feat: add globbing support for package roots #3
Conversation
There was a problem hiding this comment.
utACK
However, this leads to a change in how we handle virtual filesystems in tests
I've used the real FS in a different repo in the past for a more integrated test (it's not actually too slow). Wondering if it would make sense to swap out eventually as this logic is quite reliant on things working as they would on the real FS and subtle differences in the memory FS implementation may lead to passing tests when it would fail on the real FS or the other way around (not to be done in this PR, but maybe to capture).
Once you release and add this to the lerna-release-action, would you mind testing a few scenarios in the test repo to see if the release tooling can cope with it?
- creating a package that is a little more deeply nested
- grouping 2 or more packages
|
Also thought about using the real FS with some fixtures, but didn't want to squeeze it in here, like you said. Captured in #4. Will definitely do some testing with lerna-version-selectively before sending out the review for lerna-release-action! |
# [1.4.0](v1.3.0...v1.4.0) (2023-07-19) ### Features * add globbing support for package roots ([#3](#3)) ([92155c1](92155c1))
|
🎉 This PR is included in version 1.4.0 🎉 The release is available on: Your semantic-release bot 📦🚀 |
This allows us to use globs for package roots in
lerna.json.Part of https://github.com/ExodusMovement/exodus-hydra/issues/1474.
It utilizes Glob to ensure wide globbing support.
However, this leads to a change in how we handle virtual filesystems in tests. Glob requires an
fs-like object to function. WithVolumethis seems not to be the case and that is whycreateFsFromVolumeexists in memfs. See vol vs fs.I added a testing utility to handle this in one place and should be kept in mind when migrating other packages. I will do that most likely myself.