**** BEGIN LOGGING AT Wed Oct 30 02:59:57 2019 Oct 30 08:49:21 i have a probably stupid question: could the LCD CAPE touchscreen management be done outside from the kernel? something that runs as service, maybe? Oct 30 11:44:47 i have a probably stupid question: could the touchscreen management be a service instead of a kernel driver, so one could change it without messing with the kernel? Oct 30 11:53:04 Parduz: Hello! Oct 30 11:53:19 Hey :) Oct 30 11:53:57 happy to see you here.... i'd need some assistance :) Oct 30 11:54:09 I think that the touchscreen drivers are already completed, i.e. at least mine is done in the repo. from rcn. Oct 30 11:54:31 yep, problem is they're not so good :( Oct 30 11:54:35 Oh! Oct 30 11:54:40 What is the issue you are having? Oct 30 11:54:51 Touch screen? Oct 30 11:55:02 Cap. or Resistive? Oct 30 11:55:08 i'm using a LCD4 cape, and when you touch the screen the CPU goes up to 100% Oct 30 11:55:16 Yikes. Oct 30 11:55:33 Well, it would take a long time to make another driver. Right? Oct 30 11:55:34 'cause the driver fires too much interrupts and events. Oct 30 11:55:38 Oh. Oct 30 11:56:31 i'd like to make it more ... slow, also the cursor is "nervous". There's was a bad jitter problem in the past but it was patched Oct 30 11:56:51 Man, I think that LCD4 Cape is an older one, right? Oct 30 11:57:05 Let me go and see. Oct 30 11:57:42 it the displaytech one Oct 30 11:58:14 DT043BTFT-TS Oct 30 11:58:41 https://www.digikey.com/en/product-highlight/4/4d-systems/4dcape-43-4-3-lcd-cape-for-beaglebone-black?utm_adgroup=xGeneral&slid=&gclid=Cj0KCQjw6eTtBRDdARIsANZWjYaxGqM9um4TC7y_QRg0sWOaUhB-Iv6uA2yuaZk7PBv1D_YhbHRO39gaAn4vEALw_wcB? Oct 30 11:58:58 ... Oct 30 11:59:14 I am unfamiliar w/ that Cape you have. Is it from Circuit Co? Oct 30 11:59:31 yeah, same stuff as your link Oct 30 11:59:40 Okay. Oct 30 12:00:12 Are you using it w/ the BBB or another version of their family of boards? Oct 30 12:00:25 BBB Oct 30 12:00:45 Hmm. Oct 30 12:00:51 That seems like a nice Cape. Oct 30 12:01:04 I am sorry that it promotes 100% CPU usage. Oct 30 12:01:19 Dang. What OS do you have? Oct 30 12:01:26 debian 9 Oct 30 12:01:45 jeez! gotta go for about an hour :( Oct 30 12:01:47 sorry Oct 30 12:01:56 Okay. Oct 30 12:01:58 Have fun! Oct 30 12:34:57 Parduz: When you arrive back, this one has a datasheet: https://www.digikey.com/product-detail/en/4d-systems-pty-ltd/GEN4-4DCAPE-43T/1613-1386-ND/7360959. Oct 30 13:05:07 thanks, but i already have the documentation. What i miss is to understand what the parameters in the BB_BONE_LCD4 dts file means and do, if they are useful for my purpose... Oct 30 13:06:05 ... or if i can use the driver for the display and make and run a service for the touch management, so i don't need to tamper with the kernel. Oct 30 13:06:37 (i mean, the kernel driver for the display without the touch part) Oct 30 13:27:52 Parduz: here's a DT fragment for an LCD panel I wrote, which has a fair bit of comments regarding the properties set: https://pastebin.com/raw/xnPyKNfi Oct 30 13:28:36 I haven't checked whether the remark at the top is still true on current kernels Oct 30 13:29:27 thanks! Oct 30 13:32:59 zmatt: so those are the parameters to drive the LCD, right? Oct 30 13:33:34 there's not the touchscreen part, right? Oct 30 13:38:38 anyway, i've included all of your comments in my dts B) Oct 30 13:50:32 I have a BeagleBone Black and I would like to set the Ethernet to be only 10Mbit and turn off the higher frequency oscillators. Is that possible to do? Oct 30 13:52:00 technosam: you can limit the inteface to 10Mbit thru standard console tools, but the oscillator clocks the phy.. Oct 30 13:53:00 I left the AI on our board overnight to see if it overheated, which it didn't. \o/ But despite what I thought was a static IP config, it magically changed its eth0 IP. I'm thinking connman is the culprit? Oct 30 13:54:12 That's what I figured, but I wasn't sure how to go about proving that was the