-
Notifications
You must be signed in to change notification settings - Fork 843
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Support an optional charset parameter in Content-Type
on requests
#874
Conversation
c58dd12
to
2daade0
Compare
I'm 👍 to this (needs to be added to the normative statements json file tho). |
Awesome. I'll add it to the json file too. In that file, can I change an existing id, or will that break some tests? Specifically, I was thinking of breaking up the |
We have not yet implemented a complete test suite using this, so i think it's okay to change the ID in this instance. In the future it'll be really important that we don't, though. |
20b8717
to
f4fe07d
Compare
@tkellen Ok cool. I changed the id and pushed the normative json. |
@dgeb, you good with this? |
f4fe07d
to
cb1e217
Compare
Given that Firefox is about to fix this bug, and it's the last client that has this problem, I'm thinking we should instead merge a narrower version that allows the server to ignore |
Works for me! |
Seems like we might want to mention this in the FAQ as well? |
cb1e217
to
a1b3982
Compare
Ok, I've rebased this all to be the minimal workable change. Lmk if it's good to merge. Re FAQs: I'm thinking this problem will probably go away fast enough on it's own that it may not be worth it to give it an FAQ item (that we'd just have to remove later). But if we find that people keep asking, I'm definitely up for adding one. |
Ping @dgeb |
My only nit is in the line:
The clause "but SHOULD ignore its value" feels like an overreach. If we're allowing the |
I am not sure that any of this is strictly necessary. It seems like the spec could happily steer clear of any of this. Firefox clients could temporarily be non-compliant, and servers could helpfully provide some wiggle room in their compliance check. Very soon this won't be an issue for anyone. I don't see a reason to formalize any of this in the spec unless we consider it desirable on its own merits. |
Right, the reason to add this to the base spec is that, without it, a server can't provide wiggle room without becoming non-compliant itself. So allowing that wiggle room on servers is what this does, but it still leaves bad clients technically non-compliant.
The reason I said it should be ignored is because the charset for valid JSON can be (more reliably) determined just from the content itself. (See #889.) |
Since we're now targeting 1.1 with this change, and it is directly related to a clause in place to make the spec future proof for extensions (also targeting 1.1), maybe we can arrive at a joint solution that is inclusive of specific media type parameters and ignores others? Even if we can't, is there any harm in waiting on this until we decide the extension mechanism? |
Maybe...if we do that, we'd be saying that we'll never add another media type parameter, but maybe that's fine. It simply would simplify processing logic!
No, there's no harm in that! I'm happy to wait :) |
Thanks @ethanresnick :) |
I'm going to give this a close (and open a new PR to fix the normative-statements typo), as Firefox fixed their bug months ago and no one has brought this up as an issue since. |
A fix for #837.