**** BEGIN LOGGING AT Fri Jan 10 02:59:58 2020 Jan 10 03:43:32 Has anyone visited a maker faire here instead of being a "maker" at one? Jan 10 04:43:57 I went to one in Newcastle UK a couple of years back. Great day out. Jan 10 04:45:46 Nice. Jan 10 04:46:00 Hey. Look here: https://hackaday.com/2020/01/08/commercial-circuit-simulator-goes-free/?fbclid=IwAR2tfRNWne_1YMmLck4v7RMUoLeYkvxckfYP7USY23EDdxzK2tOOF7Cjnh0. Jan 10 04:46:20 It seems these people went out of business and now have some nice source for free. Enjoy. Jan 10 04:46:58 I have been a participant in one but I have not been to one to see the fiasco. Jan 10 04:47:10 I wanted to walk around so badly. Jan 10 04:48:13 I was stuck! Stuck. Jan 10 04:50:16 I did think I might go to maker faire paris, but it hasn't happened yet. Jan 10 04:51:05 Oh. The maker faire stuff is neat. I do not travel any longer b/c I am chicken but there is one in my town for now. I plan on being there, i.e. participant or just an onlooker. Jan 10 04:51:34 I need to get a person to hold my seat when I scurry off. Jan 10 04:52:14 There are many around the globe now. Jan 10 04:52:16 It is nice. Jan 10 04:52:44 We have a large one in New Orleans, I think. I will find out. Jan 10 04:54:20 https://paris.makerfaire.com/ happened in 2019, man. You missed it? Jan 10 04:55:20 That must have been an enormous one. Jan 10 04:57:27 Nope. No New Orleans maker faire. Sheesh. Jan 10 04:57:56 The Saints probably would not share their stadium. Jan 10 10:01:09 hello :) Jan 10 10:01:39 i'm trying to give permissions to my app to reboot the BBB, but whatever i do does'nt seems working Jan 10 10:02:37 there's someone which can tell me how can i give to my app the permission to reboot (hopefully without running it with sudo)? Jan 10 10:02:58 hmm, interesting question... I wonder if there's an easier way than by using policykit Jan 10 10:03:22 well, that or a suid root helper Jan 10 10:03:32 that would be easier but not very elegant Jan 10 10:03:52 i followed some answer on stackoverflow tampering with sudoers.d but without result Jan 10 10:04:45 oh yeah that could work too, but it feels a bit yuck to do that for a service.. sudo is meant for people, not for programs Jan 10 10:05:18 i don't care a lot about elegance, right now Jan 10 10:05:29 or "yuck" :) Jan 10 10:06:00 anyway, it did'nt worked at all, so ... :| Jan 10 10:07:25 i tried with "debian ALL=NOPASSWD: /sbin/halt, /sbin/reboot, /sbin/poweroff" Jan 10 10:08:01 and just to see what it will do, "%admin ALL = NOPASSWD: /home/debian/MyAPP/MyAPP" Jan 10 10:08:18 that last one makes no sense Jan 10 10:10:07 something i've got from https://unix.stackexchange.com/questions/18830/how-to-run-a-specific-program-as-root-without-a-password-prompt Jan 10 10:13:47 @zmatt: i followed the steps you gave accordingly,first i created the build folder and navigated into the folder and ran this command : cmake -D PAHO_ENABLE_CPACK:BOOL=OFF -D CMAKE_INSTALL_DOCDIR:PATH=share/doc/paho.mqtt.c but i got this as response CMake Error: The source directory "/home/sco/git/paho.mqtt.c/build/CMAKE_INSTALL_DOCDIR:PATH=share/doc/paho.mqtt.c" does not exist. Jan 10 10:16:18 m-blasie: you're missing the last argument Jan 10 10:16:22 the .. is not decorative Jan 10 10:16:56 Parduz: this seems to work for me: Jan 10 10:16:58 debian ALL=(root) NOPASSWD: /sbin/reboot Jan 10 10:17:38 i try it now Jan 10 10:19:12 Failed to set wall message, ignoring: The name org.freedesktop.PolicyKit1 was not provided by any .service files Jan 10 10:19:12 Failed to reboot system via logind: The name org.freedesktop.PolicyKit1 was not provided by any .service files Jan 10 10:19:12 Failed to open /dev/initctl: Permission denied Jan 10 10:19:12 Failed to talk to init daemon. Jan 10 10:19:52 Parduz: hmm Jan 10 10:19:56 that is what i got. Jan 10 10:20:18 I'll admit I only tried it while logged in via ssh, not from a service... maybe stuff breaks in that case Jan 10 10:20:33 