Speed up s3 etag retrieval

[evanwieren]

Ran rake tasks this time, and they all passed.

This modification just uses the information that is passed back from the request to populate the etag entry for an s3 object. It sets the value to whatever is passed, but leaves as nul if it is not present.

The get etag feature will pull from the object itself, or if it is not present, it will then do a head request.

[trevorrowe]

Sorry for not commenting on this pull request earlier. It is a known issue that the SDK does not hold onto etag values from the list objects response. This is problematic when you need the etag for objects you are enumerating.

Your pull request would resolve this issue, but it would be inconsistent with the SDK's strategy for caching data. I was hoping to resolve this issue by making the S3Object extend AWS::Core::Resource. This would make S3 behave like the other service interface and allow you to selectively hold onto volatile attributes.

I'll be back in the office soon and I'll try to take a look at this.

[evanwieren]

Thanks. I know there are a number of items to look at when it comes to this.

Thanks.

[AlexanderEkdahl]

Any updates regarding this feature?

[courtsimas]

+1

[yhirano]

+1

[awood45]

This has been stale for several months. I'm going to be closing this, but there are a couple options as far as bringing this change up to date:

• You can update this pull request for the current state of V1, and our current dependencies. We can review that PR. This may take longer for us to get to, as we are currently focused on new features for V2.
• You can write this feature as a plugin for V2 of the SDK, or as a third-party gem you can later integrate with V2. If you'd like to go this route and have any questions, please feel free to reach out.