Drop-in replacement to panics in proc-macros.
Motivation
Error handling in proc-macros sucks. It's not much of a choice today:
you either "bubble up" the error up to top-level of you macro and convert it to
a compile_error!
invocation or just use a good old panic. Both these ways suck:
- Former sucks because it's quite redundant to unroll a proper error handling
just for critical errors that will crash the macro anyway so people mostly
choose not to bother with it at all and use panic. I have yet to see a crate > 700
lines of code that does it, simple
.expect
is too tempting. - Later sucks because there's no way to carry out span info via panic.
rustc
will highlight the whole invocation itself but not some specific token inside it. Furthermore, panics aren't for error-reporting at all; panics are for bug-detecting (like unwrapping onNone
or out-of range indexing) or for early development stages when you need a prototype ASAP and error handling can wait. Mixing these usages only messes things up.
Solution
That said, we need a solution, but this solution must meet these conditions:
- It must be better than panics. The main point: it must offer a way to carry span information over to user.
- It must require as little effort as possible to migrate from panic. Ideally, a new macro with the same semantics plus ability to carry out span info.
This crate aims to provide such a mechanism. All you have to do is enclose all
the code inside your top-level #[proc_macro]
function in [filter_macro_errors!
]
invocation and change panics to [span_error!
]/[call_site_error!
] where appropriate:
// This is your main entry point
How it works
I must confess: I used panics as a try/catch mechanism. I've committed this sin so others may live in peace and prosperity, god save my soul.
Essentially, the [filter_macro_errors!
] macro is a
try catch
[span_error!
] and co are
throw ;
By calling [span_error!
] you trigger panic
that will be caught by [filter_macro_errors!
] and converted to compile_error!
invocation.
All the panics that wasn't triggered by [span_error!
] and co but any other reason will be resumed as is.
Panic catching is indeed slow but the macro is about to abort anyway so speed is not
a concern here. Please note that this crate is not intended to be used in any other way
than a proc-macro error reporting, use [Result
] and ?
instead.
Testing
TODO: fork https://github.com/laumann/compiletest-rs and make it understand explicit line numbers.