child_process: build the default env block in one native pass - #65325

When spawn()/spawnSync() are called without options.env,
normalizeSpawnArguments() copied process.env with a spread and then
walked the copy to build the KEY=value array uv_spawn() takes.
Spreading the process.env proxy costs one enumerator callback plus a
query and a getter interceptor per variable, each doing a linear
getenv() scan and allocating; with a couple of hundred variables that
was the single largest JS-side cost of spawning a process.

Add KVStore::Pairs() (Enumerate() + Get() by default, one
uv_os_environ() pass for the real environment, skipping hidden
variables on Windows exactly like Enumerate() does), expose it as
process_wrap.getEnvPairs(), and use it for the default-environment
case. A user supplied options.env and the permission model case keep
the existing code. On Windows the same sort/first-wins-case-insensitive
filter is applied to the pairs. The variables copyProcessEnvToEnv()
propagates are part of the real environment by definition, and its
entries cannot contain null bytes, so those steps only remain on the
options.env path.

Signed-off-by: Shelley Vohr <[email protected]>
PR-URL: #65325
Reviewed-By: Antoine du Hamel <[email protected]>
Reviewed-By: James M Snell <[email protected]>

When spawn()/spawnSync() are called without options.env,
normalizeSpawnArguments() copied process.env with a spread and then
walked the copy to build the KEY=value array uv_spawn() takes.
Spreading the process.env proxy costs one enumerator callback plus a
query and a getter interceptor per variable, each doing a linear
getenv() scan and allocating; with a couple of hundred variables that
was the single largest JS-side cost of spawning a process.

Add KVStore::Pairs() (Enumerate() + Get() by default, one
uv_os_environ() pass for the real environment, skipping hidden
variables on Windows exactly like Enumerate() does), expose it as
process_wrap.getEnvPairs(), and use it for the default-environment
case. A user supplied options.env and the permission model case keep
the existing code. On Windows the same sort/first-wins-case-insensitive
filter is applied to the pairs. The variables copyProcessEnvToEnv()
propagates are part of the real environment by definition, and its
entries cannot contain null bytes, so those steps only remain on the
options.env path.

Signed-off-by: Shelley Vohr <[email protected]>
PR-URL: #65325
Reviewed-By: Antoine du Hamel <[email protected]>
Reviewed-By: James M Snell <[email protected]>

Measured on the daily `bumpii digest --brew-upgrade` with nothing pending:
7.2 s, of which the forge fetch phase took 4.7 s because gitea.com sends 30
release bodies uncompressed (343 KB in 3.6 s; 10 in 1.9 s) and the plain and
greedy `brew outdated` listings ran one after the other (0.65 s each).

- The digest asks for FIRST_PAGE (10) releases first and fetches the 30 only
  when the installed release itself is not on the short page (`pageReaches`).
  Not "a release below it": forges list by date, and review measured a
  multi-line project undercounting 4 for 9 and an installed 3.0.0 reading
  "ahead of 2.9.20" with exit 0 under the weaker test. What the short page
  still cannot see is stated on the docblock as an accepted narrowing.
- Both `brew outdated` listings start together in every digest; the greedy
  one is read only after the plain one answered, and drained on failure.
- Three runs afterwards: 5.9 / 3.9 / 4.8 s, fetch phase 1.8–2.9 s. The JSON
  document and every count are unchanged for the live tools.json.

Also: since Node 26.11 (nodejs/node#65325) a child spawned without `env`
gets the OS environment, so tests that reassigned `process.env` never ran
their fake brew — eleven overview scenarios went red after this machine's
brew upgrade pulled 26.11.1. They mutate the object in place now; trap in
AGENTS.md.