**** BEGIN LOGGING AT Tue Nov 27 03:00:01 2018 Nov 27 03:05:12 o/ Nov 27 03:05:34 zmatt, get this Nov 27 03:05:54 I was trying to fire up the beaglebone to do some simple i2c poking on this little i2c oled unit Nov 27 03:06:26 so, once we discovered that (potentially) one of the dram cells is dead, I fired up my 1bitsy Nov 27 03:06:39 thinking I would use that to do the thing i was trying to do instead Nov 27 03:06:51 only, guess what? theres some kind of intermittent short to ground Nov 27 03:07:11 wait, you had external hardware connected to the beaglebone during all this? Nov 27 03:07:14 so, yesterday i discovered not one, but 2 separate boards were dead Nov 27 03:07:15 no no no Nov 27 03:07:18 oh ok Nov 27 03:07:28 i was just trying to get it working first before hooking it up to anything Nov 27 03:07:54 I just thought it was funny that I discovered 2 of my dev boards were dead in one day Nov 27 03:08:12 one of those "haha fuck me i guess" kinda days Nov 27 03:08:22 you have two boards with mysterious problems? maybe time to start being mindful of ESD ? :P Nov 27 03:08:40 its def not esd Nov 27 03:08:52 among other things, im in boston, its humid as all heck Nov 27 03:08:59 the air is practically conductive Nov 27 03:09:21 if anything oxidation would be more likely because of all the freaking moisture Nov 27 03:09:59 but, no, things just break around me. Thats why I got into cybersecurity afaict. I have a habit of breaking things all the time Nov 27 03:10:07 lol Nov 27 03:12:36 so, anyway, I have multiple options now but they come down to raw linux for arm on my stupid CHIP sbc, or using one of my fpga boards with a risc-v core to do this Nov 27 03:13:00 actually scratch that i also have a zynq board i have baremetal exec on so that would prolly be easier Nov 27 03:15:20 either way most of the reason im here rn is i just wanted to thank you zmatt Nov 27 03:15:37 appreciate all the time you put into helping me figure that out, even if we didnt get it resolved Nov 27 03:16:08 its a lot nicer to troubleshoot embedded problems when you have someone who knows the system really well to work with Nov 27 03:16:28 np Nov 27 03:16:43 I'm still curious what happened to your beaglebone Nov 27 03:16:55 I've never seen or heard of it failing in this way Nov 27 03:17:32 I mean, I could whip up an elf payload that tests the ram to confirm it is what it is, but its not surprising Nov 27 03:17:47 DRAM tends to fail like this, its why we have memtests Nov 27 03:17:59 to weed out the bad cells and create a blacklist of bad cells Nov 27 03:18:01 it's a bit odd for the SoC or ram chip to fail, though I guess it's not impossible... or maybe something happened to the pcb Nov 27 03:18:26 yeah, in my experience dram is one of the first things to fail, that or power supplies Nov 27 03:18:28 ehh, no that's almost never done with dram Nov 27 03:18:35 at least not in any system I've ever seen Nov 27 03:18:50 if any part of the dram fails, the device is defective and RMA'd Nov 27 03:19:22 I know dram *can* fail in this way, but I've never seen it happen Nov 27 03:19:36 Ive seen it happen at least 3 or 4 times Nov 27 03:19:52 admittedly one of those /was/ on a board a friend was trying to demo baremetal rowhammer on Nov 27 03:19:58 and there's definitely no way to compensate, remap, blacklist or whatever on any embedded system I've ever seen Nov 27 03:20:01 so thats not terribly surprising Nov 27 03:20:29 ECC (when supported, which is rare) could compensate I guess, although that's not really what it's intended for usually Nov 27 03:20:31 i had been told, and ive admittedly never verified, but i had been told that you can blacklist cells in linux Nov 27 03:20:51 or maybe using the mmu? Nov 27 03:20:53 you can probably tell linux to avoid using some part of ram Nov 27 03:21:27 i wonder if you could map a region of ram as nonread, nonwrite, nonexec in the mmu Nov 27 03:21:41 but you made it sound like it's normal practice, rather than something that requires customizing the DT to include a device-specific hack Nov 27 03:22:04 oh no its definitely not normal at all, most people definitely replace devices with this kind of failure Nov 27 03:22:23 ive just seen it done in spaces where a fast hack is needed Nov 27 03:22:40 btw, if you want to run baremetal code, I have a simple example in assembly (not because there's any good reason to use all-assembly, but someone asked for such an example a long time ago): https://github.com/mvduin/bbb-asm-demo Nov 27 03:22:47 actually, a good example of a similar issue can be found in a blog post from marcan42 from about... 