**** BEGIN LOGGING AT Sun Jan 17 02:59:57 2021 Jan 17 03:13:09 Okay. I am in ~/linux/drivers/pwm and reading it w/ a text editor. Jan 17 03:13:51 I am trying to wrap my head around it all and I am in the file pwm-tiehrpwm.c. Jan 17 03:15:08 So, it seems one has to account for the on/off factor, i.e. HIGH/LOW on a counter. Then, period and duty. Jan 17 03:16:13 Now, I will need to read about the ABI more, i.e. as I do not have a clue as to what AQCTLA or other ideas mean... Jan 17 03:19:43 How many registers does the am335x have on it for pwm? Jan 17 03:20:37 why would you need to know or care about this? it's of no relevance for using pwm outputs on linux Jan 17 03:21:51 Right... Jan 17 03:21:53 But... Jan 17 03:22:13 I am reading the output of that .c file to understand how to promote a lib for pwm. Jan 17 03:22:32 It seems that the fellows had a "good first issue" for people. Jan 17 03:22:44 On github.com for a pwm lib. Jan 17 03:22:47 Let me show you. Jan 17 03:23:13 https://github.com/beagleboard/Latest-Images/issues/49 and I want to pitch in. Jan 17 03:24:11 There was a fellow that already got to it. He knows more than me. Obviously. Jan 17 03:24:17 But... Jan 17 03:24:26 I would still like to try. Jan 17 03:27:13 I mean...I can see where it is reading from and reading to. Jan 17 03:27:34 @zmatt: Do you know of what any of those ABI functions mean? Jan 17 03:28:35 The pwm.c file is good. Right but there are so many explanations of "RED_LETTERING" and it is not easy to understand as of yet. Jan 17 03:36:22 What libs. are useful in Python to control the PWM interface API and is Python smart enough to handle it? Jan 17 03:38:29 Oh and do I need to make a wrapper from C/C++ to Python or does python or cython understand it? Jan 17 03:45:13 So, cython is just a python to c compiler. I need it the other way around. C to python. Gosh. Jan 17 03:46:03 Anyway...our day is almost over. So, w/out further ado, outie! Jan 17 04:00:56 10:00 Jan 17 04:54:44 Hello? Jan 17 12:06:59 https://pastebin.com/x7QDZyRZ is my issue. It seems my board will not connect online. Is there anything I can do on my side of things? Jan 17 12:26:39 I am using the BBBW w/ kernel 4.19.x. Jan 17 18:03:37 Did something happen w/ update/uprade w/ apt from rcn-ee.com/net or whatever? Jan 17 18:04:40 bzzt, bzzt! Jan 17 18:05:10 Anyway...waiting is my specialty (sort of). Jan 17 18:13:31 Well...it works now! Jan 17 18:13:35 Nice! Jan 17 18:14:18 Why would something not work during the wee hours of the morning and now works just fine, i.e. getting apt to update/upgrade? Jan 17 18:16:27 You guys? Jan 17 18:16:40 Heh? you guys? Jan 17 18:28:11 in stereo where available. Jan 17 18:30:06 Oh. Jan 17 18:30:08 Nice! Jan 17 18:30:18 GenTooMan! Welcome aboard, sir! Jan 17 18:36:07 umm standing on land but OK. I can see your state has a thing with swamps and bayou’s so a boat is preferable to be on then in the swamp. Jan 17 18:36:33 Swamp "Thang?" Jan 17 18:36:43 B flicks and BBBs! Jan 17 18:36:56 I cannot wait until I catch something else besides skinks. Jan 17 18:37:22 GenTooMan: I am working w/ some SDK and some servos now. Jan 17 18:37:28 W/ the BBBW. Jan 17 18:37:44 It is not so easy for me. The SDK is tricking me (as usual). Jan 17 18:38:02 So, as persistent as you know I am, I reached out! Jan 17 18:38:17 Waiting as usual over here? Jan 17 18:39:04 set the reason they don't work in the wee hours is because you probably weren't fully awake Jan 17 18:42:41 Oh. Jan 17 18:42:59 Sleepy time in the wee... Jan 17 18:43:05 For my board too. Jan 17 18:43:27 I am updating the kernel now. Jan 17 18:43:35 Hopefully, this will fix my current crisis. Jan 17 18:43:56 I am in hamper down mode, Linux style. Jan 17 18:45:03 so all the cloths are in one basket? :D Jan 17 18:45:18 Every egg has been cracked! Jan 17 18:45:30 Well, the good ones. Jan 17 18:46:22 GenTooMan: Did you see the BeagleV? Jan 17 18:46:34 RISC-V for the win! Jan 17 18:47:02 Not all that popular from what I have seen but neat nevertheless. Jan 17 18:47:28 I looked up the details and tried to find ideas. Jan 17 18:47:38 I came across two specific papers. Jan 17 18:47:55 One was about 200 pages and another one was about 82 pages. Jan 17 18:48:09 It is going to be neat to try something new. Right? Jan 17 18:51:20 FPGA next hopefully! Jan 