Introduce dereference operator for references

[darkdrag00nv2]

Closes #2803

Description

---

• Targeted PR against master branch
• Linked to Github issue with discussion and accepted design OR link to spec that describes this work
• Code follows the standards mentioned here
• Updated relevant documentation
• Re-reviewed Files changed in the Github PR explorer
• Added appropriate labels

[darkdrag00nv2]

Need to cover a few more types in the interpreter tests.

[turbolent]

Nice!

[turbolent]

CI is reporting data races: https://github.com/onflow/cadence/actions/runs/7240009930/job/19722511487?pr=2982

[codecov]

Codecov Report

Attention: 20 lines in your changes are missing coverage. Please review.

Comparison is base (77fc728) 80.15% compared to head (4736011) 80.00%.
Report is 110 commits behind head on master.

❗ Current head 4736011 differs from pull request most recent head 129d7c5. Consider uploading reports for the commit 129d7c5 to get more accurate results

Files	Patch %	Lines
runtime/sema/check_unary_expression.go	55.88%	14 Missing and 1 partial ⚠️
runtime/interpreter/value.go	72.72%	3 Missing ⚠️
runtime/interpreter/interpreter_expression.go	86.66%	1 Missing and 1 partial ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #2982      +/-   ##
==========================================
- Coverage   80.15%   80.00%   -0.16%     
==========================================
  Files         348      356       +8     
  Lines       81989    83037    +1048     
==========================================
+ Hits        65716    66431     +715     
- Misses      13952    14247     +295     
- Partials     2321     2359      +38     

Flag	Coverage Δ
unittests	80.00% <78.72%> (-0.16%)	⬇️

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.

[turbolent]

Looks good, great work!

In general I'm still not sure if this should be a magic function on references, or rather something else, like an operator.

For the functionality to be a magic function, the PR should also handle the case where a composite declares a member named the same as the magic function, as it seems like the current implementation shadows / makes it inaccessible. We might want to forbid declaring members which have that magic name. An operator would not have that problem.

[turbolent]

We might want to use an operator here instead of / in addition to the method, e.g. the unary prefix operator *, like in Rust, Go, C, C++, etc.

[darkdrag00nv2]

Sure, will update to remove the function and just have the operator.

[turbolent]

@darkdrag00nv2 Not sure if you started yet, but if not: I'm half-way done and could finish and push it up here

[darkdrag00nv2]

@turbolent Haven't started yet. You can push it. Thank you!

[turbolent]

Refactored the dereference function to the prefix * operator.

@onflow/cadence could you please have a look?

[dsainati1]

Now that dereference isn't a member function, it would be nice for * to have simple/clean behavior with respect to optional references. I.e. if we have a &Int? type, it would be nice for the * operator on that type to produce a Int?. Previously we could do this with ?. chaining, but with the change to an operator we lose this functionality. Can we add this (and tests for it)?

[turbolent]

@dsainati1 Would this simply be *(optRef!)? Do you mean that * on an optional ? of a reference type (&T?) should return T? ?

[dsainati1]

Yeah, it would be nice for the * operator to implicitly "lift" itself to operate on optional values as well as normal ones.

[turbolent]

@dsainati1 Sounds good! Let's get that added in a follow-up PR. Could you please open a feature request issue for it?

[SupunS]

LGTM!