:10:02 I stepped away for a minute. Oct 11 17:10:39 @zmatt thank you for your empathy. I have had some undesirable/unproductive responses that had me look for help elsewhere. Oct 11 17:11:10 np Oct 11 17:11:23 rd, zmatt: there's a lot of changes in the last few months around Windows driver support. Oct 11 17:11:44 dlech_ has some knowledge in the area as he contributed some key experience improvements. Oct 11 17:11:47 rd: easy solution, just install Linux on your computer! ;) Oct 11 17:11:54 @jkridner thank you for introducing me to this forum. Oct 11 17:12:19 if you can manage to run the image on https://beagleboard.org/getting-started, then it addresses a couple key Windows 10 changes. Oct 11 17:13:20 so, we'd added a "compatible id" to our USB descriptors that was meant to prevent the need to ever run our driver installer. Oct 11 17:13:50 I am not a big fan of anyone being in control of the entire ecosystem but as a manufacturing engineer, most of the ERP systems are windows based and I have a personal licence of SOLIDWORKS. That demands windows. Oct 11 17:13:50 that was quite a while ago already wasn't it? Oct 11 17:14:03 changes to Windows 10 (live patches) made this stop working. these latest images now add a "subcompatible id" that should fix the Windows 10 update issue. Oct 11 17:14:14 *facepalm* Oct 11 17:14:26 zmatt: the compatible ID was last year... Oct 11 17:14:37 the breakage was around the beginning of the year.... Oct 11 17:14:45 (hard to believe we let this issue go so long) Oct 11 17:14:58 it wasn't until the last few weeks we add the subcompatible ID. Oct 11 17:15:02 @jkrinder are you still looking at my screen sharing? Oct 11 17:15:34 jkridner|web: long live Windows Oct 11 17:15:38 in parallel, I also resigned the driver installer, which end-runs the issue by making driver selection based on VID/PID instead of the compatible/subcompatible OS descriptors. Oct 11 17:15:47 rd: your screen is full of the chat. Oct 11 17:15:59 jkridner|web: why? if the subcompatible thing works Oct 11 17:16:56 it's still fucking insane it's so hard to tell Windows to use its own fucking driver for microsoft's own fucking proprietary network protocol they implemented instead of the actual USB standard that every other OS implements Oct 11 17:17:07 because some people expect the driver installer to work anyway. :-( Oct 11 17:17:08 >< Oct 11 17:17:23 jkridner|web: just remove it, don't even mention it Oct 11 17:17:37 don't let people install a "driver" they don't need Oct 11 17:17:43 zmatt: you have no idea. we'd need to share a beer some time to discuss just HOW fucking insane. Oct 11 17:18:09 Beer....when and where? Oct 11 17:19:11 :-D Oct 11 17:19:28 * jkridner|web won't admit to opening his first beer of the day already. Oct 11 17:19:32 https://patentimages.storage.googleapis.com/c0/a9/53/29346d2929d847/US20060069789A1.pdf Oct 11 17:19:56 rd: so, on every other OS, the usb device says "I'm an usb ethernet device" and the OS says "okay!". on windows it says "I have no idea what that is" and instead you have to implement microsoft's own protocol for usb ethernet, that's similar to the standard but different. and then you still have a really hard time to somehow convince Windows to actually use Microsoft's own driver for their own ... Oct 11 17:20:02 ...protocol Oct 11 17:21:05 rd: so when you "install" a "driver", what you're actually installing is a signed text file telling windows to use its own driver Oct 11 17:21:14 -.- Oct 11 17:21:37 that's scrooey. Oct 11 17:21:47 that's mickeysloth. Oct 11 17:22:40 @zmatt I assumed that Windows has an electronic vault of drivers it approved/vetted and the installation process is just telling your PC which one to go retrieve. Oct 11 17:23:04 it does, but in this case the driver is Microsoft's own driver Oct 11 17:23:07 rd: it is kinda like that. Oct 11 17:23:27 rd: but they require some non-standard descriptors to trigger the trust. Oct 11 17:23:29 I just want to make my motor turn. Oct 11 17:23:38 hehehe Oct 11 17:23:42 maybe trust is the wrong word. Oct 11 17:23:49 time to run ubuntu in a VM with usb pass-through? ;) Oct 11 17:24:04 gah! dual boot! Oct 11 17:24:15 USB stacks on VMs are so hit-or-miss. Oct 11 17:24:17 dual boot sucks for most users I think Oct 11 17:24:30 I "trust" that if I look at https://beagleboard.org/latest-images I can feel a bit overwhelmed at first glance. Oct 11 17:24:41 rd: that page could be clearer yes Oct 11 17:24:53 nube Oct 11 17:24:53 making all the product names links to the product pages is definitely not helpful Oct 11 17:24:58 Nube Oct 11 17:25:18 a beer would probably make all this sales reachout easier :grimace: Oct 11 17:25:24 but ultimately what you need is the just the first download link Oct 11 17:25:33 patches welcome: https://github.com/beagleboard/beaglebone-getting-started Oct 11 17:25:40 I am asking for a "login" so that when I am trying to find information, I can pick up where I left off. Oct 11 17:26:38 @zmatt But I want