Strange sender log file characters

[Simo Sorce]

On Mon, 2008-09-01 at 13:31 -0700, Wayne Davison wrote:
> On Sun, Aug 31, 2008 at 10:44:11PM -0400, Matt McCutchen wrote:
> > Wayne, you might want to consider applying this even though the
> > immediate problem is solved.
> 
> Yeah, I had noticed that lack-of-negative number-handling deficiency
> when I was working on an enhancement for 3.1.0 to output numbers in
> 3-digit clusters (e.g. 1,234,567).  I hadn't thought that I needed
> special handling of INT_MIN, though, so I fixed that.  I've gone ahead
> and committed my enhancement to the 3.1.0 branch, and ported that
> improved negation fix from the trunk to the 3.0.x branch.
> 
> I'm interesting in what folks think about my making the outputting of
> clustered numbers the default for 3.1.0 (see the --human-readable option
> in the 3.1.0 manpage for a full discussion of how it works and is
> overridden).  The code attempts to intuit if it should use a comma or a
> period for the 3-digit clustering based on if sprintf() puts a period or
> a comma into a floating point number.
> 
> ..wayne..

Reading the snprintf manpage (at least on Linux). It seem you don't have
to meka eany special manipulation to get the thousands separator but
just use the "'" modifer:

  The five flag characters above are defined in the C standard. The
  SUSv2 specifies one further flag character.

  ’  For decimal conversion (i, d, u, f, F, g, G) the output is to be
     grouped with thousands’ grouping characters if the locale infor-
     mation  indicates any.  Note that many versions of gcc(1) cannot
     parse this option and will issue  a  warning.   SUSv2  does  not
     include %’F.


The advantage of using a glibc function is that it uses the correct
locale settings for the thousands separator and does not require
guessing.

Simo.

-- 
Simo Sorce * Red Hat, Inc * New York