feat(algorithms): allow calling with a PermissionName

[jyasskin]

Instead of forcing specs for simple permissions to write out a permission
descriptor. This partially fixes @annevk's objection that writing {name: "persistent-storage"} in a spec doesn't formally create a WebIDL dictionary instance, although we'll still need to formalize syntax like that for more complicated descriptors.

Preview at https://api.csswg.org/bikeshed/?url=https://raw.githubusercontent.com/jyasskin/permissions/permission-name-shorthand/index.bs#reading-current-states. I'm not sure the best place to define this. I want folks who click the link to, say, https://api.csswg.org/bikeshed/?url=https://raw.githubusercontent.com/jyasskin/permissions/permission-name-shorthand/index.bs#request-permission-to-use to find the definitions of the shorthands (so a separate "shorthands" section is out), and I don't want to make every spec write out <a for="PermissionName">request permission to use</a> to disambiguate the two definitions (so direct links to the shorthand are out). The wording doesn't fit well right next to the main definition, so I put it below.

[marcoscaceres]

LGTM, but maybe we wait for @annevk to give an ok before merging.

[annevk]

This looks okay, but doesn't really resolve the problem. It just hides it a bit more. The fact that you're using ECMAScript and IDL concepts in a place where ECMAScript and IDL are not processed (or defined to be processed) is what is problematic.

[jyasskin]

"a PermissionDescriptor with its name member set to name." seems no worse than any other time we create a WebIDL thing from prose, and is fairly close to the wording in http://heycam.github.io/webidl/#es-dictionary. Or have I missed a subtlety?

I agree that this change doesn't fix the problem of Web Bluetooth creating a descriptor in step 3 of request Bluetooth devices. I've filed whatwg/webidl#142 to try to get a good syntax for that use.

[annevk]

Ah okay, if you drop the syntax it is a lot better, yes.