something that lean "Graphical Desktop" I do not know how to code. Oct 11 17:26:55 rd: I'm find with setting some cookies as long as we have an easy way to reset them on every page they are used. Oct 11 17:26:57 "Graphical Desktop" is useless, the blue has no display output Oct 11 17:27:14 ? Oct 11 17:27:16 so using that image will just result in a slower system Oct 11 17:27:21 without any benefit Oct 11 17:27:22 cookie != login, login == keeping user data Oct 11 17:27:32 (older image, slower system, less free disk space) Oct 11 17:27:49 rd: in practice you almost always want the IoT image Oct 11 17:28:05 for both the older beaglebone and even more so for the blue Oct 11 17:28:19 zmatt: indeed. I want to kill the LXQT, but people unreasonably expect to use a GUI on Black. Oct 11 17:28:34 jkridner|web: it shouldn't be under recommended images honestly Oct 11 17:28:37 doh Oct 11 17:28:57 zmatt: but, but, that's what we put on the shipping boards based on user expectations. Oct 11 17:29:00 I really need an invite for that beagleslack? Oct 11 17:29:14 _i will soon solveeee the gui issue_ Oct 11 17:29:18 jkridner|web: recommended images should just have three big fat download links: one for beaglebone and pocketbeagle, one for x15, one for the old beagleboards Oct 11 17:29:19 and damn, have you been a chatty bunch while I was away Oct 11 17:29:20 i just have a yob Oct 11 17:29:34 @zmatt looking at the page, there is no way for me to sort the pros and cons. I bought a 32 Gb Sd card so I would have plenty of room Oct 11 17:29:37 jkridner|web: the download link should not be small print Oct 11 17:29:46 (like it is now) Oct 11 17:30:28 rd: I'm not sure if the image auto-expands to span the entire card Oct 11 17:30:29 jkridner|web: I think the expectation comes from the HDMI output ;) Oct 11 17:30:36 zmatt: want me to blow up the font size of the download link? Oct 11 17:30:51 and from the Raspi folks of course Oct 11 17:30:53 thinkfat_: well, the HDMI would still show a login screen without LXQT Oct 11 17:31:11 jkridner|web: as rd's experience shows, for some people it's just a giant sea of links and it's not clear what you should click right now Oct 11 17:31:25 I don't understand why Qt doesn't treat framebuffer output as a first-class citizen anymore. Oct 11 17:31:30 @zmatt amen Oct 11 17:31:39 jkridner|web: what do you mean? Oct 11 17:31:41 jkridner|web: it doesn't? Oct 11 17:31:45 Qt on framebuffer is really the way to go. Oct 11 17:31:47 linuxfb qpa works fine for me Oct 11 17:31:50 I don't think it does. Oct 11 17:31:52 last time I used it Oct 11 17:31:59 the documentation is deeply buried on their site. Oct 11 17:32:01 what is the first class then, eglfs? minimalegl? Oct 11 17:32:12 X Oct 11 17:32:15 X Oct 11 17:32:16 really Oct 11 17:32:18 lol Oct 11 17:32:20 no Oct 11 17:32:34 we currently use eglfs, but linuxfb also worked well for us (albeit without double-buffering) Oct 11 17:32:40 not wayland, not egl. at least that is what I saw last. Oct 11 17:32:40 no uxmux is the hype Oct 11 17:32:54 lol, really? Oct 11 17:32:58 it would be nice if they had a drm backend that didn't require EGL Oct 11 17:33:04 true dat Oct 11 17:33:09 I think uxmux is pretty cool, but I don't think the userbase is very high. Oct 11 17:33:13 haha Oct 11 17:33:14 the "framebuffer" is getting kinda old Oct 11 17:33:22 @zmatt I have had 12 pages of Beagleboard and things like Putty open at the same time. It seams to loop back on itself. Heaven forbid you click a link next to the one you intended from the list on https://beagleboard.org/latest-images Oct 11 17:33:26 new technology is new Oct 11 17:33:28 ayjay_t: are you AJ who introduced me to uxmux that you created? Oct 11 17:33:34 yes sir Oct 11 17:33:35 ayjay_t: I looked at your mouse handling. I wish I could unsee it ;P Oct 11 17:33:38 ah! Oct 11 17:33:46 I was wondering how that came out of left field. Oct 11 17:34:00 * ayjay_t is the real deal Oct 11 17:34:07 I love the idea of a lightweight framebuffer UI that is rendered HTML. Oct 11 17:34:15 I'm not a QML fan. Oct 11 17:34:28 just straight HTML is the bomb. ;-) Oct 11 17:34:31 but, you need to add JS. Oct 11 17:34:31 rd: the page makes sense to me, but I can easily understand how it could be confusing Oct 11 17:34:33 I find qml rather slick Oct 11 17:34:39 i'm totally gonna refactor it and probably build in an event loop Oct 11 17:34:44 probably event loop second stage Oct 11 17:34:50 and i'm planning on doing it today and streaming it! Oct 11 17:34:53 twitch and youtube! Oct 11 17:34:55 BUT I'VE BEEN SICK Oct 11 17:34:58 and probably better suited to UI work than HTML ;) Oct 11 17:35:12 thinkfat_: slick yes, but for me to be a fan I need to be productive in the 5 minute development chunks I have. Oct 11 17:35:21 zero-install tools FTW! Oct 11 17:35:56 well, all depends on personal taste and constraints Oct 11 17:35:57 thinkfat_: and, indeed, I'm not a UI guy. Oct 11 17:36:04 me neither Oct 11 17:36:11 QML is requires less to make some