Specify multiple configURLs #667
Conversation
|
Thanks @TallTed, done. (Some of the lines you commented on aren't new, I just changed the indentation) |
spec/index.bs
Outdated
| set to the following steps, given a <a spec=fetch for=/>response</a> |response| and |responseBody|: | ||
| 1. Let |json| be the result of [=extract the JSON fetch response=] from |response| and | ||
| |responseBody|. | ||
| 1. Set |discovery| to the result of [=converted to an IDL value|converting=] |json| |
There was a problem hiding this comment.
Why are we calling this discovery?
There was a problem hiding this comment.
That's what the code used to call it. I'm actually not sure why that name was picked.
spec/index.bs
Outdated
| 1. If |login_url| or |accounts_url| is failure, return failure. | ||
| 1. Wait for both |config| and |discovery| to be set. | ||
| 1. If |discovery| is null, return failure. | ||
| 1. If |rpOrigin| is not an [=opaque origin=], and |rootUrl|'s [=url/host=] is equal |
There was a problem hiding this comment.
This seems to mean that the well known file needs to be present now in this case (otherwise it will fail in the previous line), whereas before it could pass? Is this change intentional? Why can't we keep the existing behavior where we just do not fetch the well-known in this case?
There was a problem hiding this comment.
It was easier to specify that way and Chrome's implementation does fetch the well known file in this circumstance. However, Chrome ignores fetch failures, so you're right that I need to change this. Done.
spec/index.bs
Outdated
| |config|.{{IdentityProviderAPIConfig/login_url}}, and |globalObject|. | ||
| 1. Set |accounts_url| to the result of [=computing the manifest URL=] with |provider|, | ||
| |config|.{{IdentityProviderAPIConfig/accounts_endpoint}}, and |globalObject|. | ||
| 1. If |login_url| or |accounts_url| is failure, return failure. |
There was a problem hiding this comment.
Ah maybe an existing issue but I don't know if this is allowed to return on behalf of the algorithm since these steps are done in processResponseConsumeBody which is an algorithm of its own. So maybe modify variables as opposed to return failure?
spec/index.bs
Outdated
| set to the following steps, given a <a spec=fetch for=/>response</a> |response| and |responseBody|: | ||
| 1. Let |json| be the result of [=extract the JSON fetch response=] from |response| and | ||
| |responseBody|. | ||
| 1. Set |discovery| to the result of [=converted to an IDL value|converting=] |json| |
There was a problem hiding this comment.
That's what the code used to call it. I'm actually not sure why that name was picked.
spec/index.bs
Outdated
| |config|.{{IdentityProviderAPIConfig/login_url}}, and |globalObject|. | ||
| 1. Set |accounts_url| to the result of [=computing the manifest URL=] with |provider|, | ||
| |config|.{{IdentityProviderAPIConfig/accounts_endpoint}}, and |globalObject|. | ||
| 1. If |login_url| or |accounts_url| is failure, return failure. |
spec/index.bs
Outdated
| 1. If |login_url| or |accounts_url| is failure, return failure. | ||
| 1. Wait for both |config| and |discovery| to be set. | ||
| 1. If |discovery| is null, return failure. | ||
| 1. If |rpOrigin| is not an [=opaque origin=], and |rootUrl|'s [=url/host=] is equal |
There was a problem hiding this comment.
It was easier to specify that way and Chrome's implementation does fetch the well known file in this circumstance. However, Chrome ignores fetch failures, so you're right that I need to change this. Done.
Bug: #552
Preview | Diff