17 18:52:09 What is that fastio stuff in the kernel? Jan 17 18:55:50 rapidio, sorry. Jan 17 19:01:26 signing off for now. BBL! Jan 17 19:18:50 i'm trying to use KMS applications and Picture-in-picture features of a monitor at the same time. If I start one while picture-in-picture is not active, wait until I see the application OpenGLES rendered stream, then turn on picture-in-picture, nothing crashes and I can see the application in the small picture-in-picture monitor HUD. However, whenever I FIRST enable Picture-in-picture and THEN try to execute a KMS Jan 17 19:18:50 application, drmModeSetCrtc() returns Invalid argument, no matter which configuration I try. same behavior on two monitors with picture-in-picture features Jan 17 19:21:29 is there any obvious reason why initializing a DRM framebuffer would not be possible when the monitor connected to HDMI/DVI has PIP enabled, but enabling PIP afterward would not destroy the framebuffer? Jan 17 19:22:24 Hi people Jan 17 19:23:43 I have a question. May somebody kindly answer it. Jan 17 19:24:04 tty without accessing any GPU appears to work as expected regardless of PIP Jan 17 19:24:34 misternumberone: I'm confused.. what does picture-in-picture actually involve? you're sending two video streams to the monitor or something? (it's news to me that HDMI or DVI support this) Jan 17 19:24:44 JRAB: easy way to find out is by asking it Jan 17 19:25:08 or as the channel topic says: "don't ask to ask, just ask" Jan 17 19:27:23 I have a beaglebone black taken off an antminer. I've been able to boot it using a sdcard with debian 8 and debian 9, but no luck with debian 10. Jan 17 19:27:52 the thing on antminers isn't a beaglebone black Jan 17 19:28:20 it's some stripped-down unofficial derivative Jan 17 19:28:23 the thing on antminer IS a light version of beaglebone black. Jan 17 19:28:32 right, so it's not a beaglebone black Jan 17 19:29:07 zmatt: PIP is a feature provided by monitors that displays two input streams on the same panel at the same time yes, there is a lack of good images online of what I actually see but this illustrates it pretty well https://displaysolutions.samsung.com/static/images/feature/monitor/2018_SJ55W_feature_04_2.jpg Jan 17 19:29:40 JRAB: as for booting newer systems from sd card: on an actual BBB my assumption would be there's an old bootloader on eMMC that is now too old to understand how to boot modern images Jan 17 19:29:57 anyway, has somebody succesfully started the thing on antminers, which i call beaglebone black light, using debian 10? Jan 17 19:30:29 typical solutions to that are 1. flash to eMMC instead of booting from sd card 2. wiping eMMC entirely 3. holding the S2 button down at power-on to bypass the bootloader on eMMC Jan 17 19:31:07 there's no such thing as a "beaglebone black light", and "beaglebone" is a trademark that can't be used without license Jan 17 19:31:31 zmatt, the eMMC on "the thing" is too small to hold the operating system. Jan 17 19:31:53 the beaglebone is open hardware which means companies are free to create derivatives for their own use, but those are not "beaglebones" and cannot be called such Jan 17 19:32:02 right, so that's a problem specific for these antminer devices Jan 17 19:32:14 also, there is no S2 button on "the thing" Jan 17 19:32:46 if I try to initialize KMS+DRM applications while the input to the display is in the monitor's "SUB" configuration, it fails, but if I launch the application first and THEN use the monitor's firmware control buttons to swap the HDMI/DVI input into the SUB configuration, the KMS application continues to run and appears in the small view on the panel Jan 17 19:32:50 also, eMMC too small to hold the OS ? how large is that eMMC ? is it actually eMMC ? Jan 17 19:33:52 zmatt, it thing it isn't an eMMC, actually. Jan 17 19:34:16 so, it's not flashable. Jan 17 19:34:18 right, so this device is hardware-incompatible Jan 17 19:34:53 so, i wonder why it is ok with debian 9 and not debian 10. Jan 17 19:35:39 that will depend on the bootloader in its NAND flash (which I assume it actually has? I vaguely recall seeing devices like these with nand flash) Jan 17 19:36:12 most likely it also didn't properly boot the debian 9 images (in the sense of correctly applying device tree overlays) Jan 17 19:38:21 if