v3.2: Ordered multipart examples

[handrews]

These are the examples that go with PRs #4745 and #4744.

This includes the recently (this past week) updated multipart/related example that was provoking a lot of discussion. Otherwise, the examples are the same as in the old PR #4589.

• schema changes are included in this pull request
• schema changes are needed for this pull request but not done yet
• no schema changes are needed for this pull request

[miqui]

@handrews looks good other than the minor typo

[ralfhandl]

+1, minor nits

[handrews]

@ralfhandl @miqui I have updated the above and will mark them resolved to reduce noise when we look at this in the call, but here is a summary:

• I removed the whitespace in contentType and filed issue v3.x: Does the Encoding Object's contentType allow whitespace around commas? #4788 to track if we should say something about it. The no-whitespace version is definitely valid so we'll go with that here
• The "page" vs "pages" was a bit of a mess as I was allowing additional text/html resources, so I changed that and reworded things a bit to make it clear that this example is for a single page, even if you could also do a linked set of pages
• I changed the ref name to RFC2557NoContentId which makes it easier for me to parse- it took me a minute to realize what I'd been doing there with the other name. The referenced schema is the same- I don't actually know any other way to forbid a header except with not: {} which is 😵‍💫

[lornajane]

@handrews a note to review the wording here, and consider if we need both this example and the one above

[lornajane]

We discussed the media range in the multipart/related example - the type param is actually required so we agreed to keep it in the example this time.

[handrews]

With (more) apologies to @ralfhandl I rebased to clear conflicts with #4747, in this case it was just that both added new examples in the same place at the end of the section, so now this PR adds them after the old examples but before the nested multipart example. The commits are otherwise identical to before.