Inconsistency in AWS service model definitions

[PriyankaHariharan]

Describe the bug

Amazon's S3 Bucket shape has defined the shape of each of the members as the Name as seen below. For instance, BucketName is the name of json property retrieved from this SDK and not Name.

"members":{
        "Name":{
          "shape":"BucketName",
          "documentation":"<p>The name of the bucket.</p>"
        },
        "CreationDate":{
          "shape":"CreationDate",
          "documentation":"<p>Date the bucket was created. This date can change when making changes to your bucket, such as editing its bucket policy.</p>"
        }
      },

CreationDate is defined as "CreationDate":{"type":"timestamp"}. This extra shape definition is not present in the cloud front file.

In Amazon CloudFront, FunctionMetadata's definition defines the shape as the type and not the Name. timestamp is the data type of the property and not the Name. But FunctionArn is the json property sent by the SDK.

"members":{
        "FunctionARN":{
          "shape":"string",
          "documentation":"<p>The Amazon Resource Name (ARN) of the function. The ARN uniquely identifies the function.</p>"
        },
        "Stage":{
          "shape":"FunctionStage",
          "documentation":"<p>The stage that the function is in, either <code>DEVELOPMENT</code> or <code>LIVE</code>.</p> <p>When a function is in the <code>DEVELOPMENT</code> stage, you can test the function with <code>TestFunction</code>, and update it with <code>UpdateFunction</code>.</p> <p>When a function is in the <code>LIVE</code> stage, you can attach the function to a distribution's cache behavior, using the function's ARN.</p>"
        },
        "CreatedTime":{
          "shape":"timestamp",
          "documentation":"<p>The date and time when the function was created.</p>"
        },
        "LastModifiedTime":{
          "shape":"timestamp",
          "documentation":"<p>The date and time when the function was most recently updated.</p>"
        }
      },

Is there a reason why we use the data type in the shape for some resources but the name (introducing another shape that is defined with the type ) for others? Could you also share what resources follow which pattern?

Expected Behavior

Modifying Amazon CloudFront's FunctionMetadata to introduce another data shape for each of the members

"members":{
        "FunctionARN":{
          "shape":"FunctionARN",
          "documentation":"<p>The Amazon Resource Name (ARN) of the function. The ARN uniquely identifies the function.</p>"
        },
        "Stage":{
          "shape":"Stage",
          "documentation":"<p>The stage that the function is in, either <code>DEVELOPMENT</code> or <code>LIVE</code>.</p> <p>When a function is in the <code>DEVELOPMENT</code> stage, you can test the function with <code>TestFunction</code>, and update it with <code>UpdateFunction</code>.</p> <p>When a function is in the <code>LIVE</code> stage, you can attach the function to a distribution's cache behavior, using the function's ARN.</p>"
        },
        "CreatedTime":{
          "shape":"CreatedTime",
          "documentation":"<p>The date and time when the function was created.</p>"
        },
        "LastModifiedTime":{
          "shape":"LastModifiedTime",
          "documentation":"<p>The date and time when the function was most recently updated.</p>"
        }
      },

OR

Modifying Amazon S3 bucket to use the data type without another shape definition

"members":{
        "BucketName":{
          "shape":"string",
          "documentation":"<p>The name of the bucket.</p>"
        },
        "CreationDate":{
          "shape":"timestamp",
          "documentation":"<p>Date the bucket was created. This date can change when making changes to your bucket, such as editing its bucket policy.</p>"
        }
      },

Current Behavior

mentioned in the bug description

Reproduction Steps

check source files

S3 -

aws-sdk-net/generator/ServiceModels/s3/s3-2006-03-01.normal.json

Line 1417 in 0d41ae9

"Bucket":{

CloudFront - https://github.com/aws/aws-sdk-net/blob/master/generator/ServiceModels/cloudfront/cloudfront-2020-05-31.normal.json#L5131

Possible Solution

No response

Additional Information/Context

No response

AWS .NET SDK and/or Package version used

AWS SDK - 3.7.602.1

Targeted .NET Platform

.NET

Operating System and version

Linux

[ashishdhingra]

@PriyankaHariharan Good morning. Thanks for posting the question. If you examine the BucketName shape, it defines the actual data type string. A shape can define other attributes, for example if it is an enum or if it is required. Name member for Bucket shape could be one of case of simple type. S3 is one of the oldest services and such one of case might be expected. As such, these models are not controlled by the SDK and are pushed by the service teams. Once SDK received the models via automated process, build is triggered to update the model definitions in the SDK.

Thanks,
Ashish

[PriyankaHariharan]

thank you so much @ashishdhingra for the detailed response!

[normj]

@PriyankaHariharan In the case of S3 using BucketName as essentially an alias for string this is common across AWS models to define basically typedefs for simple types. S3 didn't do that in this case but the common reason this is done is they can place all of the constraints for the string like min/max length or a regex in one place and then refer to that typedef across the model.

As a warning be careful reading too much into the code generation for S3. All other services are generated from service models but due to legacy reasons S3 is only partially generated. Basically just the service client and inteface. All of the sub types and marshallers are handcoded.

[github-actions[bot]]

Hello! Reopening this discussion to make it searchable.