| - stites (QUIT: Read error: Connection reset by peer) (~stites@2607:fb91:dc1:d690:17dd:ae60:28fd:17ef) | 00:09 | |
| + stites (~stites@2601:19c:4180:2fa0:2f0a:960b:48ea:b731) | 00:09 | |
| - S0rin (QUIT: Ping timeout: 248 seconds) (~S0rin@user/s0rin) | 00:16 | |
| + S0rin (~S0rin@user/s0rin) | 00:17 | |
| - mtm (QUIT: Ping timeout: 248 seconds) (~mtm@c-71-228-84-213.hsd1.fl.comcast.net) | 00:18 | |
| - mjw (QUIT: Ping timeout: 246 seconds) (~mjw@gnu.wildebeest.org) | 00:29 | |
| - stites (QUIT: Ping timeout: 245 seconds) (~stites@2601:19c:4180:2fa0:2f0a:960b:48ea:b731) | 00:32 | |
| + stites (~stites@2607:fb91:dc1:d690:17dd:ae60:28fd:17ef) | 00:33 | |
| - stites (QUIT: Read error: Connection reset by peer) (~stites@2607:fb91:dc1:d690:17dd:ae60:28fd:17ef) | 00:49 | |
| + stites (~stites@130.44.147.204) | 00:50 | |
| + chomwitt (~chomwitt@2a02:587:7a0c:4d00:1ac0:4dff:fedb:a3f1) | 01:15 | |
| - XYZ (QUIT: Ping timeout: 240 seconds) (~XYZ@37-48-2-145.nat.epc.tmcz.cz) | 01:19 | |
| - chomwitt (QUIT: Ping timeout: 255 seconds) (~chomwitt@2a02:587:7a0c:4d00:1ac0:4dff:fedb:a3f1) | 01:22 | |
| - Gooberpatrol66 (QUIT: Quit: Leaving) (~Gooberpat@user/gooberpatrol66) | 01:30 | |
| - flowy (QUIT: *.net *.split) (~flowy@2a01:4f8:c0c:1a8f::1) | 02:00 | |
| - hl (QUIT: *.net *.split) (~hl@user/hl) | 02:00 | |
| - sigrid (QUIT: *.net *.split) (~sigrid@ftrv.se) | 02:00 | |
| - ethulhu (QUIT: *.net *.split) (ethulhu@nora.ethulhu.co.uk) | 02:00 | |
| - sevan (QUIT: *.net *.split) (~sevan@user/venture37) | 02:00 | |
| - natalie (QUIT: *.net *.split) (~natalie@user/natalie) | 02:00 | |
| - Asmadeus (QUIT: *.net *.split) (~asmadeus@240b:13:8c80:d300::42:13) | 02:00 | |
| - minute (QUIT: *.net *.split) (~mntirc@softboy.mntmn.com) | 02:00 | |
| - pinoaffe (QUIT: *.net *.split) (~pinoaffep@2a01:4f9:c010:a00b:1337:1337:11be:10) | 02:00 | |
| - Boostisbetter (QUIT: *.net *.split) (4a410829d7@irc.cheogram.com) | 02:00 | |
| - qbit (QUIT: *.net *.split) (~qbit@h.suah.dev) | 02:00 | |
| - yankcrime (QUIT: *.net *.split) (~nick@193.16.42.83) | 02:00 | |
| - sknebel (QUIT: *.net *.split) (~quassel@v22016013254630973.happysrv.de) | 02:00 | |
| + yankcrime (~nick@193.16.42.83) | 02:00 | |
| + flowy (~flowy@2a01:4f8:c0c:1a8f::1) | 02:00 | |
| + minute (~mntirc@softboy.mntmn.com) | 02:00 | |
| + qbit (~qbit@h.suah.dev) | 02:00 | |
| + pinoaffe (~pinoaffep@2a01:4f9:c010:a00b:1337:1337:11be:10) | 02:01 | |
| + sknebel (~quassel@v22016013254630973.happysrv.de) | 02:02 | |
| + natalie (~natalie@user/natalie) | 02:04 | |
| + hl (~hl@user/hl) | 02:05 | |
| + sigrid (~sigrid@ftrv.se) | 02:05 | |
| + Asmadeus (~asmadeus@M014009140128.v4.enabler.ne.jp) | 02:05 | |
| + sevan (~sevan@2001:470:1f1d:1d6:5a55:caff:fe24:ed4) | 02:20 | |
| + mtm (~mtm@c-71-228-84-213.hsd1.fl.comcast.net) | 02:24 | |
| - sevan (QUIT: Changing host) (~sevan@2001:470:1f1d:1d6:5a55:caff:fe24:ed4) | 02:31 | |
| + sevan (~sevan@user/venture37) | 02:31 | |
| - laumann (QUIT: Remote host closed the connection) (~quassel@user/laumann) | 03:03 | |
| + laumann (~quassel@2a0a-e5c0-2-2-0-c8ff-fe68-bef1.loves.ipv6.at.ungleich.ch) | 03:05 | |
| - mtm (QUIT: Ping timeout: 255 seconds) (~mtm@c-71-228-84-213.hsd1.fl.comcast.net) | 03:07 | |
| violet | any considerations i need to make when upgrading the SSD in the reform? | 03:26 |
|---|---|---|
| violet | or can i just use anytihng | 03:26 |
| sknebel | theres a thread on the forums where people have posted about specific models and their experiences | 03:28 |
