Logs on 2022-04-14 (liberachat/#xmonad)
| 00:43:07 | × | benin quits (~benin@183.82.204.110) (Ping timeout: 240 seconds) |
| 01:10:06 | × | Artus quits (~Artus@190.192.80.102) (Ping timeout: 250 seconds) |
| 01:45:33 | <abastro[m]> | I hate when ghc9.2.2 is unstable so my xmonad randomly crashes once a week |
| 01:57:57 | <davean> | abastro[m]: What bugs are you encountering with GHC 9.2.2? |
| 01:58:19 | <abastro[m]> | It randomly crashes, I cannot find more details |
| 01:58:34 | <abastro[m]> | Idk where to find the crash dump file, if it exists |
| 01:58:43 | <abastro[m]> | Log does not report anything in particular |
| 01:58:59 | × | jao quits (~jao@cpc103048-sgyl39-2-0-cust502.18-2.cable.virginm.net) (Remote host closed the connection) |
| 02:03:36 | × | banc quits (banc@gateway/vpn/airvpn/banc) (Ping timeout: 248 seconds) |
| 02:05:03 | × | bla quits (~bla@79.191.74.133.ipv4.supernova.orange.pl) (Ping timeout: 260 seconds) |
| 02:15:21 | × | abhixec quits (~abhixec@c-67-169-139-16.hsd1.ca.comcast.net) (Quit: leaving) |
| 02:15:30 | → | bla joins (~bla@79.191.254.193.ipv4.supernova.orange.pl) |
| 02:18:53 | → | jao joins (~jao@cpc103048-sgyl39-2-0-cust502.18-2.cable.virginm.net) |
| 02:21:54 | → | banc joins (banc@gateway/vpn/airvpn/banc) |
| 02:47:40 | × | terrorjack quits (~terrorjac@2a01:4f8:1c1e:509a::1) (Quit: The Lounge - https://thelounge.chat) |
| 02:50:06 | → | terrorjack joins (~terrorjac@2a01:4f8:1c1e:509a::1) |
| 03:00:29 | × | jao quits (~jao@cpc103048-sgyl39-2-0-cust502.18-2.cable.virginm.net) (Ping timeout: 256 seconds) |
| 03:39:56 | × | matijja quits (~matijja@193.77.181.201) (Quit: ZNC 1.8.2 - https://znc.in) |
| 03:43:12 | → | matijja joins (~matijja@193.77.181.201) |
| 05:29:48 | → | benin joins (~benin@183.82.204.110) |
| 05:58:12 | <Solid> | I again seriously doubt this is GHC's fault |
| 05:58:33 | <Solid> | and that taffybar stuff sounds like an easy fix |
| 05:58:50 | <Solid> | so how about you prepare a PR instead of complaining that it doesn't work and you know how to fix it? :> |
| 05:59:45 | <abastro[m]> | Solid: I guess how I said it was a bit confusing |
| 06:00:43 | <abastro[m]> | Taffybar one was a trivial fix of just cloning, tho it comes from breaking change by ghc. |
| 06:01:17 | <abastro[m]> | The random crash stuff is something else, a phenomenon that began occurring after updating to 9.2 2 |
| 06:02:50 | <Solid> | if you have a proper reproducer then by all means report it to GHCs issues |
| 06:03:22 | <Solid> | (as for the breaking change... there's is an automatic upgrade path, so I wouldn't even count that tbh) |
| 06:04:00 | <abastro[m]> | Wdym automatic upgrade path? |
| 06:05:48 | <Solid> | "if FlexibleContexts is on in this file but not FlexibleInstances, add FlexibleInstances" |
| 06:05:59 | <Solid> | sounds like something a computer is good at! |
| 06:20:25 | <abastro[m]> | Eh, I mean, that will generate warnings for files that does not apply |
| 06:20:44 | <abastro[m]> | Then imagine -Werror |
| 06:20:53 | <abastro[m]> | ..wait was redundant extension not a warning |
| 06:23:27 | <Solid> | if it's only a warning then it's even more "not even a breaking change" |
| 06:23:38 | <abastro[m]> | Anyway the bigger problem was that the author is not inclined to even update package on hackage to a version with a bug fixed |
| 06:24:16 | <abastro[m]> | Solid: I mean, with FlexibleInstances and without FlexibleContexts, it is error |
| 06:25:02 | <Solid> | AFAIK Hackage trustees can update packages with small fixes like that |
| 06:25:12 | <abastro[m]> | Also I was wrong with warning for redundant extension |
| 06:25:48 | <abastro[m]> | Author is alive and well, the author promised to upload the fix on hackage |
| 06:26:05 | <abastro[m]> | And that does not happen :p |
| 06:29:09 | <Solid> | use the git version in the meantime :) |
| 06:29:33 | <abastro[m]> | Yep |
| 06:29:51 | <abastro[m]> | Anyway that is more or less trivial probelm |
| 06:30:16 | × | m5zs7k quits (aquares@web10.mydevil.net) (Ping timeout: 248 seconds) |
| 06:30:19 | <abastro[m]> | Bigger one was with unidentifiable xmonad random crash, it doesn't even reproduce reliably |
| 06:30:23 | <abastro[m]> | It's just random |
| 06:30:25 | <abastro[m]> | Meh |
| 06:30:51 | → | m5zs7k joins (aquares@web10.mydevil.net) |
| 06:34:32 | → | cfricke joins (~cfricke@user/cfricke) |
| 07:19:28 | × | smashgra_ quits (~smashgrab@bare.metal.computer) (Quit: ZNC 1.7.2+deb3 - https://znc.in) |
| 07:19:46 | → | smashgrab joins (~smashgrab@bare.metal.computer) |
| 07:33:32 | <geekosaur> | Solid, Hackage trustees can only do version bounds changes in the cabal file, not actual package changes |
| 07:35:01 | <Solid> | oh I see |
| 07:36:10 | <Ether[m]> | I dont why but it seems XMonad.Layout.Fullscreen does have a way to add a binding.. |
| 07:36:43 | <geekosaur> | because it's up to the app, which often dpoes more than simply fullscreen itself |
| 07:36:54 | <geekosaur> | so you could conceivably get "fake fullscreen" |
| 07:37:48 | <Ether[m]> | i tried ", ("M-f", sendMessage $ Xmonad. Layout.MultiToggle Full) |
| 07:38:15 | <Ether[m]> | s/.// |
| 07:39:35 | <geekosaur> | (1) are yo using multitoggle? (2) its messages are all uppercase, last I checked |
| 07:39:48 | <Ether[m]> | Yes |
| 07:40:10 | <Ether[m]> | 2. Upper case for all? |
| 07:40:13 | <geekosaur> | FULL |
| 07:40:22 | <Ether[m]> | Oh yess |
| 07:40:25 | <Ether[m]> | I did FULL |
| 07:40:26 | <geekosaur> | https://hackage.haskell.org/package/xmonad-contrib-0.17.0/docs/XMonad-Layout-MultiToggle-Instances.html#t:StdTransformers |
| 07:40:31 | <Ether[m]> | No luck.. |
| 07:43:25 | <geekosaur> | sendMessage $ Toggle FULL |
| 07:45:17 | <Ether[m]> | Nope |
| 07:45:39 | <Ether[m]> | 1) using multitoggle |
| 07:46:58 | <Ether[m]> | 2) it works adding line to layout hook; $mkToggle (single FULL) |
| 07:49:28 | <Ether[m]> | But in altercations dotfiles.. it allows me to use nav 2d to change focused window.. |
| 07:49:37 | <Ether[m]> | Which is not working for me.. |
| 07:52:10 | <geekosaur> | that is MultiToggle, right from its documentation |
| 07:54:01 | <Ether[m]> | Nope, i figured it was navigation2d |
| 07:54:15 | <Ether[m]> | I had to tell it what to do with FULL Layout |
| 08:18:36 | → | coldpress_ joins (~coldpress@72.136.212.35.bc.googleusercontent.com) |
| 08:20:39 | → | haskl[error] joins (~haskl@user/haskl) |
| 08:22:34 | × | coldpress quits (~coldpress@72.136.212.35.bc.googleusercontent.com) (Quit: ZNC 1.7.2 - https://znc.in) |
| 08:22:34 | × | haskl quits (~haskl@user/haskl) (Quit: Uh oh... ZNC disconnected.) |
| 10:57:37 | × | theproffesor quits (~theprofff@user/theproffesor) (Ping timeout: 260 seconds) |
| 11:02:18 | → | theproffesor joins (~theprofff@user/theproffesor) |
