-
Notifications
You must be signed in to change notification settings - Fork 2.4k
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
shell_completions test issues #14545
Comments
Disable the shell_completions tests The shell_completions tests have a few issues that are causing some problems. See #14545 for a description. This disables the tests until those can get resolved.
This reverts commit 47f4cdc. Fixes rust-lang#14545
This was deprecated in 0.18 which is also when |
The tests were added in #14493 for #14520. As our current completions do not have tests, not having tests for these new completions is on-par. As discussed in #14493 (review), the primary focus for the tests that call shells is to make sure the integration of This could mean we drop the number of shells we verify in CI. If one works, they should all work. That could speed things up, reduce demand for custom tools, and possibly we can pick one that does a good job of threading the different failure cases. |
To add to that, we also started this experiment with enabling all shells |
test: Remove completion tests The tests are intended to spot check that shell completions are registered correctly. That is a low change, low risk area. For shell integration, we're relying on `clap_complete`s tests. For our own candidates, we should test the candidate generation directly, rather than end-to-end. This reverts parts of commit e7ca9be, reversing changes made to bd5f32b. Fixes #14545
Problem
The new shell_completions tests have a few issues that need to be resolved before they can be enabled.
fish
andzsh
randomly fail. Gate shell_completions tests on availability #14541 (comment) indicates that it might be an interactive timeout.bash
test also requires the bash-completions package to be installed.ignore
attribute so that it reports in the output when it isn't running. Currently they exit withreturn;
on macos, but that should use theignore
attribute instead.Steps
cargo test --test testsuite -- shell_completions
Possible Solution(s)
One idea I had for limiting where these tests run is to add a CI_EXTENDED environment variable. The tests would only be required to run if that environment variable is set, and then we can set that in the appropriate jobs (like
test
). Then the cargo_test macro could have something likerequires_extended_fish
or something like that. That would also be useful for the currently hard-codedhg
andlldb
.However, that doesn't help with some of the more complex issues like the bash-completions problem.
We could extend the cargo_test attribute to have
ignore_macos
to fix the macos ignore problem.Notes
No response
Version
Currently on 643a025 of master.
The text was updated successfully, but these errors were encountered: