Calls to headObject function sometimes leads to decoding errors

[robertgmelo]

Summary

As per previous discussion in Slack channel, it was requested the creation of a ticket to handle an intermittent issue when trying to access an object available on a bucket.

Here is the snippet:

let headObjectRequest = S3.HeadObjectRequest(bucket: bucket, key: key)
let headObject = s3.headObject(headObjectRequest)
...

The following error is triggered:

Error: dataCorrupted(Swift.DecodingError.Context(codingPath: [CodingKeys(stringValue: "x-amz-storage-class", intValue: nil)], debugDescription: "Cannot initialize StorageClass from invalid String value NONE", underlyingError: nil))

How to Reproduce

Pre-conditions:

• Have a file uploaded in a bucket

Steps to reproduce the behavior:

1. Try to retrieve information about the object (s3.headObject)
2. Observe the behavior

Expected behavior Retrieve an information about an object should occur without problems.

Setup (please complete the following information):

• OS: MacOS and Linux
• Version of soto: latest

Additional context In fact, looking at AWS documentation, StorageClass object is properly defined, however it seems sometimes AWS returns the NONE value in the header x-amz-storage-class.

Here is the quote from AWS documentation:

x-amz-storage-class Provides storage class information of the object. Amazon S3 returns this header for all objects except for S3 Standard storage class objects. For more information, see Storage Classes. Valid Values: STANDARD | REDUCED_REDUNDANCY | STANDARD_IA | ONEZONE_IA | INTELLIGENT_TIERING | GLACIER | DEEP_ARCHIVE | OUTPOSTS | GLACIER_IR

[robertgmelo]

Suggestion: add NONE for StorageClass

[adam-fowler]

This should be fixed in v6.0.0