thing "pretty" Oct 11 17:36:27 except their dev tool. :-) Oct 11 17:36:34 you mean, creator? Oct 11 17:36:40 yes Oct 11 17:36:55 couldn't remember the name. Oct 11 17:36:57 never used it in fact, but some of my colleagues are rather fond of it Oct 11 17:37:08 @thinkfat_ I have probably been the reason for some of the chatter. I am a nube. Been asking some questions of @jkridner on the Slack channel and it migrated over here because @zmatt is an extensive contributor. Oct 11 17:37:31 at least a little less bloaty than eclipse ;) Oct 11 17:37:39 ayjay_t: uxmux will get some real users, indeed, if you can get a few people enabled. Oct 11 17:37:58 thinkfat_: LoL, now, that is some kind of comparison! Oct 11 17:38:28 jkridner|web: yeah, the joke is on my, allright ;) Oct 11 17:38:45 jkridner|web: nevertheless, nobody rolls their own IDE any more, they all customize eclipse Oct 11 17:38:57 ayjay_t: as for your hardcoded input device paths, here's an example on finding input devices: https://pastebin.com/raw/nTNCGKej where udev++.h (https://pastebin.com/raw/vZnPu6m3) is a thin (and incomplete) libudev wrapper I hacked together to hide libudev's monstrously long function names ;-) Oct 11 17:38:59 * jkridner|web thinks of days and days wasted in conference rooms discussing CCS download size to get people to take-in how big of a deal the Eclipse download size was to on-boarding users. Oct 11 17:39:10 ah yes, those were the days Oct 11 17:39:18 now, dev.ti.com exists, for better or worse. Oct 11 17:39:52 yeah i think uxmux will be good too. i really want noobs to have a "node.js" experience with it tbh Oct 11 17:40:04 I find myself mostly visiting the parts catalogue of the TI site nowadays ;) Oct 11 17:40:08 like i want people to take the pocketbeagle/bbb with an LCD cape, front end web devs, non hardware people Oct 11 17:40:11 and be able to make something with it Oct 11 17:40:13 not so much the developer forum Oct 11 17:40:21 ayjay_t: gah! ++ features of C! Oct 11 17:40:35 ++ features of C? Oct 11 17:40:40 rd: your screen hasn't changed in like an hour. Oct 11 17:41:25 yes, your udev++ stuff. kinda cool, but "for( auto syspath : enumerate ) " Oct 11 17:41:26 @jkridner I did not know you were looking. I was going to walk through it. Oct 11 17:41:29 jkridner|web: can you drop me an invite to the slack beagle? I'd like to have a look Oct 11 17:41:30 who even knows what that does?!? Oct 11 17:41:56 jkridner|web: after it put the kids to bed, that is Oct 11 17:42:01 jkridner|web: same thing as for( let syspath of enumerate ) in javascript Oct 11 17:42:04 rd: you ask, I look for about 1/2 second and then poll every 360 seconds or so. Oct 11 17:42:16 oh that's zmatts? i'm not really into how inputs were done tbh Oct 11 17:42:32 uxmux is basically my concept/architecture- i picked all the libraries, basic features, etc Oct 11 17:42:39 jkridner|web: how is iterating over a collection unclear? Oct 11 17:42:42 keelin built it- it was his "test projecT" when he was coming to work for me Oct 11 17:42:58 zmatt: in everything you just said there. Oct 11 17:43:33 zmatt: I get about as far in c++ as classes/constructors/destructors Oct 11 17:43:37 yeah i mean, zmatt works with pretty esoteric syntax pretty comfortably Oct 11 17:43:44 i'm more a "keep it simple" Oct 11 17:43:54 so i'm with jkridner|web Oct 11 17:44:00 *shrug* Oct 11 17:44:09 @jkridner Oh yeah, Now I remember, the Debian image is the same color as all the board options. I kept clicking on the link to the Beaglebone blue and getting lost. Oct 11 17:44:12 I really like the for( auto var : container ) syntax Oct 11 17:44:25 versus for( auto var = container.begin(); var != container.end(); ++var ) Oct 11 17:44:29 * jkridner|web is all into tables of function pointers when it comes to object oriented programming. Oct 11 17:44:31 which is what it's equivalent to Oct 11 17:44:38 (mostly) Oct 11 17:44:46 zmatt: now, THAT I can read. Oct 11 17:45:12 jkridner|web: but now that you know the for( var : expr ) syntax, don't you think it's a lot cleaner? Oct 11 17:45:13 except for 'auto'... which I assume is breaking all that is good with strictly typed lanugages. Oct 11 17:45:24 isn't that redundant, actually? Oct 11 17:45:28 i thought all variables were implied auto Oct 11 17:45:39 auto meaning "on the stack" or w/e Oct 11 17:45:49 auto just means "c'mon compiler, you know the effing type, don't make me repeat it" Oct 11 17:45:54 ayjay_t: but how would you know it is a declaration without it? Oct 11 17:46:08 ayjay_t is referring to old C's usage of the auto keyword Oct 11 17:46:17 where it was a completely pointless storage class Oct 11 17:46:19 C has 'auto'? Oct 11 17:46:21 yep Oct 11 17:46:27 @jkridner How do I change the color of the Debian hyperlinks to Blue and add a carriage return for the list of boards? Oct 11 17:46:28 2011? Oct 11 17:46:28 yeah like i said, i think its implied Oct 11 17:46:38 198* Oct 11 17:46:49 a variable thats not extern or static is implied auto Oct 11 