If the macros only exist to get file and line information, you could do the same thing by using `#[track_caller]` functions combined with `std::panic::Location` to get that same info. For example, `stack_err!` could be replaced with
such that you call `.map_err(StackError::new_location("data is not a list of strings"))`. A macro is nice if you need to process format strings with arguments (though someone can call `StackError::new_location(format_args!(…))` if they want), but all of your examples show static strings so it's nice to avoid the error in that case.
The use of `std::panic::Location` also means instead of baking that into a format string you could also just have that be an extra field on the error, which would let you expose accessors for it, and you can then print them in your Debug/Display impls.
Speaking of, the Display impl really should not include its source. Standard handling for errors expects that an error prints just itself with Display because it's very common to recurse through sources and print those, so if Display prints the source too then you're duplicating output. Go ahead and print it on Debug though, that's nice for errors returned from `main()`.
Thanks for the insight, I wasn't aware of `track_caller`. I'll definitely be looking into this. I was scratching my head trying to figure out how to make file and line number usage consistent and customizable, this looks like the answer!
You're also right that this will pretty much eliminate the need for macros.
That's also a very key insight about Display vs. Debug printing. I'll be looking into that as well.
I hope to read through your crate and examples later, but if you have a chance, I’d be curious to hear your take on how Stack Error differs from my library, SNAFU [1]!
I played around a bit with SNAFU a couple of years ago, but I'm haven't worked deeply with the library so there might well be some features I'm not aware of.
I think SNAFU is more like a combination of anyhow and thiserror into a single crate, rather than Stack Error which leans more heavily into the "turnkey" error struct. Using the Whatever struct, you get some overlap with Stack Error features:
- Error message are co-located.
- Error type implement std::error::Error (suitable for library development).
- External errors can be wrapped and context can easily be added.
Where Stack Error differs:
- Error codes (and URIs) offer ability for runtime error handling without having to compare strings.
- Provides pseudo-stack by stacking messages.
Underlying this is an opinion I baked into Stack Error: error messages are for debugging, not for runtime error handling. Otherwise all your error strings effectively become part of your public interface since a downstream library can rely on them for error handling.
Dei here, from the team behind Error Stack [1] (a similarly named existing, context-aware error-handling library for Rust that supports arbitrary attachments). How does Stack Error, here, compare?
Anyhow still makes things easier for application development. The main drawback is that the resulting error type doesn't implement std::error::Error, so it's not suitable for library development (as pointed out in the anyhow documentation). Stack Error is a bit less ergonomic, but suitable for library development.
I prefer Anyhow's non-intrusiveness: "Result" is still "Result" and all I need is a "?". I agree with Stack Error's documentation that Anyhow can't help with debugging that well, but it's "good enough" in my opinion.
This is awesome! Will give it a try in my next project.
How does it keep track of filename and line number in a compiled binary? I'm fairly new to rust libraries and this doesn't quite make sense to me. I know in JS you need a source-map for minification, how does this work for a compiled language?
Rust provides file!, line! and column! macros that expands into a compile-time constants that the compiler embeds then into the executable. This way no source map at runtime is necessary as the relevant errors are constructed from those constants.
Presumably StackError just uses those macros.
But for debugging a source map is still necessary and is a part of various debug formats.
If the macros only exist to get file and line information, you could do the same thing by using `#[track_caller]` functions combined with `std::panic::Location` to get that same info. For example, `stack_err!` could be replaced with
such that you call `.map_err(StackError::new_location("data is not a list of strings"))`. A macro is nice if you need to process format strings with arguments (though someone can call `StackError::new_location(format_args!(…))` if they want), but all of your examples show static strings so it's nice to avoid the error in that case.The use of `std::panic::Location` also means instead of baking that into a format string you could also just have that be an extra field on the error, which would let you expose accessors for it, and you can then print them in your Debug/Display impls.
Speaking of, the Display impl really should not include its source. Standard handling for errors expects that an error prints just itself with Display because it's very common to recurse through sources and print those, so if Display prints the source too then you're duplicating output. Go ahead and print it on Debug though, that's nice for errors returned from `main()`.
Thanks for the insight, I wasn't aware of `track_caller`. I'll definitely be looking into this. I was scratching my head trying to figure out how to make file and line number usage consistent and customizable, this looks like the answer!
You're also right that this will pretty much eliminate the need for macros.
That's also a very key insight about Display vs. Debug printing. I'll be looking into that as well.
Thank you for the thoughtful reply.
I hope to read through your crate and examples later, but if you have a chance, I’d be curious to hear your take on how Stack Error differs from my library, SNAFU [1]!
[1]: https://docs.rs/snafu/latest/snafu/index.html
I played around a bit with SNAFU a couple of years ago, but I'm haven't worked deeply with the library so there might well be some features I'm not aware of.
I think SNAFU is more like a combination of anyhow and thiserror into a single crate, rather than Stack Error which leans more heavily into the "turnkey" error struct. Using the Whatever struct, you get some overlap with Stack Error features:
- Error message are co-located.
- Error type implement std::error::Error (suitable for library development).
- External errors can be wrapped and context can easily be added.
Where Stack Error differs:
- Error codes (and URIs) offer ability for runtime error handling without having to compare strings.
- Provides pseudo-stack by stacking messages.
Underlying this is an opinion I baked into Stack Error: error messages are for debugging, not for runtime error handling. Otherwise all your error strings effectively become part of your public interface since a downstream library can rely on them for error handling.
Dei here, from the team behind Error Stack [1] (a similarly named existing, context-aware error-handling library for Rust that supports arbitrary attachments). How does Stack Error, here, compare?
[1]: https://crates.io/crates/error-stack
I still prefer the Anyhow solution, but I like the approach here.
Anyhow still makes things easier for application development. The main drawback is that the resulting error type doesn't implement std::error::Error, so it's not suitable for library development (as pointed out in the anyhow documentation). Stack Error is a bit less ergonomic, but suitable for library development.
Isn't this strictly superior to Anyhow? What do you like more about Anyhow?
I prefer Anyhow's non-intrusiveness: "Result" is still "Result" and all I need is a "?". I agree with Stack Error's documentation that Anyhow can't help with debugging that well, but it's "good enough" in my opinion.
Result in `anyhow::Result` though. It's still a different type. Or do you literally mean you like that it is still spelt the same?
And I think you can still use `?` with this if you don't want to add any context... Not 100% sure on that though.
Might as well be my limited understanding from what I can read behind the link, to be fair.
This is awesome! Will give it a try in my next project.
How does it keep track of filename and line number in a compiled binary? I'm fairly new to rust libraries and this doesn't quite make sense to me. I know in JS you need a source-map for minification, how does this work for a compiled language?
Rust provides file!, line! and column! macros that expands into a compile-time constants that the compiler embeds then into the executable. This way no source map at runtime is necessary as the relevant errors are constructed from those constants.
Presumably StackError just uses those macros.
But for debugging a source map is still necessary and is a part of various debug formats.