Logs: liberachat/#haskell
| 2021-06-21 11:17:02 | × | alp quits (~alp@user/alp) (Ping timeout: 268 seconds) |
| 2021-06-21 11:17:09 | × | haveo quits (~haveo@sl35.iuwt.fr) (Ping timeout: 272 seconds) |
| 2021-06-21 11:17:16 | → | haveo joins (~haveo@sl35.iuwt.fr) |
| 2021-06-21 11:17:36 | × | pretty_dumm_guy quits (trottel@gateway/vpn/protonvpn/prettydummguy/x-88029655) (Ping timeout: 252 seconds) |
| 2021-06-21 11:17:55 | → | MasseR joins (~MasseR@51.15.143.128) |
| 2021-06-21 11:18:10 | × | Ariakenom quits (~Ariakenom@2001:9b1:efb:fc00:315a:5813:efa5:bde0) (Ping timeout: 250 seconds) |
| 2021-06-21 11:18:51 | → | stevenxl joins (~stevenlei@174.128.182.199) |
| 2021-06-21 11:19:45 | → | Hafydd joins (~Hafydd@user/hafydd) |
| 2021-06-21 11:20:11 | <merijn> | It's funny that people think "package take-over process is comically inefficient" is a bug as opposed to "intended by design" |
| 2021-06-21 11:20:40 | <merijn> | It's almost as if it's something we want to discourage... |
| 2021-06-21 11:21:23 | <Las[m]> | I have a Haskell program that segfaults somewhere, and I'm trying to get a stack trace for that. I'm compiling with -g and -O1, but inspecting the dump with gdb and running `bt` just gives me garbage. How can I debug this? |
| 2021-06-21 11:21:55 | <Las[m]> | The Haskell program uses raw pointers all over the place unfortunately |
| 2021-06-21 11:22:20 | <merijn> | RIP |
| 2021-06-21 11:22:31 | × | srk quits (~sorki@user/srk) (Remote host closed the connection) |
| 2021-06-21 11:22:49 | → | srk joins (~sorki@user/srk) |
| 2021-06-21 11:22:52 | <merijn> | -g just gives you c symbols for the RTS |
| 2021-06-21 11:22:57 | <merijn> | not the compiled Haskell |
| 2021-06-21 11:23:16 | <merijn> | GHC has (experimentalish) support for DWARF, but I don't think -g enables that atm? |
| 2021-06-21 11:23:34 | × | bitdex quits (~bitdex@gateway/tor-sasl/bitdex) (Quit: = "") |
| 2021-06-21 11:25:39 | <Las[m]> | FWIW I get some mangled symbol name, can I somehow use that? |
| 2021-06-21 11:25:57 | <Las[m]> | I used the -ddump thing to get some dumps but I can't find the symbol name there |
| 2021-06-21 11:26:27 | <merijn> | Las[m]: https://www.haskell.org/ghc/blog/20200403-dwarf-1.html |
| 2021-06-21 11:26:30 | <Las[m]> | > Emit debug information in object code. Currently only DWARF debug information is supported on x86-64 and i386. Currently debug levels 0 through 3 are accepted, with 0 disabling debug information production and higher numbers producing richer output. If ⟨n⟩ is omitted, level 2 is assumed. |
| 2021-06-21 11:26:32 | <lambdabot> | <hint>:1:24: error: parse error on input ‘in’ |
| 2021-06-21 11:26:41 | <Las[m]> | it's not supported on my platform... |
| 2021-06-21 11:26:45 | <merijn> | ah |
| 2021-06-21 11:26:47 | <Las[m]> | (aarch64-linux) |
| 2021-06-21 11:27:05 | <merijn> | Then you might be, as we experts say, "fucked" >.> |
| 2021-06-21 11:27:12 | <Las[m]> | can I use the mangled symbol name then somehow to at least find out where it happened? |
| 2021-06-21 11:27:38 | <merijn> | No clue, tbh |
