feat!: Disallow nonlocal value edges into FuncDefn's

[acl-cqc]

Closes #936

Test for ext edges only in this PR. (Would be good to have one for dom edges too but that can follow.)

BREAKING CHANGE: Hugr's with nonlocal ("Inter-Graph") edges that enter FuncDefns will now fail validation

[codecov]

Codecov Report

Attention: Patch coverage is 82.97872% with 8 lines in your changes are missing coverage. Please review.

Project coverage is 86.58%. Comparing base (1c422ab) to head (100ad00).
Report is 4 commits behind head on main.

Files	Patch %	Lines
hugr/src/hugr/validate/test.rs	76.66%	0 Missing and 7 partials ⚠️
hugr/src/hugr/validate.rs	94.11%	0 Missing and 1 partial ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #1061      +/-   ##
==========================================
+ Coverage   86.26%   86.58%   +0.31%     
==========================================
  Files          82       83       +1     
  Lines       17236    17594     +358     
  Branches    17236    17594     +358     
==========================================
+ Hits        14869    15234     +365     
+ Misses       1547     1532      -15     
- Partials      820      828       +8     

Flag	Coverage Δ
rust	86.58% <82.97%> (+0.31%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[mark-koch]

LGTM 👍

[acl-cqc]

I fiddled the order of things so that if you have an edge that enters just a FuncDefn, you'll get the "no edge into funcdefn" error; previously you'd have got the "missing order edge" error (and you can't actually add the order edge to the funcdefn, as you would need to do to get the "no edge into funcdefn" error, because funcdefns can't receive order edges).

The test for that can't be constructed with the builder (because the builder always tries to add the order edge, and fails), so would have to duplicate most of test_ext_edge. Which doesn't feel worth it to check we've slightly-improved an error message :(. I'm hoping this way is acceptable for being simpler anyway :)...