i'm using a file "/etc/sudoers.d/allowreboot" Jan 10 10:20:48 should i write it in the etc/sudoers file? Jan 10 10:20:55 no, the problems isn't with the sudoers line Jan 10 10:21:01 ok. Jan 10 10:21:09 actually wait, how are you invoking reboot? Jan 10 10:21:13 with "sudo reboot" ? Jan 10 10:21:18 nope Jan 10 10:21:26 I mean... sudoers is the config file for sudo Jan 10 10:21:33 it works (and asks for password) Jan 10 10:21:40 it shouldn't ask for a password Jan 10 10:21:52 and it didn't for me Jan 10 10:22:10 debian@beaglebone:~$ sudo reboot Jan 10 10:22:10 [sudo] password for debian: Jan 10 10:22:27 that's exactly what the line means: allow user "debian" to execute /sbin/reboot as root without password Jan 10 10:22:48 i'm using PuTTY logged as "debian", if this matters Jan 10 10:23:06 your shell prompt was already showing you're logged in as debian Jan 10 10:25:46 it works the same if i connect a keyboard to the BBB and open the openbox terminal Jan 10 10:26:22 or better .. it doesnt works as well Jan 10 10:28:26 Parduz: can you share (via e.g. pastebin) the output of: Jan 10 10:28:39 ls -l /etc/sudoers/allowreboot Jan 10 10:28:51 sure Jan 10 10:28:54 sudoers.d sorry Jan 10 10:29:00 cat /etc/sudoers.d/allowreboot Jan 10 10:29:08 sudo cat /etc/sudoers Jan 10 10:33:59 https://pastebin.com/tXkUYVQc Jan 10 10:34:55 okay I don't know why the last line of /etc/sudoers is commented out, but that's why your config file is being ignored Jan 10 10:35:54 it throws errors if i write it bad Jan 10 10:36:03 ? Jan 10 10:36:11 (actually, it WANTS a last empty line) Jan 10 10:36:43 if i write in it something wrong, it complains when i use reboot Jan 10 10:36:58 so i think that file is being read each time Jan 10 10:37:15 or am i missing what you mean? Jan 10 10:37:27 "it WANTS a last empty line" ? what do you mean by that? Jan 10 10:38:09 i mean that after the last charachter in the last line, it wants a \n. Jan 10 10:38:17 note that on linux the convention is that a newline is a line-terminator, not a line-separator, hence a text file is expected to end in a newline and editors will add one automatically Jan 10 10:38:31 ok Jan 10 10:38:50 editors like vim and nano will not show this as being an empty line at the end Jan 10 10:39:08 if your editor does, you may want to check if that is fixable Jan 10 10:39:18 i'm using sublimetext on windows to edit, Jan 10 10:39:45 it manages linux lines, but shows an empty line at the end. Jan 10 10:40:25 anyway, what i mean is that the file is being read each time i try "reboot". Jan 10 10:40:37 apparently it has a newline_at_eof setting Jan 10 10:40:57 oh it seems it's actually called ensure_newline_at_eof_on_save Jan 10 10:42:03 although it sounds like this still doesn't prevent it from showing this as if it's a blank line at the end Jan 10 10:42:50 oh where I'm reading this is 7 years ago, so take with grain of salt Jan 10 10:43:49 also never mind, it seems that "#includedir" (with the #) is just how sudo writes the directive, even though it also uses # as comment char Jan 10 10:43:58 that's... an interesting design choice Jan 10 10:44:21 also you still omitted the ls -l /etc/sudoers.d/allowreboot Jan 10 10:45:22 -r--r----- 1 root root 275 Jan 10 10:17 /etc/sudoers.d/allowreboot Jan 10 10:45:31 ok, should be fine Jan 10 10:46:02 ... Jan 10 10:46:15 well, NOW it rebooted without asking for password Jan 10 10:46:27 dunno what i've touched Jan 10 10:51:06 now, should this allow my app to reboot? Jan 10 10:51:59 it will allow any process running as user 'debian' to execute 'sudo reboot' Jan 10 10:52:31 ok. Jan 10 10:52:37 time to debug, then Jan 10 10:52:39 thnsk Jan 10 10:57:43 oh wtf debian, why are you still using polkit-0.105 (released 2012-04-24) even in buster... hell even still in sid Jan 10 10:59:12 oh, that's why Jan 10 11:01:18 listening Jan 10 11:01:20 :) Jan 10 11:01:38 newer versions depend on a javascript interpreter for executing policykit