GH-39119: [C++] Refactor the Azure FS tests and filesystem class instantiation#39207
GH-39119: [C++] Refactor the Azure FS tests and filesystem class instantiation#39207felipecrv merged 26 commits intoapache:mainfrom
Conversation
|
|
1d96443 to
27eae54
Compare
There was a problem hiding this comment.
No more awkward status_ being set on ctor now that Make is used to construct the environments.
There was a problem hiding this comment.
@kou I'm explicitly checking that the Status::IsCancelled() to make sure other failures don't make the tests be skipped quietly.
kou
left a comment
There was a problem hiding this comment.
+1
I couldn't review all changes carefully because there are too many changes. But I trust you. :-)
cpp/src/arrow/filesystem/azurefs.h
Outdated
There was a problem hiding this comment.
Do we plan to autodetect the backend at some point? We do that in S3 (the DetectS3Backend function).
There was a problem hiding this comment.
Now that I have the environments, I think I can remove this option completely. 🤔
There was a problem hiding this comment.
Now that I have the environments, I think I can remove this option completely. 🤔
Do you plan to do that here or in a separate PR?
There was a problem hiding this comment.
I tried to do on this one, but this is used to configure the URLs in AzureOptions, so I will do it in another PR where I'm adding more options to AzureOptions and documenting it better.
|
The CI failures look unexpected, can you perhaps rebase on main? |
…ad of ConcatAbstractPath where applicable
4023774 to
4f45c67
Compare
pitrou
left a comment
There was a problem hiding this comment.
+1. Let's see if tests are ok now.
|
After merging your PR, Conbench analyzed the 6 benchmarking runs that have been run so far on merge-commit 4e58f7c. There was 1 benchmark result indicating a performance regression:
The full Conbench report has more details. It also includes information about 2 possible false positives for unstable benchmarks that are known to sometimes produce them. |
…s instantiation (apache#39207) ### Rationale for this change This PR contains some unrelated improvements (like better docs) and some nitpicky fixes. The test refactoring makes it easier to see on which environments tests run and make them able to be instantiated with different options in the future once we extend the `AzureOptions`. ### What changes are included in this PR? - Random cleanups - Short namespace aliases - Refactoring of the tests (multiple environments, TYPED_TEST_SUITE, explicit preexisting data setup) - Cleanup of the `AzureOptions` class ### Are these changes tested? Yes. I created Azure Storage accounts to test with and without Hierarchical Namespace support. I also ran the tests in a shell without my environment variables to ensure the test-skipping behavior is correct. ### Are there any user-facing changes? Changes to `AzureOptions`, but since `AzureFileSystem` is not really used yet, these breaking changes won't be a problem. * Closes: apache#39119 Authored-by: Felipe Oliveira Carvalho <[email protected]> Signed-off-by: Felipe Oliveira Carvalho <[email protected]>
Rationale for this change
This PR contains some unrelated improvements (like better docs) and some nitpicky fixes. The test refactoring makes it easier to see on which environments tests run and make them able to be instantiated with different options in the future once we extend the
AzureOptions.What changes are included in this PR?
AzureOptionsclassAre these changes tested?
Yes. I created Azure Storage accounts to test with and without Hierarchical Namespace support. I also ran the tests in a shell without my environment variables to ensure the test-skipping behavior is correct.
Are there any user-facing changes?
Changes to
AzureOptions, but sinceAzureFileSystemis not really used yet, these breaking changes won't be a problem.