Feature/ordered float #399
Conversation
…trait No tests have been run on this yet
This includes adding a couple of unsafe blocks that, for _some_ reason are not required (or, at least not FLAGGED) when my new feature is activated... I'll look into that later. For now, though, stuff appears to build correctly
|
This is a draft PR at the moment. However, at the time of this comment all existing tests pass with and without the new feature, and all my clippy lints have been addressed. |
|
value-trait is updated to .10.1 with your ordered-float |
|
can we add the feature to the test matrix so we have the ordered float path covered? |
All tests pass with and without - as an intellectual exercise I'm looking at LazyValue too :)
Your wish is my command (which I was actually doing when you posted that :) ) |
|
great minds think alike :) |
|
The goal of solving the Eq issue in my main project has been achieved - still a couple of issues that will likely be best dealt with in that project rather than hear (the necessary unsafe and mutability differences and a couple of missing methods only used in tests) Once I get those tweaked away, I'll see if my full "droppin" goal is achieved. |
|
Awesome, I think this needs a cargo fmt to make the CI happy; otherwise good to merge :) if I can help with some thoughts on other problems feel free to ping |
This PR addresses issue #398 by implementing a feature named
ordered-floatthat activates the same-named feature invalue-traitwhich itself activates the wrapping of floats withOrderedFloatand thus allowsStaticValueto bestd::cmp::Eq. Various feature-related "tweaks" are also implemented, while trying to use as few#cfg[feature...]annotations as possible.Obvious the PR is dependent upon the
value-traitPR, but while that is pendingCargo.tomlin this PR's source branch patches thevalue-traitcrate source to be the source repo and branch of thevalue-traitPR.