| 2021-06-21 11:28:03 | <merijn> | tbh, at this point binary searching your code by throwing Debug.Trace prints everywhere might be the fastest/most accurate to figure out what fails and where |
| 2021-06-21 11:28:10 | <tomsmeding> | what _is_ the mangled symbol name? |
| 2021-06-21 11:28:24 | → | raehik joins (~raehik@cpc95906-rdng25-2-0-cust156.15-3.cable.virginm.net) |
| 2021-06-21 11:28:52 | <Las[m]> | I need to generate the coredump again for that since it was purged after a reboot |
| 2021-06-21 11:29:34 | <tomsmeding> | just curious, no need to expend effort :) |
| 2021-06-21 11:30:37 | × | is7s quits (~is7s@2a01:4b00:895f:3d00:ddd:ae00:86d4:d79c) (Quit: Client closed) |
| 2021-06-21 11:30:41 | × | maroloccio quits (~marolocci@189.15.9.54) (Quit: Client closed) |
| 2021-06-21 11:30:46 | → | alp joins (~alp@163.172.83.213) |
| 2021-06-21 11:31:01 | <Las[m]> | it says `c2Xrz_info$def` |
| 2021-06-21 11:31:32 | <tomsmeding> | right, I have no idea :p |
| 2021-06-21 11:31:37 | <merijn> | Sounds like an info pointer |
| 2021-06-21 11:31:59 | <Las[m]> | what does that mean? |
| 2021-06-21 11:32:16 | <merijn> | Las[m]: info pointers are internal meta information for GHC about datatypes |
| 2021-06-21 11:32:25 | <Las[m]> | ah |
| 2021-06-21 11:32:27 | <merijn> | and closures |
| 2021-06-21 11:32:49 | <merijn> | Like "is this thunk already evaluated?" "how big is the result?" etc. |
| 2021-06-21 11:33:02 | <merijn> | "where is the code for this?" |
| 2021-06-21 11:33:38 | <Las[m]> | I think I'll have to go the Debug.Trace route like you said... |
| 2021-06-21 11:33:44 | → | mpt joins (~tom@2a02:908:1862:49e0::3) |
| 2021-06-21 11:36:30 | × | wei2912 quits (~wei2912@112.199.250.21) (Quit: Lost terminal) |
| 2021-06-21 11:38:47 | × | haveo quits (~haveo@sl35.iuwt.fr) (Quit: leaving) |
| 2021-06-21 11:42:13 | → | haveo joins (~haveo@sl35.iuwt.fr) |
| 2021-06-21 11:46:49 | × | fizbin quits (~fizbin@c-73-33-197-160.hsd1.nj.comcast.net) (Remote host closed the connection) |
| 2021-06-21 11:47:30 | × | azeem quits (~azeem@176.201.29.233) (Ping timeout: 265 seconds) |
| 2021-06-21 11:53:33 | × | lavaman quits (~lavaman@98.38.249.169) (Remote host closed the connection) |
| 2021-06-21 11:53:46 | → | azeem joins (~azeem@176.201.29.233) |
| 2021-06-21 11:56:26 | → | brandonh joins (~brandonh@151.68.5.207) |
| 2021-06-21 12:01:44 | → | waleee joins (~waleee@2001:9b0:216:8200:d457:9189:7843:1dbd) |
| 2021-06-21 12:01:56 | × | mpt quits (~tom@2a02:908:1862:49e0::3) (Ping timeout: 250 seconds) |
| 2021-06-21 12:03:56 | <maerwald[m]> | Is there a tool to vendor 3rd party deps (not on hackage) into the library component so I can upload it on hackage? |
| 2021-06-21 12:04:27 | <merijn> | maerwald[m]: Eh "copy and paste the source into your repo"? :p |
| 2021-06-21 12:04:34 | <merijn> | maerwald[m]: iow, no |
| 2021-06-21 12:04:51 | <maerwald[m]> | Yeah, that, adjust modules in cabsl file |
| 2021-06-21 12:04:56 | <maerwald[m]> | And all the jazz |
| 2021-06-21 12:05:16 | <maerwald[m]> | Seems this can be automated |
| 2021-06-21 12:05:34 | <merijn> | Sure, just a Small Matter of Programming (TM) :) |
| 2021-06-21 12:06:25 | <maerwald[m]> | OCD lock-in... 3 months later... "done... so back to what I was working on... what was I working on?" |
