zfs-core: fix tests for openzfs 2.0 on macos

zfs-core/tests/basic.rs

@@ -316,13 +316,11 @@ fn hold_not_snap() {
  // variant. This might be a bug in our code, or some change in how lzc handles errors for
  // single results
  match e {
- // zfs 2.0.0 on linux:
- #[cfg(target_os = "linux")]
+ // openzfs 2.0.0:
  zfs_core::Error::Io { source: e } => {
  assert_eq!(e.kind(), io::ErrorKind::InvalidInput);
  }
- // zfs 1.9.4 on macos:
- #[cfg(target_os = "macos")]
+ // openzfs 1.9.4:
  zfs_core::Error::List { source: el } => {
  let mut hm = std::collections::HashMap::new();
  hm.insert(tmpfs.path().to_owned() + "/2", io::ErrorKind::InvalidInput);
@@ -336,10 +334,6 @@ fn hold_not_snap() {
  }
  }
  }
-
- _ => {
- panic!("unexpected error kind: {:?}", e);
- }
  }
 }
 
@@ -364,13 +358,14 @@ fn hold_not_exist() {
  e
  } else {
  // macos hits this for some reason
+ /*
  #[cfg(target_os = "macos")]
  {
  eprintln!("macos zfs 1.9.4 is for some reason totally cool with creating holds on non-existent snaps");
  return;
  }
+ */
 
- #[cfg(not(target_os = "macos"))]
  panic!("expected an error, got {:?}", e);
  };
 
@@ -379,7 +374,6 @@ fn hold_not_exist() {
  // linux (zfs 2.0.0) doesn't appear to return our error list, which is also concerning
  match e {
  // zfs 2.0.0 on linux:
- #[cfg(target_os = "linux")]
  zfs_core::Error::Io { source: e } => {
  assert_eq!(e.kind(), io::ErrorKind::NotFound);
  }