Add {flag?} MPI function

[wyld-sw]

This adds an MPI counterpart to the FLAG? MUF primitive. The check logic has been extracted into its own function.

Functionality

Both MPI and MUF versions check a single flag, which can be a prefix of a flag alias (regardless of object type). The check can also include any number of "!" characters on the left-hand side of the string for alternating negation.

This may be the first object-facing MPI function I've contributed, so let me know if I missed something there.

Consistency

This check now recognizes BOUND, HIDE, SETUID, and XPRESS as flag aliases. SET also was changed to recognize them.

Minor change

The code for FLAG? currently:

• allows "" as the flag to check, and eventually returns 0.
• allows "U" as the flag to check, and eventually returns 0.
• does not allow "!" as the flag to check, and aborts with "Unknown flag"

This seems inconsistent, as the first two are similar conditions to the third. I've changed the third case to return 0 (with a bonus of returning 0 more quickly in the first case). I can see a corner case where one would want to abort when checking for "!", but it doesn't seem that compelling. Feel free to disagree, and I can change it to pass the error condition back to {flag?} and FLAG?.

[tanabi]

One necessary change, one suggestion/comment. Looks really good! :)

[tanabi]

Perhaps we should note in both this code and in has_flag that any change needs to be made in both places?

Ideally, we'd unify these 2 case statements in some way since it looks like this has a lot of overlap with has_flag's code. However, there's also such a thing as optimizing yourself until you're weary, so if you rather put a TODO and a note that both places need to be changed, I wouldn't judge you :)

If you wanted to go the extra mile, though, I'd say make a str_to_flag call that has this case statement, then both has_flag and prim_set can use it (along with anywhere else this mess lives).

[wyld-sw]

str_to_flag is definitely coming - just wanted to see if I could be clever with ALL the flag checking. I can certainly add str_to_flag to this PR in the meantime so it definitely doesn't wait for a unified process.

It may have been added when I got to fully reviewing p_db.c :)

[wyld-sw]

I think both changes are made. Please review; I think everything's in order...

I'm not too thrilled about the duplicate "truewizard" check, but it seemed like the best way (right now?) to preserve the different meanings for SET and FLAG?.

do_set is on notice as needing a change, too, but it uses both the str_to_flag logic and the restricted function (while introducing its own restrictions) and will take more thought.