| 21 Oct 2024 |
 maralorn | But it might be a neccessary calculated risk. | 13:00:10 |
| maralorn | * | 13:00:17 |
 Artem | In reply to @sternenseemann:systemli.org
Solid: it's hard to give a formal requirement, but basically we'd want to be confident that new maintainers are familiar enough with the internals and established conventions so that they can handle a rotation on their own and guide drive by contributors.
You also don't have to take on such a big responsibility to help out, it already helps a lot if people follow maintenance and help out occasionally with fixing regressions or unbreaking packages. The probably best thing to start out with is taking over maintenance for specific haskell packages you are personally familiar with and/or use yourself. You'll be notified automatically when there's something to fix and you'll gain experience working with Haskell in nixpkgs. Bonus is that it benefits yourself because the package won't break in the channels on your watch.
following this advice, I opened https://github.com/NixOS/nixpkgs/pull/350301! | 18:28:52 |
 sterni | maralorn: hm there isn't a ghc-lib-parser yet | 19:35:21 |
| sterni | I guess downgrade ghc98 to 9.8.2… | 19:35:32 |
|  (artur 'manuel) changed their display name from (lambda (f l) (format nil "~a ~a")) "Artur" "Manuel" to (artur 'manuel). | 20:02:55 |
| sterni | PSA: https://github.com/NixOS/nixpkgs/pull/346720#issuecomment-2427628892 | 20:14:54 |
| maralorn | sterni: excellent | 20:25:25 |
| 22 Oct 2024 |
 chreekat | What's the deal about ghc-lib-parser? | 05:21:24 |
|  @zarel_it:matrix.org left the room. | 09:31:33 |
| maralorn | chreekat: It needs to be compatible with the used ghc version, apparently we switched so fast to 9.8.3 that there wasn’t a compatible release with that. | 11:07:45 |
| maralorn | In reply to * @sternenseemann:systemli.org
should really start on the contributing manual section… You wanna do that in the nixpkgs manual? | 11:08:20 |
| chreekat | maralorn: ah i see. Well in that case I have good news https://hackage.haskell.org/package/ghc-lib-parser-9.8.3.20241022 | 11:09:05 |
| chreekat | those folks are quick! | 11:09:20 |
| maralorn | Yeah, but not as fast as sterni … | 11:09:32 |
| chreekat | hehe | 11:09:36 |
| maralorn | Anyway, our current PR is merged already. But we can merily switch to 9.8.3 on the next one. | 11:10:11 |
| * chreekat is thinking of necroing https://github.com/NixOS/nixpkgs/issues/41340 because he also hit an "argument list is too long" issue | 11:12:12 |
| chreekat | in that issue, cdepillabout reports a problem while using buildStackProject. It looks different to the main problem of the ticket, so maybe I should open another one? In our case (I have the same set of conditions), the thing getting repeated is -I/foo/bar/baz | 11:13:11 |
| chreekat | When I split them up and uniq them, I see e.g.
177 -I/nix/store/4lssw6pk59r654j0yrfydv0kngk8zrqp-pkg-config-wrapper-0.29.2/include
177 -I/nix/store/55h89a2wp6vi5dndcyz8c5sgl2ahmr4r-postgresql-15.8/include
177 -I/nix/store/77i5izsm6i7fyzdb5h8w28rcpawwqj6q-zlib-1.3.1-dev/include
177 -I/nix/store/a8ihcrk34ls3mdrzbvqgqqkdnx8alwgn-gmp-with-cxx-6.3.0/include
177 -I/nix/store/ajqf8b7qvcz6qzamx2fv6y1qvvs2zqa8-libsodium-1.0.19/include
177 -I/nix/store/i2ib184w4f4r644ciq9miwvqkk1334zd-libsodium-1.0.19-dev/include
177 -I/nix/store/pkl664rrz6vb95piixzfm7qy1yc2xzgc-zlib-1.3.1/include
178 -I/nix/store/qzm3pxhfrh17mglh9afywg727i2b4vwd-gmp-with-cxx-6.3.0-dev/include
176 -I/nix/store/r6k305g2rn0qqkvdxfvzqg7mhh7rw679-gcc-wrapper-13.2.0/include
1 -I/nix/store/r6k305g2rn0qqkvdxfvzqg7mhh7rw679-gcc-wrapper-13.2.0/include -DSTACK_DEVELOPER_MODE_DEFAULT=False
177 -I/nix/store/vdd8ccp9x69v7kp5g9khq7j43n6dqf16-git-2.44.1/include
| 11:13:44 |
| chreekat | That's a lot of repeats | 11:13:48 |
| maralorn | In reply to @b:chreekat.net
When I split them up and uniq them, I see e.g.
177 -I/nix/store/4lssw6pk59r654j0yrfydv0kngk8zrqp-pkg-config-wrapper-0.29.2/include
177 -I/nix/store/55h89a2wp6vi5dndcyz8c5sgl2ahmr4r-postgresql-15.8/include
177 -I/nix/store/77i5izsm6i7fyzdb5h8w28rcpawwqj6q-zlib-1.3.1-dev/include
177 -I/nix/store/a8ihcrk34ls3mdrzbvqgqqkdnx8alwgn-gmp-with-cxx-6.3.0/include
177 -I/nix/store/ajqf8b7qvcz6qzamx2fv6y1qvvs2zqa8-libsodium-1.0.19/include
177 -I/nix/store/i2ib184w4f4r644ciq9miwvqkk1334zd-libsodium-1.0.19-dev/include
177 -I/nix/store/pkl664rrz6vb95piixzfm7qy1yc2xzgc-zlib-1.3.1/include
178 -I/nix/store/qzm3pxhfrh17mglh9afywg727i2b4vwd-gmp-with-cxx-6.3.0-dev/include
176 -I/nix/store/r6k305g2rn0qqkvdxfvzqg7mhh7rw679-gcc-wrapper-13.2.0/include
1 -I/nix/store/r6k305g2rn0qqkvdxfvzqg7mhh7rw679-gcc-wrapper-13.2.0/include -DSTACK_DEVELOPER_MODE_DEFAULT=False
177 -I/nix/store/vdd8ccp9x69v7kp5g9khq7j43n6dqf16-git-2.44.1/include
Huh, that’s a nice insight. | 11:14:16 |
| chreekat | I just don't have any idea where that might be happening. nixpkgs? stack? cabal? ghc? Haven't been able to repro it yet with the few things I've tried | 11:15:15 |
| chreekat | so not even sure where to open a ticket first | 11:16:18 |
| maralorn | Don’t we classically have those problem with pkg-config? | 11:16:36 |
| maralorn | But this might be a tottaly different one. | 11:16:45 |
| chreekat | this project does depend on pkg-config | 11:17:02 |
| chreekat | uh ok either i've cached some failure, or repro is very easy.
stack new --resolver lts-20.5 blah
cd blah
stack build stack
| 11:34:02 |
| maralorn | Is there an easy way to tell stack to just use all dependencies and ghc from the nix-provided environment and just build my project? Basically ignoring any snapshots? | 11:38:55 |
| chreekat | I rather doubt it. There used to be a way to say "just use the packages that come with the compiler", which probably would work, but I think that went away when casa came into the scene | 11:40:04 |