case. Is it just one oscillator for all the modes? Oct 30 14:00:13 could be the ti,charge-delay parameter a solution? reading his meaning from https://www.kernel.org/doc/Documentation/devicetree/bindings/input/touchscreen/ti-tsc-adc.txt Oct 30 14:01:03 Looks like it's ignoring /etc/network/interfaces today. Oct 30 14:01:44 Yesterday eth0 got the ip I set in there. Today it's "dynamic" and ofc ignoring the static ip. Oct 30 14:05:15 rcn-ee[m] sorry, I'm not sure why I got booted. I may have missed messages. Oct 30 14:09:21 technosam: i belive the phy has it's own dedicated oscillator, haven't look at that section in years. Oct 30 14:10:24 technosam: the phy requires a 25 MHz clock and has an integrated oscillator (using external crystal) to provide it. it is required for all operational modes Oct 30 14:12:57 (in 10BASE-T mode it uses an internal PLL to create a 20 MHz clock from it which is required for the 10BASE-T transmitter) Oct 30 14:14:29 zmatt Thanks for the info, that's great. I'm looking at the schematic, and this isn't really my strongest suit, but I'm not seeing where the clock is used at all. Could you give me a page to reference? Oct 30 14:16:31 technosam: page 9: https://github.com/beagleboard/beaglebone-black/blob/master/BBB_SCH.pdf Oct 30 14:16:53 bottom left, 25Mhz driving the phy Oct 30 14:17:37 which clock? like I just said the 25 MHz clock is generated by the phy using the external crystal (Y3) for its internal use, it is not directly used by anything else (nor even available for such use) Oct 30 14:18:19 rcn-ee[m] Ah, yep, okay, I was looking in the wrong spot. That makes a lot of sense. Great, this is very helpful Oct 30 14:18:30 technosam: what are you actually trying to do? Oct 30 14:19:04 I'm using the BBB close to some very sensitive antennae, so I'm trying to turn off / reduce as many oscillators as possible Oct 30 14:19:23 note that iirc 10BASE-T mode actually consumes more power than 100BASE-TX, hence I wouldn't be surprised if it's also noisier Oct 30 14:19:48 although the MII bus is much lower speed, so I'm not sure Oct 30 14:20:15 technosam: disable hdmi, that's the biggest noise source.. we also have cec and jtag clocks disabled in the default image.. so what kernel are you running today? Oct 30 14:20:30 rcn-ee[m]: jtag doesn't produce an external clock Oct 30 14:20:35 zmatt I wouldn't be too surprised. I was thinking there might be multiple clocks, but since it's just the one doing division, yeah. Oct 30 14:21:32 I think HDMI won't do anything if nothing is plugged in (unless you force-enable it with kernel parameters), but disabling the CEC clock may be useful Oct 30 14:21:52 rcn-ee[m] Yep, HDMI was my next question. Unfortunately for the time being I'm stuck on Ubuntu 14.04 which is 3.13 kernel. I haven't done any research, but my next question was how to disable the HDMI Oct 30 14:22:22 @zmatt From my measurements, it does look like the HDMI clock is going even with nothing plugged in Oct 30 14:22:36 it may also be possible to enable spread-spectrum mode in the internal PLLs, though whether that's useful depends on whether peak power or average power is more important Oct 30 14:22:40 zmatt: i've alway assumed jtag was running: the pin was mux as out, https://github.com/beagleboard/BeagleBoard-DeviceTrees/commit/f118e45b7b4fb12a7f7fda9abb668717ab07e1bc#diff-2c32f37207f010a8cd0c64a16d4ef44a Oct 30 14:23:13 technosam: oh god! v3.13.x runs far away.. Oct 30 14:24:27 zmatt: even with nothing plugged in hdmi makes noise, that was the old workaround for fixing small wifi adapters on the black, disable hdmi, wifi would suddenly work.. Oct 30 14:24:51 rcn-ee[m] Yeah... I'd love to change it, but for business reasons complicated to explain here, I'm kind of locked in. I'm working on trying to change, but I haven't had the time to do much work Oct 30 14:25:54 rcn-ee[m]: looks like that bone-jtag file is nowhere actually #included? regardless it makes zero sense to me, it serves absolutely