Previously, `unsigned` was parsed as the shorthand for `unsigned int`.
This commit introduces code to parse `unsigned short`, `unsigned int`,
`unsigned long`, and `unsigned long long`.

There is a comment in the code about std.c.parse` - Im not
familiar with zig internals, but it seems like this is a separate
C parsing implementation. In the long run, it probably makes
sense to merge both implementations, so this commit should be
regarded as a quick fix that doesn't address an apparently
underlying issue.