Re: Report: Inconsistency in HTTP 204 response body parsing (#5227)

Ah, I see much more consistent results when the server explicitly closes the socket after writing the data. Eliding results for non-standard ports, and extending with the protocol version used to interpret the response:

Browser     | Initial conditions   | Port | 0   | 1   | 2   | 3   | 4   | 5   | 6
------------|----------------------|------|-----|-----|-----|-----|-----|-----|-----
Chrome 59   | Independent response | 80   | 1.1 | 1.1 | 1.1 | 1.1 | 1.1 | 0.9 | 0.9
Chrome 59   | Following a 204      | 80   | 1.1 | 1.1 | 1.1 | 1.1 | 1.1 | 0.9 | 0.9
Firefox 55  | Independent response | 80   | 1.1 | 1.1 | 1.1 | 1.1 | 1.1 | 0.9 | 0.9
Firefox 55  | Following a 204      | 80   | 1.1 | 1.1 | 1.1 | 1.1 | 1.1 | 1.1 | 1.1
Edge 40     | Independent response | 80   | 1.1 | 1.1 | 1.1 | 1.1 | 1.1 | 0.9 | 0.9
Edge 40     | Following a 204      | 80   | 1.1 | 1.1 | 1.1 | 1.1 | 1.1 | 0.9 | 0.9
Safari 10.2 | Independent response | 80   | 1.1 | :x: | :x: | :x: | :x: | :x: | :x:
Safari 10.2 | Following a 204      | 80   | 1.1 | :x: | :x: | :x: | :x: | :x: | :x:

Note that the surprising Firefox behavior holds even in this case. I'm following up with a Firefox bug report--hopefully, the maintainers there will be able to shed some light on this:

https://bugzilla.mozilla.org/show_bug.cgi?id=1356614

Thanks again for your help, @MattMenke!

View on GitHub: https://github.com/w3c/web-platform-tests/pull/5227#issuecomment-294201196

Received on Friday, 14 April 2017 17:50:20 UTC