buffer: remove unreachable overflow check in atob

[haramj]

This commit resolves a TODO comment in lib/buffer.js by changing the error thrown by atob() during a potential overflow condition.

Modified lib/buffer.js to instantiate and throw new ERR_INVALID_ARG_VALUE('input', result, 'The string to be decoded is too long.'); for the overflow case (result === -3).

This change makes the atob implementation more robust and developer-friendly.
@anonrig

->

The case for a result of -3 was intended to handle a potential
buffer overflow error. However, the underlying C++ implementation using
simdutf::base64_to_binary does not perform bounds-checking and therefore
can never return an OUTPUT_BUFFER_TOO_SMALL error.

This makes the case -3: block unreachable code. This commit removes
the dead code for correctness and clarity.

[Renegade334]

simdutf::base64_to_binary can never return OUTPUT_BUFFER_TOO_SMALL, as it doesn't perform bounds-checking of its target. This should just be removed in favour of an assertion in case of an unrecognised error.

[haramj]

simdutf::base64_to_binary can never return OUTPUT_BUFFER_TOO_SMALL, as it doesn't perform bounds-checking of its target. This should just be removed in favour of an assertion in case of an unrecognised error.
@Renegade334 I've pushed a new commit. Please let me know if any other changes are needed.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 88.57%. Comparing base (b13f24c) to head (f5c78ba).
⚠️ Report is 34 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main   #60161      +/-   ##
==========================================
+ Coverage   88.53%   88.57%   +0.03%     
==========================================
  Files         703      704       +1     
  Lines      207997   208125     +128     
  Branches    40015    40012       -3     
==========================================
+ Hits       184150   184338     +188     
+ Misses      15864    15815      -49     
+ Partials     7983     7972      -11     

Files with missing lines	Coverage Δ
lib/buffer.js	100.00% <100.00%> (ø)

... and 51 files with indirect coverage changes

🚀 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.

[haramj]

@nodejs/buffer please take a look at this PR when you have some time?

[Renegade334]

I think the comment at the corresponding location in the C++ source should be changed to match:

node/src/node_buffer.cc

Lines 1470 to 1477 in be7ea27

	// Default value is: "possible overflow"
	int32_t error_code = -3;
	if (result.error == simdutf::error_code::INVALID_BASE64_CHARACTER) {
	error_code = -2;
	} else if (result.error == simdutf::error_code::BASE64_INPUT_REMAINDER) {
	error_code = -1;
	}

Just needs to comment that if the error code case is not recognised, then the function will return the "unrecognised error" value -3 to JS.

[haramj]

I think the comment at the corresponding location in the C++ source should be changed to match:

node/src/node_buffer.cc

Lines 1470 to 1477 in be7ea27

	// Default value is: "possible overflow"
	int32_t error_code = -3;
	if (result.error == simdutf::error_code::INVALID_BASE64_CHARACTER) {
	error_code = -2;
	} else if (result.error == simdutf::error_code::BASE64_INPUT_REMAINDER) {
	error_code = -1;
	}

Just needs to comment that if the error code case is not recognised, then the function will return the "unrecognised error" value -3 to JS.

Thanks, updated the C++ comments to describe -3 as the unrecognized simdutf error value returned to JS.

[haramj]

@anonrig Could you also take a look since this addresses the atob() TODO? Thanks!

[nodejs-github-bot]

CI: https://ci.nodejs.org/job/node-test-pull-request/73932/

[haramj]

It looks like the Jenkins failures are due to flaky tests rather than issues introduced by this PR.
Could someone please re-run the CI? thanks