| 2021-06-21 12:06:34 | → | machinedgod joins (~machinedg@24.105.81.50) |
| 2021-06-21 12:07:03 | <merijn> | maerwald[m]: That never happens to me >.> |
| 2021-06-21 12:08:10 | <merijn> | maerwald[m]: Can't convince the 3rd party to make a Hackage release? |
| 2021-06-21 12:08:22 | <maerwald[m]> | No |
| 2021-06-21 12:09:14 | <maerwald[m]> | My tar patches haven't been merged since 1.5 years and the other library smells like NMU needed too |
| 2021-06-21 12:09:22 | <maerwald[m]> | No time for that nonsense |
| 2021-06-21 12:09:38 | <merijn> | Doesn't vanessa actively maintain a tar library? |
| 2021-06-21 12:10:01 | <merijn> | At least, she's the only person I see shitposting about tar being terrible more than you :p |
| 2021-06-21 12:10:12 | <maerwald[m]> | Yes, I'm using that |
| 2021-06-21 12:10:21 | <maerwald[m]> | But it doesn't build everywhere |
| 2021-06-21 12:10:27 | <maerwald[m]> | Thanks to c2hs |
| 2021-06-21 12:10:46 | <merijn> | She's not including your patches? |
| 2021-06-21 12:11:10 | <maerwald[m]> | No no |
| 2021-06-21 12:11:16 | <maerwald[m]> | She's maintaining libarchive |
| 2021-06-21 12:11:34 | <maerwald[m]> | That works well for platforms where c2hs is not broken |
| 2021-06-21 12:12:18 | <merijn> | I'd argue that c2hs is broken everywhere xD |
| 2021-06-21 12:12:32 | <merijn> | It just happens to work sometimes :p |
| 2021-06-21 12:13:21 | <maerwald[m]> | For the others I need tar, which I had initially forked to bytestring based filepaths, but when supporting windows, this would require implementing AFPP first or some light wrapper. So I'm back to using a patched branch of upstream haskell tar, which is more or less unmaintained. |
| 2021-06-21 12:14:50 | <maerwald[m]> | So I don't care to fix any on that, just want to be able to make a hackage release. |
| 2021-06-21 12:16:05 | <maerwald[m]> | So need a tool that vendors all deps into the codebas and call it a day, haha |
| 2021-06-21 12:16:11 | <maerwald[m]> | Don't care |
| 2021-06-21 12:16:47 | <maerwald[m]> | Not much different from using PVP bounds |
| 2021-06-21 12:18:03 | → | Morrow joins (~Morrow@37.142.145.18) |
| 2021-06-21 12:18:57 | → | Ariakenom joins (~Ariakenom@2001:9b1:efb:fc00:8de:a8d1:83fb:c4a6) |
| 2021-06-21 12:19:01 | <merijn> | maerwald[m]: So, how about just copying the source tree into a separate subdir, copying the library entry from the original cabal file and making it an internal library in your package? |
| 2021-06-21 12:19:13 | <merijn> | Still *some* manual work, but not a lot |
| 2021-06-21 12:19:23 | → | lbseale joins (~lbseale@user/ep1ctetus) |
| 2021-06-21 12:19:33 | <maerwald[m]> | For tar yes... the other is a transitive dep |
| 2021-06-21 12:19:42 | <adamse> | how about uploading your fork to hackage under a new name? |
| 2021-06-21 12:20:00 | <maerwald[m]> | That's gonna be fun |
| 2021-06-21 12:20:14 | <merijn> | I'm sure Hecate has tons of recommendations on yak herding and shaving implements :p |
| 2021-06-21 12:20:39 | derelict_ | is now known as derelict |
| 2021-06-21 12:20:39 | <maerwald[m]> | Hackage should have had fork support |
All times are in UTC.