[WIP] Row Variables, the hard (but good?) way

[acl-cqc]

• Add RowVarOrType, used instead of Type in places where row variables are permitted
• Add TypeRowVarLen (via common TypeRowBase), FuncTypeVarLen (via common FuncTypeBase) and PolyFuncVarLen(via common PolyFuncBase).
• Function types in the type hierarchy can have variable arity; OpDefs can have variable arity; nodes (even with type variables) must have fixed arity
• A few awkward cases of extra .into() or non-working Default::default() might be better
• TypeRowElem seems badly placed and/or named - ideas welcome? There are also visibility issues.
• Not nearly enough tests

Is it worth it??

• It's a big change
• But, once all the relevant classes/types had been defined (the three pairs listed above in naming) pretty much everything did just drop neatly into place.
• However, +400 lines!! And that's without enough tests....
• See what you think....the hacky version feat!: Allow "Row Variables" declared as List<Type> #804 might be more appealing now!!

[codecov]

Codecov Report

Attention: Patch coverage is 81.46965% with 58 lines in your changes are missing coverage. Please review.

Project coverage is 85.39%. Comparing base (d98fb79) to head (34bd13b).

Files	Patch %	Lines
hugr/src/types.rs	50.94%	26 Missing ⚠️
hugr/src/types/serialize.rs	50.00%	11 Missing ⚠️
hugr/src/types/poly_func.rs	53.84%	6 Missing ⚠️
hugr/src/types/signature.rs	92.59%	4 Missing and 2 partials ⚠️
hugr/src/hugr/validate/test.rs	86.84%	1 Missing and 4 partials ⚠️
hugr/src/types/type_row.rs	95.45%	2 Missing ⚠️
hugr/src/extension/op_def.rs	93.75%	0 Missing and 1 partial ⚠️
hugr/src/types/check.rs	75.00%	0 Missing and 1 partial ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #903      +/-   ##
==========================================
- Coverage   85.49%   85.39%   -0.11%     
==========================================
  Files          78       78              
  Lines       14220    14398     +178     
  Branches    14220    14398     +178     
==========================================
+ Hits        12158    12295     +137     
- Misses       1428     1461      +33     
- Partials      634      642       +8     

Flag	Coverage Δ
rust	85.39% <81.46%> (-0.11%)	⬇️

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.

[acl-cqc]

This is to allow instantiating a TypeParam::List with something like [USIZE, var, BOOL] where var is itself a List....I'm not sure all cases here are necessary, and such an instantiation won't actually validate, so more to do here.

[doug-q]

I prefer #804 to this. Although I do agree that the approach here is more correct, it complicates the api to no benefit for uses that are "downstream" from any polymorphism.

[acl-cqc]

I prefer #804 to this. Although I do agree that the approach here is more correct, it complicates the api to no benefit for uses that are "downstream" from any polymorphism.

Fair. TBH I think the main benefit of this way is that I am a lot more confident that it is (at least mostly) correct!...

[doug-q]

Another option is to parameterise instead by a const bool (defaulting to false) which determines whether row variables are allowed. Perhaps this would strike a balance between the correctness-by-construction here and the simplicity of the other PR.

[acl-cqc]

Let's go with #804