fix(webidl): correct signed integer bounds in ConvertToInt

[deepview-autofix]

The signed integer lower bound was computed as Math.pow(-2, bitLength) - 1, which evaluates to (-2)^bitLength - 1 instead of the spec-required -2^(bitLength - 1). The step 11 overflow threshold similarly used 2^bitLength - 1 instead of 2^(bitLength - 1), causing values in [2^(bitLength-1), 2^bitLength - 2] to skip the signed wrap. Fix both expressions and update the surrounding comments to match the WebIDL spec.

[codecov-commenter]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 93.03%. Comparing base (bc0a19c) to head (90d3c3d).

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #5038      +/-   ##
==========================================
- Coverage   93.03%   93.03%   -0.01%     
==========================================
  Files         110      110              
  Lines       35793    35793              
==========================================
- Hits        33301    33299       -2     
- Misses       2492     2494       +2     

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

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

[KhafraDev]

the comments come from https://webidl.spec.whatwg.org/#abstract-opdef-converttoint, the changes to them should be reverted

[ChALkeR]

@KhafraDev No.

The comments were copy-pasted incorrectly originally due to them including formatting.

It's the main that has comments mismatching the spec, not the PR.

Please check the spec you linked

[KhafraDev]

I did, for example in the reference image, there are no parenthesis:

The comments here are meant to match the spec -- not to dictate the correct behavior necessarily, but to make it easier (mostly for me) to match what code matches the instruction in the spec. Makes it easier to refactor whenever the specs change.

[KhafraDev]

I see what you mean though

[ChALkeR]

@KhafraDev 2^bitLength - 1 - 1 (as was in main) does not correctly represent that screenshot.

Also, you already added ^, why omit parenthesis?

[KhafraDev]

Yeah it appears as though it is wrong in both cases unfortunately

[ChALkeR]

it is wrong in both cases unfortunately

WDYM?

[KhafraDev]

When refactoring (especially in the fetch spec) one of the best parts of the undici code base is the ability to copy a comment and CTRL+F to find where exactly a step was moved/changed to/from in the spec. It makes refactoring much easier from what I've found. In this case, I'd prefer to keep webidl comments the same, but they should also be readable (and the spec doesn't really ever change, due to its nature).

[ChALkeR]

and CTRL+F to find where exactly a step was moved

That already doesn't work as ^ was inserted.

[KhafraDev]

Which is why I approved, main is wrong, it's not your responsibility to fix an unrelated 'issue' in a comment