- conversion: implement additional traits for converting integer types to Version. (d1b6021)
- namespace: add prefix to std to avoid namespace conflict (02a866b)
- range: boxing
BoundSet
(#8)
- deps: bump deps
bump MSRV to 1.70.0
(#7)
- range: add
min_version
(14ab733)
- winnow: migrate to
winnow
(#3)
- serde: add optional feature:
serde
(3842e9af)
- examples: add examples (648b45c)
- deps: bump deps
- diff: add
diff
forVersion
(1dae57fa)
- range: make
<=11
work the same as in npm (#1) (6438d9f4)
remove unnecessary serde
(18f26b43)
- vector: initialize vector with capacity(f454c8e1)
- crate: Project forked.
- syscall: Reduced unnecessary cloning operations(598e3554)
- format: Include the build and prerelease when stringifying to maintain consistency (#9) (f2b2e44c)
- satisfies: Fix
.satisfies
bug for higher major/minor/path pre-release versions (#8) (ee8376e7) - range: handle partial
=
ranges, which was causing panics (#7) (f0eef040)
This is an almost full rewrite of the Range parser to make it work much more
closely to how the JS node-semver
parser works. Not by using regex,
fortunately.
As such, this is potentially a pretty breaking change, but it's a breaking change in the direction of compatibility.
Please file issues for any compatibility issues you find and we'll fix them asap!
- loose: rewrite to support loose mode better (#5) (20fb02d8)
- BREAKING CHANGE: This accepts (and rejects) some semver strings that were valid before, and I'm not comfortable just calling thos e bugs. It also vastly reduces the number of "bad" semver parses by outright throwing out bad-looking data without warning you. This is literally what the JavaScript node-semver does. And so...
- api: stop exporting anything but Range from range mod (4eeb862d)