6 months ago? Nov 27 03:22:53 it runs from internal SRAM Nov 27 03:23:18 though to be able to do a memory test, you'd still need to init the memory controller... so it's probably not extremely useful Nov 27 03:23:30 yuck, yeah Nov 27 03:23:40 that would also mean training dram, wouldnt it? Nov 27 03:23:44 nah Nov 27 03:23:57 it doesn't require more than writing a bunch of values to registers Nov 27 03:24:12 does the mmu auto-train? Nov 27 03:24:37 I think software leveling is used, i.e. values hardcoded in u-boot Nov 27 03:24:44 keep in mind it's only ddr3-800 Nov 27 03:24:49 oh yeah ok Nov 27 03:24:58 still needs to be trained though iirc, but less aggressively Nov 27 03:25:26 the zynq board i mentioned, when i was working on baremetal i had to figure out how to train that, it was terrible Nov 27 03:25:36 ended up borrowing code from xilinx Nov 27 03:25:50 its a mess to figure it out too, they do it all via microcode Nov 27 03:26:28 hardware leveling has proven to be buggy on a fair number of TI SoCs, so it makes sense they used manual leveling here, especially since it's really not all that critical yet at these low speeds Nov 27 03:27:21 here we go, xilinx's big nasty ram training blob Nov 27 03:27:22 https://github.com/hedgeberg/SnickerOS/blob/master/src/lib/ps7_init_gpl.c Nov 27 03:27:51 the good news is they document what each microcode op does pretty well, bad news is its a beast Nov 27 03:29:43 https://github.com/RobertCNelson/u-boot/blob/v2018.09/board/ti/am335x/board.c#L132-L137 I reaaaly don't understand why they're using constants defined in some other file named after the memory chip, considering these values are dependent mostly on the pcb layout Nov 27 03:30:18 yeah Nov 27 03:30:36 pcb layout and parasitic atmosphere capacitances, mostly Nov 27 03:32:00 but as far as ddr3 layouts go the beaglebone is as simple as it gets... both cmd and 16-bit data are point-to-point to a single memory chip. no branching, no fly-by, no multiple ranks Nov 27 03:32:58 sure but there has to be some length matching at least, right? Nov 27 03:33:10 sure Nov 27 03:33:41 yeah and that makes it unpredictable because of mutual capacitances and miller capacitance multipliers and blah blah blah and physics Nov 27 03:33:55 although ddr3-800... that's 1.25 ns per transfer Nov 27 03:34:25 whats time of flight from SoC bga to ram bga? Nov 27 03:35:15 I don't know the velocity factor. at lightspeed 1.25 ns would be 37.5 cm Nov 27 03:35:26 soo... yeah.... Nov 27 03:35:49 propagation times are normally pretty tiny compared to that Nov 27 03:35:57 copper on FR4 has a relative permitivitty of about uh... .75? Nov 27 03:36:44 so, epsilon for copper on fr4 comes out to about 2.3e8 Nov 27 03:36:54 wait no i beefed that up Nov 27 03:37:09 thats em propagation speed in copper on fr4 Nov 27 03:37:21 one sec while I remember why I forgot epsilon-0 Nov 27 03:37:48 right, why did i forget it was Farads/meter??? Nov 27 03:38:00 first google results says velocity factor on a pcb trace is about 0.5 c Nov 27 03:38:13 my degree was in semiconductors, im slipping. Nov 27 03:38:26 yeah, that sounds more right, depending on the pcb Nov 27 03:39:05 so yeah, 1.25 ns is give-or-take a bit less than 20 cm Nov 27 03:39:42 so, on the beaglebone the memory chip and SoC are basically lumped together Nov 27 03:40:13 pretty much Nov 27 03:40:44 yeah Nov 27 03:54:08 hedgeberg there's also pocketbeagles Nov 27 03:54:21 which are the same soc but w/o as many headers (no lcd) and i think... no ethernet Nov 27 03:54:43 there's an expander from microwavemont on tindie for ethernet Nov 27 03:59:54 i think if i were to buy a new beaglebone I'd prolly get a bbone black wireless Nov 27 03:59:56 I am having trouble with getting my BeagleBone Blue up and running. It appears to be a driver issue. The new driver helped but I am still stuck at the point where I get PuTTY to talk to my BeagleBone Blue. In one section it looks like the driver is working in another section it does not. I made some screen captures. I am trying to follow the BeagleBone Robotics Project book. This is a link to the screen captures I made. Any Nov 27 03:59:56 keywords shared or direction would be greatly appreciated. https://drive.google.com/file/d/1DFKhoKGgPd1cV2qvNMiwmpF-7XiC_Gys/view?usp=sharing Nov 27 04:00:01 also maybe a blue? Nov 27 04:00:36 RD_ can you open device manager? Nov 27 04:01:49 if so, take a screenshot with all the values collapsed and send it Nov 27 04:02:41 Not sure what you are referring to. I can go to Windows settings/devices/Bluetooth & other devices Nov 27 04:02:57 press windows key + r at the same time Nov 27 04:03:05 that will open a run dialog Nov 27 04:03:14 then type in "devmgmt.msc" and press enter Nov 27 04:03:19 that will be device manager Nov 27 04:03:22 Think I I did what your asking for in the link. Nov 27 04:03:25 no Nov 27 04:03:35 thats the windows metro device window Nov 27 04:03:45 but it doesnt give you nearly as much info as device manager Nov 27 04:04:14 Okay. Thank you. I am a manufacturing engineer so it will take me a bit. Nov 27 04:05:05 no worries, used to work IT way back. I'm used to walking people through the weird steps Nov 27 04:05:41 Ok I recognize that screen. Nov 27 04:06:16 yup, can you send a picture? Nov 27 04:07:36 Yes. Working on it. Is there an easier way than creating a file in Google Drive and making a link? Nov 27 04:07:57 yeah, you can upload to imgur Nov 27 04:08:08 https://imgur.com/upload Nov 27 04:08:29 https://drive.google.com/file/d/1U_bNMdOp6iIkszFWW5BcDLtAwcSXBHii/view?usp=sharing Nov 27 04:09:12 ok, so, see how in the first pic that you sent, in the titlebar it says "COM1"? Nov 27 04:09:21 Yes Nov 27 04:09:39 COM1 on modern windows isnt a real serial device Nov 27 04:09:50 its a virtual port for compatibility purposes Nov 27 04:10:06 LOL ....That's what the books is telling me to do...... Thank you. Nov 27 04:10:19 what that means is windows is trying to open a non-existent port Nov 27 04:10:37 look at the devices in device manager, see how they have (COM x) at the end? Nov 27 04:10:37 Well that sounds compatible..... Nov 27 04:10:44 Yes. Nov 27 04:11:01 those are the different ports you have Nov 27 04:11:14 other than COM1, those are all the possibilities for what the beaglebone is on Nov 27 04:11:27 LPT1 is a printer parallel port, so thats probably not it Nov 27 04:11:48 COM3 is for communicating with Intel AMT, so thats also probably not it Nov 27 04:11:51 The Gadget Serial (COM8) looks like it has an error on it. Nov 27 04:12:01 yeah, probably a driver issue Nov 27 04:12:13 do you have any other usb serial devices on that computer? Nov 27 04:12:24 or is the only thing plugged into it the beaglebone? Nov 27 04:12:35 Numerous usb devices Nov 27 04:12:50 ok, can you unplug everything except the beaglebone? Nov 27 04:13:18 (just for a bit, the right device is probably COM8 but this way we can be sure) Nov 27 04:14:31 I did the opposite of what you asked. ( I was already in the process) I unplugged the BBB and the Gadget Serial (COM 8) disappeared and came back when I plugged it back in. Nov 27 04:14:44 that works too Nov 27 04:14:52 :) Nov 27 04:14:57 does it still have an error after unplugging? Nov 27 04:15:04 Yes. Nov 27 04:15:13 so, its possible its not an issue since it enumerated Nov 27 04:15:22 so lets try opening putty again Nov 27 04:15:24 HMMMM Nov 27 04:15:36 and change the serial port from COM1 to COM8 Nov 27 04:15:53 enumerated meaning that it recognizes it well enough to put a number behind it? Nov 27 04:16:03 enumeration is a thing that usb devices do Nov 27 04:16:11 USB is weird, its like part network and part bus Nov 27 04:16:37 when you plug a device in, the host (your computer) has to talk to it enough to assign it an addrss and then get a bunch of details from it, decide what driver to use, etc Nov 27 04:16:40 thats enumeration Nov 27 04:17:05 so, its possible as is its functional enough for what you need and theres just a minor driver issue Nov 27 04:17:15 so, lets just go ahead and test putty with it first Nov 27 04:17:35 Not recognizing COM8 Nov 27 04:17:43 ?? Nov 27 04:17:47 what says that? putty? Nov 27 04:18:03 can you take a pic of the error and upload it Nov 27 04:18:50 Yes. Already working on it. Nov 27 04:19:22 https://imgur.com/a/fJ8zF9Z Nov 27 04:19:47 k, so the driver error is significant enough to be causing problems Nov 27 04:21:02 can you right click on the listing for it in device manager and select "update device driver"? Nov 27 04:22:18 @hedgeberg........I have a volunteer commitment I need to attend at 6:00 AM. Yes, I have opened the update driver window. It is asking me where to search. Nov 27 04:23:00 tell it to search online Nov 27 04:23:15 we'll try to get this done beforehand, dont worry Nov 27 04:23:23 how long until 6 AM there? Nov 27 04:23:40 (also ill brb, need to check something in the oven) Nov 27 04:25:00 Okay. I am searching online. It says the best driver is already installed. I am in Austin, Texas. So 7.5 hours. This is me. https://www.linkedin.com/in/rdchilders/ Nov 27 04:25:49 ok, back Nov 27 04:26:09 so, windows says its the best driver already?? Nov 27 04:26:11 hmm Nov 27 04:26:22 ok, thats always frustrating, you dont get that error much anymore Nov 27 04:26:30 This is me. https://www.linkedin.com/in/rdchilders/ Nov 27 04:27:10 that used to be much more common 4 years ago before win10 when MS expanded their repository of generic drivers Nov 27 04:27:14 Would it matter if my laptop is a few years old? I am running Windows 10. Nov 27 04:27:21 no, or it shouldnt Nov 27 04:27:37 especially if other usb devices are working Nov 27 04:27:39 It is an i7 Nov 27 04:27:39 one second Nov 27 04:30:31 ugh, theres a ton of discussion of gadget serial drivers online, but all from sketchy "free drivers!" websites Nov 27 04:30:46 where did you get your driver from? Nov 27 04:32:17 https://beagleboard.org/latest-images Nov 27 04:32:18 actually wait, you should also be able to use putty SSH and ethernet Nov 27 04:32:54 ok uh, can you hold for 15 minutes? cyber monday ends here in a few minutes and i need to buy pants Nov 27 04:33:25 wait actually that shouldnt matter Nov 27 04:33:32 can you open a command prompt for me? Nov 27 04:33:48 I was told by Jason Kridner that there were driver issues with Windows and to wait a few days. Then a new release came out which is what I am using now. Nov 27 04:34:24 yeah, we should be able to completely skip the usb serial route Nov 27 04:34:39 Command prompt being Windows key and +R? Nov 27 04:34:46 no but that works Nov 27 04:34:52 LOL Nov 27 04:34:59 type in cmd.exe there and press enter Nov 27 04:35:23 got it Nov 27 04:35:54 ok, now type ipconfig and press enter Nov 27 04:36:16 and then highlight all that text, right click to copy, and then paste it here: https://pastebin.com/ Nov 27 04:36:18 lots of stuff Nov 27 04:36:23 yup, as it should be Nov 27 04:36:37 paste it to pastebin and send it to me Nov 27 04:38:15 Hmmm Never used Pastebin. I need to make an account it looks like. Nov 27 04:38:20 no Nov 27 04:38:40 if you just use the link i sent, then paste in the big text box, and hit submit at the bottom, it will work Nov 27 04:38:43 then send me the link Nov 27 04:38:49 https://pastebin.com/uEvLjeAS Nov 27 04:39:05 ok perfect Nov 27 04:39:31 open putty, select ssh instead of serial, and then use the following settings: Nov 27 04:39:38 ip address: 192.168.7.2 Nov 27 04:39:41 port: 22 Nov 27 04:39:45 username: debian Nov 27 04:39:50 I did not see a share link so I just cut and pasted the address bar. Nov 27 04:39:50 password: temppwd Nov 27 04:41:09 Do I hit enter after port 22? (I do not see a place yet to add the user name and password) Nov 27 04:41:21 yeah, should be fine Nov 27 04:41:49 havent used putty in a fair bit of time so i dont remember 100% but i think it should work Nov 27 04:42:29 There is a warning that popped up. Nov 27 04:43:01 https://imgur.com/a/uzRlQp9 