Skip to content

Conversation

ahl
Copy link
Contributor

@ahl ahl commented Sep 27, 2024

This is kind of gross.

Omicron currently is out of sync with maghemite, see #6693 and the two maghemite pushes that require synchronization: oxidecomputer/maghemite#359 and oxidecomputer/maghemite#360. I'd like to make forward progress with the hyper v1 migration, and that's blocked on updating the omicron dependency of maghemite which pulls in old hyper (reqwest, progenitor, etc). This gets pull into other repos via omicron-common.

It's worth noting that this circular arrangement seems lousy. The "vassal crates" (crucible, propolis, maghemite, (and dendrite to a lesser degree)) depend on omicron-common, but omicron-common pulls in maghemite's mg-admin-client which in turn pulls in... lots... including progenitor, hyper, reqwest, etc.

My goal here is to have a temporary dependency on a branch of maghemite (oxidecomputer/maghemite#378); once #6693 is integrated, we can then pin the dependency on maghemite's HEAD rev.

Comment on lines +425 to +428
# mg-admin-client = { git = "https://github.com/oxidecomputer/maghemite", rev = "9e0fe45ca3862176dc31ad8cc83f605f8a7e1a42" }
# ddm-admin-client = { git = "https://github.com/oxidecomputer/maghemite", rev = "9e0fe45ca3862176dc31ad8cc83f605f8a7e1a42" }
mg-admin-client = { git = "https://github.com/oxidecomputer/maghemite", branch = "hyper-v1-no-merge" }
ddm-admin-client = { git = "https://github.com/oxidecomputer/maghemite", branch = "hyper-v1-no-merge" }
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The explanation for what we're doing here makes sense (there's a circular dependency between maghemite and omicron via omicron-common, and the maghemite main branch is broken with respect to omicron APIs) but this is also some hackery of a pretty high order.

I think it makes sense -- it moves maghemite forward as much as we can, although it is relying on a random ahl-owned branch, rather than something merged -- but it also merits some pretty clear explanation, outside of digging through git history.

Could we file an issue describing what's happening here, and how to get ourselves out of this situation?

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@ahl ahl merged commit 75214e1 into main Sep 27, 2024
18 checks passed
@ahl ahl deleted the hyper-v1-mg branch September 27, 2024 16:48
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants