Enforce constraints for unnamed enums #3884

drganjoo · 2024-10-17T23:13:48Z

Enforces Constraints for Unnamed Enums

This PR addresses the issue where, on the server side, unnamed enums were incorrectly treated as infallible during deserialization, allowing any string value to be converted without validation. The solution introduces a ConstraintViolation and TryFrom implementation for unnamed enums, ensuring that deserialized values conform to the enum variants defined in the Smithy model.

The following is an example of an unnamed enum:

@enum([
    { value: "MONDAY" },
    { value: "TUESDAY" }
])
string UnnamedDayOfWeek

On the server side the following type is generated for the Smithy shape:

pub struct UnnamedDayOfWeek(String);

impl ::std::convert::TryFrom<::std::string::String> for UnnamedDayOfWeek {
    type Error = crate::model::unnamed_day_of_week::ConstraintViolation;

    fn try_from(
        s: ::std::string::String,
    ) -> ::std::result::Result<Self, <Self as ::std::convert::TryFrom<::std::string::String>>::Error>
    {
        match s.as_str() {
            "MONDAY" | "TUESDAY" => Ok(Self(s)),
            _ => Err(crate::model::unnamed_day_of_week::ConstraintViolation(s)),
        }
    }
}

This change prevents invalid values from being deserialized into unnamed enums and raises appropriate constraint violations when necessary.

There is one difference between the Rust code generated for TryFrom<String> for named enums versus unnamed enums. The implementation for unnamed enums passes the ownership of the String parameter to the generated structure, and the implementation for TryFrom<&str> delegates to TryFrom<String>.

impl ::std::convert::TryFrom<::std::string::String> for UnnamedDayOfWeek {
    type Error = crate::model::unnamed_day_of_week::ConstraintViolation;
    fn try_from(
        s: ::std::string::String,
    ) -> ::std::result::Result<Self, <Self as ::std::convert::TryFrom<::std::string::String>>::Error>
    {
        match s.as_str() {
            "MONDAY" | "TUESDAY" => Ok(Self(s)),
            _ => Err(crate::model::unnamed_day_of_week::ConstraintViolation(s)),
        }
    }
}

impl ::std::convert::TryFrom<&str> for UnnamedDayOfWeek {
    type Error = crate::model::unnamed_day_of_week::ConstraintViolation;
    fn try_from(
        s: &str,
    ) -> ::std::result::Result<Self, <Self as ::std::convert::TryFrom<&str>>::Error> {
        s.to_owned().try_into()
    }
}

On the client side, the behaviour is unchanged, and the client does not validate for backward compatibility reasons. An existing test has been modified to ensure this.

#[test]
fn generate_unnamed_enums() {
    let result = "t2.nano"
        .parse::<crate::types::UnnamedEnum>()
        .expect("static value validated to member");
    assert_eq!(result, UnnamedEnum("t2.nano".to_owned()));
    let result = "not-a-valid-variant"
        .parse::<crate::types::UnnamedEnum>()
        .expect("static value validated to member");
    assert_eq!(result, UnnamedEnum("not-a-valid-variant".to_owned()));
}

Fixes issue #3880

github-actions · 2024-10-17T23:31:33Z

A new generated diff is ready to view.

No codegen difference in the AWS SDK
No codegen difference in the Client Test
No codegen difference in the Server Test
No codegen difference in the Server Test Python
No codegen difference in the Server Test Typescript

A new doc preview is ready to view.

…tr` methods

github-actions · 2024-10-18T00:01:38Z

A new generated diff is ready to view.

AWS SDK (ignoring whitespace)
No codegen difference in the Client Test
No codegen difference in the Server Test
No codegen difference in the Server Test Python
No codegen difference in the Server Test Typescript

A new doc preview is ready to view.

github-actions · 2024-10-18T01:05:21Z

A new generated diff is ready to view.

AWS SDK (ignoring whitespace)
No codegen difference in the Client Test
No codegen difference in the Server Test
No codegen difference in the Server Test Python
No codegen difference in the Server Test Typescript

A new doc preview is ready to view.

github-actions · 2024-10-23T15:00:11Z

A new generated diff is ready to view.

No codegen difference in the AWS SDK
No codegen difference in the Client Test
No codegen difference in the Server Test
No codegen difference in the Server Test Python
No codegen difference in the Server Test Typescript

A new doc preview is ready to view.

github-actions · 2024-10-29T09:07:09Z

A new generated diff is ready to view.

No codegen difference in the AWS SDK
No codegen difference in the Client Test
No codegen difference in the Server Test
No codegen difference in the Server Test Python
No codegen difference in the Server Test Typescript

A new doc preview is ready to view.

...n/kotlin/software/amazon/smithy/rust/codegen/client/smithy/generators/ClientEnumGenerator.kt

david-perez · 2024-10-29T11:57:27Z

...e/src/test/kotlin/software/amazon/smithy/rust/codegen/core/smithy/generators/TestEnumType.kt

@@ -10,6 +10,7 @@ import software.amazon.smithy.rust.codegen.core.rustlang.rust
 import software.amazon.smithy.rust.codegen.core.rustlang.rustTemplate
 import software.amazon.smithy.rust.codegen.core.rustlang.writable
 import software.amazon.smithy.rust.codegen.core.smithy.RuntimeType
+import software.amazon.smithy.rust.codegen.core.smithy.RuntimeType.Companion.preludeScope
 import software.amazon.smithy.rust.codegen.core.util.dq

 object TestEnumType : EnumType() {


Huh, I wonder why we have this class at all i.e. why we don't test directly against InfallibleEnumType. It feels wrong to copy over the implementations from the "real" classes to this class.

I agree with you on this. I'll see if it can be easily fixed in this PR, otherwise will raise another one for this.

InfallibleEnumType allows UnknownVariants, whereas TestEnumType panics when an unknown variant is encountered. I tried composing TestEnumType to use InfallibleEnumType internally, but InfallibleEnumType is defined in codegen-client, while TestEnumType is in codegen-core.

david-perez · 2024-10-29T14:20:51Z

...n/kotlin/software/amazon/smithy/rust/codegen/server/smithy/generators/ServerEnumGenerator.kt

+                impl #{TryFrom}<&str> for ${context.enumName} {
+                    type Error = #{ConstraintViolation};
+                    fn try_from(s: &str) -> #{Result}<Self, <Self as #{TryFrom}<&str>>::Error> {
+                        s.to_owned().try_into()


This is first converting to a heap-allocated String and only then matching on the enum values. So invalid enum values would unnecessarily be heap-allocated. We should make TryFrom<String> and TryFrom<&str> instead delegate to FromStr, which should only heap-allocate when the enum value is valid.

Unlike named enums, which can use &str for comparison and directly return an enum variant, unnamed enums need to store an owned, heap-allocated String. The TryFrom<String> implementation already receives an owned String as a parameter, so delegating to FromStr would result in an unnecessary heap allocation. Additionally, ConstraintViolation takes ownership of a heap-allocated String. Therefore, both code paths (valid and invalid enum values) require a heap-allocated String. Calling to_owned within TryFrom<&str> only shifts the allocation slightly earlier without adding any additional allocation.

...n/kotlin/software/amazon/smithy/rust/codegen/client/smithy/generators/ClientEnumGenerator.kt

github-actions · 2024-10-30T07:17:25Z

A new generated diff is ready to view.

No codegen difference in the AWS SDK
No codegen difference in the Client Test
No codegen difference in the Server Test
No codegen difference in the Server Test Python
No codegen difference in the Server Test Typescript

A new doc preview is ready to view.

github-actions · 2024-10-30T09:27:21Z

A new generated diff is ready to view.

No codegen difference in the AWS SDK
No codegen difference in the Client Test
No codegen difference in the Server Test
No codegen difference in the Server Test Python
No codegen difference in the Server Test Typescript

A new doc preview is ready to view.

github-actions · 2024-10-30T12:09:00Z

A new generated diff is ready to view.

No codegen difference in the AWS SDK
No codegen difference in the Client Test
No codegen difference in the Server Test
No codegen difference in the Server Test Python
No codegen difference in the Server Test Typescript

A new doc preview is ready to view.

github-actions · 2024-11-11T12:29:25Z

A new generated diff is ready to view.

No codegen difference in the AWS SDK
No codegen difference in the Client Test
No codegen difference in the Server Test
No codegen difference in the Server Test Python
No codegen difference in the Server Test Typescript

A new doc preview is ready to view.