no purpose for jtag Oct 30 14:26:19 the reason that cpu pin is connected to the jtag header is because in mux mode 6 it's "emu3", part of system trace output Oct 30 14:26:29 technosam: looking at my git log: Jul 3, 2014 = 3.13.11-bone12... Oct 30 14:27:40 rcn-ee[m]: jtag is entirely clocked by the jtag debugger / test controller, which is asynchronous to any on-chip clock Oct 30 14:28:33 note also that that pin is also tied to P9.41 so this can't be enabled by default Oct 30 14:30:22 technosam: is electromagnetically shielding the BBB not an option? Oct 30 14:32:11 technosam: so sadly, i didn't do the dtb-rebuilder project nor a non-hdmi varient based on the old patch: https://rcn-ee.net/deb/sid-armhf/v3.13.11-bone12/ thus you get to hack up the dtb directly and remove the hdmi stuff. here is teh cec pins https://github.com/beagleboard/BeagleBoard-DeviceTrees/commit/e34f71f6b953094478c7c44c623c3a9f22de9d76 probally won't match 3.13.x thou.. Oct 30 14:37:59 rah: negative result :( but the patch looks pretty hacky so I'm going to hook it up to the scope to verify whether the reset extender is actually working Oct 30 14:40:08 zmatt, rcn-ee[m] Walked away for a meeting and got booted again, sorry. Looking at disabling HDMI, probably going to start here: https://www.onlogic.com/company/io-hub/disabling-the-beaglebone-black-hdmi-cape/ Oct 30 14:41:00 i feel dumb but .... how do i get (using the browser) the "complete" ti_am335x_tsc.c file for the 4.14 branch from Git site? (i mean, i see the patches applied as diff but don't know how to have the "whole single file") :( Oct 30 14:41:36 technosam: that guide is for v3.8.x, with the interface that got rejected.. Oct 30 14:41:54 rcn-ee[m] Hah, good to know, thanks Oct 30 14:42:19 Parduz You should be able to click "Browse files" from the commit you want and then navigate to the file you want. Oct 30 14:43:33 technosam: it just show two patch files. Oct 30 14:44:26 and picking one shows the diff Oct 30 14:45:03 Parduz Here's the complete repo at the official 4.14 release: https://github.com/torvalds/linux/tree/bebc6082da0a9f5d47a1ea2edc099bf671058bd4 Oct 30 14:45:31 technosam: try the middle one first, then the final one.. https://gist.github.com/RobertCNelson/6898d1f782381c75ef5b7b5aa25aaafb/revisions Oct 30 14:45:39 (this was v3.13.8's dts) Oct 30 14:48:50 many thanks technosam Oct 30 14:49:24 technosam: here's anything you may have missed: https://pastebin.com/QDCkeG7g Oct 30 14:52:40 zmatt rcn-ee[m] Thank you guys so much for doing all this for me. Oct 30 14:53:53 rcn-ee[m] If I were to upgrade to the latest kernel version, would I be able to disable the HDMI from user land? If I'm going to be modding the kernel, it might make sense to just bite the bullet and upgrade. Oct 30 14:54:31 rcn-ee[m] I will also ask the hw engineer if shielding is an option, sometimes the simplest is actually best. Oct 30 14:54:44 modding the kernel is never required except for extremely ancient pre-DT kernels Oct 30 14:54:51 at most you'd need to mod the DT Oct 30 14:55:10 the official 4.14 release ti_am335x_tsc is not patched. Oct 30 14:55:28 on the latest images the DT has been modularized using u-boot overlays, which are configured in /boot/uEnv.txt Oct 30 14:55:41 Sorry for my ignorance, what is the DT? Oct 30 14:56:19 Device Tree, a datastructure that u-boot loads from storage, optionally applies patches to, and then passes to the kernel, which uses it to learn about the hardware present Oct 30 14:58:04 in theory all you'd need to use overlays is an updated u-boot and associated userspace files (/boot/uEnv.txt and the overlays) but I don't think the new system is compatible with ancient kernels Oct 30 14:58:37 oh and I just remembered you're using an unofficial ubuntu image rather than the official debian images, so I don't know if that has any impact Oct 30 15:00:58 Okay, so the idea would be that I would compile the HDMI DT object and use uEnv.txt to tell the kernel where it is Oct 30 