wherror

Fork of thiserror derive(Error) with std::panic::Location support. This library provides a convenient derive macro for the standard library's std::error::Error trait.

This fork was created to add location support (thiserror#291) and additional convenience features.

[dependencies]
wherror = "2"

Location Support

Add a field of type &'static std::panic::Location<'static> to automatically capture where errors are created:

use wherror::Error;

#[derive(Error, Debug)]
#[error("Failed at {location}: {source}")]
pub struct MyError {
    #[from]
    source: std::io::Error,
    location: &'static std::panic::Location<'static>,  // Auto-populated
}

// Location automatically captured when using `?`
fn read_file() -> Result<String, MyError> {
    let content = std::fs::read_to_string("file.txt")?;
    Ok(content)
}

Example

use wherror::Error;

#[derive(Error, Debug)]
pub enum DataStoreError {
    #[error("data store disconnected")]
    Disconnect(#[from] io::Error),
    #[error("the data for key `{0}` is not available")]
    Redaction(String),
    #[error("invalid header (expected {expected:?}, found {found:?})")]
    InvalidHeader {
        expected: String,
        found: String,
    },
    #[error("unknown data store error")]
    Unknown,
}

Details

Wherror deliberately does not appear in your public API. You get the same thing as if you had written an implementation of std::error::Error by hand, and switching from handwritten impls to thiserror or vice versa is not a breaking change.

Errors may be enums, structs with named fields, tuple structs, or unit structs.

A Display impl is generated for your error if you provide #[error("...")] messages on the struct or each variant of your enum, as shown above in the example.

The messages support a shorthand for interpolating fields from the error.

#[error("{var}")] ⟶ write!("{}", self.var)
#[error("{0}")] ⟶ write!("{}", self.0)
#[error("{var:?}")] ⟶ write!("{:?}", self.var)
#[error("{0:?}")] ⟶ write!("{:?}", self.0)

These shorthands can be used together with any additional format args, which may be arbitrary expressions. For example:

# use core::i32;
# use wherror::Error;
#
#[derive(Error, Debug)]
pub enum Error {
    #[error("invalid rdo_lookahead_frames {0} (expected < {max})", max = i32::MAX)]
    InvalidLookahead(u32),
}

If one of the additional expression arguments needs to refer to a field of the struct or enum, then refer to named fields as .var and tuple fields as .0.

# use wherror::Error;
#
# fn first_char(s: &String) -> char {
#     s.chars().next().unwrap()
# }
#
# #[derive(Debug)]
# struct Limits {
#     lo: usize,
#     hi: usize,
# }
#
#[derive(Error, Debug)]
pub enum Error {
    #[error("first letter must be lowercase but was {:?}", first_char(.0))]
    WrongCase(String),
    #[error("invalid index {idx}, expected at least {} and at most {}", .limits.lo, .limits.hi)]
    OutOfBounds { idx: usize, limits: Limits },
}

A From impl is generated for each variant that contains a #[from] attribute.

The variant using #[from] must not contain any other fields beyond the source error (and possibly a location or backtrace — see below). Usually #[from] fields are unnamed, but #[from] is allowed on a named field too.

# use core::fmt::{self, Display};
# use std::io;
# use wherror::Error;
#
# mod globset {
#     #[derive(wherror::Error, Debug)]
#     #[error("...")]
#     pub struct Error;
# }
#
#[derive(Error, Debug)]
pub enum MyError {
    Io(#[from] io::Error),
    Glob(#[from] globset::Error),
}
#
# impl Display for MyError {
#     fn fmt(&self, formatter: &mut fmt::Formatter) -> fmt::Result {
#         unimplemented!()
#     }
# }

For Box<T> fields with #[from], both From<Box<T>> and From<T> implementations are automatically generated for enhanced ergonomics:

# use std::io;
# use wherror::Error;
#
#[derive(Error, Debug)]
#[error("Error occurred")]
pub struct MyError {
    #[from]
    source: Box<io::Error>,
}
#
# fn example() -> Result<(), Box<dyn std::error::Error>> {
#     let io_error = io::Error::new(io::ErrorKind::Other, "test");
#
#     // Both work:
#     let err1: MyError = Box::new(io_error).into();
#     let err2: MyError = io::Error::new(io::ErrorKind::Other, "test").into();  // automatically boxed
#     Ok(())
# }

The Error trait's source() method is implemented to return whichever field has a #[source] attribute or is named source, if any. This is for identifying the underlying lower level error that caused your error.

The #[from] attribute always implies that the same field is #[source], so you don't ever need to specify both attributes.

Any error type that implements std::error::Error or dereferences to dyn std::error::Error will work as a source.
```
# use core::fmt::{self, Display};
# use wherror::Error;
#
#[derive(Error, Debug)]
pub struct MyError {
    msg: String,
    #[source]  // optional if field name is `source`
    source: anyhow::Error,
}
#
# impl Display for MyError {
#     fn fmt(&self, formatter: &mut fmt::Formatter) -> fmt::Result {
#         unimplemented!()
#     }
# }
```

Fields of type &'static std::panic::Location<'static> are automatically populated with the call site location when errors are created via From trait conversion. This works seamlessly with the ? operator for precise error tracking.

For errors with location fields, a location() method is generated that returns Option<&'static std::panic::Location<'static>>.

# use wherror::Error;
#
#[derive(Error, Debug)]
#[error("Parse error at {location}: {source}")]
pub struct ParseError {
    #[from]
    source: std::num::ParseIntError,
    location: &'static std::panic::Location<'static>,  // automatically detected
}
#
# fn example() -> Result<(), ParseError> {
#     let result: Result<i32, _> = "not_a_number".parse();
#     let error: ParseError = result.unwrap_err().into();
#     if let Some(location) = error.location() {
#         eprintln!("Error at {}:{}", location.file(), location.line());
#     }
#     Ok(())
# }

The Error trait's provide() method is implemented to provide whichever field has a type named Backtrace, if any, as a std::backtrace::Backtrace. Using Backtrace in errors requires a nightly compiler with Rust version 1.73 or newer.

# use std::backtrace::Backtrace;
# use wherror::Error;
#
#[derive(Error, Debug)]
#[error("Something went wrong: {msg}")]
pub struct MyError {
    msg: String,
    backtrace: Backtrace,  // automatically detected
}

If a field is both a source (named source, or has #[source] or #[from] attribute) and is marked #[backtrace], then the Error trait's provide() method is forwarded to the source's provide so that both layers of the error share the same backtrace. The #[backtrace] attribute requires a nightly compiler with Rust version 1.73 or newer.
```
# use std::io;
# use wherror::Error;
#
#[derive(Error, Debug)]
#[error("IO error occurred")]
pub enum MyError {
    Io {
        #[backtrace]
        source: io::Error,
    },
}
```

For variants that use #[from] and also contain a Backtrace field, a backtrace is captured from within the From impl.

# use std::backtrace::Backtrace;
# use std::io;
# use wherror::Error;
#
#[derive(Error, Debug)]
#[error("IO error occurred")]
pub enum MyError {
    Io {
        #[from]
        source: io::Error,
        backtrace: Backtrace,
    },
}

Errors may use error(transparent) to forward the source and Display methods straight through to an underlying error without adding an additional message. This would be appropriate for enums that need an "anything else" variant.

# use wherror::Error;
#
#[derive(Error, Debug)]
pub enum MyError {
    # /*
    ...
    # */

    #[error(transparent)]
    Other(#[from] anyhow::Error),  // source and Display delegate to anyhow::Error
}

Another use case is hiding implementation details of an error representation behind an opaque error type, so that the representation is able to evolve without breaking the crate's public API.

# use wherror::Error;
#
// PublicError is public, but opaque and easy to keep compatible.
#[derive(Error, Debug)]
#[error(transparent)]
pub struct PublicError(#[from] ErrorRepr);

impl PublicError {
    // Accessors for anything we do want to expose publicly.
}

// Private and free to change across minor version of the crate.
#[derive(Error, Debug)]
enum ErrorRepr {
    # /*
    ...
    # */
}

See also the anyhow library for a convenient single error type to use in application code.

Comparison to anyhow

Use wherror if you care about designing your own dedicated error type(s) so that the caller receives exactly the information that you choose in the event of failure. This most often applies to library-like code. Use Anyhow if you don't care what error type your functions return, you just want it to be easy. This is common in application-like code.

License

Attribution

License: MIT OR Apache-2.0