Jeremiah Senkpiel
Jeremiah Senkpiel
Here's a backtrace: ``` RUST_BACKTRACE=1 longboard GET https://futureuae.com thread 'main' panicked at 'called `Result::unwrap()` on an `Err` value: String slice should be valid ASCII', /home/jeremiah/.cargo/registry/src/github.com-1ecc6299db9ec823/http-types-2.10.0/src/headers/headers.rs:54:62 stack backtrace: 0: rust_begin_unwind at...
I suppose this should not panic, but it should be an error.
Maybe a reset would be better?
Note: the tests for this are brittle. The order of added header is non-deterministic (HashMap iterator).
The problem is we are comparing to raw output files and unless headers are always going to be sorted there isn't a way to do that just for test output...
Given this is a perf change, a benchmark is probably more appropriate. Also, regarding it being a perf change, it may be better to write this logic ourselves rather than...
We should not accept non-ASCII characters in headers (it is against the http spec as fare as I am aware), but it should be checked before we do things which...
Re-running CI now that #399 was merged
Needs a rebase it seems
Hmph. The CI duly notes that this makes interop with hyper http more difficult (obviously).