15:02:06 Parduz What do you mean by "not patched"? Oct 30 15:03:02 i mean that the pathes in https://github.com/RobertCNelson/ti-linux-kernel-dev/blob/ti-linux-rt-4.14.y/patches/drivers/ti/tsc/0002-ti_am335x_tsc-correct-formula-code-to-calculate-pres.patch are not in the repo you give me Oct 30 15:03:57 in that ti_am335x_tsc.c file there's still the old code Oct 30 15:04:02 Here's the link to a file named ti_am335x_tsc.c: https://github.com/torvalds/linux/blob/bebc6082da0a9f5d47a1ea2edc099bf671058bd4/drivers/input/touchscreen/ti_am335x_tsc.c Oct 30 15:04:11 Not sure if that's exactly what you're looking for Oct 30 15:04:39 it isn't Oct 30 15:05:37 Parduz: you can find the patched kernel tree for one of rcn's kernel releases here: https://github.com/RobertCNelson/linux-stable-rcn-ee/tree/4.14.108-ti-r120/ (replace the last part by the exact release you're looking for) Oct 30 15:10:44 oh, thanks! One day i'll understand how to find what i'm searching for :( Oct 30 15:10:51 zmatt Here's my uEnv.txt https://pastebin.com/NFdZtTH3 It looks like the HDMI stuff is already commented out Oct 30 15:11:52 rcn-ee[m] Now that I'm actually logged into the beaglebone, Is see that I'm actually on kernel 4.1.17. My apologies, I though 3.13 was with Ubuntu 14.04 Oct 30 15:16:32 m Oct 30 15:22:55 I was wondering if it's possible to underclock an AI and get it to generate less heat? I don't need all the CPU is has by any stretch ... so I'd like to investigate this possibility. Oct 30 15:32:33 zmatt :-( Oct 30 15:35:03 still struggling with the ethernet issue: hard-reset doesn't always solve the problem Oct 30 15:35:51 the techs managed to keep the BBB resettd for 200ms at poweron, but this does'nt guarantee that ethernet port is running Oct 30 15:36:11 Parduz What ethernet issue are you referring to? Oct 30 15:36:50 the eth drivers sometime (often) does'nt start up properly. Oct 30 15:37:17 it may have all led off, or the green one on (not-blinking), and so no ethernet available Oct 30 15:38:01 That doesn't sound exactly like this symptoms I ran into, but here's something that might fix it: https://wp.josh.com/2018/06/04/a-software-only-solution-to-the-vexing-beagle-bone-black-phy-issue/ Oct 30 15:39:58 uh.... Oct 30 15:40:04 reading it Oct 30 15:50:40 jeez ... that guy is very negative about the BBB as a production device. Oct 30 15:51:02 guess my boss will stay asleep this night :| Oct 30 15:53:13 Probably jinxing myself but I've never heard of any Ethernet failing to work on our Blacks. We power them externally via the header. Oct 30 15:53:31 rah: 9/344 failures :/ Oct 30 15:56:12 zmatt: do you think that "playing" with the ti,charge-delay and the ti,chan-step-*delay values could help reducing the amount of interrupts generated by the touch? Oct 30 15:56:43 i'm trying to follow what that values do in the .c files but it is hard... Oct 30 15:58:34 Ragnorock: we have them everyday.... perhaps the fact that our BBBs never get a smooth shutdown before power-off does'nt help. Oct 30 15:58:51 Ragnorok , sorry :) Oct 30 15:59:04 dunno, I mean in theory you could make the measurements themselves really sluggish but that doesn't sound like a net win. I'd assume the main problem insufficient hysteresis Oct 30 15:59:23 Ours don't either, but frankly most of them are never shut off, either, so once they work... Oct 30 16:00:03 yeah if you rarely power-cycle them you'll probably also rarely see the issue Oct 30 16:00:11 That's why I'm so worried about the AI ... we'd like to replace the Black with AI but if it's never turned off... Oct 30 16:00:51 Does the AI have this problem? Oct 30 16:01:02 zmatt: seems wothing a try... after all it's just a change in the dts instead of in the kernel.... would you suggest what values should i try? Oct 30 16:01:13 the AI won't share any problem with the BBB, they're completely unrelated designs Oct 30 16:01:21 different components, everything is different Oct 30 16:01:29 And no thoughts on underclocking it