Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

proposal: testing: verbose output should print package path before printing test output #61647

willfaught opened this issue Jul 29, 2023 · 4 comments


Copy link

This makes it clear which package is being tested as the test output is printed.

For example, at the end of

=== RUN   TestFoo
--- PASS: TestFoo (0.00s)
ok	0.138s
=== RUN   TestBar

which package are we in? We don't know from the output. For large projects, where there are too many packages to memorize, especially their lexicographic order, this is valuable when you spot output you immediately want to investigate, but it takes a long time for the full package testing to finish, at which point the actual package path is printed out.

So, it would instead be something like

=== RUN
=== RUN   TestFoo
--- PASS: TestFoo (0.00s)
--- PASS:	0.138s
=== RUN
=== RUN   TestBar
@gopherbot gopherbot added this to the Proposal milestone Jul 29, 2023
Copy link

Perhaps the new -fullpath flag added for #37708 covers your use case?
You've shown a passing test, but that's usually uninteresting.
With a failing test it would print

--- FAIL: TestXXX (0.00s) it broke
exit status 1
FAIL	0.001s

Copy link

I don't think we can change the default output, but we could add a new flag like -test.fullpath if that is not enough.

Copy link
Contributor Author

@seankhliao I think that would indeed solve the problem in most cases, but only if you think to use -fullpath in the first place. If the current verbose test output format won't be changed, then I suppose remembering to use -fullpath is no worse than remembering to use a new -printpkg flag for the proposed behavior.

One case where it wouldn't work is if t.FailNow() is used without other output, which is probably rare, but still possible:

=== RUN   TestCreate
--- FAIL: TestCreate (0.00s)
=== RUN   TestDelete
--- FAIL: TestDelete (0.00s)

I didn't understand from #37708 the reason why changing the output format would break anyone. I could understand that position for changing the JSON format, but what would break if, say, a === RUN pkg/path line was added?

I do like the symmetry of before/after package delimiters for package results, like we have for tests, but perhaps that's just me.

Copy link
Contributor Author

I just ran across a different case where this would be useful, and -fullpath doesn't help.

If a test hangs forever, you can get output like this (all tests are parallel):

=== NAME  TestIntegrateRealQueueDeleteMessage
    integration_test.go:58: 2a
--- PASS: TestIntegrateRealQueueSendMessage (0.50s)
=== CONT  TestRetryQueueDeleteMessage
--- PASS: TestIntegrateRealServiceUserActive (1.46s)
=== CONT  TestRetryQueueReceiveMessage
--- PASS: TestRetryQueueDeleteMessage (0.05s)
=== CONT  TestRetry
--- PASS: TestRetry (0.00s)
--- PASS: TestRetryQueueReceiveMessage (0.05s)

There's no clear indication of what test is hanging. Nothing is logged, so -fullpath doesn't help.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Status: Incoming

No branches or pull requests

4 participants