c27d768bb1
<!--- Thanks so much for your contribution! If this is your first time contributing, please ensure that you have read the [CONTRIBUTING](https://github.com/pulumi/pulumi/blob/master/CONTRIBUTING.md) documentation. --> # Description <!--- Please include a summary of the change and which issue is fixed. Please also include relevant motivation and context. --> Adds the first pass of matrix testing. Matrix testing allows us to define tests once in pulumi/pulumi via PCL and then run those tests against each language plugin to verify code generation and runtime correctness. Rather than packing matrix tests and all the associated data and machinery into the CLI itself we define a new Go package at cmd/pulumi-test-lanaguage. This depends on pkg and runs the deployment engine in a unique way for matrix tests but it is running the proper deployment engine with a proper backend (always filestate, using $TEMP). Currently only NodeJS is hooked up to run these tests, and all the code for that currently lives in sdk/nodejs/cmd/pulumi-language-nodejs/language_test.go. I expect we'll move that helper code to sdk/go/common and use it in each language plugin to run the tests in the same way. This first pass includes 3 simple tests: * l1-empty that runs an empty PCL file and checks just a stack is created * l1-output-bool that runs a PCL program that returns two stack outputs of `true` and `false * l2-resource-simple that runs a PCL program creating a simple resource with a single bool property These tests are themselves tested with a mock language runtime. This verifies the behavior of the matrix test framework for both correct and incorrect language hosts (that is some the mock language runtimes purposefully cause errors or compute the wrong result). There are a number of things missing from from the core framework still, but I feel don't block getting this first pass merged and starting to be used. 1. The tests can not currently run in parallel. That is calling RunLanguageTest in parallel will break things. This is due to two separate problems. Firstly is that the SDK snapshot's are not safe to write in parallel (when PULUMI_ACCEPT is true), this should be fairly easy to fix by doing a write to dst-{random} and them atomic move to dst. Secondly is that the deployment engine itself has mutable global state, short term we should probably just lock around that part RunLanguageTest, long term it would be good to clean that up. 2. We need a way to verify "preview" behavior, I think this is probably just a variation of the tests that would call `stack.Preview` and not pass a snapshot to `assert`. 3. stdout, stderr and log messages are returned in bulk at the end of the test. Plus there are a couple of calls to the language runtime that don't correctly thread stdout/stderr to use and so default to the process `os.Stdout/Stderr`. stdout/stderr streaming shows up in a load of other places as well so I'm thinking of a clean way to handle all of them together. Log message streaming we can probably do by just turning RunLanguageTest to a streaming grpc call. ## Checklist - [x] I have run `make tidy` to update any new dependencies - [x] I have run `make lint` to verify my code passes the lint check - [ ] I have formatted my code using `gofumpt` <!--- Please provide details if the checkbox below is to be left unchecked. --> - [x] I have added tests that prove my fix is effective or that my feature works <!--- User-facing changes require a CHANGELOG entry. --> - [x] I have run `make changelog` and committed the `changelog/pending/<file>` documenting my change <!-- If the change(s) in this PR is a modification of an existing call to the Pulumi Cloud, then the service should honor older versions of the CLI where this change would not exist. You must then bump the API version in /pkg/backend/httpstate/client/api.go, as well as add it to the service. --> - [ ] Yes, there are changes in this PR that warrants bumping the Pulumi Cloud API version <!-- @Pulumi employees: If yes, you must submit corresponding changes in the service repo. --> --------- Co-authored-by: Abhinav Gupta <abhinav@pulumi.com> |
||
|---|---|---|
| .. | ||
| cmd | ||
| dist | ||
| lib | ||
| scripts | ||
| stubs | ||
| .gitignore | ||
| .pylintrc | ||
| Makefile | ||
| README.md | ||
| mypy.ini | ||
| python.go | ||
| python_test.go | ||
| requirements.txt | ||
| shim_unix.go | ||
| shim_windows.go | ||
README.md
Pulumi Python SDK
The Pulumi Python SDK (pulumi) is the core package used when writing Pulumi programs in Python. It contains everything that you’ll need in order to interact with Pulumi resource providers and express infrastructure using Python code. Pulumi resource providers all depend on this library and express their resources in terms of the types defined in this module.
The Pulumi Python SDK requires Python version 3.7 or greater through official python installer
note: pip is required to install dependencies. If you installed Python from source, with an installer from python.org, or via Homebrew you should already have pip. If Python is installed using your OS package manager, you may have to install pip separately, see Installing pip/setuptools/wheel with Linux Package Managers. For example, on Debian/Ubuntu you must run sudo apt install python3-venv python3-pip.
Getting Started
The fastest way to get up and running is to choose from one of the following Getting Started guides: -aws -microsoft azure -google cloud -kubernetes
Pulumi Programming Model
The Pulumi programming model defines the core concepts you will use when creating infrastructure as code programs using Pulumi. Architecture & Concepts describes these concepts with examples available in Python. These concepts are made available to you in the Pulumi SDK.
The Pulumi SDK is available to Python developers as a Pip package distributed on PyPI . To learn more, refer to the Pulumi SDK Reference Guide.
The Pulumi programming model includes a core concept of Input and Output values, which are used to track how outputs of one resource flow in as inputs to another resource. This concept is important to understand when getting started with Python and Pulumi, and the [Inputs and Outputs] (https://www.pulumi.com/docs/intro/concepts/inputs-outputs/)documentation is recommended to get a feel for how to work with this core part of Pulumi in common cases.
The Pulumi Python Resource Model
Like most languages usable with Pulumi, Pulumi represents cloud resources as classes and Python programs can instantiate those classes. All classes that can be instantiated to produce actual resources derive from the pulumi.Resource class.