blob: unaligned UTF-16LE decode for odd-length/odd-address views

src/runtime/webcore/Blob.rs

@@ -2695,17 +2695,34 @@ impl BlobExt for Blob {
 
         if bom == Some(strings::BOM::Utf16Le) {
             let _free = (LIFETIME == Lifetime::Temporary).then(|| TemporaryBytes(raw_bytes));
-            // Mirrors Zig `bun.reinterpretSlice(u16, buf)`: drop a trailing odd byte
-            // (`@divTrunc`) so `bytemuck::cast_slice` cannot panic on slop.
+            // Mirrors Zig `bun.reinterpretSlice(u16, buf)`: `buf` may be
+            // odd-length (truncated input) or odd-address (`.slice(odd)` of a
+            // shared store) — either makes `bytemuck::cast_slice` `panic!`
+            // (uncatchable). Drop the trailing odd byte (`@divTrunc`), then
+            // borrow as `&[u16]` when already 2-aligned (the common case:
+            // file reads, whole stores) and only copy into a fresh `Vec<u16>`
+            // when `try_cast_slice` rejects on alignment.
             // +1 WTF ref; `OwnedString` releases it on scope exit (Zig: `defer out.deref()`).
             //
             // ZIG PARITY: this branch intentionally does NOT `self.detach()` for
             // `Lifetime::Transfer` — `Blob.zig` `toStringWithBytes` (UTF-16LE arm)
             // returns without detaching, unlike `toJSONWithBytes`. Any change to
             // that asymmetry belongs upstream in the Zig source, not here.
             let buf = &buf[..buf.len() & !1];
-            let out =
-                OwnedString::new(BunString::clone_utf16(bytemuck::cast_slice::<u8, u16>(buf)));
+            let unaligned: Vec<u16>;
+            let utf16: &[u16] = match bytemuck::try_cast_slice(buf) {
+                Ok(s) => s,
+                Err(_) => {
+                    unaligned = buf
+                        .as_chunks::<2>()
+                        .0
+                        .iter()
+                        .map(|c| u16::from_le_bytes(*c))
+                        .collect();
+                    &unaligned
+                }
+            };
+            let out = OwnedString::new(BunString::clone_utf16(utf16));
             return out.to_js(global);
         }
 
@@ -2920,12 +2937,29 @@ impl BlobExt for Blob {
         }
 
         if bom == Some(strings::BOM::Utf16Le) {
-            // Mirrors Zig `bun.reinterpretSlice(u16, buf)`: drop a trailing odd byte
-            // (`@divTrunc`) so `bytemuck::cast_slice` cannot panic on slop.
+            // Mirrors Zig `bun.reinterpretSlice(u16, buf)`: `buf` may be
+            // odd-length (truncated input) or odd-address (`.slice(odd)` of a
+            // shared store) — either makes `bytemuck::cast_slice` `panic!`
+            // (uncatchable). Drop the trailing odd byte (`@divTrunc`), then
+            // borrow as `&[u16]` when already 2-aligned (the common case:
+            // file reads, whole stores) and only copy into a fresh `Vec<u16>`
+            // when `try_cast_slice` rejects on alignment.
             // +1 WTF ref; `OwnedString` releases it on scope exit (Zig: `defer out.deref()`).
             let buf = &buf[..buf.len() & !1];
-            let mut out =
-                OwnedString::new(BunString::clone_utf16(bytemuck::cast_slice::<u8, u16>(buf)));
+            let unaligned: Vec<u16>;
+            let utf16: &[u16] = match bytemuck::try_cast_slice(buf) {
+                Ok(s) => s,
+                Err(_) => {
+                    unaligned = buf
+                        .as_chunks::<2>()
+                        .0
+                        .iter()
+                        .map(|c| u16::from_le_bytes(*c))
+                        .collect();
+                    &unaligned
+                }
+            };
+            let mut out = OwnedString::new(BunString::clone_utf16(utf16));
             // PORT NOTE: Zig used `defer { free; detach }`. Reshaped to compute the
             // result first, then perform the deferred work explicitly — capturing
             // `&mut self` in a scopeguard closure conflicts with later uses below.

test/js/web/fetch/blob.test.ts

@@ -1,5 +1,5 @@
 import { expect, test } from "bun:test";
-import { bunEnv, bunExe, isASAN, tempDir } from "harness";
+import { bunEnv, bunExe, isASAN, isDebug, tempDir } from "harness";
 import type { BlobOptions } from "node:buffer";
 import type { BinaryLike } from "node:crypto";
 import path from "node:path";
@@ -350,7 +350,9 @@ test("Bun.file(path, {type}).text() does not leak the duped content_type", async
   const [stdout, stderr, exitCode] = await Promise.all([proc.stdout.text(), proc.stderr.text(), proc.exited]);
   expect(stderr).toBe("");
   const { deltaMiB } = JSON.parse(stdout);
-  expect(deltaMiB).toBeLessThan(isASAN ? 400 : 40);
+  // Debug builds (bun-debug is ASAN + debug allocator) inflate RSS the same
+  // way the named bun-asan CI binary does.
+  expect(deltaMiB).toBeLessThan(isASAN || isDebug ? 400 : 40);
   expect(exitCode).toBe(0);
 });
 
@@ -471,3 +473,33 @@ test("Blob constructor copies typed array parts before later parts run user code
   expect(stdout.trim()).toBe("OK 68");
   expect(exitCode).toBe(0);
 });
+
+test("Blob.json()/.text() on odd-length/odd-aligned UTF-16LE+BOM does not abort", async () => {
+  // Odd length: stripping the 2-byte BOM keeps the length odd. Odd address:
+  // `.slice(1)` of a shared byte store hands an odd pointer straight to the
+  // u8->u16 cast. Either used to `panic!` and abort the whole process
+  // (uncatchable). Run in a subprocess: pre-fix it exits 133 with no output;
+  // fixed it drops the trailing odd byte and reads unaligned like Zig.
+  const src = `
+    const oddJson = Buffer.concat([Buffer.from([0xFF, 0xFE]), Buffer.from(JSON.stringify({ a: 1 }), "utf16le"), Buffer.from([0x20])]);
+    const oddText = Buffer.concat([Buffer.from([0xFF, 0xFE]), Buffer.from("hi", "utf16le"), Buffer.from([0x20])]);
+    const evenJson = Buffer.concat([Buffer.from([0xFF, 0xFE]), Buffer.from(JSON.stringify({ a: 1 }), "utf16le")]);
+    // Odd address: pad 1 byte then .slice(1) so the view starts at base+1.
+    const misaligned = Buffer.concat([Buffer.from([0x00, 0xFF, 0xFE]), Buffer.from("hi", "utf16le")]);
+    const j = await new Blob([oddJson]).json();
+    const t = await new Blob([oddText]).text();
+    const e = await new Blob([evenJson]).json();
+    const m = await new Blob([misaligned]).slice(1).text();
+    process.stdout.write(JSON.stringify(j) + "|" + JSON.stringify(t) + "|" + JSON.stringify(e) + "|" + JSON.stringify(m));
+  `;
+  await using proc = Bun.spawn({
+    cmd: [bunExe(), "-e", src],
+    env: bunEnv,
+    stdout: "pipe",
+    stderr: "pipe",
+  });
+  const [stdout, stderr, exitCode] = await Promise.all([proc.stdout.text(), proc.stderr.text(), proc.exited]);
+  expect(stderr).toBe("");
+  expect(stdout).toBe(`{"a":1}|"hi"|{"a":1}|"hi"`);
+  expect(exitCode).toBe(0);
+});