rules Jan 10 11:02:06 https://bugs.launchpad.net/ubuntu/+source/policykit-1/+bug/1086783 has some recent discussion Jan 10 11:03:42 of course from a safety pov it helps a lot that it'll only execute trusted javascript code, but it still seems like a rather heavyweight dependency for such a core system component Jan 10 11:05:38 interesting, I was wondering what this "duktape" thing was they mentioned, and it's an embeddable javascript engine that "can run on platforms with 160kB flash and 64kB RAM" Jan 10 11:07:07 @zmatt: oh i see, actually thought it was decorative. yeeppy it worked! Jan 10 11:07:25 m-blasie: that was a very silly thing to think Jan 10 11:08:10 it's the path to the source tree... and since the command is run in a subdirectory ("build") of the source tree, that path is the parent directory, i.e. ".." Jan 10 11:09:14 am sorry about that Jan 10 11:09:38 thanks alot you have been very helpful Jan 10 11:44:29 @zmatt: sorry to border you again, is there a special way to compile the samples or am i doing something wrong? i compile them like this "gcc program-source-code.c -o executable-file-name" and i get this fatal error: MQTTClient.h: No such file or directory #include "MQTTAsync.h" and several other samples keep pointing at different headers with same error . i manage to compile one by adding ../ Jan 10 11:44:32 before the header just like this ("../MQTTClient.h") but instead i got several undefined variables. Jan 10 11:45:22 don't modify your source code to try to work around something being installed wrong Jan 10 11:45:35 do however use instead of "MQTTAsync.h" since it's a system header Jan 10 11:47:12 although either should work Jan 10 11:48:08 yeah I can compile the examples just fine and they use "" rather than <> ... (but <> is more correct) Jan 10 11:53:38 oh god what is this horrible library... it defines global symbols with names such as "mymalloc", "isBlack", "bufchar", "ts", "state", and the prize-winner, "s" Jan 10 11:53:51 is there no mqtt library available that sucks less? Jan 10 12:02:22 okay maybe i shold just start all over by deleting the whole library and then reclone it to see if that solves my issues Jan 10 12:02:39 it just sounds like you didn't install it, or installation failed Jan 10 12:03:00 that's the only plausible reason for the compiler to be unable to find the headers Jan 10 12:06:49 i quit sure it did installed or i might have actually just been doing the wrong way. i would just start all over and follow the steps you gave me previously from here https://pastebin.com/nFsgfSh4 Jan 10 12:09:08 the last output lines from "make install" show it's installing headers into /usr/local/include/ Jan 10 12:09:24 that directory is in the compiler's search path for #include Jan 10 12:26:47 note that you'll need -lpaho-mqtt3c or -lpaho-mqtt3a (the latter seems to be for the async API) when linking programs that use the library or you'll get errors for undefined symbols. (but this is separate from your issue of not being able to locate the headers) Jan 10 12:27:03 e.g. gcc -Wall -o MQTTClient_publish /usr/local/share/doc/paho.mqtt.c/MQTTClient_publish.c -lpaho-mqtt3c Jan 10 12:30:36 yea i just tried installing that now and i got several errors for undefined variables, i actually saw this errors initially but i actually thought it did installed(am so dumb) afterwards. Jan 10 12:31:55 generally you'd use a Makefile containing LDLIBS = -lpaho-mqtt3c Jan 10 12:32:04 (or -lpaho-mqtt3a) Jan 10 12:33:41 am actually lost, how do i go by this Jan 10 12:33:47 │13:31:55 @zmatt | generally you'd use a Makefile containing LDLIBS = -lpaho-mqtt3c Jan 10 12:34:38 I mean, that's just literally it, put that line in your Makefile, but I'm definitely not in a mood to give a tutorial on makefiles Jan 10 13:01:48 @zmatt: can you please take a look at this? https://pastebin.com/YtbSt8dj Jan 10 13:02:24 I don't understand what I'm looking at Jan 10 13:03:02 this