Resource overrideMimeType not working on Edge #7085
Comments
|
Update: WhatWG is changing their mind about this logic, so maybe the other browsers won't be matching Edge after all. Still, the fact remains that this breaks on some shipping Edge versions and I don't think it hurts to reorder the logic in Resource. |
|
I'm not sure this was the same thing I was seeing, but couldn't hurt. Thanks for the write up @thw0rted, I'll open a PR just to cover our bases. |
mramato
added a commit
that referenced
this issue
Sep 27, 2018
As per the updated XHR spec, you need to call open before calling `overrideMimeType`. While they are most likely rolling back this portion of the spec, it is currently required in Edge and is harmless to keep this pay even after the spec is reverted. Fixes #7085
|
Thanks again for reporting this @thw0rted! We just merged a fix for this into master and it will be included in the 1.50 release on Monday |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
TL;DR: I found a spec-compliance issue with how Resource does XHRs.
I came here to report an issue I was having with Edge, but to my surprise it isn't Edge's fault. I have some code that makes a Resource and calls
fetchXML, treating the response as an XMLDocument. This failed in Edge, which led me to dig a bit -- I noticed that my server was returning the (definitely valid XML, not HTML) content with aContent-Typeoftext/html. I don't know why, it's not under my control, so I'm stuck with it. Firefox and Chrome return an XMLDocument anyway; Edge was returning an HTMLDocument. I thought this was the bug, until I did a bunch of extra testing.I found out that the XHR spec was recently updated to state that calling
open()must reset the "override MIME type" -- see bullet 12 in the current spec. I saw that the Resource class callsoverrideMimeTypebefore callingopen. This results in the override getting ignored in recent builds of Edge, and the problem will get worse as other browsers come into compliance with the changes.I believe the fix is as simple as moving the
opencall before the call tooverrideMimeType. This would be a PR instead of an issue, but my CLA issues remain unresolved, so I'll just point 👉The text was updated successfully, but these errors were encountered: