Martijn van Duren wrote:
> Hello tech@,
>
> I looked at the doas code and really like the concept and simplicity. I
> did found the strtogid code a little odd for two reasons:
> 1) It reads like an mangled variety on the parseuid function.
> 2) It returns -1 on error, but gid_t is defined as a __uint32_t as per
> /usr/include/sys/{,_}types.h.
>
> Option 1 makes it strange to read and option 2 could potentially cause
> an error when someone is stupid enough to assign GID_MAX as gid.
>
> To make it more consistent I reformatted strtogid to parsegid to
> resemble parseuid and solve the potential incorrect error.
agreed, but the patch doesn't apply. can you resend a fixed version?
