mirror of https://github.com/databricks/cli.git
a5e09ab28a
## Changes Add two new make commands: - make acc-cover: runs acceptance tests and outputs coverage-acceptance.txt - make acc-showcover: show coverage-acceptance.txt locally in browser Using the GOCOVERDIR functionality: https://go.dev/blog/integration-test-coverage This works, but there are a couple of issues encountered: - GOCOVERDIR does not play well with regular "go test -cover". Once this fixed, we can simplify the code and have 'make cover' output coverage for everything at once. We can also probably get rid of CLI_GOCOVERDIR. https://github.com/golang/go/issues/66225 - When running tests in parallel to the same directory there is rare conflict on writing covmeta file. For this reason each tests writes coverage to their own directory which is then merged together by 'make acc-cover'. <!-- Summary of your changes that are easy to understand -- ## Tests Manually running the new make commands. |
||
---|---|---|
.. | ||
bin | ||
build | ||
bundle | ||
help | ||
README.md | ||
acceptance_test.go | ||
script.cleanup | ||
script.prepare | ||
server_test.go |
README.md
Acceptance tests are blackbox tests that are run against compiled binary.
Currently these tests are run against "fake" HTTP server pretending to be Databricks API. However, they will be extended to run against real environment as regular integration tests.
To author a test,
- Add a new directory under
acceptance
. Any level of nesting is supported. - Add
databricks.yml
there. - Add
script
with commands to run, e.g.$CLI bundle validate
. The test case is recognized by presence ofscript
.
The test runner will run script and capture output and compare it with output.txt
file in the same directory.
In order to write output.txt
for the first time or overwrite it with the current output pass -update flag to go test.
The scripts are run with bash -e
so any errors will be propagated. They are captured in output.txt
by appending Exit code: N
line at the end.
For more complex tests one can also use:
errcode
helper: if the command fails with non-zero code, it appendsExit code: N
to the output but returns success to caller (bash), allowing continuation of script.trace
helper: prints the arguments before executing the command.- custom output files: redirect output to custom file (it must start with
out
), e.g.$CLI bundle validate > out.txt 2> out.error.txt
.