feat: run GNU testsuite

[garasubo]

Fix #5

Added CI to run GNU testsuite for this project

[jmr]

This is useful info, but having this check fail on every commit is suboptimal. Should the failures be compared to a list of expected failures? Is there a finer-grained way to control the tests run?

[garasubo]

@jmr Thank you for your review!

having this check fail on every commit is suboptimal.

I agree. uutils/coreutils repository has a similar CI.
It uploads a reference report file as a github artifact, and compare the current test result against it to detect new failures. This way, the check doesn't fail on every commit.

I can implement a similar approach here if that sounds good to you.

[jmr]

I can implement a similar approach here if that sounds good to you.

FWIW (nothing), SGTM.

I hope some form of this can get merged. It already pointed out missing features (#528). Now, if we could get some PRs reviewed/merged...

[garasubo]

@jmr Hi, I updated this PR to compare the test result againt the reference. So, we won't see the check fail on every commit. I'd appreciate if you could review it. Thanks.

[sylvestre]

can we make this list dynamic ?

[jmr]

Could use:

UTILS=($(ls "$PROJECT_DIR/src/uu/"))
...
"${UTILS[@]}"

Same for gen-test-helper.sh.

Something fancier will be needed for setarch/i386/....

[jmr]

https://github.com/uutils/coreutils/blob/ec72d08bf1799d53b11d12502eb8cabe67d89297/src/bin/coreutils.rs#L87-L101

[jmr]

#535