More i8 legalizations

[bjorn3]

• A short description of what this does, why it is needed; if the
  description becomes long, the matter should probably be discussed in an issue
  first.
  This adds several legalizations for small integers.
• This PR contains test cases, if meaningful.
• A reviewer from the core maintainer team has been assigned for this PR.
  If you don't know who could review this, please indicate so and/or ping
  bnjbvr. The list of suggested reviewers on the right can help you.

Fixes #433.
cc #466 and #1117

[abrown]

I think this PR makes sense to me but I'm not confident about how the legalization groups work (e.g. widen): @bnjbvr, any issues with this or should I approve and merge?

[abrown]

(note to self: if we do merge we should probably "squash and merge")

[bnjbvr]

I think @julian-seward1 proposed to give a look at this PR; I'll let the two of you coordinate for reviews.

[julian-seward1]

I have to say, I don't understand the legalisation machinery enough to comment meaningfully here. I would however say that IMO we should try to avoid generating operations that read or write 16 bit registers on x86_{32,64} since that puts us at risk of partial register stalls.

[bjorn3]

This makes some existing "macro" legalizations apply to small ints too. It also uses uextend to legalize brz.i8 to brnz.i32. There should be a faster way for this, but that can be implemented later.