Matthew Flatt

The `-D_GNU_SOURCE` and `-lucontext` flags should not be needed (or the `libucontext-devel` package) for building Racket CS. I imagine they're leftover from compiling Racket BC. Also: Thanks for the report!

From the template linked above, it looks like cross build sets `--enable-racket=`, but cross building in that way also requires `--enable-scheme=` to point to part of a Racket build. An...

I think you're right that the best approach here is to correct the documentation. Some notes on the history: The `#%top` form originally accessed top-level bindings more reliably. As the...

After looking at this all again, I agree with your reservations about `#%top` and local variables outside of a module. I'm trying out the change where `#%top` around a variable...

I'm ok with `AC_CHECK_MEMBERS`, but note that the `__QNX__` check would need to be earlier to continue paying attention to `d_namelen`. I don't know that anyone is using QNX. Unless...

@michaelballantyne I think the solution to the letrec bug is to introudce an extra scope for the body of a `letrec` (really, `letrec-values` and `letrec-values+syntaxes`). I expect to push that...

@rmculpepper Thanks - that helps me see the deeper issue. Here's a standalone example that fails if you comment out with `with-syntax` clause that applies `syntax-local-identifier-as-binding`: ``` #lang racket/base (require...

@eddierucker From email, it sounds like this isn't a change after all, and the difference is what `text-field%` and `editor-canvas%` do by default. But if I understand correctly, it could...

Included in v9.9.9 merge (as a continuation-marks API) in 47daa9b34016de84fd111801d9d589d15a523abe.

RISC-V is now supported.