| 12:07:42 | <abastro[m]> | geekosaur: I forgot how to debug a crash |
| 12:07:46 | <abastro[m]> | How do i do it? |
| 12:16:24 | × | cfricke quits (~cfricke@user/cfricke) (Ping timeout: 248 seconds) |
| 12:54:27 | × | benin quits (~benin@183.82.204.110) (Ping timeout: 240 seconds) |
| 13:03:36 | <geekosaur> | crash debugging is painful especially if it's in compiled code rather than FFI: debuggers realy want call stacks, but there isn't one in Haskell |
| 13:06:42 | <geekosaur> | but you can compile with -g to get debug symbols which will help somewhat |
| 13:07:48 | <geekosaur> | then you have to run your program in a debugger or attach to it from a debugger (the latter will be necessary if it's xmonad that is crashing; switch to a console vt and run gdb / lldb there) |
| 13:10:27 | <geekosaur> | from there, you're on your own: I can help a bit if it's in the FFI but if it's in the RTS all I can do is point you to #ghc |
| 13:31:25 | <geekosaur> | at some point I'll probably switch my default ghc to 9.2.2 and see if I replicate your crash locally, since I'm probably better at poking around with gdb than you are. but even so I may have to invoke #ghc if it'snot obviously FFI-related |
| 13:31:35 | <geekosaur> | and/or will have to try to come up with a reproducer |
| 13:54:30 | <abastro[m]> | Hmm |
| 13:54:31 | <abastro[m]> | geekosaur: is there any way to see if the problem was caused by FFI portion? |
| 13:55:21 | <geekosaur> | only by seeing if it has a call stack, since once you enter C there is one |
| 13:55:32 | <geekosaur> | the STG is all jumps so there is no call stack |
| 13:55:34 | <abastro[m]> | Or compiled code portion |
| 13:55:35 | <abastro[m]> | I guess I need to compile with profiler to get proper exception messages |
| 13:56:08 | <geekosaur> | are you getting messages? |
| 13:56:23 | <geekosaur> | I thought you weren't, which suggested core dump |
| 13:56:41 | <geekosaur> | sadly I'm about to be on a call, so won't be around much longer |
| 13:59:10 | <geekosaur> | but yes,if you doget exception messages then you need to compile everything with --profile (stack) or -prof -prof-all (raw ghc / cabal) |
| 14:08:39 | <abastro[m]> | Yep, I am not getting messages |
| 14:08:46 | <abastro[m]> | So I guess core dump should be somewhere |
| 14:08:48 | <abastro[m]> | I just wonder where it would be.. |
| 14:14:11 | → | jao joins (~jao@92.233.85.247) |
| 15:07:11 | <geekosaur> | check ulimit |
| 15:08:06 | <geekosaur> | most linuxes default to `ulimit -c 0` which disables core dumps. some (notably RHEL and possibly Fedora) feed it to a system daemon instead, but in that case you should get a popup (although if it's xmonad that's dumping core, that could work poorly) |
| 15:16:32 | <geekosaur> | blegh, everything is happening today. counseling calls, service calls, … |
| 15:20:34 | <abastro[m]> | Ohh. |
| 15:25:40 | <geekosaur> | note that it's process-inherited, so if you change it in a terminal it won't affect anything already running |
| 15:51:25 | <geekosaur> | aaaand we rebuild the whole world :) |
| 15:51:40 | geekosaur | just set 9.2.2 as default and is rebuilding xmonad |
| 15:57:05 | <geekosaur> | and then I'll have to do this again because I forgot -g |
| 16:00:13 | × | liskin[m] quits (~liskinmat@2001:470:69fc:105::768) (Quit: You have been kicked for being idle) |
