Re: [PATCH, HURD][RFC] hurdselect: Step7x, almost complete rewrite finished

[Samuel Thibault]

Pino Toscano, le Wed 13 Feb 2013 00:08:21 +0100, a écrit :
> Alle mercoledì 13 febbraio 2013, Svante Signell ha scritto:
> > On Tue, 2013-02-12 at 23:52 +0100, Pino Toscano wrote:
> > > Alle martedì 12 febbraio 2013, Svante Signell ha scritto:
> > > >  * change the FD_SETSIZE upper value check to larger than or
> > > >  equal
> > > > 
> > > > from larger than.
> > > > (from POSIX definition of select:
> > > > http://pubs.opengroup.org/onlinepubs/009604499/functions/select.h
> > > > tml )
> > > 
> > > It's
> > > http://pubs.opengroup.org/onlinepubs/9699919799/functions/select.ht
> > > ml
> > > 
> > >  actually, and it says:
> > > |[EINVAL]
> > > |
> > > |    The nfds argument is less than 0 or greater than FD_SETSIZE.
> > > 
> > > and not "greater or equal than", so it allows FD_SETSIZE.
> > > (And logically, if a fdset_t can contain at most FD_SETSIZE fd's,
> > > you need to allow FD_SETSIZE as maximum number of fd's.)
> > > 
> > > Please remove this change, which is wrong.
> > 
> > Then you have to convince the Linux man page and the python3.2
> > developers too. From python3.2:
> > ./Include/fileobject.h: #define _PyIsSelectable_fd(FD) (((FD) >= 0)
> > && ((FD) < FD_SETSIZE))
> 
> Feel free to open bugs about them, then.

That being said, for the python case they may prefer to stay on the safe
side, and just not attempt to select FD_SETSIZE, in the case the OS has
bugs that leaves it with the old standard.

Samuel