I often use `\w` to match "C identifier characters", i.e. `[a-zA-Z0-9_]`, but I didn't find any way to write that in "native" `rex()`.