fix(waproto): normalize high to int32 + tighten roundtrip test

Addresses 2 review comments on PR #385:

1. CodeRabbit Major — longToString/longToNumber checked
   `value.high < 0` on the raw input. When a Long-like object carries
   high as an unsigned 32-bit value (e.g. raw {low, high} from JSON
   deserialization, not constructed via Long.fromBits/fromValue which
   normalize via `| 0`), the sign-bit detection fails and the value is
   interpreted as unsigned. Example: {low: 0, high: 0xFFFFFFFF} signed
   should be -4294967296 but returned 18446744069414584320.

   Fix: `const high = value.high | 0` before the sign check. Applied
   to both helpers in WAProto/index.js (generated artifact) and
   WAProto/fix-imports.js (template). Diverges from upstream Baileys
   PR #2333 which has the same latent bug.

   Added 2 test cases in bigint-validation.test.ts for the raw
   {low, high} scenario (Long.fromBits would mask the bug because it
   normalizes internally).

2. CodeRabbit Nitpick — the encode/decode roundtrip test in
   proto-tojson-long.test.ts used safe integers (1700000000 and
   123456789), which would silently pass even if longToString fell
   back to Number(value). Switched to uint64 values above
   MAX_SAFE_INTEGER (max uint64 and 9999999999999999999) so the
   roundtrip actually exercises the BigInt fast path.

Test suite: 789/789 passing (+2 cases for the unsigned-high scenario).

