switch integration tests to new test-utils #3898
Conversation
|
A new generated diff is ready to view.
A new doc preview is ready to view. |
|
A new generated diff is ready to view.
A new doc preview is ready to view. |
| .build_uri() | ||
| .to_string() | ||
| .parse() |
There was a problem hiding this comment.
This looks weird. is this a 0.4 to 1.0 conversion? Why do we create a URI, stringify it, and then parse it as a URI?
There was a problem hiding this comment.
This is a 0.4 to 1.0 conversion. This is the exact same code we had for going from 1.0 to 0.4 which you can see removed/relocated on L209
| @@ -1,6 +1,6 @@ | |||
| [package] | |||
| name = "aws-smithy-checksums" | |||
| version = "0.60.13" | |||
| version = "0.70.0" | |||
There was a problem hiding this comment.
Not that it really matters, but I would expect the next version to be v0.61.0
There was a problem hiding this comment.
I hit this with the flexibile checksums work, 0.61.0 was already published and yanked: https://crates.io/crates/aws-smithy-checksums/versions
But I would probably expect 0.62.0
| @@ -12,7 +12,7 @@ publish = false | |||
| repository = "/~https://github.com/smithy-lang/smithy-rs" | |||
|
|
|||
| [features] | |||
| http_1x = ["dep:http-1x", "dep:http-body-1x", "aws-smithy-runtime-api/http-1x"] | |||
| http_1x = ["dep:http-body-1x", "aws-smithy-runtime-api/http-1x"] | |||
There was a problem hiding this comment.
FYI, the feature name was updated to http-1x in the main branch, as the updated nightly compiler told us that those who used this feature were specifying http-1x (and we tend to use a hyphen in feature names in our crates).
|
A new generated diff is ready to view.
A new doc preview is ready to view. |
|
A new generated diff is ready to view.
A new doc preview is ready to view. |
|
A new generated diff is ready to view.
A new doc preview is ready to view. |
|
A new generated diff is ready to view.
A new doc preview is ready to view. |
|
A new generated diff is ready to view.
A new doc preview is ready to view. |
Motivation and Context
Follow up PR to #3888
Description
aws/sdk/integration-teststo use thetest-utilfeature(s) directly fromaws-smithy-http-clientrather thanaws-smithy-runtimeRefactoraws-smithy-httpto use `http-1.ximdscredentials provider in passing while tracking down a test failureAWS_EC2_METADATA_ENABLED=falseis set by default and was causing some test failures)I started pulling at some threads and found them to be quite load bearing (e.g.
inlineable/serialization_settings.rs) and would require updating all of codegen to usehttp-1.x. ~This PR is a bit of a half step with some stop gaps put into place (e.g. the newcompatfeature ofaws-smithy-httpor inlining a function) to prevent the need to update everything in a single PR. Future PR's will continue where I left off here. ~One lesson I learned the hard way was that our integration test
Cargo.tomlhave to reflect codegen. I had updated them to all usehttp = "1"and got it all working only to realize codegen generateshttp = "0.2.9"andhttp-1x = { package = "http", version = "1"}and thus the integration tests have to look the same as that.Future
aws-smithy-httpto upgrade tohttp-body-1.x. ~http-1.xUPDATE: I reverted the update to http 1.x for
aws-smithy-http. This needs to be done and planned with the server and I don't want it to unnecessarily block hyper 1. We will pursue the remaining hyper 1 changes and then make a decision on internal use of http and http-body pre 1.x usage.By submitting this pull request, I confirm that you can use, modify, copy, and redistribute this contribution, under the terms of your choice.