fix[codegen]: fix double eval of start in range expr

[cyberthirst]

What I did

• fix of: Double Evaluation of Range's Start #3964

How I did it

• cache the start argument to the stack

How to verify it

• added a double eval test

Commit message

Commit message for the final, squashed PR. (Optional, but reviewers will appreciate it! Please see our commit message style guide for what we would ideally like to see in a commit message.)

Description for the changelog

Cute Animal Picture

[charles-cooper]

we should have tests for both 1 and 2 argument cases

[charles-cooper]

also this seems fine, but didn't we just add a fence for pop()? this has a performance penalty as well.

[cyberthirst]

we should have tests for both 1 and 2 argument cases

addressed in: 8b5b364

[codecov]

Codecov Report

Attention: Patch coverage is 61.53846% with 5 lines in your changes are missing coverage. Please review.

Project coverage is 52.70%. Comparing base (f0afaf0) to head (8b5b364).

❗ Current head 8b5b364 differs from pull request most recent head d329685

Please upload reports for the commit d329685 to get more accurate results.

Files	Patch %	Lines
vyper/codegen/stmt.py	61.53%	4 Missing and 1 partial ⚠️

Additional details and impacted files

@@             Coverage Diff             @@
##           master    #4033       +/-   ##
===========================================
- Coverage   91.12%   52.70%   -38.42%     
===========================================
  Files         106      106               
  Lines       15308    15309        +1     
  Branches     3365     3366        +1     
===========================================
- Hits        13949     8069     -5880     
- Misses        925     6603     +5678     
- Partials      434      637      +203     

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

[cyberthirst]

also this seems fine, but didn't we just add a fence for pop()? this has a performance penalty as well.

as discussed, it's a question of panicking for certain classes of contracts vs having some perf penalty

eg contracts like this wouldn't compile without the cache

@internal
@view
def bar() -> uint256:
    return 0

@external
def foo():
    for i:uint256 in range(self.bar(), 5, bound = 3):
        pass