seems like. I'm thinking I can underclock it but I'm worried about spurrious effects on other bits. Oct 30 16:01:51 zmatt: Was hoping that was the case. Oct 30 16:01:57 Ragnorok: are you still seeing excessive temperatures? Oct 30 16:02:29 Currently it's around 60C, but it has to live inside an enclosure w/ limited airflow. Oct 30 16:02:37 its lifetime specification is based on junction temperature up to 90C Oct 30 16:02:49 It also may be in a hot environment, like a vehicle in summer. Oct 30 16:03:28 I'd like to get it to run as cool as possible from the get-go. Oct 30 16:03:35 yeah, sounds reasonable Oct 30 16:04:14 I think underclocking could have a big impact on that. It does with a desktop ... why not this? Oct 30 16:04:36 I think it already gets underclocked currently? when idle anyway. check cpufreq Oct 30 16:04:46 Roit toe. Oct 30 16:05:04 it might actually be a net loss though Oct 30 16:05:24 you'd rather do the work that needs to be done twice as fast and then allow the cpu to enter idle state for longer Oct 30 16:05:38 so unless underclocking allows undervolting, I don't think it's actually a good idea Oct 30 16:06:09 Interesting point. I'll start with cpufreq and do some fiddling about. Oct 30 16:41:13 well, setting the ti,charge-delay from 0x400 to 0x4000 result in a super smooth touchscreen, 12% CPU max, and my app refreshing itself while dragging the finger across the screen Oct 30 16:42:11 ugly or not, that solves my problems! I'd like to understand how to fine-tuning this all, but for today i call the day! :) Oct 30 17:16:17 I am glad I showed Parduz that link earlier. Those Capacitive Touch Capes on DigiKey for the BBB look nice. Oct 30 17:16:48 I am serious. They have already assembled encasings/housings. Oct 30 17:16:55 Nice! Oct 30 17:17:49 4d systems? yeah they are nice. Oct 30 17:17:58 I just found them today. Oct 30 17:18:27 They even have nice datasheets that come along w/ their Capes. Oct 30 17:18:44 It is getting harder to find people willing to take time out to make datasheets. Oct 30 17:19:03 I mean, unless they are well funded and have writers. Oct 30 17:20:34 +rcn-ee[m]: Yea. the 4d Systems ones on Digikey.com. Oct 30 17:21:14 The resistive are nice but those Capacitive touch displays look like a fresh pot roast on a cold day. Oct 30 17:21:36 Hell, the Cape itself is only $14.00 or around that price. Oct 30 17:22:25 <--- leaving the page open! I have to cop it one day. Oct 30 17:24:41 For instance, does this one come w/ the Cape or is it just the Display: https://www.digikey.com/product-detail/en/4d-systems-pty-ltd/GEN4-4DCAPE-43CT-CLB/1613-1387-ND/7360960? Oct 30 17:25:00 I read about it but it seems to picture one thing and not describe things well enough. Oct 30 17:26:03 Oh well. There is only one way to find out! Oct 30 20:23:49 Hii, I currently have a problem connecting to .local:3000 and was wondering if you could help me with this. The device is connected through a miscro-usb to usb to computer Oct 30 20:34:39 RALPH92: ".local" is not a valid hostname, maybe you mean something like "beaglebone.local" ? regardless, with as little information as you have provided it is very unlikely anyone can possibly help you Oct 30 20:35:49 it helps a lot to describe your problem clearly in such a way that when someone glances at chat (like I'm doing right now) they might actually be able to identify the problem Oct 30 20:36:07 anyway, afk Oct 30 20:36:35 What other info can I provide? I am unable to connect a serial to it due to the fact I do not have a mini usb. Oct 30 20:38:14 RALPH92: you said it was "connected through micro-usb to usb to computer" so I just assumed you had a model which uses micro-usb Oct 30 20:38:23 I am able to connect to http://192.168.7.2/bone101/Support/bone101/ but not http://beaglebone.local:3000/. However I am still able to connect to Node-Red. Oct 30 20:39:08 The model is BeagleBone Black Wireless Oct 30 20:40:19 okay, so why were you saying something about