You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Right now there isn't an API contract between the two extensions, so it's difficult to share code.
We want to do this in a way that minimizes long term changes to Cursorless extension (since we want to eliminate the fork over time), or add stuff to its public API (i.e., the extension points).
After talking with pokey, we decided on:
Having the Cursorless extension build a secret dependency on the sidecar (e.g. it would look for the phillco.cursorless-sidecar extension if it existed), but only within the ide abstraction, and only if a secret setting is turned on, so it wouldn't be exposed to regular users.
Most likely, we would have the sidecar extension give the Cursorless extension its own IDE object, which Cursorless would swap in for the regular VscodeIDE; that way we have maximum flexibility to inject code as needed, without having to first ship a bunch of abstractions in the Cursorless extension. It also encourages migrating stuff to the IDE abstraction.
The advantages of this approach are that we don't have to add anything into the Cursorless public API (its extension points), and we can minimize the changes in the Cursorless fork, since it will be mostly scoped to what's in the IDE abstraction.
The text was updated successfully, but these errors were encountered:
Right now there isn't an API contract between the two extensions, so it's difficult to share code.
We want to do this in a way that minimizes long term changes to Cursorless extension (since we want to eliminate the fork over time), or add stuff to its public API (i.e., the extension points).
After talking with pokey, we decided on:
phillco.cursorless-sidecar
extension if it existed), but only within theide
abstraction, and only if a secret setting is turned on, so it wouldn't be exposed to regular users.IDE
object, which Cursorless would swap in for the regularVscodeIDE
; that way we have maximum flexibility to inject code as needed, without having to first ship a bunch of abstractions in the Cursorless extension. It also encourages migrating stuff to theIDE
abstraction.The advantages of this approach are that we don't have to add anything into the Cursorless public API (its extension points), and we can minimize the changes in the Cursorless fork, since it will be mostly scoped to what's in the
IDE
abstraction.The text was updated successfully, but these errors were encountered: