Email List: Xaustin-group-lX
[All Lists]

Re: timestamp resolution

To: austin-group-l@xxxxxxxxxxxxx
Subject: Re: timestamp resolution
From: Geoff Clare <gwc@xxxxxxxxxxxxx>
Date: Fri, 4 May 2007 09:47:10 +0100
References: <463A5496.9060402@xxxxxx> <17978.22890.470593.629189@xxxxxx> <463A940F.7080700@xxxxxx> <17978.39101.864550.456251@xxxxxx>
wollman+austin-group@xxxxxx wrote, on 03 May 2007:
>
> >> Independently of this, I don't like #2.  Normally, the _POSIX_* limits
> >> specify minimum maxima (or maximum minima, I suppose); this is
> >> something completely different.  Compare NAME_MAX.
> 
> > Nonsense.  There is precedence for this kind of macro: _POSIX_VDISABLE.
> 
> > And even if no, there is no reason whatsoever to not do it.
> 
> You were claiming that adding this macro was for consistency.  I
> pointed out that naming it _POSIX_something was not particularly
> consistent with the existing practice.  _POSIX_VDISABLE is an oddball,
> currently the only one of its kind (the other _POSIX_* constants that
> correspond to _PC_* keys are all either limits or identify optional
> interfaces).

The _PC_* keys associated with options are not that much different
from _PC_VDISABLE.  They all return some piece of information that
isn't a limit, it's just that the value returned for _PC_2_SYMLINKS,
_PC_ASYNC_IO, etc. is a yes/no indication whereas a wider range of
values is returned for _PC_VDISABLE.

Also, the values returned for _PC_REC_INCR_XFER_SIZE and
_PC_REC_XFER_ALIGN are not really "limits", despite being associated
with constants in <limits.h>.

-- 
Geoff Clare <g.clare@xxxxxx>
The Open Group, Thames Tower, Station Road, Reading, RG1 1LX, England

<Prev in Thread] Current Thread [Next in Thread>