TestMediaWithoutFileName test does not allow changes in content type between media upload and download

[anoadragon453]

The tests:

• Can download without a file name locally
• Can download without a file name over federation

that are a part of TestMediaWithoutFileName both have a check for the Content-Type header value that was provided during initial upload of the test media. They ensure this value does not change between uploading and downloading media and its associated metadata.

Synapse will however augment the content type of text files in order to specify UTF-8 as their encoding. The reasoning behind this is at matrix-org/synapse#7043.

While the spec doesn't explicitly state that the Content-Type response header when downloading a file needs to exactly match what was provided during upload, it does state that it should be the content type of the file, and so homeservers should at least try to get it right.

The original sytest this was taken from, Can download without a file name locally, did not check Content-Header. I personally think it is valuable to do so, but we may want to make an exception for this modification, as it does seem to be beneficial and not against the spec, however vague it might be.

[richvdh]

both have a check for the request's Content-Type header value.

which request? The one Complement sends to the server to fetch the media? That seems to make no sense?

[anoadragon453]

both have a check for the request's Content-Type header value.

which request? The one Complement sends to the server to fetch the media? That seems to make no sense?

No sorry, the Content-Type header of the initial upload media request.

[kegsay]

Maybe just check the mime-type on the Content-Type header and not a literal string match?

[anoadragon453]

Right, looks like we can use http.DetectContentType for this.

[anoadragon453]

http.DetectContentType returns text/plain; charset=utf-8 lol.

I'll just do a split on the first string before ;.

[turt2live]

fwiw this is a conversation that should happen in the spec at some point - I have fairly strong opinions that the server should not be doing certain things to the header (like interpreting the body to pick a different header) or changing charsets on people, however the spec does indeed do a poor job of explaining what the requirements are. In general, the server is expected to be largely a passthrough for the content-type header, but whether that makes sense or not is a spec concern.

Ideally complement doesn't stray away from the spec like sytest did/does, so it would be good to solve this for everyone and not patch over it.

[anoadragon453]

@turt2live Good point. Is there a spec issue for this already?

[turt2live]

@anoadragon453 matrix-org/matrix-spec-proposals#2701