feat: Implement proposal-arraybuffer-base64

[magic-akari]

• close feat: Implement proposal-arraybuffer-base64 #4616

[github-actions]

Test262 conformance changes

Test result	main count	PR count	difference
Total	52,862	52,862	0
Passed	49,471	49,534	+63
Ignored	2,249	2,187	-62
Failed	1,142	1,141	-1
Panics	0	0	0
Conformance	93.59%	93.70%	+0.12%

Fixed tests (63):

test/built-ins/Uint8Array/fromHex/length.js (previously Ignored)
test/built-ins/Uint8Array/fromHex/descriptor.js (previously Ignored)
test/built-ins/Uint8Array/fromHex/nonconstructor.js (previously Ignored)
test/built-ins/Uint8Array/fromHex/odd-length-input.js (previously Ignored)
test/built-ins/Uint8Array/fromHex/name.js (previously Ignored)
test/built-ins/Uint8Array/fromHex/results.js (previously Ignored)
test/built-ins/Uint8Array/fromHex/illegal-characters.js (previously Ignored)
test/built-ins/Uint8Array/fromHex/string-coercion.js (previously Ignored)
test/built-ins/Uint8Array/fromHex/ignores-receiver.js (previously Ignored)
test/built-ins/Uint8Array/prototype/toBase64/receiver-not-uint8array.js (previously Ignored)
test/built-ins/Uint8Array/prototype/toBase64/length.js (previously Ignored)
test/built-ins/Uint8Array/prototype/toBase64/omit-padding.js (previously Ignored)
test/built-ins/Uint8Array/prototype/toBase64/descriptor.js (previously Ignored)
test/built-ins/Uint8Array/prototype/toBase64/option-coercion.js (previously Ignored)
test/built-ins/Uint8Array/prototype/toBase64/nonconstructor.js (previously Ignored)
test/built-ins/Uint8Array/prototype/toBase64/alphabet.js (previously Ignored)
test/built-ins/Uint8Array/prototype/toBase64/detached-buffer.js (previously Ignored)
test/built-ins/Uint8Array/prototype/toBase64/name.js (previously Ignored)
test/built-ins/Uint8Array/prototype/toBase64/results.js (previously Ignored)
test/built-ins/Uint8Array/prototype/setFromHex/length.js (previously Ignored)
test/built-ins/Uint8Array/prototype/setFromHex/descriptor.js (previously Ignored)
test/built-ins/Uint8Array/prototype/setFromHex/nonconstructor.js (previously Ignored)
test/built-ins/Uint8Array/prototype/setFromHex/writes-up-to-error.js (previously Ignored)
test/built-ins/Uint8Array/prototype/setFromHex/subarray.js (previously Ignored)
test/built-ins/Uint8Array/prototype/setFromHex/detached-buffer.js (previously Ignored)
test/built-ins/Uint8Array/prototype/setFromHex/throws-when-string-length-is-odd.js (previously Ignored)
test/built-ins/Uint8Array/prototype/setFromHex/name.js (previously Ignored)
test/built-ins/Uint8Array/prototype/setFromHex/results.js (previously Ignored)
test/built-ins/Uint8Array/prototype/setFromHex/illegal-characters.js (previously Ignored)
test/built-ins/Uint8Array/prototype/setFromHex/target-size.js (previously Ignored)
test/built-ins/Uint8Array/prototype/setFromHex/string-coercion.js (previously Ignored)
test/built-ins/Uint8Array/prototype/setFromBase64/length.js (previously Ignored)
test/built-ins/Uint8Array/prototype/setFromBase64/trailing-garbage-empty.js (previously Ignored)
test/built-ins/Uint8Array/prototype/setFromBase64/descriptor.js (previously Ignored)
test/built-ins/Uint8Array/prototype/setFromBase64/option-coercion.js (previously Ignored)
test/built-ins/Uint8Array/prototype/setFromBase64/nonconstructor.js (previously Ignored)
test/built-ins/Uint8Array/prototype/setFromBase64/alphabet.js (previously Ignored)
test/built-ins/Uint8Array/prototype/setFromBase64/detached-buffer.js (previously Ignored)
test/built-ins/Uint8Array/prototype/setFromBase64/name.js (previously Ignored)
test/built-ins/Uint8Array/prototype/setFromBase64/results.js (previously Ignored)
test/built-ins/Uint8Array/prototype/setFromBase64/illegal-characters.js (previously Ignored)
test/built-ins/Uint8Array/prototype/setFromBase64/whitespace.js (previously Ignored)
test/built-ins/Uint8Array/prototype/setFromBase64/string-coercion.js (previously Ignored)
test/built-ins/Uint8Array/prototype/toHex/receiver-not-uint8array.js (previously Ignored)
test/built-ins/Uint8Array/prototype/toHex/length.js (previously Ignored)
test/built-ins/Uint8Array/prototype/toHex/descriptor.js (previously Ignored)
test/built-ins/Uint8Array/prototype/toHex/nonconstructor.js (previously Ignored)
test/built-ins/Uint8Array/prototype/toHex/detached-buffer.js (previously Ignored)
test/built-ins/Uint8Array/prototype/toHex/name.js (previously Ignored)
test/built-ins/Uint8Array/prototype/toHex/results.js (previously Ignored)
test/built-ins/Uint8Array/fromBase64/length.js (previously Ignored)
test/built-ins/Uint8Array/fromBase64/descriptor.js (previously Ignored)
test/built-ins/Uint8Array/fromBase64/option-coercion.js (previously Ignored)
test/built-ins/Uint8Array/fromBase64/nonconstructor.js (previously Ignored)
test/built-ins/Uint8Array/fromBase64/alphabet.js (previously Ignored)
test/built-ins/Uint8Array/fromBase64/name.js (previously Ignored)
test/built-ins/Uint8Array/fromBase64/results.js (previously Ignored)
test/built-ins/Uint8Array/fromBase64/illegal-characters.js (previously Ignored)
test/built-ins/Uint8Array/fromBase64/whitespace.js (previously Ignored)
test/built-ins/Uint8Array/fromBase64/string-coercion.js (previously Ignored)
test/built-ins/Uint8Array/fromBase64/ignores-receiver.js (previously Ignored)
test/staging/Uint8Array/fromBase64/invalid-options.js (previously Ignored)
test/staging/sm/TypedArray/prototype-constructor-identity.js (previously Failed)

