You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I've been thinking about this recently as the discussion of breaking changes comes up regularly in PR reviews. I wanted to start bringing things to the table to make sure we were headed in the right direction for a 1.0 release. I think things will need to be resolved a little better with libc before we get there, but I'm expecting this to be about a year away unless things speed up in the refactoring department.
I think we should aim towards a minimal 1.0 release, so we wouldn't have any sort of goal in terms of features, but really a goal in terms of API quality & stability. At the bare minimum this is:
Upstream all constants & ffi declarations
Provide extensive documentation for all APIs.
Decide on common interface types (TimeValLike vs Duration, CString versus String, etc.).
Finalize how we want to expose constants (do we make people use libc:: types or re-export them from nix?, Do we rename things nicely or use the original names?).
Do we really want to have an RFC process or are just issues here good enough?
I think once that's all done we could do a 1.0 release.
There's also the option of trying to focus all efforts towards this and refusing PRs that don't move towards these goals (as in add features). I think we're a ways out from this, but it might be a good idea as we get closer for a final push.
The text was updated successfully, but these errors were encountered:
I'd also like to address #338 as part of this. I think the consensus is correct. Just need to modify CONVENTIONS.md and do a general audit of the codebase.
I've been thinking about this recently as the discussion of breaking changes comes up regularly in PR reviews. I wanted to start bringing things to the table to make sure we were headed in the right direction for a 1.0 release. I think things will need to be resolved a little better with
libc
before we get there, but I'm expecting this to be about a year away unless things speed up in the refactoring department.I think we should aim towards a minimal 1.0 release, so we wouldn't have any sort of goal in terms of features, but really a goal in terms of API quality & stability. At the bare minimum this is:
TimeValLike
vsDuration
,CString
versusString
, etc.).libc::
types or re-export them fromnix
?, Do we rename things nicely or use the original names?).I think once that's all done we could do a 1.0 release.
There's also the option of trying to focus all efforts towards this and refusing PRs that don't move towards these goals (as in add features). I think we're a ways out from this, but it might be a good idea as we get closer for a final push.
The text was updated successfully, but these errors were encountered: