develooper Front page | perl.perl5.porters | Postings from June 2013

Re: [perl #117539] Fix incorrect assumptions about gids & uids

Thread Previous
Leon Timmermans
June 3, 2013 12:05
Re: [perl #117539] Fix incorrect assumptions about gids & uids
Message ID:
On Mon, Jun 3, 2013 at 1:17 PM, Tony Cook via RT
<> wrote:
>> I think the �Uid_t_size > IVSIZE� case is crazy and can be safely
>> removed. I can't imagine any platform having bigger uid_t than
>> pointers.
> That isn't a new check - it was added in 1999 by jhi - see
> 887d29384f0bc4b6197573ce19ff42abfe67fa51.
> Brian's patch simply reproduces it.

I realize that, but I still suspect it's wrong.

That patch introduces it for «st_size, st_uid, st_gid», though it
makes a lot of sense for st_size, I don't see how it does for [ug]id.
This seems like accidental cruft to me.

>> Also, I find sv_setuid/sv_setgid rather confusing names, given the
>> setuid/setgid syscalls. I'd prefer to see it being called something
>> else (sv_setid may be a good one, I can't imagine the signedness of
>> uids not matching gids).
> POSIX provides two types, Configure probes for both types.
> I consider it unlikely for them to be different, but both POSIX and
> Configure treats them different, so treating them as the same time feels
> fragile to me.

Fair enough.


Thread Previous Perl Programming lists via nntp and http.
Comments to Ask Bjørn Hansen at | Group listing | About