| 21 Oct 2024 |
 Fabián Heredia | just remove the --only-throws flag | 17:20:18 |
 emily | yeah, so it has removed nodejs_16 but left nodejs-16_x = nodejs_16, e.g. | 17:21:33 |
| emily | I guess that's a reason to not do that kind of thing | 17:21:39 |
| emily | those seem to be the only cases, perhaps you could clean up the broken Node aliases in another commit? | 17:22:25 |
| Fabián Heredia | done | 17:25:17 |
 Alyssa Ross | Eval broke | 17:30:40 |
| Alyssa Ross | https://github.com/NixOS/nixpkgs/actions/runs/11445435644/job/31842486226?pr=350290 | 17:30:48 |
| Alyssa Ross | gnuradio-with-packages is still in aliases.nix, but depends on gnuradio_3_7, which was removed.
| 17:31:12 |
| Alyssa Ross | * gnuradio-with-packages is still in aliases.nix, but depends on gnuradio3_7, which was removed.
| 17:31:23 |
| emily | gah. should have pored over the diff more carefully. or figured out how you're meant to test aliases.nix eval | 17:31:42 |
| emily | will fix | 17:31:47 |
| Alyssa Ross | or waiting for CI :P | 17:31:53 |
| Alyssa Ross | * | 17:31:58 |
| emily | (but also how do you test the latter? nixpkgs-review won't do it, right?) | 17:32:01 |
| emily | In reply to @qyliss:fairydust.space
or waited for CI :P yeah, sorry :( | 17:32:08 |
| emily | I wish CI was faster so I didn't feel the need to replicate it locally all the time | 17:32:31 |
| Alyssa Ross | In reply to @emilazy:matrix.org
(but also how do you test the latter? nixpkgs-review won't do it, right?) No idea, although I guess I've just learned that nixpkgs-vet does it. | 17:32:33 |
| Alyssa Ross | nixpkgs-vet is fast :P | 17:32:38 |
| emily | right. not yet used to having nixpkgs-vet around | 17:33:07 |
| emily | (how come it can do eval checks so much quicker?) | 17:33:14 |
| emily | also lol at us having this expression lying around that points to a throw in the first place…
gnuradio-with-packages = gnuradio3_7.override {
extraPackages = lib.attrVals [
"osmosdr"
"ais"
"gsm"
"nacl"
"rds"
"limesdr"
]
gnuradio3_7Packages;
}; # Added 2020-10-16
| 17:33:40 |
| Fabián Heredia | https://github.com/NixOS/nixpkgs/pull/350291 | 17:34:07 |
| emily | In reply to @qyliss:fairydust.space
No idea, although I guess I've just learned that nixpkgs-vet does it. (the nixpkgs-vet author is for some reason opposed to having it as a first-class package in Nixpkgs, so unfortunately it is not at all convenient to run locally) | 17:34:12 |
| emily | In reply to @fabianhjr:matrix.org
https://github.com/NixOS/nixpkgs/pull/350291 ah, I was just about to do this 😅 | 17:34:25 |
| emily | I'll be more patient this time. | 17:34:34 |
| emily | I guess having aliases that point to other throwing aliases is a pretty bad antipattern since it breaks our tooling | 17:36:02 |
| emily | I've introduced some of those, I wonder if we can programmatically rule them out in future. | 17:36:13 |
| emily | Fabián Heredia: I'm thinking we should just revert and do this again if the whack-a-mole with the eval issues continues beyond another push | 17:39:56 |
| emily | I'll open a revert | 17:40:38 |
| emily | merged the revert, sorry about that. better to clean up the strays without time pressure | 17:45:02 |