17:46:55 ^ Oct 11 17:46:56 "Debian" doesn't have hyperlinks. Do you mean something on latest-images webpage? Oct 11 17:47:11 but in c++ it means derive type from rvalue? Oct 11 17:47:17 in C++ they replaced that useless storage class by a type specified instead Oct 11 17:47:18 yes Oct 11 17:47:21 and that type could change at runtime or? Oct 11 17:47:25 no Oct 11 17:47:25 @jkridner https://beagleboard.org/latest-images Oct 11 17:47:26 okay Oct 11 17:47:28 so its like go Oct 11 17:47:29 zmatt: really?!? guess I lose whatever C language cred I thought I had. Oct 11 17:47:44 golang is all about that implying type Oct 11 17:47:49 jkridner|web: in C it means "automatically allocated", as in allocated on stack Oct 11 17:47:58 zmatt: any feedback on https://pastebin.com/QSSGaumR ? Oct 11 17:48:10 ayjay_t: in C it can be especially nice when you have atrociously long or complicated types to deal with, like iterators Oct 11 17:48:27 oh I just realizes I misexplained the for( var : expr ) syntax, I'm so sorry! Oct 11 17:49:01 it means: for( auto i = expr.begin(); i != expr.end(); ++i ) { var = *i; ... } Oct 11 17:49:15 alt_init=flasher Oct 11 17:49:18 (where var is a variable declaration) Oct 11 17:49:32 zmatt: I'd assumed the assignment. Oct 11 17:50:01 still think you are missing a detail, but I got the point. Oct 11 17:50:12 and, yes, I really didn't get that it was an iterator. Oct 11 17:50:25 I mean, yes, it is a for loop... Oct 11 17:50:26 yeah, for example expr.end() is computed once and cached, not computed every loop iteration Oct 11 17:50:32 details like that Oct 11 17:50:45 also expr doesn't have to have begin() and end() methods Oct 11 17:50:47 but, I didn't know if the iterator was something defined within the class. Oct 11 17:51:10 it also works if begin(expr) and end(expr) are valid Oct 11 17:51:19 or if expr is a c-style array Oct 11 17:51:24 the methods used for defining the iterator are totally implied and I wouldn't know where to find them Oct 11 17:51:46 begin() and end() are the standard Oct 11 17:52:29 k, again, that was non-obvious for me and I would have used 'auto' as the search term, which would have been useless. Oct 11 17:52:37 https://pastebin.com/raw/ZPqb3hqA works too Oct 11 17:52:42 well, I'd've learned that was part of C syntax. :-) Oct 11 17:52:47 (ditto with auto x instead of int x) Oct 11 17:53:09 nah, googling 'c++ auto' would find you the right thing Oct 11 17:53:59 in this case I could have written for( char const *syspath : enumerate ) but I'm lazy and like to use auto Oct 11 17:54:16 jkridner|web: thoughts on macrofab or tempo automation? i'm trying to reach out ot them to find out if there is a market for my product Oct 11 17:54:24 so nervis doing all this reach out Oct 11 17:54:31 there's also a school of thought that feels you should use "Almost Always Auto" Oct 11 17:54:42 not sure if I quite agree with it, but it's there Oct 11 17:55:28 c++ has such a larger technical surface area than c Oct 11 17:55:33 @jkridner do I just cut and paste the Debian image onto the Micro SD card and insert it into the Beaglebone Blue? Oct 11 17:55:38 c++ is a fucking monster Oct 11 17:55:41 yeah its too big Oct 11 17:55:46 i mean it. Oct 11 17:55:50 also heavily burdened by legacy shit Oct 11 17:58:43 rd, zmatt: regarding fixing the links on https://beagleboard.org/latest-images , what else do you need besides making the actual links larger? Blue? Underlined? Oct 11 17:59:28 just make it clear what's the actual link. right now there's a big heading containing a link that doesn't yield the download, and then a line of smallprint containing a ton of links, one of which is a the download Oct 11 17:59:30 ayjay_t: I really like Circuithub. Just had a very positive experience with Macrofab for the "Nebula One" boards. Oct 11 18:00:25 ayjay_t: I know Jeff McAlvay at Tempo Automation and they've always ended up being too expensive for me. Oct 11 18:00:45 at one point, they were building pick-and-place machines using BeagleBone Blacks.... Oct 11 18:00:49 Make the Actual links blue and add a carriage return. Change the text to say this image has been tested for the following products. Oct 11 18:00:58 then investors had them pivot to a prototyping CM. Oct 11 18:01:20 guess I never got over being sad they didn't bring their Black-based PnP to market. Oct 11 18:01:30 I mean, Squink is *kinda* cool. Oct 11 18:01:40 and there are other Black-based PnP machines.... Oct 11 18:01:47 but, nothing is super practical. Oct 11 18:02:16 for a pnp? Oct 11 18:02:20 Just got to see Nat Sidele(sp?) of Sparkfun talk about practical cheap PnP machines at Open Hardware Summit (couple of weeks ago). Oct 11 18:02:53 tbf temp has a really weird marketing presence, i guess they are prototyping CM Oct 11 18:03:04 and got really excited about the MIT portable fab lab stuff. Oct 11 18:03:05 macrofab i feel like has a strong sell as "we can do whatever you need for assembly" Oct 11 18:03:20 ayjay_t: they are really flexible and affordable. Oct 