shell: add experimental_test_shell_command for arbitrary shell-driven tests
#17640
Conversation
tdyas
requested review from
Eric-Arellano,
stuhood,
chrisjrn and
benjyw
and removed request for
stuhood
|
Example from pants-go-testing: tdyas/pants-go-testing@64a0d49 |
tdyas
force-pushed
the
shell_test_support
branch
from
9cc7c4e to
679b5a9
Compare
| @@ -393,6 +403,42 @@ class ShellCommandRunTarget(Target): | |||
| ) | |||
|
|
|||
|
|
|||
| class ShellCommandTestTarget(Target): | |||
| alias = "experimental_test_shell_command" | |||
There was a problem hiding this comment.
This name is a bit of a mouthful. I'm wondering about naming it experimental_generic_test? The "shell" name here (and in experimental_shell_command) is a bit misleading, since this can run arbitrary processes, it's not limited to just shell constructs. And it's also easily confused with "tests shell code".
There was a problem hiding this comment.
I don't have a particular care as to what the name is. Please feel free to just tell me what the name should be.
There was a problem hiding this comment.
Also if called generic_test, should it be in a separate generic backend instead of the shell backend?
There was a problem hiding this comment.
We can leave as-is for now, since we have to rename experimental_shell_command anyway before we de-experimentalize it.
Add an
experimental_test_shell_commandtarget type to enable defining arbitrary tests using shell code.Note: This target is not a unit test of shell sources like
shunit2_tests; rather, it is a test defined in shell which could test anything which can be packaged and consumed by the other shell targets (e.g.,experimental_shell_commandandexperimental_run_shell_command). It is basicallyexperimental_run_shell_commandbut for thetestgoal instead of therungoal.