you want to boot from SD card maybe you can manage to reconfigure its boot mode to bypass/ignore the NAND and just boot from SD card directly, but that depends on whether the relevant sysboot pins are accessible so you can override them (on beaglebones they're P8.31-46) Jan 17 19:38:23 the boot loader looks for an uEnv.txt file at the root of the sdcard. if it is present, "the thing" boot from the sdcard. it works that way for any debian version. so, i don't think it's a bootloades issue. Jan 17 19:38:56 the variables defined in /boot/uEnv.txt and the bootloader's actions based upon them have changed over time Jan 17 19:39:14 so even if it manages to understand enough to boot the system, it doesn't necessarily boot it _correctly_ Jan 17 19:39:27 i'll take a look at that. Jan 17 19:39:53 note also that using the wrong DT (one for the beaglebone black rather than one for this hardware) means it's pretty much pure luck whether or not you're causing hardware damage Jan 17 19:41:30 well, i've been using "those things" from quite long time without any issue. so, i've been lucky. Jan 17 19:42:02 that's possible, it depends on what pins it uses for what and whether or not that results in a drive conflict Jan 17 19:43:40 changing the sysboot config to bypass NAND is probably your best bet to move forward, since you probably won't be able to update the bootloader in NAND (at the very least it would require building a custom one) Jan 17 19:44:07 I'll take a closer look at the uEnv.txt variables. thanks for the enlightenment. Jan 17 19:44:36 I mean there's not much to "look at" ... the bootloader has seen huge changes since debian 8 Jan 17 19:45:18 I see. Jan 17 19:45:29 so there was already incompatibility between the bootloader and debian 9, but not severe enough to completely prevent booting the system (though no doubt enough to result in a system that doesn't work properly as it was designed to) Jan 17 19:45:57 and based on reports I've seen it does seem like debian 10 images are not bootable at all anymore by these ancient bootloaders Jan 17 19:46:21 the sd card image includes a suitable bootloader, the problem will be getting this device to use that instead of the one on NAND Jan 17 19:46:34 No good news, then. Jan 17 19:46:57 I don't suppose you have a schematic of this thing? :P Jan 17 19:47:18 You suppose well. Jan 17 19:48:16 well that sucks :) of course it's still possible to determine everything you need to know, but not having a schematic means that process will be annoying and time-consuming Jan 17 19:49:04 basically, since these boards are not supported by the beagleboard.org images nor by their manufacturer, you're kinda on your own Jan 17 19:49:26 I'm gonna stick with debian 9. Jan 17 19:49:36 (I hope you got them cheap) Jan 17 19:49:59 but i'll keep digging about deb10. Jan 17 19:50:29 does it have P9/P8 expansion headers like the BBB ? Jan 17 19:50:59 i got them for free. someone's farm leftover. Jan 17 19:51:25 yes, it has p9/p8. Jan 17 19:51:49 free sounds like the right price for them, considering the headache they cause ;) Jan 17 19:52:35 i run some fun experiments on them. nothing serious. Jan 17 19:53:13 okay so you could use P8.31-46 in gpio mode to determine 1. the current sysboot config (which will be set by pull-up/down resistors on the board) 2. if they are indeed connected to the P8 header, in which case you'll be able to override them that way Jan 17 19:54:25 got that Jan 17 19:54:47 see https://goo.gl/Jkcg0w#gid=1847985463 for which pin is what on the BBB, and the Boot tab ( https://goo.gl/Jkcg0w#gid=750891223 ) to interpret the sysboot values Jan 17 19:56:34 nice sheets Jan 17 19:56:41 thanks Jan 17 19:59:15 i'm gonna take a look at them. taking off. thanks. Jan 17 20:02:29 zmatt: not sure who pdp7 is but i read the discussion clip and definately agree Jan 17 20:02:48 last time i heard pdp7 it was a old dec machine Jan 17 20:02:50 haha Jan 17 20:03:24 buzzmarshall: https://twitter.com/pdp7 Jan 17 20:03:35 also, shhhh, he's here ;) Jan 17 20:03:44 lol Jan 17 20:04:09 i agree tho i think things need to be done diffently and not everything needs to be mainstream Jan 17 20:04:41 been looking greybus as well for awhile for something else Jan 17 20:05:04 i like the udev and tree thinking Jan 17 20:08:30 yeah and human-written DTs have an advantage that automatic discovery doesn't: I know what the attached devices are, what their purpose is, and can give them identification that is meaningful, which can then be used by udev to create symlinks Jan 17 20:08:43 i never really thought that much about gpio stuff till i read what you said Jan 17 20:09:24 neither have the gpio subsystem maintainers apparently ;P Jan 17 20:10:23 most of the older embedded crap ive messed with over the years tended to rely on propriatary stuff and with the advent of device trees i just never really thought much about it outside of my own narrow needs Jan 17 20:11:33 as well systemd is still a ongoing learning thing and maybe this would help push me more in that direction Jan 17 20:12:14 but you make great arguments in what i read and can see how it would make things so much easier Jan 17 20:12:35 specially if your making pcbs and altering things Jan 17 20:12:35 systemd itself doesn't really have much to do with any of this though... udev may be part of the systemd umbrella project but it's still a fairly separate thing Jan 17 20:12:45 k Jan 17 20:12:55 (and udev existed before systemd) Jan 17 20:14:10 ya i realize that but systemd with its ability to control services differently especially in parallel mode i could see making thing easier Jan 17 20:15:01 for years i just had no real reason to move off the old init's other then being forced to follow along with the mainstream distro's Jan 17 20:15:18 but i never really pushed to learn it very deeply Jan 17 20:15:47 the main interaction between systemd and udev is that you can have services depend on devices, to ensure the service isn't started until the device is actually usable (kernel has detected device, probed the driver, udev has performed the applicable actions for it) Jan 17 20:16:30 and conversely udev can request a systemd unit to be started when a device is detected Jan 17 20:16:53 ya thats where i can see the use Jan 17 20:18:29 since the dtb advent ive always had it in the back of my mind to find a way to add intellence to the idea of it similar in concept to how usb enumerates its devices Jan 17 20:18:45 just never had a real reason to deeply pursue the idea Jan 17 20:19:44 now that you make me think about gpio's i can see alot of real use of something along them lines Jan 17 20:19:53 specially with the embedded stuff Jan 17 20:22:07 and not just gpios, for any type of DT-declared platform device that creates a device node for userspace use probably benefits from a convenient symlink Jan 17 20:22:32 being i do mostly all my own thing or just within a small group of us things like the userland/kernel split was never really much of a concern so i tended not to worry about the separation that dev's in the public world are constrained with Jan 17 20:22:47 e.g. our software uses /dev/spi/dsp and doesn't care what chipselect of what spi bus it's attached to (which actually varies between hardware revisions of our pcb) Jan 17 20:23:46 i caught that point in the discussions which is a real good point Jan 17 20:23:58 specially for peeps like you and what you do Jan 17 20:23:59 (for that purpose I have a symlink = "spi/dsp"; property on it in DT, which is used by an udev rule to create that symlink... that udev rule is also part of current beagleboard.org images) Jan 17 20:27:57 after going thru the discussion it helped explain what a kernel patches i had seen on various git's Jan 17 20:28:42 i will definately put more thought into this whole road as i think even for my own purposes i can see nothing but positives in it Jan 17 20:29:50 i never worry bout upstream contributions and meeting their qualifications and this is something i would definately like to take a shot at Jan 17 20:30:16 and personally i think things like gpios need to have more control in userland Jan 17 20:31:05 yeah, in embedded systems it's very normal to want to essentially connect a wire into software Jan 17 20:32:23 and its everywhere these days and only getting worse Jan 17 20:33:14 to be honest as much as i live in my own sheltered area i kinda had taken it for granted it was somewhat already fixed up as you suggested and i was just not up to date on stuff Jan 17 20:37:47 thanx for enlightening me tho in the subject as its got me thinking Jan 17 21:41:05 yes I am here ;) Jan 17 21:41:36 I just pretend to be a DEC minicomputer on the internet Jan 17 21:43:15 do you have any DEC computers? Jan 17 22:00:51 funny how mini computer used to refer to something that was easily over a 1000 lbs Jan 17 22:01:13 they were tiny compared to what came before Jan 17 22:01:30 true Jan 17 22:02:09 see also: everything named "new" something Jan 17 22:02:35 they only took up part of a room rather then the whole room or building Jan 17 22:05:29 no, I just picked the name back in college because I liked uniux history :) Jan 17 22:05:49 though I did get to see a PDP-7 in Seattle at the computer history museum Jan 17 22:05:58 I do have a couple of dec machines Jan 17 22:06:02 nothing that old though Jan 17 22:06:02 they have one running first edition of Unix Jan 17 22:18:04 ive only seen one in pieces once a long time ago as it was being scraped Jan 17 22:18:32 did last year pitch out a couple of old dec alphas i had laying around from the old days Jan 17 22:18:52 them and a couple of old sparcs Jan 17 22:19:08 never had a sparc Jan 17 22:20:27 they were in a couple of old sun workstations Jan 17 22:20:52 I've used sun/sparc machines Jan 17 22:20:59 never owned one Jan 17 22:22:28 most probably wouldnt, unless you were around old corporate crap or maybe in univeritys in the old days Jan 17 22:22:59 I picked up some dec alphas at uni Jan 17 22:23:04 and elsewhere Jan 17 22:23:30 the EE department was all dec Jan 17 22:23:38 compsci was sun Jan 17 22:23:43 they were quite the thing back then as some of the early risc based units Jan 17 22:24:12 for a little while, the fastest x86 was one emulated on an alpha Jan 17 22:24:58 ya they were all the rage when they came out Jan 17 22:25:22 it was a nice architecture Jan 17 22:25:47 well, once they added the bwx instructions Jan 17 22:25:56 sparse mapping was no fun at all Jan 17 22:26:57 i find it truly amazing at how far computer tech has come in a relatively short time Jan 17 22:27:10 and every year things just keep gettin smaller Jan 17 22:27:28 mru: lol, it initially didn't have byte/halfword loads/stores Jan 17 22:27:34 that's rough Jan 17 22:29:04 ew, it was a 32-bit arch that initially didn't even support 16-bit memory access, yet it defined "word" to mean 16 bits? wtf Jan 17 22:29:06 for normal memory you'd load the containing word and extract the desired byte using a special instruction Jan 17 22:29:22 alpha was always 64-bit Jan 17 22:29:46 the real fun was dealing with mmio regs smaller than 32 bits Jan 17 22:29:46 64-bit even Jan 17 22:30:19 that's where the sparse mapping came in Jan 17 22:30:46 like the component identifier in ARM peripherals... one byte per register Jan 17 22:31:48 (because APB also doesn't have byte-enable lines, hence identification registers were limited to the narrowest width supported by APB, i.e. 8-bit, even though I've never seen APB instantiated at any width other than 32-bit) Jan 17 22:32:12 and then there was the trick for byte-reversing a word Jan 17 22:32:55 load/store at address | 3 or something? Jan 17 22:33:09 or that might have been powerpc Jan 17 22:33:15 using the vax float compat instructinos Jan 17 22:33:29 vax being mixed endian Jan 17 22:33:31 I think I was happier not knowing Jan 17 22:40:42 of course, it all pales in comparison to r-pi Jan 17 22:53:45 lol Jan 17 23:49:36 mru: yeah especially the first rpi, where the AXI interface on the arm core didn't have semantics quite matching the interconnect, which meant that under certain circumstances two sequential load instructions could get their data results swapped Jan 17 23:53:46 (because the CPU could issue a second load tagged with the same ID before the first one has completed and simply required in-order completion for those, while the interconnect had no way of ensuring this thus whichever read-response was received first would complete the load that was issued first.) Jan 17 23:54:19 I do not envy whoever had to debug that Jan 17 23:54:36 hehehe Jan 17 23:55:09 workaround: insert memory barrier between loads from different peripherals Jan 17 23:55:28 the weirdest hw quirk I've debugged was a case of conditional branches sometimes going the wrong way Jan 17 23:55:40 branch predictor and interrupts not getting along Jan 17 23:56:01 sounds "fun" Jan 17 23:56:59 it's