.changelog/4329788.md

github-actions · 2024-11-12T10:49:22Z

A new generated diff is ready to view.

No codegen difference in the AWS SDK
No codegen difference in the Client Test
No codegen difference in the Server Test
No codegen difference in the Server Test Python
No codegen difference in the Server Test Typescript

A new doc preview is ready to view.

github-actions · 2024-11-12T13:56:11Z

A new generated diff is ready to view.

No codegen difference in the AWS SDK
No codegen difference in the Client Test
No codegen difference in the Server Test
No codegen difference in the Server Test Python
No codegen difference in the Server Test Typescript

A new doc preview is ready to view.

Allow unnamed enum implementation to override FromStr and `FromForS…

651538a

…tr` methods

drganjoo force-pushed the fahadzub/unnamed-enum-constrained branch from f2aa580 to 651538a Compare October 17, 2024 23:43

New test added that has commented out recursive structs

6deae2f

Merge branch 'main' into fahadzub/unnamed-enum-constrained

bfdf606

drganjoo force-pushed the fahadzub/unnamed-enum-constrained branch from 17ca030 to bfdf606 Compare October 29, 2024 08:29

Merge branch 'main' into fahadzub/unnamed-enum-constrained

e410ad8

drganjoo marked this pull request as ready for review October 29, 2024 08:42

drganjoo requested review from a team as code owners October 29, 2024 08:42

david-perez requested changes Oct 29, 2024

View reviewed changes

Client SDK should not validate unknown variant

029a6e3

Changelog added

042fe5c

drganjoo force-pushed the fahadzub/unnamed-enum-constrained branch from fbcefa0 to 042fe5c Compare October 30, 2024 11:51

drganjoo requested a review from david-perez October 30, 2024 14:29

drganjoo added server Rust server SDK needs-sdk-review labels Oct 31, 2024

Merge branch 'main' into fahadzub/unnamed-enum-constrained

8226826

david-perez approved these changes Nov 11, 2024

View reviewed changes

.changelog/4329788.md Show resolved Hide resolved

Update changelog

2ceceb7

Fix merge issue

879f350

Velfi approved these changes Nov 13, 2024

View reviewed changes

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Enforce constraints for unnamed enums #3884

Enforce constraints for unnamed enums #3884

drganjoo commented Oct 17, 2024 •

edited

Loading

github-actions bot commented Oct 17, 2024

github-actions bot commented Oct 18, 2024

github-actions bot commented Oct 18, 2024

github-actions bot commented Oct 23, 2024

github-actions bot commented Oct 29, 2024

david-perez Oct 29, 2024

drganjoo Oct 29, 2024

drganjoo Oct 31, 2024

david-perez Oct 29, 2024

drganjoo Oct 29, 2024 •

edited

Loading

github-actions bot commented Oct 30, 2024

github-actions bot commented Oct 30, 2024

github-actions bot commented Oct 30, 2024

github-actions bot commented Nov 11, 2024

github-actions bot commented Nov 12, 2024

github-actions bot commented Nov 12, 2024

Enforce constraints for unnamed enums #3884

Are you sure you want to change the base?

Enforce constraints for unnamed enums #3884

Conversation

drganjoo commented Oct 17, 2024 • edited Loading

Enforces Constraints for Unnamed Enums

github-actions bot commented Oct 17, 2024

github-actions bot commented Oct 18, 2024

github-actions bot commented Oct 18, 2024

github-actions bot commented Oct 23, 2024

github-actions bot commented Oct 29, 2024

david-perez Oct 29, 2024

Choose a reason for hiding this comment

drganjoo Oct 29, 2024

Choose a reason for hiding this comment

drganjoo Oct 31, 2024

Choose a reason for hiding this comment

david-perez Oct 29, 2024

Choose a reason for hiding this comment

drganjoo Oct 29, 2024 • edited Loading

Choose a reason for hiding this comment

github-actions bot commented Oct 30, 2024

github-actions bot commented Oct 30, 2024

github-actions bot commented Oct 30, 2024

github-actions bot commented Nov 11, 2024

github-actions bot commented Nov 12, 2024

github-actions bot commented Nov 12, 2024

drganjoo commented Oct 17, 2024 •

edited

Loading

drganjoo Oct 29, 2024 •

edited

Loading