fix(.github/codecov): Ignore aws-lc-rs-fips for codecov

[dongcarl]

An attempt to fix https://github.com/quinn-rs/quinn/actions/runs/11562272211/job/32183071070 on main and (unreleased) 0.5.6

[dongcarl]

It seems from the CI failure that code coverage builds for Windows for aws-lc-fips-sys is more complicated than expected.

More resources:

• https://github.com/aws/aws-lc/blob/main/BUILDING.md#build-prerequisites (seems to recommend ninja, nasm, cmake, and libunwind)

@mxinden @Ralith @djc I guess a decision has to be made if full coverage for aws-lc-fips-sys is worth the build system troubles

[djc]

What goal are you trying to achieve here?

[dongcarl]

@djc Just to fix https://github.com/quinn-rs/quinn/actions/runs/11562272211/job/32183071070

[djc]

@mxinden @Ralith @djc I guess a decision has to be made if full coverage for aws-lc-fips-sys is worth the build system troubles

I think we should definitely fix it so coverage builds can succeed (I'm surprised the main CI failures don't generate any notifications for me...). I think rustls has this working so it shouldn't be too hard to fix?

https://github.com/rustls/rustls/blob/main/.github/workflows/build.yml#L51

[dongcarl]

@djc Should be good to go now. Checks are passing: https://github.com/quinn-rs/quinn/actions/runs/12588265064

I can drop the last commit (meant to be temporary just for testing) if this looks good.

[gretchenfrage]

Yeah just copying what Rustls is doing makes sense, and the second commit which makes the coverage CI run in PRs combined with the fact that CI passed for this PR seems to indicate that the change works. Thank you for working on this irksome thing!

(We must remember not to merge until the second commit is removed.)

[djc]

So looking at this more, maybe we should exclude fips from the features used for coverage? The feature has no effect on Quinn itself, so it should not change anything about coverage. The disadvantage would be that we sort of need to keep features in sync manually, but IMO that might be better than showing everything down for a relatively niche use case.

[dongcarl]

So looking at this more, maybe we should exclude fips from the features used for coverage? The feature has no effect on Quinn itself, so it should not change anything about coverage. The disadvantage would be that we sort of need to keep features in sync manually, but IMO that might be better than showing everything down for a relatively niche use case.

Yeah that was my initial thinking. Let me know what you guys decide and I can adjust this PR.

[Ralith]

+1 in favor of excluding it, since our CI is getting pretty big already.

[codecov-commenter]

⚠️ Please install the to ensure uploads and comments are reliably processed by Codecov.

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 77.89%. Comparing base (f5b1ec7) to head (19273f5).
Report is 3 commits behind head on main.

❗ Your organization needs to install the Codecov GitHub app to enable full functionality.

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #2023      +/-   ##
==========================================
+ Coverage   77.43%   77.89%   +0.46%     
==========================================
  Files          69       71       +2     
  Lines       18167    18722     +555     
==========================================
+ Hits        14067    14584     +517     
- Misses       4100     4138      +38     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[dongcarl]

Updated to include all features except for aws-lc-rs-fips and rustls-aws-lc-rs-fips by parsing cargo metadata with jq. Let me know if this looks good.

Note that the last commit needs to be dropped.

[gretchenfrage]

I think this makes sense, although I wish there was a way built in to cargo to do this. But it looks like there might not be.

[djc]

Can we just get the flattened version of this, without all the shell/jq magic? That stuff is unfamiliar (at least to me) and thus harder to maintain, and I don't think we change our features all that often.