Skip to content

Don't start an engine for dependencies #542

@doorgan

Description

@doorgan

Currently using Go to definition on a dependency module or function causes the expert to start an engine for that dependency. This is unnecessary because the engine for the project already handles the indexing of dependency code

Questions for monorepos:

  • Should this be the case for dependencies that are also a project in a monorepo or umbrella?

Let's say there's projects A and B in a monorepo, and A depends on B. If opening a file in B should start an engine, but then:

  • What engine should handle the "Find references" request? If B, uses of A won't be reported. If A then those would be reported, but the more "correct" behavior would be for B to handle it instead?

  • If we decide to not start an engine for B, only for A, then this becomes tricker or wasteful depending on the order files are opened. If you open a file in B an engine gets started, then you open a file in A and another engine is started, both know about B, so we need to select one again?

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions