Description

This aim to fix a particular type of circular dependency that causes tricky deadlocks, deadlocks with forked imports stack

A -> B means: B is imported by A and B has A in its stack
A ... B means: A is waiting for B to ssrLoadModule()

H -> X ... Y
H -> X -> Y ... B
H -> A ... B
H -> A -> B ... X

Forked deadlock description:

[X] is waiting for [Y] to resolve
 ↑                  ↳ is waiting for [A] to resolve
 │                                    ↳ is waiting for [B] to resolve
 │                                                      ↳ is waiting for [X] to resolve
 └────────────────────────────────────────────────────────────────────────┘

This may seems a traditional deadlock, but the thing that makes this special is the import stack of each module:

[X] stack:
	[H]

[Y] stack:
	[X]
	[H]

[A] stack:
	[H]

[B] stack:
	[A]
	[H]

Even if [X] is imported by [B], [B] is not in [X]'s stack because it's imported by [H] in first place then it's stack is only composed by [H]. [H] forks the imports stack and this make hard to be found.

Fix description

Vite, when imports [X], should check whether [X] is already pending and if it is, it must check that, when it was imported in first place, the stack of [X] doesn't have any module in common with the current module; in this case [B] has the module [H] is common with [X] and i can assume that a deadlock is going to happen.

What is the purpose of this pull request?

• Bug fix
• New Feature
• Documentation update
• Other

Before submitting the PR, please make sure you do the following

• Read the Contributing Guidelines.
• Read the Pull Request Guidelines and follow the Commit Convention.
• Check that there isn't already a PR that solves the problem the same way to avoid creating a duplicate.
• Provide a description in this PR that addresses what the PR is solving, or reference the issue that it solves (e.g. fixes #123).
• Ideally, include relevant tests that fail without this PR but pass with it.