mini-usb? the black-wireless uses micro-usb, not mini-usb Oct 30 20:41:16 RALPH92: http://192.168.7.2:3000/ <- that should work if bone101 worked.. Oct 30 20:41:21 The Black-wireless has both. I believe micro-usb for the beaglebone.local connections and mini-ubs for serial connections through Putty/other consoles Oct 30 20:41:55 what host OS and browser are you using? beaglebone.local should work in any browser on Mac, as well as on Linux (provided libnss-mdns is installed), and I think it should work in Chrome on Windows, but it won't work in general on windows Oct 30 20:42:03 the black-wireless does not have mini-usb Oct 30 20:42:04 RALPH92: nope just one.. the other is hdmi Oct 30 20:42:18 it replaced mini-usb by micro-usb Oct 30 20:42:50 the micro-hdmi port looks vaguely like mini-usb I guess (not really, but it's about the same size) Oct 30 20:44:38 Using Windows10/Chrome. Oct 30 20:44:59 micro-hdmi: https://i.ebayimg.com/images/g/cmMAAOSw-CpX-Ley/s-l300.jpg Oct 30 20:45:02 mini-usb: https://konsolowo.pl/6778-large_default/mini-usb-socket-for-ps3-controller.jpg Oct 30 20:45:05 anyway afk Oct 30 20:45:31 Yes I believe you're right, it is a mini-HDMI, Oct 30 20:46:01 It was working fine yesterday but I'm unable to connect to the cloud9IDE Oct 30 20:48:36 RALPH92: try the ip address.. Oct 30 20:49:28 http://192.168.7.2:3000/ does not work while http://192.168.7.2 works Oct 30 20:52:52 RALPH92: if you login over ssh ( ssh debian@192.168.7.2 ), then run "journalctl | grep cloud" we could see what it's doing. Oct 30 20:55:29 Im using SSH over Putty and having trouble with the password Oct 30 20:55:31 I have tried: Oct 30 20:55:48 [blank], BeagleBone, root, admin, debian, temppwd Oct 30 20:57:17 it's temppwd Oct 30 21:01:35 back Oct 30 21:02:25 Oct 24 22:04:44 beaglebone systemd[1]: Listening on cloud9.socket. Oct 30 21:02:26 standalone Oct 30 21:02:55 RALPH92: so you asked "What other info can I provide?" .. the fact that you can reach the BBB, just not on port 3000 was pretty crucial information since until you said so it sounded like a general networking problem, the fact that it used to work but ceased to is also pretty important Oct 30 21:03:52 RALPH92: did you just try to paste multiple lines btw or was there really just a single log entry? never try to paste multi-line output into chat, use a paste service like pastebin.com Oct 30 21:04:50 of course in any situation where something used to work but no longer does the question is automatically: what did you do? Oct 30 21:04:53 RALPH92: it should look like this: http://paste.debian.net/1112102/ Oct 30 21:05:46 rcn-ee[m]: it "should" have an error traceback? :P Oct 30 21:05:59 http://paste.debian.net/1112103/ Oct 30 21:06:51 zmatt: yeah that last part is buster's nodejs v10 running a v8 only cloud9... buggy as heck.. ;) Oct 30 21:07:26 RALPH92: yeap, sorry cloud9ide is dead.. did you upgrade nodejs? Oct 30 21:07:42 yeeeeep Oct 30 21:08:16 RALPH92: so, instead of wasting a lot of people's time, you could have opened up with the question "after upgrading nodejs, I can no longer connect to the cloud9 IDE" Oct 30 21:08:19 :P Oct 30 21:08:27 What version of nodejs did you install? Oct 30 21:08:37 ( i might have a build... ) Oct 30 21:09:02 12.13.0 Oct 30 21:09:32 zmatt so far you've only wasted mine by me having to read your text Oct 30 21:09:36 Oh sorry, that's humm.. Oct 30 21:09:46 RALPH92: basically any time you upgrade nodejs to a newer version, expect stuff to break Oct 30 21:10:10 So background.. the git checkout of cloud9 offically works with v8... got it sorta working on v10.. haven't tested v12.. Oct 30 21:10:24 the problem amazon bought them and they stopped development... Oct 30 21:10:33 RALPH92: ah yes, people rude as fuck to people trying to help you, a guaranteed recipe for success Oct 30 21:10:39 RALPH92: anything in the first digit of nodejs is a breaking change by definition Oct 30 21:10:46 that sentence failed Oct 30 21:10:49 *being rude Oct 30 21:11:16 