RFC: Superseding public/private dependencies #3516
Conversation
epage
added
the
T-cargo
|
Really excited to see this moving forward, thanks for making this happen! I advocated for an approach like this (cutting out to the resolver parts at first) because I think it'll realize 95% of the value at 20% of the complexity. An alternative suggested phasing in path: in current edition (and earlier?), all dependencies are public by default but you can annotate them as non-public. In 2024 edition (and later?), all dependencies are private by default and you must explicitly make them public. Naming nit: feels to me like going with full-on |
| - `Cargo.toml`: instead of `pub` (named after the [Rust keyword](https://doc.rust-lang.org/reference/visibility-and-privacy.html), we could name the field `public` (like [RFC 1977]) or name the field `visibility = "<public|private>"` | ||
| - The parallel with Rust seemed worth pursuing | ||
| - `pub` could be seen as ambiguous with `publish` | ||
| - While `visibility` would offer greater flexibility, it is unclear if we need that flexibility and if the friction of any feature leveraging it would be worth it |
There was a problem hiding this comment.
Moving @djc's comment here for a more focused conversation
Naming nit: feels to me like going with full-on public makes more sense in this context than mirroring the pub from function/type syntax.
There was a problem hiding this comment.
btw what decision we make for naming here will impact #3487
There was a problem hiding this comment.
We talked about this in the cargo team meeting and the tie breaker for pub is that it allows our no-MSRV-bump approach to work. public is already reserved from the previous RFC's implementation.
There was a problem hiding this comment.
What do you mean by the "no-MSRV-bump"? If public is reserved from the previous RFC (and assuming that there is substantial overlap between this RFC's meaning and that one's), how does it make sense to use a different keyword now?
There was a problem hiding this comment.
The original plan had the lint on-by-default with the pub field being treated as an unused field without -Z, much like [lints].
However, since we switched to allow-by-default, we can re-evaluate this part. Treating this field as unused without -Z does mean it can be adopted more quickly though.
There was a problem hiding this comment.
Going for the worse name here feels like prioritizing short-term MSRV support over long-term readability, given how many years we've waited to use this is delaying it another year for just those that provide extreme MSRVs such a problem? (And if the hard-error is removed now, that delay would be reduced relative to when it is stabilized).
There was a problem hiding this comment.
We talked about this in the cargo team meeting and decided to force an MSRV bump and switch to public.
rfcbot
added
finished-final-comment-period
|
The final comment period, with a disposition to merge, as per the review above, is now complete. As the automated representative of the governance process, I would like to thank the author for their work and everyone else who contributed. This will be merged soon. |
|
Huzzah! The @rust-lang/cargo team has decided to accept this RFC. To track further discussion, subscribe to the tracking issue here: |
To implement rust-lang/rfcs#3516, we need to decouple the resolver's behavior from the unstable flag. Since the code path is now dead, I went ahead and removed it.
To implement rust-lang/rfcs#3516, we need to decouple the resolver's behavior from the unstable flag. Since the code path is now dead, I went ahead and removed it.
To implement rust-lang/rfcs#3516, we need to decouple the resolver's behavior from the unstable flag. Since the code path is now dead, I went ahead and removed it.
To implement rust-lang/rfcs#3516, we need to decouple the resolver's behavior from the unstable flag. Since the code path is now dead, I went ahead and removed it.
To implement rust-lang/rfcs#3516, we need to decouple the resolver's behavior from the unstable flag. Since the code path is now dead, I went ahead and removed it.
To implement rust-lang/rfcs#3516, we need to decouple the resolver's behavior from the unstable flag. Since the code path is now dead, I went ahead and removed it.
To implement rust-lang/rfcs#3516, we need to decouple the resolver's behavior from the unstable flag. Since the code path is now dead, I went ahead and removed it.
fix(resolver): Remove unused public-deps error handling To implement rust-lang/rfcs#3516, we need to decouple the resolver's behavior from the unstable flag. Since the code path is now dead, I went ahead and removed it.
#1977 has languished for 6 years, mostly due to the cargo resolver design section.
This updates the RFC by
cfgdependencies, workspace inheritance, renamed dependencies, etc)Pub-aware version resolution was such a large part of the previous RFC, that we felt this needed the wider input of an RFC, rather than addressing these in stabilization FCP.
Rendered