bucket.grantRead/Arn.format might be able to emit more efficient statements

[eladb]

Here's an example of an output emitted by bucket.grantRead:

{
    "Fn::Join": [
        "",
        [
            "arn",
            ":",
            { "Ref": "AWS::Partition" },
            ":",
            "s3",
            ":",
            "",
            ":",
            "",
            ":",
            { "Ref": "SampleAssetS3BucketE6B2908E" }
        ]
    ]
}                  

[rix0rrr]

Well that is just silly Why is that not an { "Fn::GetAtt": [ "SampleAssetS3BucketE6B2908E", "Arn" ]}?

And also, do we really care?

[RomainMuller]

It makes the template artificially bigger... Which can become annoying for large applications.

[eladb]

It's a kind of aesthetic optimization which we can do if we want, but it's not super important.

[eladb]

Closing for now, I believe we are doing some more optimizations now and the outputs are require nice