fixing the superfluous generation of enormous amounts of temporary tuple types with UUIDs for type parameter names.

[jurgenvinju]

No description provided.

[codecov]

Codecov Report

Merging #1872 (976b864) into main (7c15369) will decrease coverage by 1%.
Report is 2 commits behind head on main.
The diff coverage is 100%.

@@           Coverage Diff           @@
##              main   #1872   +/-   ##
=======================================
- Coverage       49%     49%   -1%     
+ Complexity    6159    6143   -16     
=======================================
  Files          660     660           
  Lines        58668   58668           
  Branches      8543    8543           
=======================================
- Hits         28924   28902   -22     
- Misses       27556   27577   +21     
- Partials      2188    2189    +1     

Files	Coverage Δ
...mpl/interpreter/matching/TypedVariablePattern.java	79% <100%> (ø)
...rascalmpl/interpreter/result/AbstractFunction.java	49% <100%> (ø)
...rg/rascalmpl/interpreter/result/NamedFunction.java	66% <100%> (ø)
...g/rascalmpl/interpreter/result/RascalFunction.java	80% <ø> (ø)
...g/rascalmpl/values/RascalFunctionValueFactory.java	56% <ø> (ø)

... and 11 files with indirect coverage changes

[DavyLandman]

@jurgenvinju shall we add a test where curry calls itself? so to make sure the scheme works correctly even if the function is calling itself (aka the same location)

[jurgenvinju]

Yes. I hope that works :-)

[DavyLandman]

after that is done, I think this is fine to merge.

I think this would be nice to pickup before we make a new rascal release.

[DavyLandman]

I'm wondering about the performance impact of this. ISourceLocation::toString is not the fastest operation to perform. quite some string concats.

Is there an alternative that would be unique enough? perhaps the hashcode?

[jurgenvinju]

I really don't want to debug the situation where a hash collision blows this up :-) Remember this only happens for generic functions.

[jurgenvinju]

test bool selfApplyCurry() {
    &S(&U) curry(&S(&T, &U) f, &T t) = &S (&U u) { 
      return f(t, u); 
    };

    int addition(int i, int j) = i + j;

    func = curry(curry, addition);

    assert int(int)(int) _ := func;

    func2 = func(1);

    assert int(int) _ := func2;

    return func2(1) == 2;
}

it was already one of the tests