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
This commit is contained in:
Renato Alcara
2026-04-25 15:06:16 -03:00
parent 5d7840a66c
commit f0ee205f96
4 changed files with 46 additions and 11 deletions
+6 -2
View File
@@ -21,10 +21,13 @@ try {
'\t// Fast path: convert Long {low, high} directly via native BigInt\n' +
'\t// BigInt.toString() is a native C++ operation, much faster than Long\'s pure JS division loops\n' +
'\tif (value && typeof value.low === "number" && typeof value.high === "number") {\n' +
'\t\t// Normalize high to signed int32 so the sign-bit check works for inputs\n' +
'\t\t// where high is stored unsigned (e.g. raw {low, high} JSON).\n' +
'\t\tconst high = value.high | 0;\n' +
'\t\tconst lo = BigInt(value.low >>> 0);\n' +
'\t\tconst hi = BigInt(value.high >>> 0);\n' +
'\t\tconst combined = (hi << 32n) | lo;\n' +
'\t\tif (!unsigned && value.high < 0) {\n' +
'\t\tif (!unsigned && high < 0) {\n' +
'\t\t\treturn (combined - (1n << 64n)).toString();\n' +
'\t\t}\n' +
'\t\treturn combined.toString();\n' +
@@ -41,10 +44,11 @@ try {
'\t}\n' +
'\t// Fast path: convert Long {low, high} directly via native BigInt\n' +
'\tif (value && typeof value.low === "number" && typeof value.high === "number") {\n' +
'\t\tconst high = value.high | 0;\n' +
'\t\tconst lo = BigInt(value.low >>> 0);\n' +
'\t\tconst hi = BigInt(value.high >>> 0);\n' +
'\t\tconst combined = (hi << 32n) | lo;\n' +
'\t\tif (!unsigned && value.high < 0) {\n' +
'\t\tif (!unsigned && high < 0) {\n' +
'\t\t\treturn Number(combined - (1n << 64n));\n' +
'\t\t}\n' +
'\t\treturn Number(combined);\n' +
+6 -2
View File
@@ -15,10 +15,13 @@ function longToString(value, unsigned) {
// Fast path: convert Long {low, high} directly via native BigInt
// BigInt.toString() is a native C++ operation, much faster than Long's pure JS division loops
if (value && typeof value.low === "number" && typeof value.high === "number") {
// Normalize high to signed int32 so the sign-bit check works for inputs
// where high is stored unsigned (e.g. raw {low, high} JSON).
const high = value.high | 0;
const lo = BigInt(value.low >>> 0);
const hi = BigInt(value.high >>> 0);
const combined = (hi << 32n) | lo;
if (!unsigned && value.high < 0) {
if (!unsigned && high < 0) {
return (combined - (1n << 64n)).toString();
}
return combined.toString();
@@ -35,10 +38,11 @@ function longToNumber(value, unsigned) {
}
// Fast path: convert Long {low, high} directly via native BigInt
if (value && typeof value.low === "number" && typeof value.high === "number") {
const high = value.high | 0;
const lo = BigInt(value.low >>> 0);
const hi = BigInt(value.high >>> 0);
const combined = (hi << 32n) | lo;
if (!unsigned && value.high < 0) {
if (!unsigned && high < 0) {
return Number(combined - (1n << 64n));
}
return Number(combined);
+25 -2
View File
@@ -19,10 +19,11 @@ function longToStringNew(value: any, unsigned?: boolean): string {
if (typeof value === 'string') return value
if (typeof value === 'number') return String(value)
if (value && typeof value.low === 'number' && typeof value.high === 'number') {
const high = value.high | 0
const lo = BigInt(value.low >>> 0)
const hi = BigInt(value.high >>> 0)
const combined = (hi << 32n) | lo
if (!unsigned && value.high < 0) {
if (!unsigned && high < 0) {
return (combined - (1n << 64n)).toString()
}
@@ -50,10 +51,11 @@ function longToNumberNew(value: any, unsigned?: boolean): number {
if (typeof value === 'number') return value
if (typeof value === 'string') return Number(value)
if (value && typeof value.low === 'number' && typeof value.high === 'number') {
const high = value.high | 0
const lo = BigInt(value.low >>> 0)
const hi = BigInt(value.high >>> 0)
const combined = (hi << 32n) | lo
if (!unsigned && value.high < 0) {
if (!unsigned && high < 0) {
return Number(combined - (1n << 64n))
}
@@ -158,6 +160,27 @@ describe('BigInt vs Long equivalence validation', () => {
it('object without low/high', () => {
expect(longToStringNew({ foo: 'bar' })).toBe('[object Object]')
})
// Defensive: a raw {low, high} JSON object can carry `high` as an
// unsigned 32-bit value (Long.fromBits would normalize via `| 0`,
// but plain JSON deserialization does not). Without `value.high | 0`
// the sign-bit check fails and the result is interpreted as unsigned.
// Upstream Baileys PR #2333 has this latent bug; InfiniteAPI fixes it.
it('signed sign-bit detection on raw unsigned high (-1 as {low: -1, high: 4294967295})', () => {
const raw = { low: -1, high: 0xffffffff }
// signed: -1
expect(longToStringNew(raw, false)).toBe('-1')
// unsigned: max uint64
expect(longToStringNew(raw, true)).toBe('18446744073709551615')
})
it('signed sign-bit detection on raw unsigned high (-4294967296 as {low: 0, high: 4294967295})', () => {
const raw = { low: 0, high: 0xffffffff }
// signed: -4294967296 (high word = -1 after normalization)
expect(longToStringNew(raw, false)).toBe('-4294967296')
// unsigned: 0xFFFFFFFF00000000 = 18446744069414584320
expect(longToStringNew(raw, true)).toBe('18446744069414584320')
})
})
describe('fuzz: random Long values', () => {
+9 -5
View File
@@ -90,17 +90,21 @@ describe('proto serialization', () => {
expect(json.message?.imageMessage?.fileLength).toBe('1')
})
it('roundtrips encode/decode with Long fields preserving values', () => {
it('roundtrips encode/decode with Long fields preserving values (> MAX_SAFE_INTEGER)', () => {
// Use uint64 values above Number.MAX_SAFE_INTEGER (2^53 - 1) so the
// roundtrip actually exercises the BigInt fast path on decode->toJSON.
// Safe integers would silently work even if longToString fell back to
// Number(value).
const original = proto.WebMessageInfo.fromObject({
key: {
remoteJid: '123@s.whatsapp.net',
id: 'ABC123',
fromMe: false
},
messageTimestamp: 1700000000,
messageTimestamp: Long.fromString('18446744073709551615', true),
message: {
imageMessage: {
fileLength: 123456789
fileLength: Long.fromString('9999999999999999999', true)
}
}
})
@@ -109,7 +113,7 @@ describe('proto serialization', () => {
const decoded = proto.WebMessageInfo.decode(encoded)
const json = decoded.toJSON()
expect(json.messageTimestamp).toBe('1700000000')
expect(json.message?.imageMessage?.fileLength).toBe('123456789')
expect(json.messageTimestamp).toBe('18446744073709551615')
expect(json.message?.imageMessage?.fileLength).toBe('9999999999999999999')
})
})