Overridden default function argument not visible in the Text representation

[radeusgd]

Our new textual representation of functions shows its non-default arguments, which is aiding with understanding errors of argument mismatches (e.g. #7359).

After some further practical usage of this new feature, I've noticed that if I set an argument which has a default value, it is not displayed in the function representation.

E.g.

from Standard.Base import all

foo a b c=3 d=4 =
    a + b + c + d

main =
    IO.println (foo b=100)
    IO.println (foo c=200)

Actual behaviour

yields

no-default-shown.foo[no-default-shown.enso:3-4] self=no-default-shown a=_ b=100
no-default-shown.foo[no-default-shown.enso:3-4] self=no-default-shown a=_ b=_

Expected behaviour

while I'd expect

no-default-shown.foo[no-default-shown.enso:3-4] self=no-default-shown a=_ b=100
no-default-shown.foo[no-default-shown.enso:3-4] self=no-default-shown a=_ b=_ c=200

or even better:

no-default-shown.foo[no-default-shown.enso:3-4] self=no-default-shown a=_ b=100 c=<default> d=<default>
no-default-shown.foo[no-default-shown.enso:3-4] self=no-default-shown a=_ b=_ c=200 d=<default>

---

After seeing this in practice, I'm sure we want to display the argument when its overridden, regardless if it has a default value or not. Otherwise there is no way to tell if the argument is set or not.

Additionally, I know I initially voted that there is no need to display defaulted arguments. I'm not yet that sure. I think after all it may be useful to include them in the textual representation, e.g. as shown above as x=<default>.

Either way, if the argument has a provided value - then its 'defaulted' status should no longer matter - it should always be displayed (like the c=200 above).

[enso-bot]

Jaroslav Tulach reports a new STANDUP for yesterday (2023-09-07):

Progress: - merged: #7762

• display overriden default args: Display overriden non-default values #7762
• sorting numpy array
• bug for GraalPy: numpy numbers aren't reported as numbers oracle/graalpython#354
• ugly workaround: StackOverflowError in sort called on numpy array. #7677 (comment)
• investigating & support: Remove \ from IR.Type.Set but support & #6165
• rewritting ReadArgumentCheckNode to have a hierarchy
• draft PR: Support runtime checks of intersection types #7769 It should be finished by 2023-09-07.

Next Day: Bugfixing