stable-2015-09-17 vs. beta-2015-10-02
stable-2015-09-17 vs. nightly-2015-10-06
Lot’s of reported regressions. I haven’t looked at them yet.
Had a quick look through to find the relevant issues:
Beta:
- pod/immeta/rcon/elf/nue - read_exact - https://github.com/rust-lang/rust/issues/28126
- chipmunk/chipmunk-sys) - extern fn with 6+ args - https://github.com/rust-lang/rust/issues/28559
- aurelius/sprs - lifetime problems
- cuckoofilter - deprecated method - removed in https://github.com/rust-lang/rust/pull/27684
Nightly:
- gfx-pipeline/backend - feature gating import problem - https://github.com/rust-lang/rust/issues/28075
- graph - task log is empty? crate does nothing anyway.
- pinyin - lifetime problem
- cudd-sys/ao - invalid attributes - https://github.com/rust-lang/rust/pull/28650
Maybe https://github.com/rust-lang/rust/pull/28560 be backported to beta?
Other than that, I didn’t see any surprises looking through the regressions.
Awesome! Thank you for doing the analysis @Eljay @eefriedman.
The cuckoofilter breakage was because of a bug that allowed it to import in a use list an unstable feature, which is fixed by https://github.com/rust-lang/rust/pull/28364
is this related to any known issue?
Even though it was allowed because of a bug, fixing said bug is a breaking change. I’d personally expect such changes not to happen, considering the "commitment to stability".
It's just a missing bound. Not sure why this particular formulation triggers an error, as opposed to an RFC 1214 warning. Reduces to:
use std::ops::Deref;
pub struct CsMat<N, DataStorage> where DataStorage: Deref<Target=[N]> {
_data: DataStorage,
}
pub struct ChunkOuterBlocks<'a, N> {
_mat: CsMat<N, &'a [N]>,
}
fn main(){}
Understood, thanks.
This topic was automatically closed 90 days after the last reply. New replies are no longer allowed.