I'm excited to announce that we released SNAFU 0.5.0 today. The biggest user-facing change is that more usages of SNAFU will now cause the compiler to fail! 
Of course, what I really mean is that error-reporting has been greatly improved for incorrect usages of the SNAFU attributes. This has all been driven by a new contributor, tjkirch. They started innocently enough, chiming in on an issue. Some back-and-forth showed that SNAFU wasn't reporting a misused attribute. Shortly thereafter, they took on the job of improving the messages and then submitted a plethora of pull requests to fix things.
I'd like to heartily thank tjkirch for their contributions; working with them has been a breeze, and I hope that other people will have the pleasure of working with them in the future!
Other highlights
Since the previous post on 0.2.1:
-
#[snafu(source(from))]can be used to perform transformations of the error type. -
#[snafu(source)]and#[snafu(backtrace)]can be used to mark fields as a source or backtrace, if they aren't automatically detected (or not use them if they've been mistakenly detected). - Support for futures 0.1, standard library futures, and futures 0.3 has been added.
- Documentation comments are used as default
Displayimplementations. - The default
Displayimplementation includes the underlying cause.
Contributors
Since 0.2.1, the following excellent human beings have contributed code:
What's next?
The top-of-mind things we are looking at for 0.6:
- Better support for environments without the standard library.
- End-user control of whether backtraces are enabled or not.
- Upgrading syn/quote/proc-macro2 to 1.x versions and requiring Rust 1.31+.
Please come participate if you are interested in these or other topics!