Jacob Finkelman
Jacob Finkelman
When were done with all the significant changes to our API, we should look at whether it would be more appropriate to past `&mut self` where we currently use `&self`....
Is there a way to avoid the clones and allocation in https://github.com/mpizenberg/elm-test-rs/blob/31596f874a6e9340350de3e7c783a5cde9ca396b/src/solver.rs#L439? In Cargo we use `Rc` https://github.com/rust-lang/cargo/blob/07162dbaa84aaad97d6f3be5069265e784179bcd/src/cargo/core/resolver/dep_cache.rs#L35, so that a cache hit is and Rc clone instead of a...
Our production user is hitting slow cases in ways that are hard for us to optimize separately. We should have clear instructions for how to reproduce the slow cases they...
Our implementation of range just gained a new `simplify` method (cc #156). It would not be too hard to make this a method on DependencyProvider and so available to the...
I was looking at the error message created by the test case for #135. The case I was looking at was: ```rust let mut dependency_provider = OfflineDependencyProvider::::new(); // root depends...
Most breaking changes in our API involve changes to the dependency provider. We can cut new breaking changes if we need to, but we would all prefer to avoid it....
PubGrub has good error messages, in fact that's one of its main selling points. When resolution fails you get a rich object with all the information to explain why that...
Mostly a follow-up to this comment: https://github.com/pubgrub-rs/pubgrub/pull/108#issuecomment-1038437430 But we should provide some way for a user to run "all the tests we can think of" on their types. Not only...
There is a really cool part of the algorithm that makes things more efficient and makes for better error messages. Unfortunately it is not clear when it is correct to...