import-in-the-middle@1.11.1 broke Instrumentation#disable() functionality for ESM by introducing support to hook the same module multiple times (https://github.com/nodejs/import-in-the-middle/pull/153) - this makes tests fail in #5045 which bumps the lock file to that version.
This PR is to unblock updating by skipping the unwrap tests for ESM instrumentation. Going forward, we'll have to decide if the unwrapping-during-runtime is worth the complexity and maintenance overhead it introduces to every instrumentation package. Going through use-cases it seems like disabling/unwrapping is mostly used for testing - since the release of import-int-the-middle@1.11.1 we also had 0 user reports that this functionality is broken, which leads me to believe that this feature is not of significant enough value to a significant portion our end-users.
I'll open a follow-up issue to discuss this in-depth.
Short description of the changes
all changes from #5045
Skips ESM unwrap tests in @opentelemetry/instrumentation
Which problem is this PR solving?
import-in-the-middle@1.11.1
brokeInstrumentation#disable()
functionality for ESM by introducing support to hook the same module multiple times (https://github.com/nodejs/import-in-the-middle/pull/153) - this makes tests fail in #5045 which bumps the lock file to that version.This PR is to unblock updating by skipping the unwrap tests for ESM instrumentation. Going forward, we'll have to decide if the unwrapping-during-runtime is worth the complexity and maintenance overhead it introduces to every instrumentation package. Going through use-cases it seems like disabling/unwrapping is mostly used for testing - since the release of
import-int-the-middle@1.11.1
we also had 0 user reports that this functionality is broken, which leads me to believe that this feature is not of significant enough value to a significant portion our end-users.I'll open a follow-up issue to discuss this in-depth.
Short description of the changes
@opentelemetry/instrumentation