aws-s3-deploy fails - Feels like not deleting all S3 files before upload.

[Jason Redwood]

This is an intermittant issue, but it does look like it occurs more often when redeploying (uploading from bitbucket pipeline to S3) an older build over the top of a newer build.

It looks like the files are not fully or properly being deleted in the S3 bucket before the upload occurs and I often find the files in the s3 bucket are not accurate with the build

The aws-s3-deploy:1.1.0 config I am using is

```

- pipe: atlassian/aws-s3-deploy:1.1.0

    variables:

        AWS_ACCESS_KEY_ID: $AWS_ACCESS_KEY_ID

        AWS_SECRET_ACCESS_KEY: $AWS_SECRET_ACCESS_KEY

        AWS_DEFAULT_REGION: $AWS_DEFAULT_REGION

        S3_BUCKET: $AWS_PROD_S3_BUCKET

        LOCAL_PATH: "output"

        DELETE_FLAG: "true"

```

As mention, the issues dont seem to occur when incrementally deploying later builds, or at least much much less often, the issue seems to occur when deploying a later build, then redeploying an earlier build later on.

When the issue occurs, to resolve it, I login to AWS and manually delete all the files, then do the deployment and everything works as expected.

Is anyone else experiencing this issue or does someone know what I might be doing wrong?

[Igor Stoyanov]

@Jason Redwood hi . Thanks for your question. This pipe based on aws s3 sync where DELETE_FLAG is option:
`--delete (boolean) Files that exist in the destination but not in the source are deleted during sync`. 
To test your case you could try directly call `aws s3 sync` command inside pipelines script, and investigate the behaviour.

Regards, Igor