mirror of
https://github.com/vercel-labs/scriptc.git
synced 2026-10-26 06:34:20 +08:00
Chris Tate
9f2e9b00c9
Each limitations partial held its whole text on one line (map-and-set.mdx was a single 1,080-character line), so any two edits to the same partial conflicted. The partials now put one sentence per line, with a bullet's continuation lines indented two spaces. Markdown joins the lines, so the rendered pages are unchanged. scripts/limitations-partials.mjs reads a labeled bullet's indented continuation lines as part of the bullet when it splits and joins the page, and AGENTS.md states the convention.
17 lines
1.8 KiB
Plaintext
17 lines
1.8 KiB
Plaintext
Typed-array reads at an index that is not an integer in the array's range return `undefined`, as in Node.js, and invalid numeric indexed writes are ignored, while the assignment expression retains its original value.
|
|
A non-null assertion on such a read (`bytes[i]!`) yields `NaN` instead of `undefined`.
|
|
|
|
Ordinary arrays track holes separately from present `undefined` values.
|
|
Missing reads and empty `pop()`/`shift()` return `undefined`, indexed writes can grow the array, and length growth creates holes.
|
|
|
|
Keyed reads of a `Record` or an index signature return `undefined` for a missing key, as in Node.js, with or without `noUncheckedIndexedAccess`.
|
|
A missing array element or key that reaches storage that cannot hold `undefined`, such as a class field typed `number`, throws a `TypeError`.
|
|
|
|
Runtime traps cannot be caught.
|
|
User-thrown values and supported exceptions, including JSON parse, checked-cast, filesystem, regex, and `String.prototype.repeat` count errors, can be caught; hard traps such as out-of-range reads of statically typed ordinary arrays terminate the process.
|
|
|
|
Deep recursion in Linux and macOS executables throws Node's catchable `RangeError: Maximum call stack size exceeded`, on the main thread, in worker threads, and inside async functions and generators.
|
|
The depth at which it is thrown differs from Node: the main thread uses the process stack limit (usually 8 MiB), worker threads get an 8 MiB stack, and async functions and generators run on 512 KiB stacks, so native code usually recurses deeper than Node on threads and somewhat less deeply inside async bodies.
|
|
Recursion inside runtime library operations, such as serializing deeply nested data, is not checked.
|
|
Windows executables, WASI builds and library builds do not check the stack, and exhausting it there terminates the process.
|