RALPH92: give this pkg a try... http://repos.rcn-ee.net/debian/pool/main/c/c9-core-installer/c9-core-installer_3.1.5000.git20180223-0rcnee8~buster+20190718_armhf.deb Oct 30 21:11:16 zmatt: take a breath, a break and put on the mental ignore filter Oct 30 21:11:32 sadly, you'll have to manually unzip it... give me a sec.. Oct 30 21:13:21 RALPH92: sorry if what I said earlier came across as rude btw, but it's a common problem for people to omit a lot of essential details from their question / problem statement, forcing those who try to help them to fish the details out of them. I'm just trying to make it clear what sort of information to include to expediate getting help in the future Oct 30 21:14:27 if something used to work, you upgraded or changed something, and now that thing no longer works... that's absolutely something you should always be up front about Oct 30 21:14:37 RALPH92: NM: it'll extract it fine: https://github.com/rcn-ee/repos/blob/master/debian-10-c9-core-installer/suite/buster/debian/postinst#L46-L51 Oct 30 21:14:47 you just need to force dpkg to install it.. Oct 30 21:15:44 it might still not work, but i thought nodejs promissed after v10 there would be an api.. Oct 30 21:16:43 n-api was introduced in v8 actually, but it only helps for modules that have been ported over to use it Oct 30 21:17:24 and c9 is development is dead, so yeah we are stuck. ;) Oct 30 21:17:27 (which is a bigger change than the typical fixes needed from release to release, but it should avoid future breakage) Oct 30 21:17:37 well it only matters for binary modules anyway Oct 30 21:18:43 and it doesn't seem like binary incompat is the issue here, that would have resulted in a different error... though honestly I have a hard time guessing what could cause the error RALPH92 showed Oct 30 21:18:49 i'm wonder if might be useful to fork nodejs for cloud9, it has it's own special build, then users can upgrade nodejs anytime they want.. (it would be nodejs v8) Oct 30 21:19:01 and would never break cloud9 Oct 30 21:19:26 no need to "fork" presumably, it would just need a private nodejs installation Oct 30 21:20:03 rcn-ee[m]: concerning anything nodejs-ish that is network reachable, forkins is an awful idea unless you want to hire like 2 or 3 folks taking care of securtiy backporting Oct 30 21:20:10 ^ yeah exactly that.. Oct 30 21:20:38 but isn't v8 still lts'ed officially? Oct 30 21:20:54 because then probably some proper scripting inspired by nvm is fine Oct 30 21:21:09 if cloud9 is reachable from untrusted network, your problems are bigger than the potential for security issues in nodejs itself Oct 30 21:21:14 LetoThe2nd: i agree.. we have that problem with cloud9, it's pretty much abondoned by amazon, it works best with v6/v8.. v10 is broken.. Oct 30 21:21:31 anytime a user touches nodejs's version it breaks.. Oct 30 21:21:38 rcn-ee[m]: I noticed there were still commits on github relatively recently Oct 30 21:21:41 zmatt: soooo right Oct 30 21:21:58 is there no alternative to cloud9 ? Oct 30 21:23:32 There are committs as of Oct 17, 2018.. we are stuck on a Dec x, 2017 commit, as they removed our "special" standalone build.. Sadly for install size, there is no good alternative.. Oct 30 21:30:01 I use a debian VM and delete cloud9. This isn't an arduino. lol Oct 30 21:31:23 Ragnorok: actually, making it behave like an arduino wouldn't even be *THAT* complicated Oct 30 21:31:40 Ragnorok: https://www.arduino.cc/en/Main/ArduinoBoardTre ;) i have a beta here in the lab.. Oct 30 21:32:09 (ignore the coming soon in the picture, it's a dead project) Oct 30 21:32:11 No, but still. I peronally see no ROI to cloud9. Oct 30 21:32:40 I heard of those, but use 'bones instead. Oct 30 21:33:00 it's not bad for education and labs. Oct 30 21:33:16 just use plain old C Oct 30 21:33:25 less trouble, portable, and it just works without the bloat Oct 30 21:33:58 ds2: ehm, I don't see how that statement makes any sense in this context Oct 30 