Compilation failure on nightly 1.45.0 2020-05-24

[timvisee]

I just started seeing a compilation error in time-macros-impl v0.1.1 when using the latest Rust nightly, which I'm using in ffsend-api.

This may just be a temporary nightly regression, I'm not sure. I figured it'll be useful to track this here anyway.

Here's the Rust version that was used:

rustc 1.45.0-nightly (46e85b432 2020-05-24)
cargo 1.45.0-nightly (500b2bd01 2020-05-18)

Here is the error that comes up:

 error: expected identifier
  --> /usr/local/cargo/registry/src/github.com-1ecc6299db9ec823/time-macros-impl-0.1.1/src/lib.rs:84:20
   |
84 |               pub fn $name(input: proc_macro::TokenStream) -> proc_macro::TokenStream {
   |                      ^^^^^
...
91 | / impl_macros! {
92 | |     time: Time,
93 | |     offset: Offset,
94 | |     date: Date,
95 | | }
   | |_- in this macro invocation
   |
   = note: this error originates in a macro (in Nightly builds, run with -Z macro-backtrace for more info)
error: aborting due to previous error
error: could not compile `time-macros-impl`.

See the error on the bottom of this CI log here:
https://gitlab.com/timvisee/ffsend-api/-/jobs/566079187#L735

[jhpratt]

I'll look into this when I'm on my laptop, but aftet quick look at the source code there, I see absolutely nothing wrong with how I've implemented that macro.

If this is in fact a regression that isn't on me, I'll file an issue on rust-land/rust pointing to this.

[timvisee]

It appears someone already opened an issue for it a day ago: rust-lang/rust#72545

[jhpratt]

Looking at that issue, it appears as though it was an issue with proc macro hack, which has been fixed.

I believe the solution to this is to run cargo update, which will pull in the appropriate version of proc macro hack. If that doesn't work, let me know, and I'll see if publishing a new version myself helps.

[xStrom]

cargo update -p proc-macro-hack fixes the issue, no need for a new release.

[jhpratt]

Thanks for confirmation @xStrom! In that case, I'll just close this issue, leaving it as a signpost for anyone running into this.