| violet | oh i should check that out | 03:28 |
| violet | i guess il just get the transcend one that seems fine enough | 03:35 |
| + Gooberpatrol66 (~Gooberpat@user/gooberpatrol66) | 03:37 | |
| - truck (QUIT: Server closed connection) (truck@shell.suomiscene.fi) | 03:37 | |
| + truck (truck@shell.suomiscene.fi) | 03:38 | |
| - vagrantc (QUIT: Ping timeout: 246 seconds) (~vagrant@2600:3c01:e000:21:7:77:0:50) | 04:38 | |
| + mtm (~mtm@c-71-228-84-213.hsd1.fl.comcast.net) | 05:13 | |
| - GNUmoon (QUIT: Remote host closed the connection) (~GNUmoon@gateway/tor-sasl/gnumoon) | 05:38 | |
| + GNUmoon (~GNUmoon@gateway/tor-sasl/gnumoon) | 05:38 | |
| + Boostisbetter (4a410829d7@irc.cheogram.com) | 06:37 | |
| + vagrantc (~vagrant@2600:3c01:e000:21:7:77:0:50) | 07:24 | |
| - wstlr (QUIT: Server closed connection) (~wstlr@504e1bef.host.njalla.net) | 07:26 | |
| + wstlr (~wstlr@504e1bef.host.njalla.net) | 07:26 | |
| + ethulhu (ethulhu@nora.ethulhu.co.uk) | 08:24 | |
| - laumann (QUIT: Changing host) (~quassel@2a0a-e5c0-2-2-0-c8ff-fe68-bef1.loves.ipv6.at.ungleich.ch) | 08:48 | |
| + laumann (~quassel@user/laumann) | 08:48 | |
| noam | For what it's worth I picked up a random NVMe drive at a local store and it JustWorked, but I also never thought about it | 09:13 |
| noam | And that's hardly evidence that _any_ would work | 09:13 |
| josch | there are some that definitely do not work (Samsung 970 Evo) | 09:14 |
| josch | you could leave a comment with the working one here: https://community.mnt.re/t/confirmed-working-nvme-drives/224 | 09:15 |
| + mjw (~mjw@gnu.wildebeest.org) | 09:33 | |
| + ggoes_ (~gregf@ool-44c1517f.dyn.optonline.net) | 10:05 | |
| - ggoes (QUIT: Ping timeout: 246 seconds) (~gregf@fsf/staff/ggoes) | 10:08 | |
| * ggoes_ -> ggoes | 10:08 | |
| + XYZ (~XYZ@89-24-49-225.nat.epc.tmcz.cz) | 10:38 | |
| - ec0 (QUIT: Ping timeout: 244 seconds) (~ec0@vps-446f4f39.vps.ovh.ca) | 10:57 | |
| + ec0 (~ec0@vps-446f4f39.vps.ovh.ca) | 11:00 | |
| - mjw (QUIT: Ping timeout: 245 seconds) (~mjw@gnu.wildebeest.org) | 11:10 | |
| - jfred (QUIT: Server closed connection) (sid534649@libera/sponsor/jfred) | 11:30 | |
| + jfred (sid534649@libera/sponsor/jfred) | 11:30 | |
| * mark_ -> mjw | 11:32 | |
| - XYZ (QUIT: Ping timeout: 255 seconds) (~XYZ@89-24-49-225.nat.epc.tmcz.cz) | 11:45 | |
| grimmware | It strikes me that a community wiki would be a good idea for stuff like that | 11:47 |
| josch | or better maintanance of the handbook? i've had https://source.mnt.re/reform/reform-handbook/-/issues/2 open for a long time now because a *lot* of things changed with respect to how the reform is booted since the handbook was written | 11:49 |
| + XYZ (~XYZ@89-24-49-225.nat.epc.tmcz.cz) | 11:51 | |
| minute | josch: yes, and i'm sorry for that! we have a project of overhauling the handbook coming very soon. also because the first print is sold out | 12:16 |
| minute | josch: we will need to rewrite the handbook to accomodate a bunch of changes, incl. the new modules etc | 12:17 |
| minute | also, 14x LS1028A modules are here now | 12:17 |
| josch | wohoo! \o/ | 12:21 |
| grimmware | minute: A community wiki would probably be really helpful not only for having e.g. information on what third party hardware is compatible but as a good point to crowdsource errata and contributions for the handbook in between iterations | 12:21 |
| grimmware | I appreciate that's additional overhead though. | 12:21 |
| minute | grimmware: yes, there were several people who wanted to set one up but none of them kind of went through with it | 12:22 |
