feat: implement no_std #17
Conversation
| match segments.last() { | ||
| Some(syn::PathSegment { ident, .. }) => ident == "Cow", | ||
| _ => false, | ||
| } |
There was a problem hiding this comment.
I wonder if this accidentially adds support for beef.
Well, it's no problem. Before getting rid of the duck typing there is really no way to tell with #14.
Is it clear to you what is the segments.last() == None case? if not, could add unreachable!() or segments.last().expect("type must always have at least one segment"). Or maybe could use the error reporting facilities of syn, which haven't been used so far.
There was a problem hiding this comment.
Erroring here is not meaningful. It will error downstream, so it's not wrong to use false.
Regarding beef, I suspect it will support beef. I wouldn't say this is bad; I think it's well-known to not name types after standard types unless it's planned for complete compatibility.
Uses the original path rather than hardcoding
std. This provides more flexibility for imports.