The s3boto3 backend overrides custom Content-Encoding object parameters

[aldencolerain]

When a custom Content-Encoding is specified in the AWS_S3_OBJECT_PARAMETERS setting it is overwritten by the encoding calculated by mimetypes. It looks like this bug was introduced in PR #122.

django-storages/storages/backends/s3boto3.py

Line 421 in adefe32

_type, encoding = mimetypes.guess_type(name)

This could be easily fixed by only updating the Content-Encoding to the guessed encoding only if it wasn't already present in object parameters. See below:

django-storages/storages/backends/s3boto3.py

Lines 428 to 433 in adefe32

	if self.gzip and content_type in self.gzip_content_types:
	content = self._compress_content(content)
	parameters.update({'ContentEncoding': 'gzip'})
	elif encoding:
	# If the content already has a particular encoding, set it
	parameters.update({'ContentEncoding': encoding})

This is an issue when you want to upload a gz file but you don't want the browser to automatically attempt to decompress it, you could work around with content-disposition but it would be better to be able to control Content-Encoding and not have it overridden by the guessed encoding.