2025/10/08

Newest at the top

2025-10-08 15:24:08 +0200MelodyOwO(~MelodyOwO@user/MelodyOwO) MelodyOwO
2025-10-08 15:23:49 +0200ZLima12(~zlima12@user/meow/ZLima12) (Ping timeout: 260 seconds)
2025-10-08 15:21:35 +0200comerijn(~merijn@77.242.116.146) (Ping timeout: 256 seconds)
2025-10-08 15:21:12 +0200Googulator77(~Googulato@2a01-036d-0106-03fa-5967-6f2a-1e2d-f05a.pool6.digikabel.hu) (Quit: Client closed)
2025-10-08 15:21:03 +0200ZLima12_(~zlima12@user/meow/ZLima12) ZLima12
2025-10-08 15:21:01 +0200Googulator76(~Googulato@2a01-036d-0106-03fa-5967-6f2a-1e2d-f05a.pool6.digikabel.hu)
2025-10-08 15:15:04 +0200 <tomsmeding> oh and it's also possible to manually add cost-centre annotations with CCS pragmas
2025-10-08 15:13:45 +0200fp(~Thunderbi@2001:708:150:10::72df) (Remote host closed the connection)
2025-10-08 15:10:43 +0200Googulator53(~Googulato@2a01-036d-0106-03fa-5967-6f2a-1e2d-f05a.pool6.digikabel.hu) (Quit: Client closed)
2025-10-08 15:10:38 +0200Googulator77(~Googulato@2a01-036d-0106-03fa-5967-6f2a-1e2d-f05a.pool6.digikabel.hu)
2025-10-08 15:08:50 +0200 <tomsmeding> the terminology is a little ambiguous around this topic
2025-10-08 15:08:45 +0200fp(~Thunderbi@2001:708:150:10::72df) fp
2025-10-08 15:08:25 +0200 <tomsmeding> yeah I got that, sorry
2025-10-08 15:08:10 +0200 <haskellbridge> <Morj> I think i misspoke, by annotations way above i meant compiler annotations inserted by profiling flags, not HasCallStack inserted by hand
2025-10-08 15:07:09 +0200guest3860(~username@2a04:27c0:0:d::e018) (Ping timeout: 250 seconds)
2025-10-08 15:07:05 +0200 <tomsmeding> and those are not necessarily the same
2025-10-08 15:06:36 +0200 <tomsmeding> so if you have HasCallStack _and_ profiling, you may get two backtraces for every printed exception :p
2025-10-08 15:06:15 +0200 <tomsmeding> Morj: and HasCallStack is also used for error reports
2025-10-08 15:05:17 +0200 <tomsmeding> in a strict language this is all moot because you have a normal call stack to work from
2025-10-08 15:04:57 +0200 <tomsmeding> people realised that and made it work :p
2025-10-08 15:04:45 +0200 <tomsmeding> but yeah the cost-centre annotations for execution time profiling are very nice for error reporting too
2025-10-08 15:04:45 +0200 <haskellbridge> <Morj> HasCallStack is location only
2025-10-08 15:04:38 +0200 <haskellbridge> <Morj> I know, I'm talking about error location/backtrace reporting, that's what I'm after
2025-10-08 15:04:00 +0200 <tomsmeding> HasCallStack is separate from profiling
2025-10-08 15:03:57 +0200 <haskellbridge> <Morj> Cabal with their docs again :/
2025-10-08 15:03:46 +0200 <haskellbridge> <Morj> I thought profiling is about execution time, not about error reporting, but it makes sense that annotations can work for both
2025-10-08 15:02:28 +0200 <tomsmeding> but that doesn't have the granularity that you're asking for
2025-10-08 15:01:53 +0200 <tomsmeding> Morj: there is this https://cabal.readthedocs.io/en/stable/cabal-project-description-file.html#cfg-field-profiling-det…
2025-10-08 15:01:26 +0200 <tomsmeding> not that I know of
2025-10-08 15:01:26 +0200 <haskellbridge> <Morj> For now I have a vim macro that will insert HasCallStack everywhere and recompile
2025-10-08 15:01:03 +0200 <haskellbridge> <Morj> Btw is there by chance a ghc option that will automatically collect evaluation stack in IO only?
2025-10-08 15:01:00 +0200poscat(~poscat@user/poscat) poscat
2025-10-08 15:00:58 +0200 <tomsmeding> https://hackage.haskell.org/package/base-4.21.0.0/docs/GHC-Stack.html#t:CallStack
2025-10-08 15:00:42 +0200 <tomsmeding> bwe: the low-tech alternative to all this is to add HasCallStack constraints to functions that call `last`
2025-10-08 15:00:35 +0200poscat(~poscat@user/poscat) (Remote host closed the connection)
2025-10-08 15:00:30 +0200fp(~Thunderbi@2001:708:150:10::72df) (Remote host closed the connection)
2025-10-08 15:00:19 +0200 <tomsmeding> :)
2025-10-08 15:00:11 +0200 <bwe> tomsmeding: Well, I just eliminated any occurence of last and I am done. That's simpler than restarting repl with profiling options…
2025-10-08 14:59:34 +0200 <tomsmeding> bwe: nice
2025-10-08 14:59:28 +0200 <bwe> tomsmeding: 9.12.2
2025-10-08 14:59:28 +0200 <tomsmeding> I consider this smart_filter stuff essential to even consider participating in a room like this :p
2025-10-08 14:58:00 +0200humasect(~humasect@dyn-192-249-132-90.nexicom.net) (Ping timeout: 265 seconds)
2025-10-08 14:57:32 +0200 <haskellbridge> <Morj> Lowercase morj is my irc nick, but I only use it from one PC which is not this one
2025-10-08 14:56:58 +0200 <tomsmeding> I did that without thinking when Morj didn't autocomplete, but that was because you're over the bridge lol
2025-10-08 14:56:41 +0200fp(~Thunderbi@2001:708:150:10::72df) fp
2025-10-08 14:56:34 +0200 <haskellbridge> <Morj> I did think there already exist some that are reasonable, it's an old protocol
2025-10-08 14:56:33 +0200 <tomsmeding> that @tell is a lambdabot thing that sends that to you in a private message when you next join on irc
2025-10-08 14:56:19 +0200 <haskellbridge> <Morj> Thanks for the info tоmsmeding and mаuke
2025-10-08 14:56:15 +0200 <tomsmeding> oh brainfart
2025-10-08 14:56:06 +0200 <haskellbridge> <Morj> I really want to know how that at-tell is supposed to look without a bridge