| minute | if it were run by community member(s) then we wouldn't be the bottleneck | 12:23 |
| grimmware | yeah that makes sense | 12:23 |
| grimmware | welp I might come back round to it when I get my pocket but realisitically I'm not going to maintain a thing I don't have any use for yet | 12:23 |
| josch | doesn't gitlab provide some wiki thing? | 12:23 |
| minute | just someone with server space and admin experience would need to step up, and appoint some other(s) for community management / making sure that there's no vandalism etc | 12:24 |
| minute | josch: yes, we can also check how that works, but we would need to funnel people through our gitlab account approval | 12:24 |
| erle | i think wikis are where data goes to die lol | 12:24 |
| josch | yeah, there is this "add wiki" button in projects on source.mnt.re -- no idea how good that is | 12:24 |
| violet | realy appreciate the power port being on the side instaed of the back, its like one of my biggest qualms with my thinkpad. | 12:24 |
| violet | constant paranoia about flexing th connector too hard | 12:25 |
| minute | erle: except for the arch wiki! | 12:26 |
| josch | (and gentoo wiki in the past before it broke) | 12:26 |
| grimmware | man I love the arch wiki | 12:26 |
| josch | so, just add a reform article on onto the arch wiki and the problem will solve itself? ;) | 12:27 |
| violet | i could definithely sysadmin a wiki. if someone else handled running it from a community point of view | 12:27 |
| grimmware | hahah | 12:27 |
| erle | josch genius idea | 12:27 |
| violet | also gentoo wiki is still good | 12:27 |
| violet | i always check both arch and gentoo wiki to cross reference | 12:28 |
| violet | gentoo wiki is often more realistic about the limitations of software than the arch wiki i find | 12:28 |
| violet | and stability of the software | 12:28 |
| violet | the thing is that i feel like the forum+handbook combination already works well enough for most things youd want out of a wiki, for something as small in scope as the reform | 12:30 |
| violet | like its not *small* small but compared to something like a full linux distro you know | 12:30 |
| grimmware | yeah that's fair | 12:30 |
| violet | i usually just use duckduckgo to search the forum | 12:31 |
| josch | now with the imx8mq AB, there is even more SoM variety (even if unintentional) | 12:31 |
| violet | in my experience i find the git fjorge wikis to be a lot harder to navigate than a mediawiki though so im kinda negative on those | 12:32 |
| violet | and they dont get indexed as well by search engines i think | 12:32 |
| violet | like mediawiki is noot actually that hard to run, used to admin one for work | 12:33 |
| violet | the setup is just annoying | 12:33 |
| violet | but anyways if theres demand for it i have the server space i can put towards it in my friend's colocated hardware | 12:34 |
| violet | but i havent found myself itching for one myself | 12:35 |
| erle | if your wiki for a project that uses git is not in git you have yourself found a data graveyard btw | 12:36 |
| erle | been there, done that | 12:36 |
| violet | that doesnt reflect my experiences | 12:37 |
| violet | my main concern in that regard in the context of mnt is that theres so many repos | 12:37 |
| violet | you need a way for people to know which repo has the wiki, ifyou attach the wiki to a repo | 12:38 |
| violet | i feel like readthedocs is not a bad solution though | 12:39 |
| violet | always happy to see a readthedocs | 12:39 |
| violet | and then the wiki data is all in git iirc | 12:39 |
| josch | but before setting up a readthedocs you might as well just fix up the handbook instead as both are centrally maintained | 12:41 |
| violet | yeah that is true | 12:41 |
| violet | and i really appreciate the handbook being available offlinee | 12:42 |
| - vagrantc (QUIT: Quit: leaving) (~vagrant@2600:3c01:e000:21:7:77:0:50) | 12:53 | |
| grimmware | sounds like doubling down on the handbook is probably the way to go. | 12:53 |
| grimmware | like I say, I don't have a reform yet so everything I'm saying here is super speculative | 12:54 |
| grimmware | also sorry for asking what colour the bikeshed should be ;) | 12:56 |
| josch | minute: seems that https://github.com/rbz-embedded-logics/uboot_ls1028 is 404? | 13:02 |
| - eschaton (QUIT: Server closed connection) (~eschaton@li541-49.members.linode.com) | 13:07 | |
| + eschaton (~eschaton@li541-49.members.linode.com) | 13:07 | |
| josch | i'm currently perusing the ls1028a-dev branch of reform-system-image | 13:08 |
| josch | i wanted to create a reform-ls1028a-uboot script in the same way as i have for reform-a311d-uboot but then got that 404 :( | 13:10 |
| - tarxvf (QUIT: Server closed connection) (~tarxvf@mail.tarxvf.tech) | 13:19 | |
| + tarxvf (~tarxvf@mail.tarxvf.tech) | 13:20 | |
| - XYZ (QUIT: Remote host closed the connection) (~XYZ@89-24-49-225.nat.epc.tmcz.cz) | 13:25 | |
| + XYZ (~XYZ@89-24-49-225.nat.epc.tmcz.cz) | 13:28 | |
| - stites (QUIT: Ping timeout: 240 seconds) (~stites@130.44.147.204) | 13:31 | |
| + stites (~stites@2607:fb91:dc8:870c:e6db:e0fc:e90f:d41c) | 13:32 | |
| minute | we need to solve this terrible caam problem in 6.5 | 13:42 |
| minute | josch: sorry, i guess it's private, i will clone it to source.mnt.re | 13:44 |
| josch | oooh okay :) | 13:44 |
| josch | indeed, the first thing that turns up in google if i search for the caam_jr probe error is the thread in community.mnt.re XD | 13:49 |
| minute | > you can echo y > /sys/module/cryptomgr/parameters/notests to turn off testing noise. the crypto testing is in bad shape even on x86_64 – it used to misbehave if compiled into the kernel but work ok if loaded as a module. | 13:56 |
| minute | i wonder if we can turn off crypto testing in the kernel build | 13:56 |
| josch | minute: if it's a module option, can it not be added to /etc/modprobe.d/reform.conf | 14:02 |
| josch | options cryptomgr notests=y | 14:02 |
| - mtm (QUIT: Ping timeout: 245 seconds) (~mtm@c-71-228-84-213.hsd1.fl.comcast.net) | 14:02 | |
| minute | josch: ah, do we know it's a loadable module? | 14:04 |
| minute | (didn't check yet) | 14:04 |
| josch | sorry, i don't know that out of the top of my head | 14:04 |
| minute | josch: https://source.mnt.re/reform/reform-ls1028a-uboot | 14:05 |
| minute | josch: https://source.mnt.re/reform/reform-ls1028a-atf | 14:05 |
| minute | josch: i'm pretty sure we should be able to use mainline uboot, but we need the custom atf because there we set the memory timings | 14:05 |
| minute | josch: on ls1028a, all DDR4 stuff is done in hardware and some code in atf just gives the hardware the parameters for the DDR4 chips | 14:06 |
| minute | josch: see https://source.mnt.re/reform/reform-ls1028a-atf/-/commit/f92edf5175b20fc82d7d17b4461ce3e1b9689254 | 14:06 |
| josch | mainline u-boot would be nice -- something to evaluate | 14:07 |
| josch | though probably sensible to start with a known-to-work fork first | 14:08 |
