Add resource option PreCheckCallback #1310
Conversation
|
Diff for pulumi-azuread with merge commit 539459f |
|
Diff for pulumi-random with merge commit 539459f |
|
https://go.dev/play/p/-eNaTRqOVhu perhaps consider a way to override it without making changes to this codebase? It seems Go language has the facility to do this via embedded structs. |
|
IN addition, FWIW there's PreConfigureCallback machinery that is allowed to modify the data at Pulumi layer before it drops to TF; not sure if that is sufficient for the use case you have. |
I don't see how this is applicable to this case. Where would this come into play? Which interface would I be hijacking here?
|
|
Ok, my bad you're right about PreConfigureCallback belonging to a different part of the system. Regarding overriding via Go embedding I guess it can be too tricky to be useful if we rely on calling into Main/ having multiple layers. So let's discuss making a change to this codebase to support customizing Check. One sec. |
|
Giving it a second read, I think this can be useful to have, I wanted to reach out for something like this in other work. Perhaps let's call it PreCheckCallback? Add some quick tests to make sure it's covered, similar to PreConfigureCallback. |
iwahbe
changed the title
|
Diff for pulumi-random with merge commit b97edc4 |
|
Diff for pulumi-azuread with merge commit b97edc4 |
pkg/tfbridge/provider_test.go
Outdated
| callback := func(_ context.Context, config, _ resource.PropertyMap) (resource.PropertyMap, error) { | ||
| delete(config, "conflictingProperty2") | ||
| config["arrayPropertyValues"] = resource.NewArrayProperty([]resource.PropertyValue{ | ||
| resource.NewStringProperty("v1"), |
There was a problem hiding this comment.
This is tough to follow, do we have an assert somewhere (where)? that observes that the effect of this callback has been actually performed? resource.NewStringProperty("v1") I looked around for "v1" but cannot find it easily.
For an example of how to make asserts more obviously visually correlated to the property being tested, have a look at TestPreConfigureCallback in this file.
There was a problem hiding this comment.
testCheckFailures returns no errors. Without the callback, it returns 3 errors. We don't have a direct assertion that the callback run, but we observe its results.
|
This is great just needs a little more test TLC. |
iwahbe
force-pushed
the
iwahbe/add-resource-option-XCustomCheck
branch
from
a90f474
to
0d807d9
Compare
|
Diff for pulumi-random with merge commit 011045c |
|
Diff for pulumi-azuread with merge commit 011045c |
iwahbe
force-pushed
the
iwahbe/add-resource-option-XCustomCheck
branch
from
0d807d9
to
e10adff
Compare
|
Diff for pulumi-azuread with merge commit 6266eaa |
|
Diff for pulumi-random with merge commit 6266eaa |
iwahbe
force-pushed
the
iwahbe/add-resource-option-XCustomCheck
branch
from
e10adff
to
272e881
Compare
|
Diff for pulumi-azuread with merge commit ea090d2 |
| return config, nil | ||
| } | ||
|
|
||
| for _, f := range factories { |
There was a problem hiding this comment.
By-reference loop variable capture :)
There was a problem hiding this comment.
The test isn't parallelized. I can change that (272e881), but the by-ref isn't actually a risk here.
There was a problem hiding this comment.
Yeah I wanted to enable a linter that flags these but we can deal with it then.
There was a problem hiding this comment.
Have you seen golang/go#56010?
| @@ -550,6 +521,104 @@ func TestProviderCheckV2(t *testing.T) { | |||
| assert.Equal(t, "", failures[2].Property) | |||
| } | |||
|
|
|||
| func TestProviderCheck(t *testing.T) { | |||
| testFailures := map[string]func(*testing.T, []*pulumirpc.CheckFailure){ | |||
| "v1": testCheckFailuresV1, | |||
There was a problem hiding this comment.
This is great. I think V1 is not in prod anymore so in the future we can avoid any work pertaining to testing it, and start deprecating/removing.
iwahbe
force-pushed
the
iwahbe/add-resource-option-XCustomCheck
branch
from
272e881
to
b26fd13
Compare
|
Diff for pulumi-azuread with merge commit 3b96d9e |
* Add GRPC test to replicate 2633 * Test that we set "tags" on creation TF does, and if we assume that "tags" is set correctly, we then get the expected diff. * Ensure valid diffs by overriding check * Remove `.XComputeInput` This fix takes effect during Check, which occurs before Diff. We don't need to do anything special for Diff now. * Depend on pulumi/pulumi-terraform-bridge#1310 * Turn tests back on * Consume upstream changes The previous fix works on all resources... except our s3legacy bucket because it was not updated to use upstream's new tags strategy. I have changed our fork to use the new strategy. The change is here: https://github.com/pulumi/terraform-provider-aws/compare/patched-v5.9.0...patched-v5.9.0-with-modern-s3legacy-tags?expand=1 * Explain .Mappable() has the same shape as NewTagConfig expected * Template out YAML Tests * Accommodate aws:cognito:UserPool * Switch to environment for pf * Add a test for aws:s3:BucketV2 We are seeing tests fail for aws:cognito:UserPool but I know this works for aws:s3:BucketV2. These are both SDKv2 resources. Since they work differently I have ensured that both are tested. * Skip aws:cognito:UserPool test This resource is broken upstream, so we skip the test here. * Remove outdated tests I have left skeleton code for GRPC based testing, as I often find it useful when debugging locally. I have noted that it does not run in CI, and is just there for local development. * Don't pass empty tags properties
Add the ability to override the check behavior for resources.