question

Mike-E avatar image
1 Vote"
Mike-E asked ·

Why Does Azure Domain Verification Not Follow the HTTP Specification?

Hello,

I am wanting to comment on this thread:
https://social.msdn.microsoft.com/Forums/en-US/61ef2463-c132-4246-8748-c3090e0deed0/publisher-domain-verification-fails-because-quotverification-of-publisher-domain-failed-error?forum=WindowsAzureAD

But it's saying it is now found on here, and doing a search is not returning anything, so I am now posting here. :)

I am trying to apply the answer but I am not sure how to do this? I am hosting the file in GitHub Pages. Is there a configuration there to apply perhaps? I looked around and did not see anything.

Thank you for any assistance you can provide!

EDIT: Also, to be sure, this is the error message I get: "The server returned an unexpected content type header value. [nc186]"

EDITx2: It seems this issue is caused by GitHub Pages serving pages in a more detailed manner according to the HTTP specification than what Azure Domain Verification does for its parsing.

In particular, GitHub Pages emits the charset parameter of the Content-Type header and Azure Domain Verification does not recognize or further allow this, resulting in an unnecessary error.

https://www.w3.org/International/articles/http-charset/index

So, why is this the case?

not-supportedazure-active-directoryazure-ad-domain-services
· 1
10 |1000 characters needed characters left characters exceeded

Up to 10 attachments (including images) can be used with a maximum of 3.0 MiB each and 30.0 MiB total.

A big shout out to the introduced &#39's after editing this post. I can see why we're running into the problems that we are having here. 🙄

0 Votes 0 ·
DSPatrick avatar image
0 Votes"
DSPatrick answered ·

The thread was not migrated. None of the MSDN/TechNet threads will be migrated to QnA, but the discussions are moving as MSDN/TechNet forums are closing down. That forum has been archived and moved to read-only status so it isn't possible to comment the thread.

You can read on here about the migration process details.

https://docs.microsoft.com/en-us/teamblog/msdn-technet-migration

(please don't forget to mark helpful replies as answer)


· 6 ·
10 |1000 characters needed characters left characters exceeded

Up to 10 attachments (including images) can be used with a maximum of 3.0 MiB each and 30.0 MiB total.

Thank you for the effort @DSPatrick but that is hardly the answer I am looking for here. 😁

0 Votes 0 ·

You can provide feedback about the migration process at page bottom here.

https://docs.microsoft.com/en-us/teamblog/msdn-technet-migration






0 Votes 0 ·

The question is not about migration (which is indeed an annoyance) but of applying the answer provided in the original question. :)

0 Votes 0 ·
Show more comments
shashishailaj avatar image
0 Votes"
shashishailaj answered ·

Hello @Mike-EEE ,

As far as I know , giihub pages does not provide ability to set per file or per repository Content-type value. The details about can be checked on here. I checked this and found that Github pages always returns application/json; charset=utf-8 however the expected content-type value is application/json and this is the reason it is not working for you. Github pages does not allow configuring this as far as I am aware. I tried testing the same in one repo and you can see in the curl output below.

4621-gitmimetype.jpg


As an alternative I would suggest you to host your static website on Azure Storage blob container where this would work and you would be able to complete published domain verification without an issue as far as I think. Please check the tutorial for the same.

Hope this helps. In case the information provided in this post is helpful , please do accept it as answer so that it is helpful for other members of the community who have similar questions. In case you have any further queries , feel free to let us know and we will be happy to help.

Thank you.


gitmimetype.jpg (60.2 KiB)
· 9 ·
10 |1000 characters needed characters left characters exceeded

Up to 10 attachments (including images) can be used with a maximum of 3.0 MiB each and 30.0 MiB total.

Thank you for your reply @shashishailaj but I am having trouble understanding why I have to go through the effort of learning an entirely new hosting arrangement when I have followed the instructions exactly as prescribed without any indication that GitHub Pages is not supported. The file is there and it proves that I have control over the resource... why is this not enough?! Seems like a bug to me. 💯

0 Votes 0 ·

This is the wrong venue for GitHub problems. I'd suggest starting a new issue here.

https://github.com/MicrosoftDocs/feedback/issues/new/choose


0 Votes 0 ·

More information for you @DSPatrick. Why is Azure not conforming to specification?

https://www.w3.org/International/articles/http-charset/index

1 Vote 1 ·
Show more comments

The problem isn't GitHub Pages @DSPatrick ... if you read @abraunegg-9545's answer GitHub Pages is serving up content in a valid manner according to specification. The problem is that Azure is not properly parsing the header.

0 Votes 0 ·
Show more comments
abraunegg-9545 avatar image
1 Vote"
abraunegg-9545 answered ·

Hi all,

Please can Microsoft fix this by ignoring the charset if it is sent?

I have attempted to get this resolved by:

  • Raised a case with Azure Docs, as the documentation for this issue is not correct - see https://github.com/MicrosoftDocs/azure-docs/issues/46080 - get advised to open a case with Github

  • Raising a case with jekyll - they say raise a case with Github - see https://github.com/jekyll/jekyll/issues/8006

  • Raised a case Github. After 3 week delay in response, Github support says - Github uses mime-db project which aggregates MIME types from the Apache and Nginx projects. Raise a case with mime-db for .well-known mime types.

Now this is really not the right solution here.

What should happen as the easiest solution point is that any .well-known requests, the Microsoft Azure process should ignore the any charset that is sent. This will fix the problem for everyone where they have zero control of the web server - ie - if you are using GitHub pages to present the content - you have zero control.

Please can you look into this issue and come up with something workable.

· 2 ·
10 |1000 characters needed characters left characters exceeded

Up to 10 attachments (including images) can be used with a maximum of 3.0 MiB each and 30.0 MiB total.

Exactly... why on earth is this being so particular about a response header?! Didn't this process used to be done via DNS records? What happened to those?

The file is there, that should be enough proof to determine ownership and control of the resource. Quite amazing and particular -- to put it politely! -- that it simply does not accept this as the required proof -- especially since we're following the directions as given, and even more especially since this used to be completed with a simple entry in the DNS.

Please team, put this pretty on this issue!

1 Vote 1 ·

Totally agree

1 Vote 1 ·