sqlite: bind undefined to NULL - #65709

Omitting a named parameter binds NULL, but passing `undefined` for that
same parameter threw ERR_INVALID_ARG_TYPE. Bind `undefined` to NULL so
the two forms agree. This matches the conversion already applied to a
user-defined function's `undefined` return value, as well as SQLite's
own WASM oo1 API.

Fixes: #61824
Refs: #61472
Refs: #62008
Co-authored-by: mike-git374 <217764531+mike-git374@users.noreply.github.com>
Assisted-by: claude:opus-5
Signed-off-by: Trevor Burnham <trevorburnham@gmail.com>
PR-URL: #65709
Reviewed-By: Trivikram Kamat <trivikr.dev@gmail.com>

Passing undefined for a parameter threw while omitting the same named
parameter bound NULL, so the two forms disagreed. This also forced
callers using query builders like Knex, which emit undefined for columns
missing from a multi-row insert, to sanitize bindings by hand.

Ports nodejs/node#65709.

Notable changes:

crypto:
  * (SEMVER-MINOR) add crypto.parsePKCS12() (Brian Muenzenmeyer) #65627
doc:
  * add araujogui to collaborators (Guilherme Araújo) #66090
ffi:
  * (SEMVER-MINOR) load libraries from a mounted VFS (Matteo Collina) #65909
fs:
  * (SEMVER-MINOR) add openAsBlobSync (greenhead) #65644
net:
  * (SEMVER-MINOR) support sending net.BoundSocket to threads and child processes (Guy Bedford) #64725
perf_hooks:
  * (SEMVER-MINOR) implement SlidingWindowHistogram (James M Snell) #65825
  * (SEMVER-MINOR) implement qrde analysis support in Histogram (James M Snell) #65806
sqlite:
  * (SEMVER-MINOR) bind undefined to NULL (Trevor Burnham) #65709
src,lib:
  * (SEMVER-MINOR) add util.markPromiseAsHandled (James M Snell) #65805
test:
  * (SEMVER-MINOR) expand histogram test coverage (James M Snell) #65825
util:
  * (SEMVER-MINOR) implement util.throttle (James M Snell) #65899
  * (SEMVER-MINOR) implement debounce (James M Snell) #65899

PR-URL: #66163

Notable changes:

crypto:
  * (SEMVER-MINOR) add crypto.parsePKCS12() (Brian Muenzenmeyer) #65627
doc:
  * add araujogui to collaborators (Guilherme Araújo) #66090
ffi:
  * (SEMVER-MINOR) load libraries from a mounted VFS (Matteo Collina) #65909
fs:
  * (SEMVER-MINOR) add openAsBlobSync (greenhead) #65644
net:
  * (SEMVER-MINOR) support sending net.BoundSocket to threads and child processes (Guy Bedford) #64725
perf_hooks:
  * (SEMVER-MINOR) implement SlidingWindowHistogram (James M Snell) #65825
  * (SEMVER-MINOR) implement qrde analysis support in Histogram (James M Snell) #65806
sqlite:
  * (SEMVER-MINOR) bind undefined to NULL (Trevor Burnham) #65709
src,lib:
  * (SEMVER-MINOR) add util.markPromiseAsHandled (James M Snell) #65805
test:
  * (SEMVER-MINOR) expand histogram test coverage (James M Snell) #65825
util:
  * (SEMVER-MINOR) implement util.throttle (James M Snell) #65899
  * (SEMVER-MINOR) implement debounce (James M Snell) #65899

PR-URL: #66163

* fix(sqlite): undefined binds NULL, like Node 26.10

node:sqlite refused an explicitly-passed `undefined` until v26.10.0, where
nodejs/node#65709 made it bind NULL so that passing a parameter as `undefined`
agrees with omitting it — which bound NULL already, on both implementations.
The release documented the rule in doc/api/sqlite.md and dropped `undefined`
from the "unsupported data types" list this spec is ported from.

`@gjsify/sqlite` carried exactly the asymmetry Node removed: validateBindValue()
returned early for `null` alone, while sqlLiteral() below it already rendered
`undefined` as NULL and an omitted parameter already bound NULL. It now accepts
`undefined` on both legs, unconditionally.

`undefined` leaves the unsupported list but is not deleted: it moves into a test
of its own that asserts the stored value reads back `null` and that an explicit
`undefined` and an absent argument land on the same value. The Node leg runs
native node:sqlite (runtimes.node is "none") and the hosts disagree across the
26.10 line, so the boundary is named rather than left silent.

* test(sqlite): which sqlite, not which host

The `undefined` gate asked `process.versions`, so it answered "which host am I
on" — but `test:gjs-on-node` builds these same specs with
`--alias node:sqlite=@gjsify/sqlite` and runs OUR implementation on a Node host.
Host and implementation come apart there, and only the implementation decides
what `stmt.run(1, undefined)` does. Measured, one probe per leg:

  node-gi (aliased)  host_builtin=false  node=24.19.0  gjs=undefined
  native node        host_builtin=true   node=24.19.0  gjs=undefined
  gjs                host_builtin=false  node=20.0.0   gjs=1.88.1

Row 1 is the red: Node 24 reads as "refuses", so the test demanded a throw from
an implementation that binds NULL — `Expected [anonymous function] to throw an
exception`, on "sqlite suite as node-gi consumer" and "node-gi consumer
harness". Row 3 is the trap the old gate already knew about and patched with an
`IS_GJS` check; rows 1 and 3 are the same defect, and Bun and Deno would be two
more.

So the gate asks the module instead. `[native code]` is what ECMA-262 renders
for a function with no ECMAScript source: true means the binding came from the
host, false means it came from a bundle of our TypeScript. The version is read
only after that answer, and `IS_GJS` is gone because it is subsumed. It cannot
mask a regression in the direction that matters — were `@gjsify/sqlite` to go
back to refusing `undefined`, the probe stays false and the test still demands
NULL. Verified by breaking `validateBindValue` back to `value === null`: gjs
89/90 and node-gi 64/65, both on `InvalidArgTypeError: Provided value cannot be
bound to SQLite parameter 2.`

Green after: gjs 90 pass / 178 assert, node 24.19.0 90 / 176,
node-gi on node 24.19.0 65 / 135.