hardly ever a hw bug, but when it is, it's always "fun" Jan 17 23:57:57 mru: I also wonder how something like this is found: https://liktaanjeneus.nl/a15-erratum-798870.html Jan 17 23:58:42 by running the wrong code Jan 17 23:59:14 while the L2 memory system has exactly the right sort of congestion Jan 17 23:59:39 said code would be creating those conditions Jan 18 00:01:56 perhaps Jan 18 00:03:04 I'm glad the last cortex-a8 revision (r3p2, the one used in the AM335x) doesn't have any serious errata Jan 18 00:04:44 I ran across a few fun ones in whatever rev the omap3530 had Jan 18 00:04:51 r1p2 or something Jan 18 00:05:31 there was one where stale branch predictor data could make it flip between arm and thumb mode Jan 18 00:05:59 solution: use only one or the other in your code Jan 18 00:06:10 meaning use only arm Jan 18 00:06:21 since the dynamic linker trampolines are always arm regardless Jan 18 00:06:45 basically the only one that's userspace-triggerable for r3p2 is that vcvt.f32.u32 (convert unsigned int to float) with argument 0xffffff01 in "fast mode" (Default-NaN, flush denormals to zero) with rounding mode set to "round to +infinity" gives the wrong result Jan 18 00:06:56 yeah older versions had some nasty stuff Jan 18 00:08:44 dynamic linker trampolines always arm? you sure? Jan 18 00:08:55 since that sounds dumb Jan 18 00:09:40 armv7 allows even the exception vectors to be in thumb mode... there's no actual reason to ever execute a single instruction in legacy arm mode on an ARMv7 core Jan 18 00:10:42 they were then, at any rate Jan 18 00:11:52 and then there was the one where a wide thumb branch instruction straddling a page boundary could go do strange things Jan 18 00:17:29 wow it seems PLT stubs are still ARM-mode... I wonder if there's a flag to fix that Jan 18 00:17:40 what's the harm? Jan 18 00:19:21 it just seems weird to continue to generate a mix of old ARM code and Thumb2... and Thumb2 may be more compact Jan 18 00:19:39 for those particular code sequences, probably not Jan 18 00:21:06 hm, I guess not Jan 18 00:21:29 and there's no penalty for mode switching Jan 18 00:21:55 what if someone wants dynamic linking on a cortex-M ? Jan 18 00:22:14 then they're screwed Jan 18 00:22:48 is fdpic still a thing? Jan 18 00:23:12 ? Jan 18 00:24:27 an elf hack to sort of make dynamic loading work on mmuless machines Jan 18 00:24:29 ah a pointer to global data... like old Mac OS shared library Jan 18 00:25:13 (for the same reason: address space shared between multiple applications that may want to use the same shared library) Jan 18 00:26:33 I've also wanted that for targeting the dual cortex-m4 subsystem on the am572x and its dual cortex-m3 predecessor, since although that subsystem does actually have an MMU, it's shared between the cores and doesn't distinguish between them Jan 18 00:27:10 (but I never succeeded in convincing gcc to do that) Jan 18 00:58:12 if you do not use the emmc for the OS Jan 18 00:58:17 can you use it in applications Jan 18 01:01:40 what would stop you? Jan 18 01:02:17 lack of knowledge Jan 18 01:02:18 lol Jan 18 01:02:41 it could be used as a fast ssd sort of Jan 18 01:02:42 ? Jan 18 01:07:40 where is your OS? Jan 18 01:07:54 on an SD Jan 18 01:08:15 you'll have /dev/mmcblk0 and mmcblk1 Jan 18 01:08:22 one is the SD, the other eMMC Jan 18 01:08:25 I forget which is which Jan 18 01:08:30 0 is sd, 1 is eMMC Jan 18 01:09:51 I would just format mmcblk1 Jan 18 01:10:48 I am thinking I want to uses it like a circular buffer Jan 18 01:11:13 also I hate virtualbox Jan 18 01:11:23 thing randomly stops working Jan 18 01:11:50 that's true of computers in general Jan 18 01:12:59 how do I block apt from install packages for other architectures Jan 18 01:13:12 I really hate this damned machine / I wish that they would sell it / It never does quite what I want / but only what I tell it Jan 18 01:13:50 that sums it up Jan 18 01:18:24 dpkg: error: cannot remove architecture 'i386' currently in use by the database Jan 18 01:22:39 crap may of messed up my system Jan 18 01:22:50 i just purged all my i386 packages Jan 18 01:37:41 is there a way I can get all the config-pin items etc that is customized on the debian image on another systemd based distribution like archlinux? **** ENDING LOGGING AT Mon Jan 18 02:59:57 2021