Alan Somers
Alan Somers
When raising the MSRV to 1.50.0 or above: * Make `ipv4addr_to_libc` a `const fn`, since `Ipv4Addr::octets` is const in that version.
When updating the MSRV to 1.48.0 or later, we should: * Add doc aliases. #1673
The CHANGELOG file is by far our biggest source of merge conflicts. It would be great if we could fix that. One approach would be that taken by https://github.com/saltstack/salt ....
`sys::socket::addr::sys_control::SysControlAddr` is not exposed in Nix's public API. That's probably an oversight. It should probably be exposed just like `NetlinkAddr`.
Once we upgrade the MSRV to 1.52.0 or later, we should: - Add `#[cfg_attr(target_os = "fuchsia", allow(rustdoc::broken_intra_doc_links))]` to `sys::signal::kill` to fix rustdoc complaints.
Compute the required space at compile time rather than at runtime.
With the latest nightly compiler, Travis's Linux builds (but not my local linux builds) fail with this error: ``` ../sysdeps/pthread/aio_suspend.c:102: do_aio_misc_wait: Assertion `status == 0 || status == -11' failed...
Beginning with Rust 1.41.0, the build fails on Fuchsia. It triggers this error: ``` error: the type `libc::sigevent` does not permit zero-initialization --> src/sys/signal.rs:855:36 | 855 | let mut sev...
You may have noticed that my responsiveness has been lagging the past few months. COVID has given me less time to work on Nix, though TBH I've been burning out,...
After working fine for years, sys::test_aio::test_aio_suspend has begun to reliably fail in CI, when run on OSX. The trigger seems to be the recent MSRV upgrade to 1.40.0. Unfortunately, none...