| 14 Oct 2024 |
 hellwolf | fwiw, I am wrapping up the THSH project which mixes Haskell and shells, some tool I always wanted to have. | 17:06:15 |
 emily | maralorn: just a ping to check about progress on release note/announcement/survey stuff? | 17:30:39 |
 maralorn | emily: Well, frankly the policy is stuck in internal discussion. | 17:55:59 |
| emily | I thought the survey/request for feedback was to get data to help inform the policy 😅 | 17:57:34 |
| emily | but ok, good to know | 17:57:36 |
| emily | dropping or relaxing the LLVM deps for old GHCs is still a possible backstop even if they need to stay for now I guess? since from previous discussions it seems like the conservative bounds were not really related to code changes given the use of textual IR etc. | 18:04:33 |
| maralorn | Well, the problem is that I wrote that policy in an attempt to shortcut the survey thing, in the hope that it would reach the same people. | 18:06:47 |
| maralorn | But apparently we need a survey because it is too hard to agree on a policy without the data. | 18:07:06 |
| maralorn | The main thing holding me from writing a survey is actually knowing which tool to use. | 18:07:47 |
| maralorn | * The main thing holding me from writing a survey is actually knowing which tool/server to use. | 18:08:01 |
| maralorn | In reply to @emilazy:matrix.org
dropping or relaxing the LLVM deps for old GHCs is still a possible backstop even if they need to stay for now I guess? since from previous discussions it seems like the conservative bounds were not really related to code changes given the use of textual IR etc. I will refrain from further commenting on llvm related details because I am setting the discussion up for failure with my cluelessness. | 18:08:58 |
| emily | I hear CIVS has only had one data loss incident since 2006 😅 | 18:09:49 |
| emily | there's always something boring like Google Forms, though I imagine some people may object to that | 18:10:40 |
 alexfmpe | In reply to @maralorn:maralorn.de
I write a lot of bash scripts in Haskell for my nixos systems, with template haskell to ensure presence of called programs at compile time. 😄 Do you use https://hackage.haskell.org/package/which or something else? | 18:39:47 |
| hellwolf | huh, and is ghc922 going to be 9.22 or 9.2.2 :) | 19:07:13 |
| hellwolf | * huh, and is ghc924 going to be 9.24 or 9.2.4 :) | 19:07:28 |
| maralorn | In reply to @alexfmpe:matrix.org
Do you use https://hackage.haskell.org/package/which or something else? I recently switched to that yeah. In most places I use shh. | 19:08:24 |
| maralorn | In reply to @hellwolf:matrix.org
huh, and is ghc924 going to be 9.24 or 9.2.4 :) Grave Mistakes have been made. | 19:08:55 |
| maralorn | In reply to @hellwolf:matrix.org
huh, and is ghc924 going to be 9.24 or 9.2.4 :) * | 19:09:05 |
| alexfmpe | I blame cabal macros | 19:10:33 |
| alexfmpe | turns out 920 is 9.20, not 9.2.0 | 19:11:06 |
| alexfmpe | if we ever get to 9.99, there will be hell to pay | 19:11:53 |
| alexfmpe | or 99.9 ? | 19:12:05 |
| alexfmpe | guess 12.2 is the first plausible "uh oh" ? | 19:12:46 |
 Tristan Ross | In reply to @maralorn:maralorn.de
Well, the problem is that I wrote that policy in an attempt to shortcut the survey thing, in the hope that it would reach the same people. Yeah, a policy based on a survey would be best. Has there been any updates on getting a survey out? | 19:13:21 |
| hellwolf | Also, this is my new observation:
(haskell.compiler.${ghcVer}.override { enableDocs = false; }) triggers a cache miss, I guess the nix expression is not optimal in that docs disabled shouldn't cause a rebuild ghcVer from the previous ghcVer
| 19:31:42 |
| hellwolf | * Also, this is my new observation:
(haskell.compiler.${ghcVer}.override { enableDocs = false; }) triggers a cache miss, I guess the nix expression is not optimal in that docs disabled shouldn't cause a rebuild ghcVer from the previous ghcVer.
Downloading everything is still better than rebuilding.. so I fallback
| 19:32:16 |
| hellwolf | * Also, this is my new observation:
(haskell.compiler.${ghcVer}.override { enableDocs = false; }) triggers a cache miss, I guess the nix expression is not optimal in that docs disabled shouldn't cause a rebuild ghcVer from the previous ghcVer.
Downloading more from cache is still better than rebuilding.. so I fallback
| 19:32:36 |
| hellwolf | * Also, this is my new observation:
(haskell.compiler.${ghcVer}.override { enableDocs = false; }) mises the cache, I guess the nix expression is not optimal in that docs disabled shouldn't cause a rebuild ghcVer from the previous ghcVer.
Downloading more from cache is still better than rebuilding.. so I fallback
| 19:32:57 |
| hellwolf | * Also, this is my new observation:
(haskell.compiler.${ghcVer}.override { enableDocs = false; }) mises the cache -> I guess the nix expression is not optimal in that docs disabled shouldn't cause a rebuild ghcVer from the previous ghcVer.
Downloading more from cache is still better than rebuilding.. so I fallback
| 19:33:01 |