| 16:00:30 | <Solid> | or because it won't crash ;) |
| 16:00:41 | <geekosaur> | that remains to be seen |
| 16:00:45 | → | liskin[m] joins (~liskinmat@2001:470:69fc:105::768) |
| 16:01:00 | <geekosaur> | have to restart, then attach to it from a vt |
| 16:01:19 | <geekosaur> | wonder if I need -prof too |
| 16:01:28 | <geekosaur> | think I'll wait on that one |
| 16:14:54 | → | edun joins (~qbt@user/edun) |
| 16:29:13 | → | benin joins (~benin@183.82.204.110) |
| 16:30:07 | × | edun quits (~qbt@user/edun) (Quit: Leaving.) |
| 16:30:33 | → | edun joins (~qbt@user/edun) |
| 16:32:23 | <geekosaur> | here goes nothing |
| 16:32:43 | <geekosaur> | (now up in 9.2.2, will wait to see if it crashes before I do the gdb attach dance) |
| 16:34:24 | × | edun quits (~qbt@user/edun) (Client Quit) |
| 16:35:17 | <geekosaur> | abastro[m], about how long does it take for crashes to happen? |
| 16:35:26 | → | edun joins (~qbt@user/edun) |
| 16:36:06 | <geekosaur> | also one advantage of the xmonad under mate setup is I can recover from a crash |
| 16:37:23 | <abastro[m]> | Well, it is unpredictable |
| 16:37:31 | <abastro[m]> | Perhaps you could try reloading via mod+q, I guess |
| 16:39:13 | × | edun quits (~qbt@user/edun) (Client Quit) |
| 16:39:35 | → | edun joins (~qbt@user/edun) |
| 16:50:22 | × | edun quits (~qbt@user/edun) (Quit: Leaving.) |
| 17:04:06 | × | Czernobog quits (~Czernobog@user/czernobog) (Read error: Connection reset by peer) |
| 17:04:56 | → | Czernobog joins (~Czernobog@user/czernobog) |
| 17:24:55 | → | steve__ joins (~steve@ool-182c2b80.dyn.optonline.net) |
| 17:26:30 | → | syg joins (~syg@flamebot.net) |
| 17:26:37 | → | redgloboli_ joins (~redglobol@user/redgloboli) |
| 17:27:56 | → | rieper|net joins (~riepernet@sxbeta1.geo.uni-leipzig.de) |
| 17:28:38 | → | globe_ joins (~globe@stonebay32.com) |
| 17:28:55 | → | scardina1 joins (~supreme@customer-212-237-101-39.ip4.gigabit.dk) |
| 17:33:56 | × | theproffesor quits (~theprofff@user/theproffesor) (*.net *.split) |
| 17:33:56 | × | haskl[error] quits (~haskl@user/haskl) (*.net *.split) |
| 17:33:56 | × | rieper quits (~riepernet@sxbeta1.geo.uni-leipzig.de) (*.net *.split) |
| 17:33:56 | × | syg_ quits (~syg@flamebot.net) (*.net *.split) |
| 17:33:56 | × | steve_ quits (~steve@ool-182c2b80.dyn.optonline.net) (*.net *.split) |
| 17:33:57 | × | scardinal quits (~supreme@customer-212-237-101-39.ip4.gigabit.dk) (*.net *.split) |
| 17:33:57 | × | deepy quits (deepy@user/deepy) (*.net *.split) |
| 17:33:57 | × | redgloboli quits (~redglobol@user/redgloboli) (*.net *.split) |
| 17:33:57 | × | Benzi-Junior quits (~BenziJuni@dsl-149-64-179.hive.is) (*.net *.split) |
| 17:33:57 | × | globe quits (~globe@stonebay32.com) (*.net *.split) |
| 17:37:15 | → | theproffesor joins (~theprofff@user/theproffesor) |
| 17:37:15 | → | haskl[error] joins (~haskl@user/haskl) |
| 17:37:15 | → | deepy joins (deepy@user/deepy) |
| 17:37:15 | → | Benzi-Junior joins (~BenziJuni@dsl-149-64-179.hive.is) |
| 18:41:15 | <geekosaur> | nice confusing boolean blindness in one of my own modules :( |
| 18:41:22 | <geekosaur> | not even documented |
| 18:57:42 | <geekosaur> | I think this is one of those modules I whipped up on the fly for someone and then minimally packaged. |
| 18:58:10 | <geekosaur> | also, no crash yet |
| 18:58:21 | <geekosaur> | abastro[m], can you put your config somewhere? |
