This library provides eyre::Report
, a trait object based
error handling type for easy idiomatic error handling and reporting in Rust
applications.
This crate is a fork of anyhow
with support for customized
error reports. For more details on customization checkout the docs on
eyre::EyreHandler
.
The heart of this crate is its ability to swap out the Handler type to change what information is carried alongside errors and how the end report is formatted. This crate is meant to be used alongside companion crates that customize its behavior. Below is a list of known crates that export report handlers for eyre and short summaries of what features they provide.
stable-eyre
: Switches the backtrace type fromstd
's tobacktrace-rs
's so that it can be captured on stable. The report format is identical toDefaultHandler
's report format.color-eyre
: Captures abacktrace::Backtrace
and atracing_error::SpanTrace
. Provides aHelp
trait for attaching warnings and suggestions to error reports. The end report is then pretty printed with the help ofcolor-backtrace
,color-spantrace
, andansi_term
. Check out the README oncolor-eyre
for details on the report format.simple-eyre
: A minimalEyreHandler
that captures no additional information, for when you do not wish to captureBacktrace
s with errors.jane-eyre
: A report handler crate that exists purely for the pun of it. Currently just re-exportscolor-eyre
.
We recommend users do not re-export types from this library as part their own public API for libraries with external users. The main reason for this is that it will make your library API break if we ever bump the major version number on eyre and your users upgrade the eyre version they use in their application code before you upgrade your own eyre dep version1.
However, even beyond this API stability hazard, there are other good reasons to
avoid using eyre::Report
as your public error type.
- You export an undocumented error interface that is otherwise still accessible
via downcast, making it hard for users to react to specific errors while not
preventing them from depending on details you didn't mean to make part of
your public API.
- This in turn makes the error types of all libraries you use a part of your public API as well, and makes changing any of those libraries into an undetectable runtime breakage.
- If many of your errors are constructed from strings you encourage your users to use string comparision for reacting to specific errors which is brittle and turns updating error messages into a potentially undetectable runtime breakage.
-
Use
Result<T, eyre::Report>
, or equivalentlyeyre::Result<T>
, as the return type of any fallible function.Within the function, use
?
to easily propagate any error that implements thestd::error::Error
trait.use eyre::Result; fn get_cluster_info() -> Result<ClusterMap> { let config = std::fs::read_to_string("cluster.json")?; let map: ClusterMap = serde_json::from_str(&config)?; Ok(map) }
-
Wrap a lower level error with a new error created from a message to help the person troubleshooting understand the chain of failures that occurred. A low-level error like "No such file or directory" can be annoying to debug without more information about what higher level step the application was in the middle of.
use eyre::{WrapErr, Result}; fn main() -> Result<()> { ... it.detach().wrap_err("Failed to detach the important thing")?; let content = std::fs::read(path) .wrap_err_with(|| format!("Failed to read instrs from {}", path))?; ... }
Error: Failed to read instrs from ./path/to/instrs.json Caused by: No such file or directory (os error 2)
-
Downcasting is supported and can be by value, by shared reference, or by mutable reference as needed.
// If the error was caused by redaction, then return a // tombstone instead of the content. match root_cause.downcast_ref::<DataStoreError>() { Some(DataStoreError::Censored(_)) => Ok(Poll::Ready(REDACTED_CONTENT)), None => Err(error), }
-
If using the nightly channel, a backtrace is captured and printed with the error if the underlying error type does not already provide its own. In order to see backtraces, they must be enabled through the environment variables described in
std::backtrace
:- If you want panics and errors to both have backtraces, set
RUST_BACKTRACE=1
; - If you want only errors to have backtraces, set
RUST_LIB_BACKTRACE=1
; - If you want only panics to have backtraces, set
RUST_BACKTRACE=1
andRUST_LIB_BACKTRACE=0
.
The tracking issue for this feature is rust-lang/rust#53487.
- If you want panics and errors to both have backtraces, set
-
Eyre works with any error type that has an impl of
std::error::Error
, including ones defined in your crate. We do not bundle aderive(Error)
macro but you can write the impls yourself or use a standalone macro like thiserror.use thiserror::Error; #[derive(Error, Debug)] pub enum FormatError { #[error("Invalid header (expected {expected:?}, got {found:?})")] InvalidHeader { expected: String, found: String, }, #[error("Missing attribute: {0}")] MissingAttribute(String), }
-
One-off error messages can be constructed using the
eyre!
macro, which supports string interpolation and produces aneyre::Report
.return Err(eyre!("Missing attribute: {}", missing));
-
On newer versions of the compiler (e.g. 1.58 and later) this macro also supports format args captures.
return Err(eyre!("Missing attribute: {missing}"));
NOTE: tests are currently broken for no_std
so I cannot guarantee that
everything works still. I'm waiting for upstream fixes to be merged rather than
fixing them myself, so bear with me.
In no_std mode, the same API is almost all available and works the same way. To depend on Eyre in no_std mode, disable our default enabled "std" feature in Cargo.toml. A global allocator is required.
[dependencies]
eyre = { version = "0.6", default-features = false }
Since the ?
-based error conversions would normally rely on the
std::error::Error
trait which is only available through std, no_std mode will
require an explicit .map_err(Report::msg)
when working with a non-Eyre error
type inside a function that returns Eyre's error type.
The eyre::Report
type works something like failure::Error
, but unlike
failure ours is built around the standard library's std::error::Error
trait
rather than a separate trait failure::Fail
. The standard library has adopted
the necessary improvements for this to be possible as part of RFC 2504.
Use eyre
if you don't think you'll do anything with an error other than
report it. This is common in application code. Use thiserror
if you think
you need an error type that can be handled via match or reported. This is
common in library crates where you don't know how your users will handle
your errors.
This crate does its best to be usable as a drop in replacement of anyhow
and
vice-versa by re-exporting
all of the renamed APIs with the names used in
anyhow
, though there are some differences still.
As part of renaming Context
to WrapErr
we also intentionally do not
implement WrapErr
for Option
. This decision was made because wrap_err
implies that you're creating a new error that saves the old error as its
source
. With Option
there is no source error to wrap, so wrap_err
ends up
being somewhat meaningless.
Instead eyre
intends for users to use the combinator functions provided by
std
for converting Option
s to Result
s. So where you would write this with
anyhow:
use anyhow::Context;
let opt: Option<()> = None;
let result = opt.context("new error message");
With eyre
we want users to write:
use eyre::{eyre, Result};
let opt: Option<()> = None;
let result: Result<()> = opt.ok_or_else(|| eyre!("new error message"));
NOTE: However, to help with porting we do provide a ContextCompat
trait which
implements context
for options which you can import to make existing
.context
calls compile.
Unless you explicitly state otherwise, any contribution intentionally submitted for inclusion in this crate by you, as defined in the Apache-2.0 license, shall be dual licensed as above, without any additional terms or conditions.
Footnotes
-
example and explanation of breakage https://github.com/yaahc/eyre/issues/30#issuecomment-647650361 ↩