…scripts (#33779)

## Summary

Follow-up to #33760. Converts the ESM JS sources in `ext/io`, `ext/os`,
and `ext/net` to IIFE-wrapped lazy-loaded scripts (`lazy_loaded_js`),
eliminating ESM module resolution overhead.

These three extensions have no cross-extension ESM import dependencies
(only `core.loadExtScript` calls to already-converted extensions like
`ext/web`), so no `internals` bridges are needed.

Converting these also unblocks clean conversion of:
- `ext/fs` (depends on `ext/io`)
- `ext/fetch` (depends on `ext/net`)

Adapt enqueuePerformanceEntry/registerExtraEntryType to lazy-loaded
script pattern from denoland#33760: 15_performance.js now returns its exports
from an IIFE; consumers (perf_hooks.js, http2.ts) load via
core.loadExtScript instead of static ESM import.

Co-authored-by: Divy Srivastava <[email protected]>

Follow-up to #33760. Converts `ext/ffi/00_ffi.js` from ESM to an
IIFE-wrapped lazy-loaded script (`lazy_loaded_js`).

Signed-off-by: Bartek Iwańczuk <[email protected]>

Follow-up to #33760 and #33779. Converts all 8 ESM JS sources in
`ext/fetch` to IIFE-wrapped lazy-loaded scripts (`lazy_loaded_js`).

- `20_headers.js`, `21_formdata.js`, `22_body.js`, `22_http_client.js`,
`23_request.js`, `23_response.js`, `26_fetch.js`, `27_eventsource.js`
- `26_fetch.js` imports from `ext/telemetry` (TypeScript ESM, can't be
converted yet) -- bridged via `internals.__telemetry` /
`internals.__telemetryUtil` set in the telemetry source files
- All other cross-extension deps use `core.loadExtScript()` (ext/web,
ext/webidl, ext/net are already `lazy_loaded_js`)

This also eliminates the `internals.__fetchRequest/Response/Headers`
bridges from #33778 since `ext/cache` can now use `core.loadExtScript()`
for ext/fetch directly.

…zy-loaded scripts (#33778)

Follow-up to #33760. Converts the ESM JS sources in `ext/cache`,
`ext/canvas`, and `ext/crypto` to IIFE-wrapped lazy-loaded scripts
(`lazy_loaded_js`), eliminating ESM module resolution overhead.

- **ext/canvas/02_surface.js** - simple re-export from `core.ops`
- **ext/cache/01_cache.js** - ext/fetch dependencies (Request,
toInnerRequest, toInnerResponse, getHeader) bridged via `internals`
since ext/fetch is still ESM
- **ext/crypto/00_crypto.js** - `kKeyObject` symbol bridged via
`internals` from `ext/node`; updated `ext/node` consumers to use
`core.loadExtScript()`

…34877)

## Summary

A `node:worker_threads` Worker that holds a "refed" transferable object
such as a `MessagePort` must not be terminated on idle. This covers the
use cases from #23169:

- a worker pool using a dedicated `MessagePort` for IPC, with a minimum
number of workers always kept alive, and
- a worker that is set up, idle, and waiting for work.

## Root cause

The runtime idle-termination check `hasMessageEventListener()` in
`runtime/js/99_main.js` consults `messagePort.refedMessagePortsCount` to
decide whether a node worker should stay alive:

```js
return (event.listenerCount(globalThis, "message") > 0 &&
  !globalThis[messagePort.unrefParentPort]) ||
  messagePort.refedMessagePortsCount > 0;
```

That counter is a mutable module-level `let` in
`ext/web/13_message_port.js`. When ext/web was converted to lazy-loaded
IIFE scripts in #33760, the module's real ESM `export {
refedMessagePortsCount }` (a **live binding**) became a plain property
on the returned object literal, capturing a one-time snapshot of `0`.
The refed-port branch of the idle check was therefore dead, and a worker
kept alive only by a refed `MessagePort` (with no active "message"
listener op promise) was incorrectly terminated on idle.

## Fix

Restore the live-binding semantics by exposing `refedMessagePortsCount`
as a getter on the module's return object.

## Tests

Adds two spec tests under `tests/specs/node/worker_threads/`:

- `refed_messageport_idle` — a worker receives delayed work over a refed
transferred `MessagePort` after an idle period.
- `refed_messageport_holder` — a worker kept alive purely by a refed
`MessagePort` it does **not** listen on. This isolates the fix: it fails
on `main` (worker terminates on idle) and passes with the change.

Closes #23169

Closes denoland/divybot#429

Co-authored-by: divybot <[email protected]>
Co-authored-by: Divy Srivastava <[email protected]>

…d_js (#34997)

related: #33760, #34061

…33780)

Follow-up to denoland#33760. Converts `ext/ffi/00_ffi.js` from ESM to an
IIFE-wrapped lazy-loaded script (`lazy_loaded_js`).

Signed-off-by: Bartek Iwańczuk <[email protected]>

…and#33784)

Follow-up to denoland#33760 and denoland#33779. Converts all 8 ESM JS sources in
`ext/fetch` to IIFE-wrapped lazy-loaded scripts (`lazy_loaded_js`).

- `20_headers.js`, `21_formdata.js`, `22_body.js`, `22_http_client.js`,
`23_request.js`, `23_response.js`, `26_fetch.js`, `27_eventsource.js`
- `26_fetch.js` imports from `ext/telemetry` (TypeScript ESM, can't be
converted yet) -- bridged via `internals.__telemetry` /
`internals.__telemetryUtil` set in the telemetry source files
- All other cross-extension deps use `core.loadExtScript()` (ext/web,
ext/webidl, ext/net are already `lazy_loaded_js`)

This also eliminates the `internals.__fetchRequest/Response/Headers`
bridges from denoland#33778 since `ext/cache` can now use `core.loadExtScript()`
for ext/fetch directly.

…zy-loaded scripts (denoland#33778)

Follow-up to denoland#33760. Converts the ESM JS sources in `ext/cache`,
`ext/canvas`, and `ext/crypto` to IIFE-wrapped lazy-loaded scripts
(`lazy_loaded_js`), eliminating ESM module resolution overhead.

- **ext/canvas/02_surface.js** - simple re-export from `core.ops`
- **ext/cache/01_cache.js** - ext/fetch dependencies (Request,
toInnerRequest, toInnerResponse, getHeader) bridged via `internals`
since ext/fetch is still ESM
- **ext/crypto/00_crypto.js** - `kKeyObject` symbol bridged via
`internals` from `ext/node`; updated `ext/node` consumers to use
`core.loadExtScript()`