looks like a makefile of paho.mqtt.c itself, why are you modifying that? Jan 10 13:03:10 this makes no sense Jan 10 13:03:23 it would add the library as a dependency to itself Jan 10 13:10:05 that single LDLIBS line is the only paho-mqtt related line your makefile needs Jan 10 13:10:11 anyway, afk Jan 10 13:12:05 am actually getting you wrong, i thought you'd meant i added that line to the make file Jan 10 13:12:08 13:34:38 @zmatt | I mean, that's just literally it, put that line in your Makefile, but I'm definitely not in a mood to give a tutorial on makefiles Jan 10 13:17:40 i have been on this all day and it seem to be so easy with your explanation so far but i just cant get it right(i feel so stupid right now) Jan 10 14:02:41 m-blasie: m-blasie https://pastebin.com/raw/tFhpu6sL Jan 10 14:03:39 this is a rather minimal makefile, in practice dependency tracking would be nice, but it gets the job done Jan 10 14:19:53 in my own projects I have a "common.mk" file containing project-independent definitions: https://pastebin.com/raw/JB09pYp0 which I'd then include into the Makefile like this: https://pastebin.com/raw/9rNrKgWH Jan 10 14:22:53 or, since there are four variants of this library, I may be inclined to specify the library per program (so you can have multiple programs built from the same makefile that use different library variants, e.g. if you want to experiment with both MQTTAsync and MQTTClient to see which you prefer): https://pastebin.com/raw/t2bUTaiS Jan 11 02:07:57 Is this fan the correct one for cooling the mcu on the BBAI: https://www.digikey.com/product-detail/en/digi-key-electronics/X15FANKIT/X15FANKIT-ND/5822502? Jan 11 02:09:28 I am actually trying to find one that fits the screw holes in the heatsink. I measured but who knows these days? Jan 11 02:12:12 you mean "soc", not "mcu" :P Jan 11 02:12:18 I went to the wiki. The wiki...oh. Jan 11 02:12:22 Yea. SoC! Jan 11 02:12:54 Are we sure it is not the MCU? Jan 11 02:13:11 there's no mcu on the BBAI Jan 11 02:13:15 Oh. Jan 11 02:13:17 Sheesh. Jan 11 02:13:39 and microcontrollers generally don't need a heatsink, let alone a fan :P Jan 11 02:14:13 So, what is the cpu on the SoC called and is this the issue the AI heat? Jan 11 02:14:30 Or...are people kind of over it? Jan 11 02:14:49 I don't think it's been positively identified which part of the SoC is generating the most heat... I think it's just everything combined Jan 11 02:14:59 Oh. Jan 11 02:15:00 Okay. Jan 11 02:15:01 regardless, you're cooling the entire SoC anyway Jan 11 02:15:05 Right! Jan 11 02:15:12 So, that fan should work, heh? Jan 11 02:15:37 no idea, didn't look Jan 11 02:15:45 It was on the wiki as it works for alpha testers but that it did not fit the three pin plug. Jan 11 02:16:11 I was just going to get some 10 x 10mm fan and put it into vdd_5v and GND. Jan 11 02:16:29 Or is it 15 x 15mmm? Jan 11 02:16:35 I need to look again. Jan 11 02:18:09 The mechanical is not on the fan. This means, measure once and cut twice. Jan 11 02:18:10 Ha. Jan 11 02:18:56 It is either 15 x 15 or 10 x 10. Either way, I need to order the fan, i.e. as I want to try to get keras and tensorflow on the AI. Jan 11 02:19:46 I just need to get it on the board first. I see keras works on Ubuntu 16.x + but might not be available on Debian. Jan 11 02:20:53 Buster only, I see. Jan 11 02:20:57 the faq links to this one: https://www.coolerguys.com/collections/usb-fans/products/25mm-25x25x10-usb-fan Jan 11 02:21:03 and element14 links to https://www.element14.com/community/view-product.jspa?fsku=2856031&nsku=56AC4837&COM=noscript Jan 11 02:21:27 Oh. I found the github.com wiki for the AI links to the digikey site I mentioned here. Jan 11 02:21:31 and json is working on a fan-cape https://17376924468624948387.googlegroups.com/attach/55bfa4abfb0c5/image1.jpeg?part=0.1&view=1&view=1&vt=ANaJVrFZrqmDBTXObr1o5nOTfoG38E5pijP_Ef9GJv1Jw-t6SMbylOKg4rd9OipG2QfB50GHwAUIXw2rkqOvzbY5And-SNqBd-XUii79esZl9wgO-kZEMKQ Jan 11 02:21:35 *jason Jan 11 02:21:45 Aw. Jan 11 02:22:50 I see this Cape is available too: https://aisler.net/MAYERMAKES/playground/gerbers2-6b6f9cdf-1fcf-456e-8bb5-4b4728565f6a/board?timetravel=273748. Jan 11 02:25:16 That photo shows a large fan for the AI. That should move some air. Jan 11 02:25:51 I started to look for liquid cooled adapters. I came close but no cigar. Jan 11 02:26:10 It was dedicated to that Rpi only. So, I just did not bother. Jan 11 02:27:29 Every time I see that photo on google groups from rcn, I think of all that heat from the AI and how a enormous fan will do it well. Jan 11 02:29:08 For now, I got some cheap-o, depot fans and I have been afraid to connect them. They are usb fans. One fellow on google groups mentioned usb fans. So, I got two very-very inexpensive usb fans. Jan 11 02:29:17 How would you connect them? Cut the cord and plug Jan 11 02:29:22 'em in? Jan 11 02:30:10 My nice fan had three wires. I could actually control it instead of it just running. Jan 11 02:33:41 I mean...even if people wanted to make some mechanical S-Shaped screws w/ threaded ends on both ends, adding larger fans might be available. Even if that cuts into the niceness and complexity of Cape making and funding. Jan 11 02:34:17 I will get the cape but please make it cheap or liquid cooled. Jan 11 02:38:11 Is there going to be a temp. sensor on the new fan cape? Jan 11 02:43:17 https://www.robotshop.com/en/beaglebone-black-adapter-arduino.html is something I just found. Odd? Jan 11 02:45:57 why would there be? that sounds kinda pointless Jan 11 02:46:54 So, people can say when it reaches 108 degrees, hit that fan on! Jan 11 02:47:06 say = tell the 'puter Jan 11 02:49:09 the BBAI already has a bunch of temperature sensors, both inside the SoC and on the board Jan 11 02:49:33 What the heck? Are you serious? So, I can use the... Jan 11 02:49:45 So, I can use the AI to tell temp. on the board? Jan 11 02:49:46 a temperature sensor on the cape could basically only measure the air temperature, which isn't very useful Jan 11 02:49:53 Oh! Jan 11 02:49:56 That makes sense now. Jan 11 02:50:11 try this script for example to show temperature measurements: https://pastebin.com/Cj1hQuUr Jan 11 02:50:15 I could put the sensor on the board. W/ kapton tape. Jan 11 02:50:16 Okay. Jan 11 02:52:23 temp.sh: 4: temp.sh: shopt: not found <<<< Does that answer make sense? Jan 11 02:54:09 Do I save it as a .sh file or a .bsh file? Jan 11 02:55:13 I've never heard of a ".bsh file" Jan 11 02:55:30 but the first line ensures it'll get executed by bash when run Jan 11 02:55:36 Okay. Jan 11 02:55:39 Me neither. Jan 11 02:55:58 So, did you see the answer? Jan 11 02:56:05 ? Jan 11 02:56:05 4! Jan 11 02:56:17 I saved it as temp.sh. Jan 11 02:56:28 Should I just save it as temp? Jan 11 02:56:56 as the title of the paste shows I called it log_thermal.sh but filename is completely irrelevant, you can call it anything you want Jan 11 02:57:05 Right. Jan 11 02:57:07 Okay. Jan 11 02:57:13 How should I run it? Jan 11 02:57:17 sh temp.sh? Jan 11 02:57:20 no Jan 11 02:57:23 Oh. Jan 11 02:57:28 My bad. Jan 11 02:58:02 as the first line shows, it's a bash script, not an sh script. if you just make the file executable (chmod a+x) then you can simply run it with ./whateveryoucalledit Jan 11 02:58:18 Oh. Okay. I did not think that far down the line yet. Jan 11 02:58:19 or you can run it with bash whateveryoucalledit then you don't need to chmod it Jan 11 02:58:29 bash. Nice! Jan 11 02:58:29 whichever you prefer Jan 11 02:58:31 Okay. Jan 11 02:59:04 uptime=1146.16 idle=1693.20 cpu=36.600 gpu=37.000 core=37.400 dspeve=35.800 iva=37.000 loadavg=(0.09 0.30 0.53) Jan 11 02:59:08 Does this sound right? Jan 11 02:59:50 it sounds like your board is not even lukewarm, why are you trying to add a fan? Jan 11 03:00:49 B/c...I have a full blast fan on right now. **** ENDING LOGGING AT Sat Jan 11 03:00:52 2020