Treat "NaN" as a canonical numeric index string - #1643
Open
andreasrosdal wants to merge 3 commits into
Open
Conversation
JS_AtomIsNumericIndex1() short-circuits on the first character and only
lets through digits, '-' and "Infinity", so the string "NaN" never reaches
the ToNumber/ToString round-trip that implements CanonicalNumericIndexString.
An XXX comment noted this.
CanonicalNumericIndexString("NaN") is NaN, not undefined, so on an
integer-indexed exotic object "NaN" is an out-of-bounds index rather than a
regular property. Before this change:
const ta = new Int8Array(4);
ta.NaN = 1;
ta.NaN // 1, expected undefined
"NaN" in ta // true, expected false
Object.getOwnPropertyNames(ta) // includes "NaN"
Object.defineProperty(ta, "NaN", {value: 7}) // expected TypeError
"Infinity" already behaved correctly; "NaN" is now handled the same way.
"-NaN" is still rejected, because it is not the result of ToString(NaN).
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_014mv33YvfHz7t9mmkituBnn
Checks all four integer-indexed exotic operations ([[Get]], [[HasProperty]], [[Set]], [[DefineOwnProperty]]) plus [[Delete]] against a table of canonical numeric index strings and of lookalikes that are not canonical, so the prototype chain shortcut is exercised in both directions. Also pins that ordinary objects, arrays, strings and ArrayBuffers keep "NaN" as a normal property. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01K6eRbuuuCujKgQkrHgvMrc
An integer index that is out of bounds is not a property, so [[Set]] drops the write and reports success. OP_put_ref_value relies on JS_PROP_NO_ADD to notice that a `with` binding disappeared between the reference being taken and the store, and the integer-indexed path returned before that check ever ran, so the store silently succeeded instead of raising the ReferenceError that SetMutableBinding step 3 calls for. This was already reachable through Infinity, -0 and the rest; recognising "NaN" as a canonical numeric index string just adds one more spelling, and turns test262's set-mutable-binding-binding-deleted-with-typed-array-in- proto-chain-strict-mode from passing by accident into failing. Check JS_PROP_NO_ADD before the index is dropped, which fixes both spellings, and drop the non-strict half of that pair from the expected failures. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01K6eRbuuuCujKgQkrHgvMrc
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
JS_AtomIsNumericIndex1()short-circuits on the first character and only lets through digits,-and"Infinity", so the string"NaN"never reaches the ToNumber/ToString round-trip that implements CanonicalNumericIndexString. AnXXXcomment noted this.CanonicalNumericIndexString("NaN")isNaN, notundefined, so on an integer-indexed exotic object"NaN"is an out-of-bounds index rather than a regular property. Before this change:"Infinity"already behaved correctly;"NaN"is now handled the same way."-NaN"is still rejected, because it is not the result ofToString(NaN).Verified against
test/built-ins/TypedArray+TypedArrayConstructors(43034 tests): unchanged, 38 pre-existing failures before and after.🤖 Generated with Claude Code
https://claude.ai/code/session_014mv33YvfHz7t9mmkituBnn
Generated by Claude Code