Newest at the top
| 2026-02-11 23:36:19 +0100 | <perryprog> | true* |
| 2026-02-11 23:36:17 +0100 | <perryprog> | so ture |
| 2026-02-11 23:36:16 +0100 | <perryprog> | oh |
| 2026-02-11 23:36:11 +0100 | <perryprog> | tomsmeding, that's do_posix_spawn from process? https://github.com/haskell/process/blob/98101f82543b3a28e9f5192d758a508881f8b464/cbits/posix/posix… |
| 2026-02-11 23:35:38 +0100 | <tomsmeding> | debugging that function should not require a GHC with debug symbols |
| 2026-02-11 23:35:35 +0100 | merijn | (~merijn@62.45.136.136) merijn |
| 2026-02-11 23:35:30 +0100 | <tomsmeding> | perryprog: if you switch to frame 1 in lldb in that crash, you should be in do_posix_spawn, which is a normal C file from the 'process' library |
| 2026-02-11 23:35:03 +0100 | <perryprog> | maybe I just compile ghc with debug symbols |
| 2026-02-11 23:32:18 +0100 | CiaoSen | (~Jura@2a02:8071:64e1:da0:5a47:caff:fe78:33db) (Ping timeout: 252 seconds) |
| 2026-02-11 23:31:12 +0100 | oskarw | (~user@user/oskarw) (Ping timeout: 252 seconds) |
| 2026-02-11 23:29:58 +0100 | <tomsmeding> | though I guess it might be hidden inside something |
| 2026-02-11 23:29:50 +0100 | mange | (~mange@user/mange) mange |
| 2026-02-11 23:29:37 +0100 | <tomsmeding> | (https://hackage-content.haskell.org/package/process-1.6.26.1/src/cbits/posix/posix_spawn.c) |
| 2026-02-11 23:29:34 +0100 | <tomsmeding> | I don't see any such indirect call in the source of do_posix_spawn |
| 2026-02-11 23:29:14 +0100 | <tomsmeding> | or rather, a call to a null pointer |
| 2026-02-11 23:29:06 +0100 | <tomsmeding> | I just realised that in your lldb backtrace, it showed that the crash is not simply a dereference of a null pointer, it's a _jump_ to a null pointer |
| 2026-02-11 23:28:37 +0100 | <perryprog> | I mean, it's all in my .ghcup so it's all stuff I've deleted before |
| 2026-02-11 23:28:26 +0100 | <EvanR> | egyptian filename style |
| 2026-02-11 23:28:00 +0100 | fgarcia | (~lei@user/fgarcia) fgarcia |
| 2026-02-11 23:27:28 +0100 | CiaoSen | (~Jura@2a02:8071:64e1:da0:5a47:caff:fe78:33db) CiaoSen |
| 2026-02-11 23:27:09 +0100 | <perryprog> | ahh okay wasn't sure if that was the case; in that case yes |
| 2026-02-11 23:27:00 +0100 | <tomsmeding> | can you find that file on your machine somewhere? |
| 2026-02-11 23:26:37 +0100 | <tomsmeding> | it's "process" but without vowels; cabal/ghc/something removes vowels from filenames in some situations |
| 2026-02-11 23:25:37 +0100 | <perryprog> | and it's not linked to by ghc or libHSrts |
| 2026-02-11 23:24:33 +0100 | <perryprog> | weird, I can't actually fine what libHSprcss /is/. HLS has some dylib with that in the name but besides that I can't see it anywhere. |
| 2026-02-11 23:24:29 +0100 | merijn | (~merijn@host-cl.cgnat-g.v4.dfn.nl) (Ping timeout: 252 seconds) |
| 2026-02-11 23:22:09 +0100 | oskarw | (~user@user/oskarw) oskarw |
| 2026-02-11 23:22:08 +0100 | oskarw` | (~user@user/oskarw) (Remote host closed the connection) |
| 2026-02-11 23:20:04 +0100 | merijn | (~merijn@host-cl.cgnat-g.v4.dfn.nl) merijn |
| 2026-02-11 23:18:43 +0100 | ljdarj | (~Thunderbi@user/ljdarj) (Ping timeout: 260 seconds) |
| 2026-02-11 23:13:45 +0100 | Googulator | (~Googulato@84-236-65-138.pool.digikabel.hu) |
| 2026-02-11 23:13:21 +0100 | Googulator | (~Googulato@84-236-65-138.pool.digikabel.hu) (Quit: Client closed) |
| 2026-02-11 23:09:17 +0100 | merijn | (~merijn@host-cl.cgnat-g.v4.dfn.nl) (Ping timeout: 244 seconds) |
| 2026-02-11 23:04:43 +0100 | merijn | (~merijn@host-cl.cgnat-g.v4.dfn.nl) merijn |
| 2026-02-11 22:53:41 +0100 | merijn | (~merijn@host-cl.cgnat-g.v4.dfn.nl) (Ping timeout: 252 seconds) |
| 2026-02-11 22:46:41 +0100 | merijn | (~merijn@host-cl.cgnat-g.v4.dfn.nl) merijn |
| 2026-02-11 22:45:33 +0100 | target_i | (~target_i@user/target-i/x-6023099) (Quit: leaving) |
| 2026-02-11 22:44:31 +0100 | oskarw | (~user@user/oskarw) (Ping timeout: 246 seconds) |
| 2026-02-11 22:43:33 +0100 | oskarw` | (~user@user/oskarw) oskarw |
| 2026-02-11 22:42:05 +0100 | <tomsmeding> | the concrete dependency tree including versions can be found in dist-newstyle/cache/plan.json |
| 2026-02-11 22:41:25 +0100 | <perryprog> | it's not two versions of the same project—I haven't been able to get the main project (oama) working at all, but the thing that doesn't work works fine in isolation. (Will try other suggestions soon™) |
| 2026-02-11 22:40:56 +0100 | brioche | (~username@user/brioche) () |
| 2026-02-11 22:40:46 +0100 | <brioche> | :q |
| 2026-02-11 22:40:37 +0100 | <tomsmeding> | perryprog: this is a stupid suggestion, but what about taking your copy of the project that _does_ work, and copying it over the one that doesn't work |
| 2026-02-11 22:40:05 +0100 | <tomsmeding> | s/built-tool/build-tool/ |
| 2026-02-11 22:39:56 +0100 | <tomsmeding> | from skimming the scrollback I feel like this is irrelevant |
| 2026-02-11 22:39:47 +0100 | <tomsmeding> | and cabal freeze does this |
| 2026-02-11 22:39:42 +0100 | <tomsmeding> | in a 'constraints:' clause in cabal.project you can put 'any.' before a package name to apply the constraint to all targets being built, including built-tool dependencies |
| 2026-02-11 22:36:43 +0100 | <EvanR> | I think tomsmeding brought this up recently |
| 2026-02-11 22:36:36 +0100 | <EvanR> | can't that be encoded in the freeze file? |