docs: add HP OAuth compatibility - #230

Thank you — and the same apology as on #223: the delay was ours.

The eleven required checks never ran, because Require approval for first-time contributors was left at GitHub's default and the workflows sat waiting for a
button nobody was watching. Nothing to do with your change. Approved now, CI is
running.

This one is the better of the two, and specifically because of what it refuses
to claim:

You marked it partial and then named the models that do not support OAuth.
M478-M479 and M428-M429f. An HP row reading "OAuth available" would be true for
FutureSmart 5.7 Enterprise and actively harmful to somebody holding a LaserJet
Pro, who would wait for firmware that is not coming instead of planning around
it. Naming the exclusions is worth more than the inclusions here, and most
submissions leave them out.

The firmware floor is in the note rather than the status. FutureSmart 5.7
and newer - that is the detail somebody needs before they start, and it is not
in the status field where it would get lost.

Two small things:

• data/devices.json lost its trailing newline, and the entry before yours got
  its closing brace pulled onto the same line - "...repointed." },. The
  JSON parses, so this is cosmetic, but the diff is noisier than the change.
• This and docs: add Lexmark OAuth compatibility #223 both edit data/devices.json, so whichever merges second will
  need a rebase. No action needed yet; I will say which way round once CI is
  green on both.

This and #223 are the first contributions to this repository from anybody
outside it. Both are going in.

Merged. Both of yours are in — the compatibility list is at 17 entries and HP
and Lexmark are two of the largest gaps it had.

Three things happened on your branch that you did not do, so you are not left
wondering:

I pushed a merge commit to it. #223 landed first and both pull requests
edited data/devices.json, so this one conflicted. Rather than ask you to
rebase I resolved it here: main's file, plus your HP entry, then the generator
re-run so the table and the new page match. Your entry is unchanged. The
trailing newline is back too.

The red device-table check was my bug, not yours. A validator I added the
day before imports yaml, and the pip install pyyaml in that job sat in a
step below it, so it died before reaching the install. Fixed in #233. Nothing
in your change ever caused it.

Workflows no longer wait on approval for you. Once #223 merged you stopped
being a first-time contributor, so CI on this one started by itself.

If you want a third, the same gap exists for Sharp, Toshiba and Synology, and
data/devices.json is the only file that needs touching. The rule that matters
is the one you already followed twice: a vendor's own published page as
evidence, and a status that does not claim more than the page does.