ringboard-x11 fails to restart syscalls after interruption

[vi]

ringboard-x11 &Error: an I/O error occurred
│
╰─▶ Interrupted system call (os error 4)
    ╰╴Failed to wait for epoll events.

I think EINTR should be just restarted in most of the cases.

Maybe this was caused by suspend-to-ram, maybe by something else.

[SUPERCILEX]

I'm a little bit worried about getting stuck in an infinite loop because of this: https://stackoverflow.com/questions/46502099/epoll-wait-returns-eintr-infinitely. But in my testing it seemed fine? I wish that post had figured out the root cause cuz I feel like it's a bug in their code rather than something that can actually happen. Hopefully.

[vi]

I typically assume that any syscall that waits for something can spuriously wake with EINTR. Sometimes I also handle unexpected EAGAINs (by sleeping a bit before resuming the loop).

[vi]

Note: as of commit 1625939 ringboard-server crashes when I try to start ringboard-x11.

$ RUST_LOG=debug RUST_BACKTRACE=1 target/debug/ringboard-server
[2024-08-21T20:00:49Z INFO  ringboard_server] Starting Ringboard server v0.5.4.
[2024-08-21T20:00:49Z INFO  ringboard_server] Using database in "/home/vi/.local/share/clipboard-history".
[2024-08-21T20:00:49Z WARN  ringboard_server::startup] Unclean shutdown detected, forcibly claiming server lock.
[2024-08-21T20:00:49Z INFO  ringboard_server] Acquired server lock.
[2024-08-21T20:00:49Z WARN  ringboard_server::allocator] Reconstructing allocator free lists.
[2024-08-21T20:00:49Z DEBUG ringboard_server::reactor] Detected cgroup v1 which is unsupported.
[2024-08-21T20:00:49Z INFO  ringboard_server::reactor] Server event loop started.
[2024-08-21T20:00:52Z DEBUG ringboard_server::reactor] Handling accept completion.
thread 'main' panicked at server/src/reactor.rs:342:21:
assertion failed: client < u32::from(MAX_NUM_CLIENTS)
stack backtrace:
   0: rust_begin_unwind
             at /rustc/6de928dce9472b864f4e5d590dd7aa2075cb7551/library/std/src/panicking.rs:662:5
   1: core::panicking::panic_fmt
             at /rustc/6de928dce9472b864f4e5d590dd7aa2075cb7551/library/core/src/panicking.rs:74:14
   2: core::panicking::panic
             at /rustc/6de928dce9472b864f4e5d590dd7aa2075cb7551/library/core/src/panicking.rs:148:5
   3: ringboard_server::reactor::run
             at /mnt/bkel/vi/src/clipboard-history/server/src/reactor.rs:342:21
   4: ringboard_server::run
             at /mnt/bkel/vi/src/clipboard-history/server/src/main.rs:110:13
   5: ringboard_server::main
             at /mnt/bkel/vi/src/clipboard-history/server/src/main.rs:54:5
   6: core::ops::function::FnOnce::call_once
             at /rustc/6de928dce9472b864f4e5d590dd7aa2075cb7551/library/core/src/ops/function.rs:250:5
note: Some details are omitted, run with `RUST_BACKTRACE=full` for a verbose backtrace.

Value of nearby result variable is Ok(34).

[SUPERCILEX]

Fixed in 3676439. Thanks for catching that!