I want I understand this one better. We spend a lot of effort doing the opposite and now you are undoing it with no real explanation…

We spend a lot of effort doing the opposite

How so?

Either way, I don't see a strong reason for the use crate::vello/kurbo/parley pattern, and it's applied very inconsistently throughout the project. If it's supposed to be the desired coding style, then it's not documented anywhere.

To be clear, I'm not removing any re-rexports, just changing internal import patterns.

I am on my phone not the computer. Maybe do a got blame in some of the lines invoking vello and Kurbo.

After doing some git blames, I see:

• Reorganize Masonry modules #848: Replaced a lot of uses of crate::{Point, Size, etc} with crate::kurbo::{Point, Size, etc}.
• masonry: Use Kurbo via Vello. #223: Replaced kurbo::{Point, Size, etc} with either crate::kurbo::{Point, Size, etc} or vello::kurbo::{Point, Size, etc} depending on files. The rationale was that we wanted to stop importing kurbo directly, which doesn't really apply here.

I think neither are "incompatible" with this PR, or give a direction that this PR goes against.

Overall, I think we just haven't officially decided on a standard between crate::kurbo::{xxx} and vello::kurbo::{xxx}, and we've used both throughout the codebase. I strongly prefer the latter.

I guess I don’t actually care enough to argue about it. It is all bad anyway.

We can always agree on an import format after the crate split.