Luca Barbato

Thank you, I was away and I just read the code :)

Yes :) I tested with the current master and the list is shorter: ``` warning: unsupported x86 llvm intrinsic llvm.x86.avx2.pabs.w; replacing with trap warning: unsupported x86 llvm intrinsic llvm.x86.avx2.psign.w; replacing...

passing ``` -I `llvm-config --includedir` ``` should fix it (where the `llvm-config` is the one you used to build c2rust)

There is [rust-demangle](https://github.com/alexcrichton/rustc-demangle/) that could be used for a more accurate demangling.

Please notice that mangling rules [will change soon](https://github.com/rust-lang/rfcs/pull/2603#issuecomment-458410463)

Looks good! I'd avoid adding the `_` maybe though.

Wouldn't make more sense having sqlx use gluesql as one of the supported backend? With sled 1.0 it could be a compelling alternative to sqlx+sqlite.

I started https://github.com/launchbadge/sqlx/discussions/2692 so maybe we'll see if the idea has enough traction.

Because it was originally written when nom itself was only a set of very clever and convoluted macros. If you want to convert some to functions you are very welcome...

In theory you may wrap a dmabuf in a FrameBuffer, but Packet as it is assumes you are copying to its own storage, Ideally one could extend it similarly.