| minute | yep | 14:08 |
| minute | also the kernel used in that process is antique | 14:08 |
| minute | (5.8?) | 14:08 |
| josch | your latest screenshots were 5.12 i think | 14:09 |
| minute | so i had some in-between versions for local testing | 14:09 |
| minute | ideally we update to the newest stuff soon. | 14:09 |
| josch | yup :) | 14:09 |
| josch | README says that the first partition has to be fat32 but that probably is not a hard requirement as u-boot can read ext2 just fine, no? | 14:10 |
| minute | not sure if the script has fatload hardcoded | 14:10 |
| josch | ah right, moving to flash-kernel boot.scr is another task | 14:11 |
| minute | josch: during QA here we just tried `options cryptomgr notests=y`, but it doesn't work. cryptomgr is also not in lsmod. as a workaround on the testing SD card we've put in `blacklist caam_jr` | 14:18 |
| - stites (QUIT: Read error: Connection reset by peer) (~stites@2607:fb91:dc8:870c:e6db:e0fc:e90f:d41c) | 14:44 | |
| + stites (~stites@130.44.147.204) | 14:44 | |
| - jacobk (QUIT: Ping timeout: 246 seconds) (~quassel@64.189.201.150) | 14:53 | |
| + jacobk (~quassel@129.110.242.224) | 15:05 | |
| unixpoet | grimmware: chartreuse, clearly | 15:09 |
| grimmware | okay so we know what colour community documentation should be. | 15:16 |
| + mark_ (~mjw@gnu.wildebeest.org) | 15:18 | |
| + marty_mcfly88 (~marty@2603-6000-8c01-0f3e-d548-2094-b4d8-68d6.res6.spectrum.com) | 15:32 | |
| + mtm (~mtm@c-71-228-84-213.hsd1.fl.comcast.net) | 16:09 | |
| - XYZ (QUIT: Remote host closed the connection) (~XYZ@89-24-49-225.nat.epc.tmcz.cz) | 16:12 | |
| + chomwitt (~chomwitt@2a02:587:7a0c:4d00:1ac0:4dff:fedb:a3f1) | 16:15 | |
| josch | minute: is that module not ever needed or is it needed when setting up luks encryption? Because then it would not be ideal if reform-migrate copies that blacklisting over into a partition that is supposed to be decrypted later, no? | 16:56 |
| minute | josch: it is generally needed, it's just a workaround for us during testing, it's not affecting the user install | 17:05 |
| josch | ah kk | 17:05 |
| minute | josch: i'm thinking to prepare the a311d uboot repo now. how do you imagine it to be? at first just a copy of the current script hat yields flash.bin? | 17:13 |
| josch | minute: exactly like the reform-a311d-uboot branch of reform-system-image: https://source.mnt.re/josch/reform-system-image/-/tree/reform-a311d-uboot?ref_type=heads | 17:14 |
| josch | that branch already built a flash.bin for a311d | 17:14 |
| minute | ah great. | 17:15 |
| - bleb (QUIT: Server closed connection) (~cm@user/bleb) | 17:18 | |
| + bleb (~cm@user/bleb) | 17:18 | |
| minute | josch: https://source.mnt.re/reform/reform-a311d-uboot/ | 17:19 |
| josch | minute: nice! now you have to somehow un-stuck the CI pipeline :) | 17:36 |
| minute | josch: will do! but first, one question about the label based mounting. currently, it defeats the boot device priority feature in uboot, which boots in order: usb0, mmc1, mmc0 | 17:44 |
| josch | oh? | 17:45 |
| minute | josch: but i used reform-flash-rescue (well, a subset of it) to install the "unified" image to flash (mmc0). the problem is that this then also has the reformsdboot and -root labels | 17:45 |
| josch | how does it defeat that? | 17:46 |
| josch | ah | 17:46 |
| minute | josch: so a linux booted from sd card will then mount its root partition from emmc | 17:46 |
| minute | josch: because it finds the disk label there first | 17:46 |