Nov 27 04:43:12 yeah, hit yes Nov 27 04:43:51 thats a thing ssh does, basically it tries to verify that both the client and the server have talked to each other before, and if they havent then it tells you so you dont accidentally connect to a malicious host Nov 27 04:44:34 I figured it was safe. Something might have timed out. https://imgur.com/1LR0yEn Nov 27 04:44:47 did you hit yes? Nov 27 04:44:58 anyway, try the connection again Nov 27 04:45:19 I am reloading 192.168.7.2 Nov 27 04:45:25 cool Nov 27 04:46:05 Holy shit. I got a Login as: Nov 27 04:46:10 yup Nov 27 04:46:21 enter the useername there Nov 27 04:46:28 and it will prompt you for the password Nov 27 04:46:32 and then you should be good to go Nov 27 04:47:28 Oh my goodness. THANK YOU! I have been beating my head against a wall for months. Nov 27 04:47:41 Glad it helped! Nov 27 04:47:46 quick info on what we did: Nov 27 04:47:56 I take it you are on the East Coast time zone. Nov 27 04:48:06 basically, ssh is the "older brother" to serial Nov 27 04:48:12 (yeah, I am EST) Nov 27 04:48:22 (EDT, I guess, since daylight savings) Nov 27 04:48:22 Are you on LinkedIn? Nov 27 04:48:39 no, I'm paranoid about giving out my identity Nov 27 04:48:44 LOL Nov 27 04:48:54 if you ever need to reach me you can get me on twitter: Nov 27 04:48:58 twitter.com/hedgeberg Nov 27 04:49:03 always happy to help Nov 27 04:49:10 anyway, ssh is like the older brother to serial Nov 27 04:49:21 I understand. LOL I got fed up with the toxic nature of twitter and killed my account. Nov 27 04:49:39 its a remote terminal, so instead of having to have wires plugged in to something it can all be over the network Nov 27 04:50:03 the beaglebone actually enumerates as 2 different usb devices, 1 is a usb-serial device, and 1 is a usb-ethernet device Nov 27 04:50:29 basically, the usb-ethernet device creates a little isolated network that just the beaglebone and your laptop are on, and thats what lets you use ssh Nov 27 04:51:04 ah Nov 27 04:51:06 if you havent used linux much before,I recommend taking some time to look into ssh, its pretty neat. you can also use it to transfer files back and forth between your laptop and the beaglebone Nov 27 04:51:21 Okay. Nov 27 04:51:45 if you want to do that, you need to google "putty scp" because it doesnt come with it by default iirc? Nov 27 04:52:24 this is the same way youd connect to it if you were to use the real ethernet port and plug it into, like, a router Nov 27 04:53:08 when you log on via the usb-ethernet-ssh method, you should be able to use the command "ip addr" to find the ethernet adapter's ip address Nov 27 04:53:37 and then so long as youre on the same network and the beaglbone is on, you can ssh without being right next to it. its pretty handy and way less finicky than serial Nov 27 04:53:48 I know nothing of Linux. There is a Intro to Linux course I signed up for because it was about 90% off for black Friday. https://www.edx.org/course/introduction-to-linux Nov 27 04:54:24 good call, its a good thing to learn about. teaches you a lot about how computers work because a lot of the stuff that windows or OSX hide from the user, linux doesnt Nov 27 04:55:12 its a steep learning curve, but imo it's worth it Nov 27 04:55:27 Thank you, Good Night. Nov 27 04:55:35 no problem at all! I'm glad it helped! Nov 27 04:55:51 Have a good night, and have fun at the volunteering thing Nov 27 04:56:22 I am wanting to do vision system and CNC stuff with BBB. Nov 27 04:56:52 CNC stuff is really fun Nov 27 04:57:10 ever written code before? you may need to do a little programming to make it work, depending Nov 27 05:01:17 Back about 1995 I took a Pascal course and dropped out of a C++ course. (It was a night class geared toward people who already new how to use C++ but needed the validation of community college) Nov 27 05:03:02 I have a few friends that know how to program. Nov 27 05:03:06 Good Night. Nov 27 05:03:17 If youre going to try to learn nowadays, I'd try with python. its a lot simpler. Nov 27 05:03:38 I failed at my first c++ learning attempt too, python was the first language i was able to