vhd2vl: fix the tests #172443
vhd2vl: fix the tests #172443
Conversation
ofborg
bot
added
10.rebuild-darwin: 0
|
The ZHF announcement PR says: "Please ping @NixOS/nixos-release-managers on the PR and add the 0.kind: build failure label to the pull request. If you're unable to because you're not a member of the NixOS org please ping @dasJ, @tomberek, @jonringer, @Mic92" I am unable to add this build label, so I therefore am issuing this ping. |
| @@ -19,12 +19,8 @@ stdenv.mkDerivation rec { | |||
| }; | |||
|
|
|||
| patches = lib.optionals (!stdenv.isAarch64) [ | |||
There was a problem hiding this comment.
Did you check aarch64?
Checking now...
There was a problem hiding this comment.
Yeah, the weird behavior (which predates this PR) is confirmed: the patch to the tests must be applied on x86_64 and must not be applied on aarch64.
Apparently the parenthesization algorithm of iverilog's pretty-printer depends on the architecture. Wut.
However figuring out why this hack (which predates this PR) is necessary is probably not something to be done during ZHF.
The testing scheme for vhd2vl is sensitive to subtle shifts in iverilog's parenthesization choices, meaning that the golden test outputs require constant maintenance. The patch previously applied in order to deal with this situation is no longer sufficient, so a patch which is sufficient has been added. Also, the `buildTargets` and `checkTarget` attributes have been set, so future benign failures of this sort can be dealt with through `doCheck=false` in a pinch.
ghost
requested a review
from
@Mindavi the following just occurred to me: in light of the fact that this package produces nondeterministically-parenthesized outputs, and has not been updated since 2018, perhaps this would have been a good time to start the process of culling this package from the herd by allowing it to be marked as broken. I realized that ZHF is essentially nixpkgs' mechanism for managing this culling process. Perhaps your comment was a hint to this effect. If I interfered with that process, I apologize. I have no particular use or need for this package. Going forward, I'll take a moment to consider whether packages still belong in nixpkgs, before fixing them. |
|
No worries, someone else might find this package useful. But generally, if they're broken for a while, it may mean that there are no users and it can at some point be removed. As long as someone is interested enough (even out of curiosity) to patch a package up, it can stay. If nobody cares for an extended period of time, it's probably not an important package. But yeah, releases can be a good moment to clean out any old non-important packages, that's definitely true. I'd say that most people would love to keep everything, it mainly doesn't make sense from a maintenance standpoint to keep supporting everything indefinitely. Thanks for fixing things! ✨ |
ZHF: #172160
Description of changes
The testing scheme for vhd2vl is sensitive to subtle shifts in
iverilog's parenthesization choices, meaning that the golden test
outputs require constant maintenance.
The patch previously applied in order to deal with this situation is
no longer sufficient, so a patch which is sufficient has been added.
Also, the
buildTargetsandcheckTargetattributes have been set,so future benign failures of this sort can be dealt with through
doCheck=falsein a pinch.Things done
sandbox = trueset innix.conf? (See Nix manual)nix-shell -p nixpkgs-review --run "nixpkgs-review rev HEAD". Note: all changes have to be committed, also see nixpkgs-review usage./result/bin/)nixos/doc/manual/md-to-db.shto update generated release notes