| josch | i totally forgot that reform-flash-rescue literally just dd's the image to emmc | 17:46 |
| minute | yeah, i guess the fs needs to be touched up afterwards | 17:46 |
| josch | i'll prepare a merge request for reform-flash-rescue | 17:47 |
| josch | the fs needs its labels changed and /etc/fstab its entries | 17:47 |
| minute | exactly | 17:47 |
| josch | ah, reform-boot-config is also imx8mq specific | 17:48 |
| josch | minute: https://source.mnt.re/reform/reform-tools/-/merge_requests/50 | 17:55 |
| josch | at the end of reform-flash-rescue, reform-boot-config will write /dev/mmcblk... entries into /etc/fstab, overwriting the LABEL=reformsdboot entries | 17:56 |
| minute | josch: ahhhh, i think i didn't use that | 17:57 |
| josch | what did you use? | 17:57 |
| minute | josch: basically only the dd part | 17:57 |
| josch | minute: if you just dd, then /etc/fstab on emmc will be wrong | 17:58 |
| minute | josch: yeah, true | 17:59 |
| josch | you at least need to run reform-boot-config --emmc emmc afterwards | 17:59 |
| minute | will try to do that now after the fact | 17:59 |
| josch | minute: alternatively, the reform-system-image pipeline could build another set of images targeted at the emmc, then you could just dd without modifying anything afterwards | 18:12 |
| - marty_mcfly88 (QUIT: Remote host closed the connection) (~marty@2603-6000-8c01-0f3e-d548-2094-b4d8-68d6.res6.spectrum.com) | 18:16 | |
| minute | josch: yeah, but that's quite resource intensive | 18:20 |
| minute | josch: strangely my last reform image build timed out after 2h... weird | 18:20 |
| - jacobk (QUIT: Ping timeout: 246 seconds) (~quassel@129.110.242.224) | 18:20 | |
| minute | your unified build will help i guess | 18:21 |
| + jacobk (~quassel@129.110.242.224) | 18:22 | |
| - jacobk (QUIT: Ping timeout: 246 seconds) (~quassel@129.110.242.224) | 18:34 | |
| - mjw (QUIT: Killed (NickServ (GHOST command used by mark_!~mjw@gnu.wildebeest.org))) (~mjw@2001:1c06:2488:1400:4fd:39a7:74ac:7bae) | 18:47 | |
| * mark_ -> mjw | 18:47 | |
| + mark_ (~mjw@2001:1c06:2488:1400:4fd:39a7:74ac:7bae) | 18:47 | |
| minute | Seattle MNT Reform meetup on Sep. 19! https://community.mnt.re/t/next-seattle-reform-meetup-september-19/1652 | 18:54 |
| violet | yoooo lets go | 19:13 |
| + vagrantc (~vagrant@2600:3c01:e000:21:7:77:0:50) | 19:25 | |
| + jacobk (~quassel@utdpat242016.utdallas.edu) | 19:49 | |
| - jacobk (QUIT: Ping timeout: 246 seconds) (~quassel@utdpat242016.utdallas.edu) | 20:00 | |
| + jacobk (~quassel@129.110.242.173) | 20:13 | |
| - leonardo (QUIT: Server closed connection) (leonardo@user/leonardo) | 20:56 | |
| + leonardo (leonardo@user/leonardo) | 20:56 | |
| + varke (~varke@149.102.240.81) | 20:56 | |
| * varke -> v4rke | 20:57 | |
| - jacobk (QUIT: Ping timeout: 255 seconds) (~quassel@129.110.242.173) | 20:58 | |
| - v4rke (QUIT: Quit: WeeChat 3.8) (~varke@149.102.240.81) | 21:12 | |
| - Boostisbetter (QUIT: Ping timeout: 252 seconds) (4a410829d7@irc.cheogram.com) | 21:24 | |
| + Boostisbetter (4a410829d7@irc.cheogram.com) | 21:50 | |
| + cwebber (~user@user/cwebber) | 22:06 | |
| + jacobk (~quassel@64.189.201.150) | 23:10 | |
| - stites (QUIT: Ping timeout: 245 seconds) (~stites@130.44.147.204) | 23:44 | |
| + stites (~stites@2607:fb91:dc9:4b54:4710:ea16:819c:2384) | 23:44 | |
| - stites (QUIT: Ping timeout: 252 seconds) (~stites@2607:fb91:dc9:4b54:4710:ea16:819c:2384) | 23:49 | |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!