[v24.x] deps: cherry-pick 5ba9200 from V8 upstream

[juanarbol]

Original commit message:
Fix crash when pausing in for-of loop header

Given a for-of loop:

for (const each of subject) {

The bytecode generator emits the iterator.next call + done check +
assigning to each all into the source position of const each.

The pseudo-desugared code looks something like:

var tmp;
loop {
  var result = iterator.next();
  if (result.done) break;
  tmp = result.value;

  PushBlockContext;
  const each = tmp;

  // rest of the loop.
}

This is a problem, as the parser starts the block scope already on
the const each. If the scope requires a context we can pause on
bytecode that has or has not pushed the block context yet, while
the source position looks the same.

The recent addition of per-script unique scope IDs lets us fix
this problem in the debugger: We can check if the scope ID of
the runtime scope matches the parser scope. If not, the context
was not pushed yet.

The debugger already has a HasContext helper. We extend it to
also check for matching scope IDs and then use HasContext where
we would read variable values off the context. If the context was
not pushed yet, we report them as 'unavailable'.

R=[email protected]

Fixed: 384413079,399002824
Change-Id: Ia2d0008d574e7eaf6c06b640053df696014d37f8
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/6507402
Reviewed-by: Leszek Swirski [email protected]
Commit-Queue: Simon Zünd [email protected]
Cr-Commit-Position: refs/heads/main@{#100029}

Refs: v8/v8@5ba9200
Fixes: #60580
PR-URL: #60620

[nodejs-github-bot]

Review requested:

• @nodejs/gyp
• @nodejs/security-wg
• @nodejs/v8-update

[targos]

This doesn't look like a cherry-pick of v8/v8@5ba9200, which should already be on main anyway since it landed in V8 13.8

[juanarbol]

This doesn't look like a cherry-pick of v8/v8@5ba9200, which should already be on main anyway since it landed in V8 13.8

Whoops, that's right, wrong base branch.

[targos]

Still not a cherry-pick. I highly recommend you use @node-core/utils to prepare it: https://github.com/nodejs/node-core-utils/blob/main/docs/git-node.md#git-node-v8-backport-sha

[juanarbol]

Still not a cherry-pick. I highly recommend you use @node-core/utils to prepare it: https://github.com/nodejs/node-core-utils/blob/main/docs/git-node.md#git-node-v8-backport-sha

yes yes, I'm just a mess with git! Thanks!

Also, is it possible to send my "fork" of v8 to node core utils? I don't want to fetch that again, I've my own tree.

Seems ready now, sorry for the noise.

[targos]

Also, is it possible to send my "fork" of v8 to node core utils?

Yes, with --v8-dir=path/to/your/fork

[juanarbol]

Yes, with --v8-dir=path/to/your/fork

Thanks! PTAL :)

[aduh95]

$ curl -L https://github.com/v8/v8/commit/5ba9200cd04602c486ff5a5fc4f9e0d4ab19606b.patch | head -3
From 5ba9200cd04602c486ff5a5fc4f9e0d4ab19606b Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Simon=20Z=C3=BCnd?= <[email protected]>
Date: Fri, 2 May 2025 10:44:42 +0000

$ curl -L https://github.com/nodejs/node/pull/60620/commits/d8e7ab5d77e85ff0257314acfd43546ae1a4ff2f.patch | head -3
From d8e7ab5d77e85ff0257314acfd43546ae1a4ff2f Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Juan=20Jos=C3=A9=20Arboleda?= <[email protected]>
Date: Fri, 7 Nov 2025 14:26:58 -0500

Is it expected that the date and author do not match?

[juanarbol]

Is it expected that the date and author do not match?

Yes, it is expected

node/doc/contributing/maintaining/maintaining-V8.md

Line 200 in 622c372

replace yourself as the author. `git commit --amend --reset-author`. You may