11 18:03:33 yeah they are good like that, circuithub's a bit pricier i think Oct 11 18:03:39 Circuithub is really turn-key via the web interface and pretty afforable too. Oct 11 18:03:50 rd: you saying hello to me? Oct 11 18:03:50 moreso than macro? Oct 11 18:04:04 is macro more manual? Oct 11 18:04:14 rd: no need to format to use etcher.io on a microSD.... Oct 11 18:04:25 hmm maybe in part sourcing Oct 11 18:04:27 rd: I've seen a lot of people do that recently. etcher.io handles all microSD formatting. Oct 11 18:04:48 i personally feel really strongly about autopogo being important with QC but i'm still trying to reach out and find out from people if QC is a problem Oct 11 18:05:03 i feel like with all these people using macrofab and circuithub etc i mean, if everyone's coming to market i guess there is no problem? Oct 11 18:05:05 ayjay_t: in my experience, yes, but also more flexible from the get-go. CircuitHub is pretty flexible too, but more automated on the front-end. Oct 11 18:05:17 or are they doing QC with some other solution thats maybe more expensive? Oct 11 18:05:18 i don't know Oct 11 18:05:27 ayjay_t: both are amazing. Oct 11 18:05:45 I hear good things about Tempo too, but I've not had anything that really fit. Oct 11 18:05:48 fab lab: http://www.fabfoundation.org/index.php/mobile-fab-lab/index.html Oct 11 18:05:53 I feel they are better for "I want to pay to get it fast" Oct 11 18:06:02 i see Oct 11 18:06:14 and, I work with other manufacturers that give me more balance than that. GHI and Avnet in particular. Oct 11 18:06:19 GHI for "I need it fast" Oct 11 18:06:29 Avnet gives me scale. Oct 11 18:06:54 @jkridner the website is not pointing me to where I need to go. I am just fumbling around in the dark. I reformatted just to erase what was there and try to get back to a known good situation. Oct 11 18:07:24 rd: which website? etcher.io? beagleboard.org/getting-started? beagleboard.org/latest-images? Oct 11 18:07:48 rd: did you not get the idea you needed to download Etcher via https://etcher.io ? Oct 11 18:08:37 gosh Oct 11 18:10:20 etcher.io is not mentioned on the https://beagleboard.org/latest-images page. I am going back to the book and try and find a known good situation. I am getting hungry and cranky now because I never stopped to eat this morning. Oct 11 18:11:23 rd, zmatt: does the first image link on https://beagleboard.org/latest-images look good now? Oct 11 18:11:44 rd: k, chat again tomorrow? Oct 11 18:12:31 @jkridner oh hell yea. Can you add a carriage return? Oct 11 18:14:27 rd: k Oct 11 18:14:41 @jkridner you might consider routing people to the correct image by asking what board they are working with at the beginning of this page. https://beagleboard.org/getting-started#update Oct 11 18:14:46 reverted for a minute while I fix. Oct 11 18:21:12 looks like andrew seddon is a ycombinator founder too Oct 11 18:21:29 i'll poop my pants if i get funded Oct 11 18:45:46 jkridner|web: circuithub is kind of affordable if you do, like, 100's of boards Oct 11 18:47:12 as opposed to more or less thinkfat_ ? Oct 11 18:47:18 jkridner|web: i put a project up there that I developed, really like how you can select and approve parts for each BOM item Oct 11 18:48:25 I have installed quartus 13.0.1 sp1 web edition since a while. Now I wanted to use ModelSim the first time but it is complaining about a missing licence file. But I can not find a file like license.dat. Do you know something about that problem? Oct 11 18:48:48 I can not use a higher Quartus version because I need support for Cyclone II. Oct 11 18:49:00 are you sure you're in the right channel? Oct 11 18:49:02 NTQ are you in the right chna.. Oct 11 18:49:03 yeah Oct 11 18:49:23 of course not xD Oct 11 18:49:24 oops Oct 11 18:49:37 ayjay_t: I found that once you cross the 100, assembly cost roughly equals component cost, if you stick to surface mount stuff only Oct 11 18:49:59 ayjay_t: and the ratio gets better with each 100 boards you add Oct 11 18:50:20 interesting Oct 11 18:50:21 through-hole costs a premium Oct 11 18:50:40 so my business model requires custom boards in 10, 20 board runs for each customer Oct 11 18:50:52 and all assembly is with the same components but in different combinations XD Oct 11 18:51:05 so i'd need my own pnp Oct 11 18:51:12 with order integration Oct 11 18:51:34 ayjay_t: how many components for each board? Oct 11 18:51:43 unique or total? Oct 11 18:51:52 ayjay_t: total Oct 11 18:51:53 probably no more than 15 unqiue max Oct 11 18:52:02 but it could be like 50-100 total Oct 11 18:52:14 it depends on the order Oct 11 18:52:59 small count of unique parts is always good because pnps have limited number of feeders Oct 11 18:53:40 yeah i actually think the business model is feasable tbh Oct 11 18:53:52 but i'm pretty much relying on china for cost effective small-quantity 2 layer/4 layer board runs Oct 11 18:54:11 i'm not that intimidated by my own pnp but i could never make my own boards Oct 11 18:54:46 if you