actually learn wella Nov 27 05:03:43 yeah! night! Nov 27 05:04:30 Thank you for the advice. The tools are more user friendly now. (brain fart..... debugger tool I think) Nov 27 06:31:49 Hi Nov 27 06:32:49 I want to control 4 stepper motors through beaglebone black Nov 27 12:14:29 mike121: the latest (2018-10-07) stretch-iot image seems to boot fine (from sd card) on my BBB rev A5A (with XAM3359AZCZ100) Nov 27 12:17:39 i had feeling you would say that :). hopefully when i get the j1 it will tell me something Nov 27 12:25:34 did you have a 2014 debian image flashed previously. I wonder if i gradually reflash to intermediate image it might then take the new version Nov 27 13:20:27 mike121: that doesn't really make any sense Nov 27 13:23:46 when booting with S2 held down, the contents of eMMC should be completely ignored Nov 27 13:30:53 as far as I can tell u-boot also doesn't access eMMC in this case, or at least doesn't mention doing so Nov 27 13:44:53 i've no clue. when the button is pressed (to boot from sd), does any data get put onto the eMMC from the new sd image, I assume something must get loaded into memory somewhere ? Nov 27 14:25:55 stuff gets loaded from sd card into ram. eMMC is completely irrelevant in all this Nov 27 14:26:29 the beaglebone can fully operate without any eMMC at all (indeed the original BeagleBone White didn't have any eMMC) Nov 27 14:26:43 (neither does the pocketbeagle) Nov 27 15:02:09 i hope my J1 will tell me something then. Nov 27 15:37:35 m Nov 27 15:37:37 , Nov 27 17:40:12 pthhh Nov 27 18:13:47 hey guys is there any video or something for setuping beaglebone black? Nov 27 18:15:44 the instructions on beagleboard.org are quite good. If you … Nov 27 18:15:46 screw this Nov 27 18:26:13 * KotH hugs tbr Nov 27 18:27:09 tbr: the right way to deal with turkish people is to first insult their mother. after that you will have their full attention Nov 27 18:27:14 tbr: and only 5 minutes left to live Nov 27 18:46:25 KotH: rrrright :) Nov 27 18:47:11 hasario:https://www.youtube.com/watch?v=MSbJ45OnajE Nov 27 18:47:19 * tbr idly ponders if yomama.tr exists Nov 27 18:47:36 let me know if you can boot yours from sd card :) Nov 27 18:47:51 mike121: he's loooong gone (right before I said what I had half written he left, that made me post it Nov 27 18:49:31 so he did, 2 mins i'll look for that next time Nov 27 18:49:52 a sad fact around here and on IRC help channels in general Nov 27 18:52:29 we really need a big notice on the web chat page containing something along the lines of what I sometimes say to people here who ask a "can I ask a question?"-type question Nov 27 18:55:37 e.g. "just ask your questions and have patience. plenty of knowledgeable people here, but usually people will only glance at chat occasionally" Nov 27 18:55:48 or something Nov 27 18:59:49 that's the age old why bother. Nov 27 19:00:17 every irc channel has problem it seems. Nov 27 19:00:35 because I'm pretty sure a fair number of people who come here via the web chat have no idea that that's how it works Nov 27 19:51:32 i think its more so people are torn between this and something like SO, most forum posts are answered (lets pretend) in like 4 hours, so its easy to quickly toss a question out and if there is no quick response, just post it else where. not saying its right, just a matter of choices and likely turn around time Nov 27 19:51:53 the ask about asking tho.... well... maybe it has to do with checking activity before forming a well thought out question? Nov 27 20:43:55 yeah zmatt is right Nov 27 20:44:11 any of you guys played around with raspberry pi, how does that compare, guess quite similar ? just a different chipset? Nov 27 20:44:25 yeah i think most of us probably has Nov 27 20:44:36 the comparison is pretty straightforward Nov 27 20:44:49 i like the BBB because it's open and has a larger family of things based off the same chipset Nov 27 20:46:03 bbb lost it's hardware ethernet port on the new models - a minus in my book. Nov 27 20:46:25 wonder why rpi has suck a bigger followiong (certainly on reddit), better marketing i assume Nov 27 20:46:30 such Nov 27 20:46:41 