This project welcomes contributions and suggestions. Most contributions require you to agree to a Contributor License Agreement (CLA) declaring that you have the right to, and actually do, grant us the rights to use your contribution. For details, visit cla.microsoft.com.
This project has adopted the Microsoft Open Source Code of Conduct. For more information see the Code of Conduct FAQ or contact [email protected] with any additional questions or comments.
These libraries follow the Azure SDK Design Guidelines for .NET and share a number of core features such as HTTP retries, logging, transport protocols, authentication protocols, etc., so that once you learn how to use these features in one client library, you will know how to use them in other client libraries. You can learn about these shared features in the Azure.Core README.
Please do not edit any of the code in the /Generated
folders directly. If
you need to update a swagger file or change the generator, you can regenerate
by running the \sdk\storage\generate.ps1
script:
PS C:\src\azure-sdk-for-net\sdk\storage> .\generate.ps1
>>> VERBOSE: Emitting file BlobRestClient.cs
>>> VERBOSE: Emitting file FileRestClient.cs
>>> VERBOSE: Emitting file QueueRestClient.cs
The generator uses AutoRest which requires node.js. It is recommended to use the beta version as it increases the max usable memory:
npm install -g autorest@beta
There's known flakiness that results in an ERROR: Did not find any input files to generate!
message that can be safely ignored -- just run the
generator one more time.
Please ensure all tests pass with any changes and additional tests are added to exercise any new features that you've added.
We use NUnit 3 as our testing framework.
Azure.Core.TestFramework's testing framework provides a set of reusable primitives that simplify writing tests for new Azure SDK libraries.
We also have common test code in our
projects' /Shared
folders that provides helpful Storage specific
infrastructure.
Our live tests require access to a variety of Storage accounts making use of
different service features. We specify all of them via a
TestConfigurations.xml
file. It includes:
TargetTestTenant
- the default Storage account for our testsTargetSecondaryTestTenant
- A Storage account with Read Access Geo-Redundant Storage enabledTargetPremiumBlobTenant
- A Storage account using Premium SSDsTargetPreviewBlobTenant
- A Storage account with access to preview featuresTargetOAuthTenant
- A Storage account configured to authenticate with Azure Active DirectoryTargetHierarchicalNamespaceTenant
- A Storage account with hierarchical namespace enabled and configured to authenticate with Azure Active Directory (OAuth access)
If you want to run live tests against your own account, you can edit the
TestConfigurationsTemplate.xml
file and rename it to TestConfigurations.xml
in the same folder. The build
will automatically copy it to each test project. If you're working with
multiple enlistments or want to change between multiple configuration files,
you can set the AZ_STORAGE_CONFIG_PATH
environment variable to a
configuration file path that will take precedence over your local
TestConfigurations.xml
file.
We expose all of our APIs with both sync and async variants. To avoid writing
each of our tests twice, we automatically rewrite our async API calls into
their sync equivalents. Simply write your tests using only async APIs and
call InstrumentClient
on any of our client objects. The test framework will
wrap the client with a proxy that forwards everything to the sync overloads.
Please note that a number of our helpers will automatically instrument clients
they provide you. Visual Studio's test runner will show *TestClass(True)
for
the async variants and *TestClass(False)
for the sync variants.
Our testing framework supports recording service requests made during a unit
test so they can be replayed later. You can set the AZURE_TEST_MODE
environment variable to Playback
to run previously recorded tests, Record
to record or re-record tests, and Live
to run tests against the live service.
Properly supporting recorded tests does require a few extra considerations.
All random values should be obtained via this.Recording.Random
since we use
the same seed on test playback to ensure our client code generates the same
"random" values each time. You can't share any state between tests or rely on
ordering because you don't know the order they'll be recorded or replayed. Any
sensitive values are redacted via the the configuration in the constructor of StorageTestBase
.
The easiest way to run the tests is via Visual Studio's unit test runner.
You can also run tests via the command line using dotnet test
, but that will
run tests for all supported platforms simultaneously and intermingle their
output. You can run the tests for just one platform with dotnet test -f net6.0
or dotnet test -f net462
.
The recorded tests are run automatically on every pull request. Live tests are
run nightly. Contributors with write access can ask Azure DevOps to run the
live tests against a pull request by commenting /azp run net - storage - tests
in the PR.
Our samples are structured as unit tests so we can easily verify they're up to date and working correctly. These tests aren't recorded and make minimal use of test infrastructure to keep them easy to read.