have 100 parts to place, the machine better be quick Oct 11 18:54:57 not expecting high volume tho Oct 11 18:55:11 you're not wrong though Oct 11 18:55:43 I was thinking along the lines of "liteplacer", but that is really not a fast machine Oct 11 18:55:44 what do you think a "long estimate" is, or "too much time" for a total of 500 parts, like 100 components on 5 boards Oct 11 18:56:03 I have no idea, really. Oct 11 18:56:19 I'm just beginning to think about the manufacturing side of things myself Oct 11 18:57:08 i'm trying to sell everyone on using me to make bed-of-nails in-circuit testers with the sitara/beaglebone as a backend Oct 11 18:57:14 I guess it all depends on how much time you can allow to completely manufacture a board Oct 11 18:57:18 its exhausting :-( Oct 11 18:57:22 yeah Oct 11 18:57:23 i don' tknow Oct 11 18:57:29 i guess having a lot of orders is a good problem to have? Oct 11 18:57:40 if you can keep the pace Oct 11 18:58:03 if you cannot saturate the channel it's money lost Oct 11 18:59:33 I think for my kind of volume I'm good with cheap pcbs, quick hands and a reflow oven ;) Oct 11 18:59:44 _maybe_ Oct 11 18:59:47 thats what i do now Oct 11 18:59:53 QC is really an issue in that case tho, no? Oct 11 18:59:57 i mean, even more so than in pnp Oct 11 19:00:54 I'm thinking of batches of 10s or 20s, not more Oct 11 19:01:38 and they're all the same, not built-to-order Oct 11 19:03:40 yeah, 20 is ok but its not fun Oct 11 19:03:48 depends on how many/size of components Oct 11 19:03:54 placing components isn't fun, in my mind Oct 11 19:03:57 good tools make it a lot better Oct 11 19:04:02 no, fun is something else entirely Oct 11 19:04:29 i have dental instruments, all kinds of chemicals (flux, paste), different sized solder beads, decent sodlering station, special tweezers, i'm good with wick Oct 11 19:04:31 the fun is doing the design, then the first couple of boards till you see that it works as expected Oct 11 19:04:39 tbh i hate assembly Oct 11 19:05:02 i can spend 60 hours on design, 150 hours on programming, and i know its only 10 hours of assembly to get the prototype to bridge these things, and i'm curing my mother the whole time Oct 11 19:05:16 cursing* Oct 11 19:05:22 for your two left hands Oct 11 19:05:33 and 10 thumbs Oct 11 19:06:55 I kind of like assembly, I like the manual labour, but ti's best kept in small doses ;) Oct 11 19:07:45 i've built some stupid pcbs Oct 11 19:07:58 i made my own chameleon mini Oct 11 19:08:31 it's a 13.56 NF modem w/ OOK and PSK in discrete parts driven by an atmel :-( Oct 11 19:08:38 it was brutal Oct 11 19:09:18 i made a couple of them Oct 11 19:09:19 i hated it Oct 11 19:09:36 tough stuff Oct 11 19:10:37 its actually an interesting open source project Oct 11 19:10:45 it's very easy to follow how the modem works Oct 11 19:10:56 i mean, if your comfortable with digital electronics Oct 11 19:11:32 well, even the analog stuff nowadays is frequently digital ;) Oct 11 19:13:19 hmm yah Oct 11 19:54:06 on the other hand, high-speed digital is analog stuff ;) Oct 11 20:02:20 zzz! Oct 11 20:03:15 Hey...I checked online at bbb.io. I went to the latest images section. Oct 11 20:03:37 Someone made a bright BLUE in the regular, discreet black! Oct 11 20:04:08 Have people been complaining about not finding black on a screen? Oct 11 20:05:09 I was just wondering why people went through such a hassle...end here. Oct 11 20:05:21 LOL, I hadn't looked yet at what jkridner had done... it's definitely hard to miss the download links now. it's also hideous, but hey Oct 11 20:05:31 Hhahahah. Oct 11 20:05:43 I did not want to type it but yep, hideous. Oct 11 20:06:07 I saw you guys discussing it earlier. Oct 11 20:06:12 I checked it out. Oct 11 20:06:29 BLUE and BLUE again! Oct 11 20:06:56 contrast and the BBB! Oct 11 20:07:54 this post by Linus Torvalds suddenly comes to mind... https://lwn.net/Articles/190241/ Oct 11 20:08:19 I saw one online about him. He took a break from Linux b/c he was angry. Oct 11 20:08:29 He kept treating people poorly. Oct 11 20:08:40 * zmatt shrugs Oct 11 20:08:47 zmatt knows! Oct 11 20:08:50 Hhahaha. Oct 11 20:09:00 I very obliquely picked up on it but I haven't really followed the issue Oct 11 20:09:28 Me neither. Is he back yet? I wanted to find out but I do not see the info. Oct 11 20:09:31 Oh well. Oct 11 20:10:14 gregkh will be doing the upcoming 4.19 release, at the very least, after that, who knows Oct 11 20:10:35 Ut oh! Oct 11 20:11:00 I guess people need breaks. Who-d a thought it? Oct 11 20:11:14 * vagrantc suspects a good deal more is needed Oct 11 20:11:50 Hey vagrantc: Is that b/c of the millions or billions of people pestering? Oct 11 20:12:26 * vagrantc suspects it's because of https://en.wikipedia.org/wiki/Founder%27s_syndrome amoung other things Oct 11 20:12:58 Funny and oh-so-real. Oct 11 20:13:16 Like that one fellow, people have issues w/ Linux, life, and the BBB. Oct 11 20:13:55 hint, hint. Anyway, maybe he can start to trust