Adopt thiserror for libcgroups
#1872
Conversation
Signed-off-by: squili <74628488+squili@users.noreply.github.com>
Signed-off-by: squili <74628488+squili@users.noreply.github.com>
Signed-off-by: squili <74628488+squili@users.noreply.github.com>
Signed-off-by: squili <74628488+squili@users.noreply.github.com>
Signed-off-by: squili <74628488+squili@users.noreply.github.com>
Signed-off-by: squili <74628488+squili@users.noreply.github.com>
Codecov Report
Additional details and impacted files@@ Coverage Diff @@
## main #1872 +/- ##
==========================================
- Coverage 68.89% 68.13% -0.77%
==========================================
Files 122 122
Lines 13717 14011 +294
==========================================
+ Hits 9450 9546 +96
- Misses 4267 4465 +198 |
Signed-off-by: squili <74628488+squili@users.noreply.github.com>
Signed-off-by: squili <74628488+squili@users.noreply.github.com>
Signed-off-by: squili <74628488+squili@users.noreply.github.com>
Signed-off-by: squili <74628488+squili@users.noreply.github.com>
Signed-off-by: squili <74628488+squili@users.noreply.github.com>
| /// Parses a file that is structed according to the nested keyed format | ||
| /// # Example | ||
| /// ``` | ||
| /// use libcgroups::stats::parse_device_number; |
There was a problem hiding this comment.
Do we need to keep the example here?
There was a problem hiding this comment.
The example was triggering doctets, which then failed because the function was made crate-only
| #[error("failed to read {path}: {err}")] | ||
| Read { err: std::io::Error, path: PathBuf }, | ||
| #[error("failed to create dir {path}: {err}")] | ||
| CreateDir { err: std::io::Error, path: PathBuf }, |
There was a problem hiding this comment.
I love this WrappedIoError idea ❤️
If the dirs cannot be created in the libcgroup crates, does this result in the failure to create a new cgroup?
I believe it is more user-friendly to display a message such as "failed to create the cgroup" or something similar. WDYT?
There was a problem hiding this comment.
That's what errors higher up in the chain are for - they can be added to the higher-level error adts with more user-friendly messages
| @@ -1,6 +1,6 @@ | |||
| use std::fmt::Display; | |||
|
|
|||
| #[derive(Hash, PartialEq, Eq, Debug, Clone)] | |||
| #[derive(Hash, PartialEq, Eq, Debug, Clone, Copy)] | |||
There was a problem hiding this comment.
This enum is essentially just a fancy number, so the performance of passing by reference and copying are the same
There was a problem hiding this comment.
Specifically, this allows us to do *controller_type instead of controller_type.clone() to turn a reference into an owned value
|
It looks like that lint said the error. Please check them 🙏 |
Signed-off-by: squili <74628488+squili@users.noreply.github.com>
Signed-off-by: squili <74628488+squili@users.noreply.github.com>
Signed-off-by: squili <74628488+squili@users.noreply.github.com>
As prescribed in #1377