Closed nc7s closed 1 year ago
Could you rewrite the commit to be
refactor!: …
to indicate it's a breaking change? Thanks for taking a look at my other comment as well.
Oh I thought it's only an internal change so not breaking, but sure.
You don't seem convinced, so I wasn't convinced anymore either as I just looked at the diff before. Now I checked more thoroughly and can confirm that human.rs
is a publicly accessible module that publicly exposed types of an external crate that is now a different one.
It's not that I'm not convinced; my initial thoughts writing the change were "it's only replacing dependencies, no publicly visible changes", and my brain when replying was still booting up from sleep. You are right about human.rs
.
With #26 merged, would this PR be superseded? To me it seems like it can be closed (or needs work). Thank you.
With #26 merged, would this PR be superseded? To me it seems like it can be closed (or needs work). Thank you.
Indeed. humansize is by name only for sizes, it would be inappropriate to jury rig it here anyway. Sorry for the delay - caught in work.
Currently compound_duration and human_format are used. They don't see much use like the replacements, and the former has a 32-bit problem.
Another reason is that I, as a member of the Debian Rust team, hope to introduce less Rust crates into Debian, especially the less used ones. Packaging is more or less a manual process, unlike cargo publish which is automagic.