# Description
This PR implements generating the generic variant of a go SDK from
Pulumi schemas. Currently the idea is to generate a directory `x` inside
the root directory of the go SDK which will contain the same SDK except
using generics and generating far less code than its current counter
part.
Also implements an enum option `$.language.go.generics` which can be set
to the following:
- `none` is the default which maintains the current behavior that
generates legacy types without generics
- `side-by-side` generates the generics sdk variant alongside the
current sdk under directory `x`
- `generics-only` generates only the new sdk with generics at the root
of the package
Still a bunch of things to do:
- [x] Generating `InvokeResult]Output` type from `Output[InvokeResult]`
and generating accessor methods for it
- [x] Generating default values for types and using the `pulumix`
subpackage to do so
- [x] Generating generic SDK variants for all test schemas we have and
making sure they compile (currently only testing
`simple-resource-schema` as shown below)
- [x] Account for plain inputs for components
- [x] Combine pulumix.Join with pulumix.Apply to generate resource
accessor methods
- [x] Problem with `GPtrOutput[T]` and `ArrayOutput[T]` being unwrapped
to `Output[*T]` and `Output[[]T]`
- [x] Remove excess untyped container types from generated enums
- [x] Fix default values for resource methods with lifted single return
value
- [x] Secret properties
Currently the following test schemas have opted for `generics:
"side-by-side"`:
- [x] output-funcs
- [x] simple-enum-schema
- [x] secrets
- [x] simple-plain-schema
- [x] plain-and-default
## Checklist
- [ ] 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. -->
13136: Add explicit package versioning to Go codegen r=guineveresaenger a=guineveresaenger
<!---
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
This PR adds logic to explicitly set a Go SDK version as well as pluginDownloadURL in the SDK itself.
This will allow us to publish tags with explicit versions, ensure the engine references the correct plugin version, and will allow for the provider to use the correct version plugin much more reliably.
This PR adds a new `pulumiVersion` file and refactors the `pulumiUtilities` file to live alongside the new version file in an `internal` folder for utilities. This allows all resources in the provider to call on `internal.PkgVersion` without generating a mod-level utilities file alongside.
I have tested this against a flat structure SDK (pulumi-docker) and a modular SDK (pulumi-okta).
I have also edited the `schema.json` files in the unit tests to reflect the new utilities path.
The changes in the PR will allow us to:
1. Read a version into `pulumiVersion.go` during codegen
2. Commit and tag that commit as `sdk/v<Version>`
3. Have the exact same SDK that was built and tested be available on github
4. Clean up the CI step that explicitly tags the Go SDK.
Fixes#13012
## Checklist
- [ ] I have run `make tidy` to update any new dependencies
- [x] I have run `make lint` to verify my code passes the lint check
- [x] 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: Guinevere Saenger <guinevere@pulumi.com>
Fixes https://github.com/pulumi/pulumi/issues/12738https://github.com/pulumi/pulumi/pull/11834 turned on the prealloc
linter and changed a load of slice uses from just `var x T[]` to `x :=
make([]T, 0, preallocSize)`. This was good for performance but it turns
out there are a number of places in the codebase that treat a `nil`
slice as semnatically different to an empty slice.
Trying to test that, or even reason that through for every callsite is
untractable, so this PR replaces all expressions of the form `make([]T,
0, size)` with a call to `slice.Prealloc[T](size)`. When size is 0 that
returns a nil array, rather than an empty array.
Enable the prealloc linter, which identifies slices
with a known capacity, but are not preallocated, which
results in unnecessary allocations and memcpys.
* Add test case
* Fix tests
* Add test dependencies correctly
* Feed through error handling
* Include test output
* Get types to line up
* Add remaining test files
* Update changelog
* Correctly find type paths
* Handle transitive objects
* Handle required fields
* Add feature flag for go
* Add required+default test case
* Don't `<any>` cast known types.
* Add more flags.
I realize this should really wait for PR#8400 to merge.
* Add plain object to env-helper test
This test fails right now. My next problem is fixing it.
* Handle plain types
* Handle function inputs
* Fix the indentation
* Handle output types correctly
* Remove unnecessary `!`
* Add test case
* Fix tests
* Add test dependencies correctly
* Feed through error handling
* Include test output
* Get types to line up
* Add remaining test files
* Update changelog
* Correctly find type paths
* Handle transitive objects
* Handle required fields
* Add required+default test case
* Don't `<any>` cast known types.
* Add plain object to env-helper test
This test fails right now. My next problem is fixing it.
* Handle plain types
* Handle function inputs
* Fix the indentation
* Handle output types correctly
* Remove unnecessary `!`
* Start on `genPlainObjectDefaultFunc`
* Add missing change to fix test
* Run tests with merge
* Refactor out assign
* Merge in next _index.md diff
* Change method name to `Defaults`
* Handle enums correctly
* Another attempt at _index.md
* Make module generation deterministic
* Add checks for old values
* Insert defaults in resources
* Fix docs generation
Credit to @praneetloke
* Progress on adding defaults to Resource arguments
* Handle resource argument defaults
* Don't create defaults if disableObjectDefaults
* Rename test folder
* Add test for disable flag
* Fix disable test
* Update docs
* Abstract out nil comparisons
* Use reflection to test for empty values
* Simplify Ptr and pulumi.Any type handling
* Remove unused function
* Apply defaults to functions
* Update new test with master codegen
* Tests + nil check