[magic-akari]

Note

data-encoding gives us first-error boundaries (read/written/position), but stop-before-partial requires the last spec-acceptable full-chunk boundary, which is not always the same.

For setFromBase64, the spec requires early return once maxLength is reached (and trailing garbage after that must be ignored), while data-encoding has no output-cap-aware early-stop decode API.

Also, simple 4-character truncation is insufficient: whitespace affects read, padding legality depends on exact = placement, and strict/loose/stop-before-partial have different tail rules.

At the moment, we do not support stop-before-partial semantics.

If we decide to go with data-encoding (instead of simdutf), one realistic long-term path might involve forking data-encoding and add the API we need.

[codecov]

Codecov Report

❌ Patch coverage is 3.13253% with 402 lines in your changes missing coverage. Please review.
✅ Project coverage is 56.19%. Comparing base (6ddc2b4) to head (2ff9f60).
⚠️ Report is 650 commits behind head on main.

Files with missing lines	Patch %	Lines
core/engine/src/builtins/typed_array/builtin.rs	0.00%	402 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #4630      +/-   ##
==========================================
+ Coverage   47.24%   56.19%   +8.95%     
==========================================
  Files         476      548      +72     
  Lines       46892    60504   +13612     
==========================================
+ Hits        22154    34001   +11847     
- Misses      24738    26503    +1765     

☔ 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.

[jedel1043]

Thank you, this looks nice. Just to be clear, what would be the full list of missing things from data-encoding's side?

I ask this because I plan to open a discussion with the data-encoding maintainers to know if they're interested in adding those features.

[magic-akari]

The main missing feature is LastChunkHandlingOptions. Currently, data-encoding only offers two padding modes: None and Some, both of which enforce strict padding validation.

However, the ECMA-262 spec requires three distinct modes: strict("="), loose("="), and StopBeforePartial.

Our current implementation works around this limitation by checking for the presence of = in the input beforehand to toggle between Some("=") and None. Having native loose("=") support in data-encoding would eliminate this hack.

There's another subtle issue: ECMA-262 disallows concatenated padded inputs, but data-encoding with padding enabled accepts them. A proper loose mode shouldn't just be a binary switch—it needs to reject concatenated padded inputs to match the spec.

As for StopBeforePartial, we should probably pull some test cases from Test262 to ensure our implementation is correct.

[jedel1043]

Thank! I'll open an issue with them to see if they're open to implementing the features.