21:33:58 ds2: all fixed: https://github.com/RobertCNelson/boot-scripts/commit/e6fd0a097a08ace6a035ffd3ed3e803b38149d22 Oct 30 21:35:05 zmatt:eh? Oct 30 21:35:16 rcn-ee[m]: how do I pick it up w/o a full reflash? Oct 30 21:35:36 ds2: unless it's a reply to something long ago unrelated to the present discussion, in which case never mind Oct 30 21:36:53 ds2: cd /opt/scripts/ ; git pull ; sudo cp ./distro/buster/nginx/default /etc/nginx/sites-enabled/default Oct 30 21:36:53 zmatt: cloud9 is a pit for "javascript" coding... abandoning that pit avoids the whole problem with cloud9 Oct 30 21:37:11 ds2: cloud9 is an IDE, it supports multiple languages Oct 30 21:37:17 it's just *written* in javascript Oct 30 21:37:47 so, with your remark, are you offering to write a functionally equivalent browser-based IDE in "plain old C" ? Oct 30 21:37:55 zmatt: it is mostly used for that...unless you are telling me that can call up clang/gcc Oct 30 21:38:19 use the brower based SSH client and vi Oct 30 21:38:29 nothing to write Oct 30 21:38:39 probably? I haven't used it in ages, I know it supports python so I don't see why it wouldn't support c/c++ Oct 30 21:38:43 if you insist on browser stuff, that is. Oct 30 21:39:08 zmatt: I'll concede your point when you show me a Beagle related tutorial using that AND C :D Oct 30 21:39:37 a tutorial will use javascript for the same reason it uses the cloud9 ide: it's easier Oct 30 21:42:09 rcn-ee[m] Hey thanks for the debian packge! It works now, however it still says the node version is 12.13 but the IDE is up and running. Oct 30 21:43:40 RALPH92: awesome, yeah it's just the cloud9ide blob, no nodejs binary included.. the biggest bug is the "debug" doesn't work.. as it was built for v8, but v10 they removed the old debug option.. Oct 31 00:07:50 RALPH92: did you use nvm? i'm wondering about installing that by default, then setting up a version for each of the default nodejs packages we do.. Oct 31 01:16:08 2 points for BBAI - image does not go nuts trying to upgrade itself w/o permission and a monitor/kb/mouse is not needed (compared to Jetson Nano) Oct 31 01:26:41 ds2: does Nvidia still require you to sign away your soul before giving access to the SoC's TRM ? Oct 31 01:29:19 zmatt: I donno, I don't have the TRM Oct 31 01:29:36 I'd guess if you want it officially, sure Oct 31 01:29:55 just like the TRM for certain TI processors Oct 31 01:30:23 i just want something a little bit more open then the movidis stuff Oct 31 01:40:14 well TI has public TRMs for most SoCs, including the AM5729, even if some bits are missing (mainly crypto stuff and stuff encumbered by third-party rights) Oct 31 01:41:27 last time I was interested in getting some developer docs from nvidia the agreement contained such obnoxious things that I was like nope, fuck you guys Oct 31 01:41:41 it was a while ago though Oct 31 01:56:59 ds2: yep you still need a developer agreement to download the TRM, and the agreement is still just as offensive as it was previously Oct 31 02:07:22 zmatt: keyword there is "most" unless you want to navigate the twins Oct 31 02:07:58 unless I what? Oct 31 02:09:23 afaik any TI SoC that's generally available for purchase (which is kind of a prerequisite to care) also has a public TRM... you have a counterexample? Oct 31 02:17:49 if you limit to generally available, nope. Oct 31 02:21:38 I tend to care somewhat less about documentation for SoCs I can't use anyway... I still like having them of course... but I don't see how this is relevant in the context of nvidia jetson vs bbai Oct 31 02:22:37 nvidia wants you to agree to give up your firstborn child and your left kidney, TI lets you download the datasheet, TRM, errata, and application notes without even requiring registration Oct 31 02:24:05 it's something I personally value quite a bit Oct 31 02:47:20 depends on what you want to do, right now, I want to run some NN's. for that, there is no registration **** ENDING LOGGING AT Thu Oct 31 02:59:58 2019