outer elements in his journey to acceptance of chairs. Oct 11 20:14:15 And yes. Chairmen or chairladies. Oct 11 20:14:48 You know...like board members but people that have been accepted. Oct 11 20:14:54 Not hired! Oct 11 20:16:30 https://www.zdnet.com/article/linus-torvalds-takes-a-break-from-linux/ is what I found. Oct 11 20:23:49 zmatt: I guess if someone read that forks were entering their eyes over software, they would be disturbed a bit. Maybe. Oct 11 20:24:38 Oh well. Oct 11 20:24:40 Break time! Oct 11 20:48:39 can someone suggest any resource like article,book or video or whole course to start with linux kernel development Oct 11 20:49:28 how many years of experience you have with C programming on linux in general? Oct 11 20:50:02 yeah i am the guy from yesterday Oct 11 20:50:15 gsoc guy Oct 11 20:50:33 so still trying to get started Oct 11 20:50:38 why do you come here asking for advice if you're going to ignore it anyway? Oct 11 20:51:14 based on your background, you received the advice to simply avoid picking a project that has kernel programming as dependency Oct 11 20:51:30 i m not ignoring was wanted to get a something to start Oct 11 20:51:40 from which i ca build upon Oct 11 20:52:29 dude if that's was a option i would have done that Oct 11 20:52:34 believe me Oct 11 20:53:06 i get it u guys are pro but u all once were newbie Oct 11 20:53:50 u done a lot of hardwork worked different ways to get to this lvl Oct 11 20:54:54 now u are here maybe u can help someone else to get what they wanted Oct 11 20:55:31 but if u dont want to i get Oct 11 20:55:52 i guess i wil have to find another wy Oct 11 20:56:01 I'm browsing the gsoc ideas list... most of them do not involve kernel programming Oct 11 20:56:13 i get this from me i will not stop until i get want i want Oct 11 20:56:21 good for you Oct 11 20:56:33 dude there is certain hardware requirements Oct 11 20:56:48 the one which i qualify need it Oct 11 20:57:12 so dont judge ppl from your side Oct 11 20:57:55 i guess pro do get cocky overtime Oct 11 20:58:01 sad to see Oct 11 20:58:07 lol Oct 11 20:58:29 the problem btw isn't that you're "noob", the problem is that you seem to lack autonomous learning skills Oct 11 20:58:34 anyone out there had much experience getting the Mikro cape for click boards working with their LTE click board and GPS Click board...struggling to make them work, but I am very much a newbie Oct 11 20:58:41 e.g. if you want to get started with linux kernel development, http://lmgtfy.com/?q=linux+kernel+development+getting+started Oct 11 20:58:45 plenty of resources Oct 11 20:58:57 you won't even need a beaglebone for it Oct 11 20:59:23 oh, he left Oct 11 21:00:55 lmgtfy <- lol Oct 11 21:01:10 CoffeeBreakfast_: never seen it before? Oct 11 21:01:39 never.. Oct 11 21:04:59 zmatt: Do you remember my project for an arb. waveform generator with bbb -> spi -> ad5676? Oct 11 21:05:28 150KSa/s Oct 11 21:06:07 I vaguely recall one or more people doing something with pru and ad Oct 11 23:19:45 CoffeeBreakfast_: did you have something to ask or were you just testing my memory? because in the latter I can also simply tell you it's not great ;P Oct 11 23:27:10 Hey, so I want to try some USB sniffing with my BBB, anyways, googling shows some tools to try, and they all say do something along the lines of `mount -t gadgetfs gadgetfs /mnt/test-usb/` or `mount -t gadgetfs none /mnt/test-usb/` to start the USB device stuff, with the latest debian (well, took a new BBB and did apt-get update/upgrade) both of those mount commands say there is no such thing as gadgetfs, even after modprobing gadgetfs Oct 11 23:27:22 is there a trick to get gadgetfs working on BBB? Oct 11 23:27:54 hmm, I've vaguely heard of gadgetfs Oct 11 23:28:32 not sure what you mean by "usb sniffing" though. the bbb can't do anything I'd call by that name Oct 11 23:29:33 it says here that gadgetfs was replaced by functionfs Oct 11 23:29:46 which is a name that sounds more familiar to me Oct 11 23:30:51 gadgetfs should still work too though Oct 11 23:30:53 it has a host and device USB port from what everything says, there are tools that will pass through data between the two so you can sniff USB with it, lots of tools to do it, but first I need to get the device port working... Oct 11 23:30:55 both are compiled as modules Oct 11 23:31:04 yea, it modprobes fine Oct 11 23:31:05 ah like that Oct 11 23:31:07 just doesn't actually work Oct 11 23:31:49 I'm wondering if the kernel config is missing the driver... Oct 11 23:32:00 I did a menuconfig and none of the gadgetfs drivers are listed... Oct 11 23:32:25 I don't know where or how you did menuconfig, but it's definitely included, I just checked Oct 11 23:32:30 also, otherwise your modprobe would have failed Oct 11 23:32:43 wait what do you mean? Oct 11 23:33:09 "none of the gadgetfs drivers" Oct 11 23:33:13 gadgetfs is a driver Oct 11 23:33:38 yes, one second, I'll find the config names I'm unsure of Oct 11 23:34:38 nearly all gadget-related stuff seems to be enabled Oct 11 23:35:15 there are no checked "USB Peripheral Controller" Oct 11 23:37:28 not sure if I need that checked... Oct 11 23:37:57 probably because musb is a dual-role controller (host/device) Oct 11 23:38:45 it's listed in the main usb menu ("Inventra Highspeed Dual Role Controller") Oct 11 23:39:24 ahhh Oct 11 23:40:54 maybe gadgetfs is poorly maintained, given that it has a replacement that was already introduced in 2010 Oct 11 23:41:33 (functionfs) Oct 11 23:42:02 yea, that's what I'm thinking... Oct 11 23:42:28 that's just a guess though, I don't actually know Oct 11 23:44:54 see also https://github.com/gadgetd/gadgetd/wiki/Motivation for some information about the current state of usb gadgets in linux userspace Oct 11 23:46:06 that project looks dead though Oct 11 23:47:09 I can confirm I'm getting the gadgetfs error Oct 11 23:48:00 yea, I'm looking, lots of dead projects for this :/ Oct 11 23:51:15 hmm, try preventing the setup of the default usb gadget Oct 11 23:51:44 I found some old mailing list post that seems to indicate that if another usb gadget is already active, gadgetfs can't be mounted Oct 11 23:51:53 ohhh Oct 11 23:51:58 hrm... Oct 11 23:52:12 I'm assuming you're connecting to the beaglebone via ethernet? Oct 11 23:52:22 (and not usb networking :P ) Oct 11 23:53:30 ethernet, but powering it off my desktop USB, and my desktop does say it's presenting itself as a device... Oct 11 23:53:32 I think that's it Oct 11 23:53:45 there's some service named "generic-board-something" or something, check /etc/systemd/system/generic* Oct 11 23:54:03 use sudo systemctl disable generic-board-whatever to disable it and reboot Oct 11 23:54:10 brb Oct 12 00:08:52 yup, that does work, just need to `rmmod g_ether` to get gadgetfs working Oct 12 00:09:05 huh, g_ether was modprobed? Oct 12 00:09:09 what image are you using? Oct 12 00:09:50 because g_ether is not used in any standard beaglebone image Oct 12 00:10:55 I dunno, I bought it, plugged it in, did apt-get update/upgrade, and removed/added a handful of packages Oct 12 00:11:09 cat /etc/dogtag Oct 12 00:11:18 I dug around, it's modprobed by /opt/scripts/boot/am335x_evm.sh which seems to be a boot script Oct 12 00:11:29 BeagleBoard.org Debian Image 2016-01-24 Oct 12 00:11:40 holy shit that's a museumpiece Oct 12 00:12:13 lol Oct 12 00:12:40 since it sounds like you're comfortable with a debian system, you may just want to reflash it to the latest stretch-console image Oct 12 00:13:24 https://elinux.org/Beagleboard:BeagleBoneBlack_Debian#Stretch_Snapshot_console Oct 12 00:14:32 yea, I want to, but I'm out of SD cards, last card fell inside the SD card reader built into my desktop, and I don't really feel like disassembling it Oct 12 00:15:02 so I think I need to buy more SD cards Oct 12 00:15:17 you can try to see if this project works: https://github.com/ravikp7/node-beagle-boot Oct 12 00:15:50 it seems to want to replace the old brittle bbblfs tool Oct 12 00:16:20 now that will work Oct 12 00:17:04 since your beaglebone is an oldie, bbblfs might work too. in that case, just don't use its scripts, just use the "usb_flasher" tool (which doesn't flash anything but just boots the bbb into usb mass storage mode) Oct 12 00:17:41 once it appears as a block device I'm assuming you know how to take it from there Oct 12 00:18:57 yea Oct 12 00:24:31 if dd'ing an image to it, don't forget to also grow the partition ( echo ',+' | sfdisk -N1 $dev ) and filesystem ( e2fsck -f $partdev && resize2fs $partdev ) to span the whole device Oct 12 00:25:11 (that sfdisk syntax is so clear and intuitive :P ) Oct 12 01:20:00 meh i ended up reaching out to all the mfrs Oct 12 01:39:01 hello? Oct 12 01:40:03 what a strange question. have you tried googling "hello" to see if you get any useful results? Oct 12 01:43:44 ok bye Oct 12 02:34:32 rofl why zmatt Oct 12 02:34:43 why punch the noobs Oct 12 02:43:05 given how much rd was involved earlier in the discussion about communication, I had expected better from him by now :P Oct 12 02:43:41 didn't intend to scare him away Oct 12 02:49:13 cold logic often looks like noob abuse. Oct 12 02:49:54 but it's not....but dig deal. Oct 12 02:52:26 well I kinda expected some sort of reaction from him, allowing me to clarify that "hello?" is a weird thing to say here (except if you're in active conversation with someone and they suddenly fall silent) Oct 12 02:54:47 too often, people put a ? mark at the end of a statement that is not a question. Oct 12 02:55:05 it's a statement. Oct 12 02:55:25 hello. Oct 12 02:55:31 simple. Oct 12 02:55:54 "hello" is also not terribly useful to say on irc, but at least it doesn't sound like you expect a response (although people often do) Oct 12 02:56:36 also FUCK. I'm so glad we have two independent ways to apply updates to our devices, since I just discovered I fucked one of them the last time we used it Oct 12 02:56:57 (rushed updates)-- **** ENDING LOGGING AT Fri Oct 12 03:00:00 2018