Replies: 4 comments 7 replies
|
Hello @fibsifan and sorry for the delay. I do think this looks like a bug to me. I've got a fix locally that seems to fix the test as well. The only concern I'd have is this hasn't changed in possibly forever :) That said, I'd find it hard to argue when provided |
|
I've filed RESTEASY-3675 for this issue. Nice find! |
|
Given the available variants of
For the last two specifically, This is in accordance with RFC4647, if I'm reading it correctly. I would personally argue that |
|
It took longer than expected, but this is now resolved see https://redhat.atlassian.net/browse/RESTEASY-3675. This fix will be in RESTEasy 6.2.17.Final and 7.0.3.Final. |
Uh oh!
There was an error while loading. Please reload this page.
Hello,
I've found a behaviour in ServerDrivenNegotiation that was unexpected to me. I'm unsure how to correctly address it though.
I created a PR #4738 to illustrate the Idea. When requesting with an Accept-Language Header, the order of the available options influences the outcome, if the options contain a mixture of languages with and without country variant. The disabled test behaves not as I would have expected.
I tried to understand RFC 2295 and RFC 2296 and figure out whether it provides insight in what the expected behaviour would be, but it looks like it doesn't - or I didn't read carefully enough, it's a lot of text.
But maybe someone here can help me understand whether I'm wrong or this behaviour is a bug?
All reactions