[PATCH] Portability patches

[Jeremy Allison]

On Thu, Oct 09, 2014 at 07:36:27PM +0200, Ralph Böhme wrote:
> On Thu, Oct 09, 2014 at 10:33:45AM -0700, Richard Sharpe wrote:
> > On Thu, Oct 9, 2014 at 10:20 AM, Ralph Böhme <rb at sernet.de> wrote:
> > > On Thu, Oct 09, 2014 at 09:56:52AM -0700, Richard Sharpe wrote:
> > >> On Thu, Oct 9, 2014 at 9:28 AM, Ralph Böhme <rb at sernet.de> wrote:
> > >> > Hi all,
> > >> >
> > >> > attached please find two simple fixes for non-portable code that
> > >> > exploded on my FreeBSD 10 testbox (the non-portable code, not the
> > >> > fixes ;)
> > >>
> > >> Hmmm, on the CentOS 6.5 (and thus RHEL 6.x) box I am currently using,
> > >> it says of %m that:
> > >>
> > >>        m      (Glibc  extension.)   Print output of strerror(errno).  No argu-
> > >>               ment is required.
> > >
> > > where does it say that? In man sscanf? The only reference to this text
> > > I can find is from man printf: <http://linux.die.net/man/3/printf>
> > >
> > >> Isn't that going to cause more portability problems?
> > >
> > > To reiterate:
> > >
> > > a    scanf glibc extension, eg FreeBSD and Solaris don't have it
> > > m    In SUSv4 [1], but not in SUSv3 [2], supported by glibc since 2.7, FreeBSD
> > >
> > > Looks like at least Solaris 11.1 doesn't have m yet, maybe it's in 11.2.
> > >
> > > -Ralph
> > >
> > > [1] <http://pubs.opengroup.org/onlinepubs/9699919799/functions/sscanf.html>
> > > [2] <http://pubs.opengroup.org/onlinepubs/009695399/functions/sscanf.html>
> > 
> > D'oh. I looked in the printf man page not the scanf. Objection removed.
> 
> :) This one actually took me quite a lot of time to figure out. Can't
> remember ever having used sscanf in any code and I hate format
> conversion specifiers, all of them, in scanf *and* in printf. ;)

Yeah, that's 'cos they're a rich source of security bugs :-).