Update docs on rev attribute

[chharvey]

I noticed a couple of bugs with the now-reintroduced rev attribute, and wanted to document them here.

1. The .rev IDL attribute for HTMLAnchorElement has a broken documentation link. I think the solution would be to add it into this paragraph, and fix the link to point to it:

The IDL attributes download, target, rel, hreflang, and type, must reflect the respective content attributes of the same name.

2. Under the section Links created by <a> and <area> elements, there is documentation on the [rev] content attribute, which applies to a and link. This is misleading because link is not in the section title. Perhaps the title should be changed to “Links created by <a>, <area>, and <link> elements”.

3. The [rel] content attribute for a and area is documented in this same section, but [rel] for link is documented under its own section: The link Element. If the title of the section is changed to include “<link>”, perhaps documentation for link[rel] should be moved here.

4. Lastly, not a bug, but a question: Should the [rev] content attribute be allowed on area elements? Since area is very similar to a, I feel like it should be consistent.

[chaals]

Thank you for picking these up.

I have assigned myself to them, somewhat hopefully, but if you have time to make a Pull Request you are very welcome to do so. (I am happy to help with that...)

Comments on your detailed observations:

1. rev should be in that list I think, subject to testing whether that happens.
2. and 3. Yes, we should clean that up one way or another.
3. Yes, it should, IMHO.

[chaals]

This is a bit more complicated than I thought. I propose to do the following (but slip to the next milestone - instead of being now a couple of days overdue, that means within 4 weeks):

• test link rev="stylesheet" href="some-potentiall-open-doc" which should not work in browsers and should be forbidden to work by spec. If something like that does work (it seems unlikely), file browser bugs.
• test area rev="something" ... for RDFa processors like Google's and Yandex' structured data validator,
• Note that links with "rel" only create hyperlinks and do not create links to external resources (like stylesheets).
• Refactor the documentation, but probably by moving the description of rev for link elements up to the description of link elements, and if the attribute is processed by e.g. RDFa processors, make the relevant changes (which is s/link/area/ in the section mentioned, I think). The section on link types should probably also note which types are not allowed on rev (basically those that lead to creating a Link to an external resource`.

[chaals]

OK, it turns out that rev does reflect, so I am going to commit that.

having it on area at least breaks in that unlike rel it doesn't seem to reflect there, so allowing it would mean making tweaking the editorial stuff noted in some more ways.

I haven't checked if RDFa processors actually handle rev on area.

So I will commit the bugfix, and open new issues for

1. editorial cleanup
2. whether rev should be allowed on a

[chaals]

opened #1335, #1336 to deal with the aspects of this not covered by #1328

[chharvey]

Summary for newcomers:

Item 1 is closed by #1328
Items 2 and 3 are closed by #1335
Item 4 is opened in #1336 (as of this time not closed yet)

Thanks