rpi suffers from vendor lockdown. can't buy the soc for own designs, closed bootloader, sucky docs, etc Nov 27 20:46:53 in terms of raw following rpi beats anything Nov 27 20:47:27 a lot of it is just barely above ordinary consumer though Nov 27 20:48:11 despite best effort the amount of outdated and misleading blogposts about the BBB will never be the same as for the rpi Nov 27 20:48:35 Snert_: how is the ethernet implemented in newer models? Nov 27 20:48:55 rpi has poe options which is a + in my book Nov 27 20:49:07 vagrantc: wireless only. Nov 27 20:49:11 ah Nov 27 20:49:13 a supposedly b0rken poe design Nov 27 20:49:19 how common are the bbb and rpi in real world apps, are they just mainly for hobbyist. like you mention knowing the chipsets well you can easily transfer that knowledge to a real world device ? Nov 27 20:49:30 so not the beagleboneblack, just the other beaglebone* Nov 27 20:49:33 tbr: not with the navolabs poe board. Nov 27 20:50:23 mike121: i mean, electronics engineering is a gagantic field Nov 27 20:50:32 people use bbb in commercial and industrial products Nov 27 20:50:35 mike121: rpi is ok for toys and marketing demos… Nov 27 20:50:45 rpi has industrial version Nov 27 20:50:49 which is certified as such Nov 27 20:50:58 that doesn't mean its the best choice Nov 27 20:51:03 can you buy the SoM in numbers by now? Nov 27 20:51:21 rofl i wouldn't know Nov 27 20:51:23 i'm not buying it Nov 27 20:51:58 I mean you certainly can do proper embedded linux engineering with an rpi, btdt (was forced to at money point) Nov 27 20:52:45 it's just incredibly easy to be led down the path of horrible kludges that someone came up with and messes around in /dev/mem under the hood… Nov 27 20:56:33 so what are the good things to learn if looking for a job. IOT ? knowing about BT / WIFI etc Nov 27 20:57:29 knowing about creating device drivers, yocto... ? Nov 27 20:58:24 https://twitter.com/thisissmithj/status/971269954812235777 - scrolled by today Nov 27 21:01:10 thing i've found is if you have no production experience then its tricky. i was looking for jobs in golang / node / react.js which i've never done , just self taught. Nov 27 21:02:13 also that market there are tonnes of people looking. Nov 27 21:04:09 pgmmrs are dime-a-dozen these days Nov 27 21:04:19 just ask india. Nov 27 21:04:34 broad knowledge won't get you the job, but it will ensure you keep it and your career goes well Nov 27 21:05:22 it's not often you see people involved in hiring who can appreciate broad knowledge, most actually rate it negatively Nov 27 21:23:02 some people have like a walking dictionary brain, so ask them and they can talk and tell you all about it. Nov 27 21:29:39 Hi guys, on this setup guide does anyone know why the caveat "Due to the way the environment sets the partition mmcblk1" ? Like what is the difference with Debian 7.5 that makes it unnecessary to hold down the S2 button? I know it gives the wrong dev name, but why? https://elinux.org/BeagleBone_Black_Extracting_eMMC_contents#autorun.sh Probably more of a debian question now that i think of it but don't know if anyone here would Nov 27 21:41:22 @sa Nov 27 21:41:44 sarahweefolk: moving to u-boot overlays broke many of the old image's u-boot assumptions.. Nov 27 21:47:23 rcn-ee: ahhh, I see Nov 27 21:49:29 rcn-ee: thank you :) Nov 27 21:50:37 For a long time, (pre u-boot overlays) we could get it to boot with "almost" any old random u-boot and also worked around a few u-boot bugs over the years... Nov 27 21:51:00 But once we moved to u-boot overlays, if u-boot didn't support the new feature, bootup would just lock up.. Nov 27 22:02:45 rc-nee: so this just because the image the manufacturer is using is old, but the u-boot is new, or vice versa? It tracks with their serial numers (GHI electronics fyi) I wonder if they'll tell me, or if they'll let me order only specific SN ranges :p Nov 27 22:04:40 they do builds in batches.. it takes a long time for all re-seller's to sell all boards before "new" batches come in with newer "latest" images.. Nov 27 22:06:30 if 7 years ago we would have swapped the eMMC/microSD mmc lines no one would have noticed this issue.. Nov 27 22:06:40 it's just because eMMC is first.. Nov 27 22:07:31 @rcn-ee: right, digikey would have no idea. guess it just goes in the work instruction. if it don't work, push the damn button. thank you so much! Nov 27 22:08:36 rcn-ee: oh, lol just figured out who i'm talking to hahah Nov 27 22:10:29 the other issue.. it also depends when the boards come in from GHI.. new boards "might" be on the front of the shelf or back of the shelf... Nov 27 22:14:24 rcn-ee: true. sourcing isn't really my deal so these obvious common sense things don't come to my mind Nov 27 23:15:54 ohaiyou Nov 27 23:16:10 zmatt get this Nov 27 23:16:33 told coworker i bought the bad bone from that it was dead Nov 27 23:16:49 they swapped it for another of the 5 they have apparently b/c they felt bad? Nov 27 23:17:00 sd card flasher works immediately Nov 27 23:17:42 hashtag-rip Nov 27 23:49:57 do ensure that that beaglebone gets physically destroyed in some way to ensure it will never waste anyone's time and sanity again Nov 28 00:58:57 hahaha you could send it to me Nov 28 01:53:21 lol, I wondered if maybe the boost-beast library was any good for opening a websocket Nov 28 01:54:03 neh? Nov 28 01:54:11 their minimalistic async client example ( https://www.boost.org/doc/libs/1_68_0/libs/beast/example/websocket/client/async/websocket_client_async.cpp ) takes, on my laptop, _40_ seconds to compile Nov 28 01:54:14 libwebsocket... Nov 28 01:54:26 ooh puppy Nov 28 01:54:32 and that's with all headers needed already in disk cache Nov 28 01:54:44 just pure compilation time of one source file Nov 28 01:55:08 wow i hate their readme Nov 28 01:55:43 ? Nov 28 01:56:10 its all the branding inside the readme Nov 28 01:56:16 rubs me the wrong wow Nov 28 01:56:32 which readme? Nov 28 01:56:37 https://github.com/boostorg/beast Nov 28 01:57:02 this is like the myspace of opensource Nov 28 01:57:21 if it keeps going this way pretty soon we're gonna github stories and audio autoplay Nov 28 01:57:52 *shrug* most of them seem to be links to talks, which some people may find informative Nov 28 01:58:27 regardless, this compilation time can go fuck itself Nov 28 01:59:12 "libwebsocket" doesn't yield anything btw, did you mean libwebsockets? Nov 28 01:59:22 ah yeah Nov 28 02:02:06 i haven't used it Nov 28 02:02:10 i've just seen it around Nov 28 02:02:36 what... the... fuck... is this bullshit Nov 28 02:02:51 this is one of the two source files that makes up the "minimal ws client echo" example: https://libwebsockets.org/git/libwebsockets/tree/minimal-examples/ws-client/minimal-ws-client-echo/protocol_lws_minimal_client_echo.c?h=v3.1-stable Nov 28 02:03:08 this is a minimal echo example? Nov 28 02:05:18 yeah it looks like each example is a couple kb lol Nov 28 02:05:46 apparently you need to read the websockets RFC to use this library XD Nov 28 02:06:06 yeah, well I already have, and I guess I'll just roll my own code, given that the protocol is pretty trivial Nov 28 02:06:53 (code can be kept simple by not supporting extensions, since no relevant extensions are actually supported in the real world anyway) Nov 28 02:07:03 ah yah Nov 28 02:07:14 i don't know if extensions and subprotocols are the same but Nov 28 02:07:41 the only extension I've ever seen implemented is the "per-message deflate" extension, and it is recommended to disable that one Nov 28 02:07:57 no, subprotocol is not the same thing Nov 28 02:08:01 * ayjay_t shrugs Nov 28 02:08:12 it sort of a trivial issue one way or the other Nov 28 02:08:28 although its unfortunate nobodies built out a proper make-and-go implementation Nov 28 02:09:07 maybe there is, but the time spent searching and evaluating could also be spent coding Nov 28 02:10:21 right but if you find one that has an active community than you get long range benefits from that no Nov 28 02:14:15 also, most implementations appear to be servers rather than clients Nov 28 02:16:40 yeah 100% Nov 28 02:16:56 you'll probably use c++ Nov 28 02:17:01 i'm guessing Nov 28 02:17:16 yes Nov 28 02:18:29 anyway, time for zZ **** ENDING LOGGING AT Wed Nov 28 03:00:01 2018