| 18:58:25 | <geekosaur> | @where paste |
| 18:58:25 | <lambdabot> | Help us help you: please paste full code, input and/or output at e.g. https://paste.tomsmeding.com |
| 19:29:43 | → | mohab joins (~mohab@156.223.10.66) |
| 19:30:22 | × | mohab quits (~mohab@156.223.10.66) (Client Quit) |
| 19:30:52 | → | mohab joins (~mohab@156.223.10.66) |
| 19:31:00 | × | mohab quits (~mohab@156.223.10.66) (Client Quit) |
| 21:43:37 | × | jao quits (~jao@92.233.85.247) (Ping timeout: 256 seconds) |
| 21:56:07 | × | geekosaur quits (~geekosaur@xmonad/geekosaur) (Killed (NickServ (GHOST command used by allbery_b))) |
| 21:56:07 | → | allbery_b joins (~geekosaur@xmonad/geekosaur) |
| 21:56:08 | × | allbery_b quits (~geekosaur@xmonad/geekosaur) (Remote host closed the connection) |
| 21:56:32 | → | geekosaur joins (~geekosaur@xmonad/geekosaur) |
| 22:01:08 | <geekosaur> | welp, I think I reproduced the xmonad crash… can't tell for certain but there's an X error of a bad colormap parameter |
| 22:02:22 | <geekosaur> | hm, lot of noise in the session log |
| 22:12:40 | <geekosaur> | bleh. restarted okay but now I don't have logs… |
| 22:12:47 | <geekosaur> | guess I have to wait for debugging |
| 22:14:05 | <geekosaur> | which means I also have to wait to play with vbox since that's what's crashing it here (which may mean it's unrelated_ |
| 22:14:07 | <geekosaur> | ) |
| 22:33:30 | <abastro[m]> | geekosaur: https://github.com/Abastro/pulpmonad/blob/master/app/Main.hs |
| 22:38:58 | <geekosaur> | hm. that caused me to get a nasty suspicion, but no, it's not the debugStack |
| 22:39:40 | <geekosaur> | I am reliably able to crash xmonad on 9.2.1 (but not 8.10.7) by starting VirtualBox |
| 22:39:54 | <geekosaur> | otherwise it's been stable |
| 22:40:34 | <geekosaur> | and I don't see much of anything else that would do it (mixing forkProcess with threads is a problem, but you don't appear to have anything that uses threads) |
| 22:51:21 | <geekosaur> | I'm seeing a lot of user error (Error in function getWindowAttributes) but I think that happens anyway |
| 22:52:17 | <geekosaur> | and some BadDrawable errors which might be related if it's trying to draw a border. wonder if we need to userCode that path or otherwise guard against exceptions if a window goes away while being managed |
| 23:06:05 | <geekosaur> | switching back to 8.10.7 so I can mess with vbox (or reproduce the problem anyway) |
| 23:06:49 | <geekosaur> | would like to know what it is about vbox and the second window it appears to open that I have yet to identify (because xmonad crashes when it pops) that is doing this |
| 23:07:04 | <geekosaur> | even when xmonad doesn't crash there's a lot of border flickering before it settles down |
| 23:08:25 | <geekosaur> | I can't be the only one using vbox with xmonad |
| 23:08:38 | <geekosaur> | (haven't done this before, I usually use vmware) |
| 23:25:03 | × | banc quits (banc@gateway/vpn/airvpn/banc) (Ping timeout: 256 seconds) |
| 23:35:08 | <abastro[m]> | Oh right, I have a much simpler repo with more frequent crashes |
| 23:58:17 | <geekosaur> | back on 8.10.7 and the crashes have stopped… along with the BadDrawable errors |
| 23:58:42 | <geekosaur> | wonder if it's a timing issue |
| 23:59:06 | <geekosaur> | sadly vbox is still not behaving |
All times are in UTC on 2022-04-14.