Re: Bug in _VFPRINTF_R

[Samuel Vinson <samuelv at laposte dot net>]

Yes you are right, I forgot to mention len = 0.
I'm sorry.

Samuel

Matt Gessner a écrit :
> First, apologies because the first email I posted in response was sent before I finished editing it.
>
>   
> > I'm sorry Eric, but I gave you a specific example :
> >     printf("%.*s\n",len,(char *)NULL) /* len = 0 */
> >
> > and not printf("%s", (char *)NULL)
> >
> >     
>
> Actually, you did NOT use that as your example.
>
> You posted:
>
> *******************************************************************
> Hi,
>
> In newlib printf, snprintf (perhaps others string functions) use 
> _VFPRINTF_R.
> So if you call :   
>     printf("%.*s\n",len,(char *)NULL)   
>     snprintf(ptr, 40, "%.*s", len, (char *)NULL)
> You obtain :
>     (null)
>     ptr = "(null)"
>
> *******************************************************************
>
> Never did you mention anywhere that len == 0.
>
> That drastically changes the meaning of the code you originally posted, and while I don't have the code in front of me, the len == 0 case is not going to be processed in the precise section where you originally posted your change.
>
> FWIW, I think your suggested behaviour makes sense, for snprintf, but not for printf.
>
>   
> > In my example, I think the result should be empty an not (null)
> >
> > Samuel
> >
> > Eric Blake a écrit :
> >     
> > > -----BEGIN PGP SIGNED MESSAGE-----
> > > Hash: SHA1
> > >
> > > According to Samuel Vinson on 9/15/2007 1:35 AM:
> > >
> > >       
> > > > I never see that with different libc implementation, I used
> > > > (linux/gcc/libc/glibc, windows/msvcrt, solaris/libc...)
> > > >
> > > >         
> > > It may not match msvcrt or solaris libc, but it does match glibc - I'm
> > >       
> > not
> >     
> > > sure how you tested it.
> > >
> > >
> > >       
> > > > Perhaps POSIX and C99 allow your choice, but this is not portable in
> > > > different platforms.
> > > >
> > > >         
> > > This is not portable, period.  Calling printf("%s", (char*)NULL) invokes
> > > undefined behavior, and is a bug in your program, not in newlib.
> > >
> > > - --
> > > Don't work too hard, make some time for fun as well!
> > >
> > > Eric Blake             ebb9@byu.net
> > > -----BEGIN PGP SIGNATURE-----
> > > Version: GnuPG v1.4.5 (Cygwin)
> > > Comment: Public key at home.comcast.net/~ericblake/eblake.gpg
> > > Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org
> > >
> > > iD8DBQFG68gr84KuGfSFAYARAr2PAKCR56DlvOvhKZXMeDn+ecmYO7fXEQCeMEUV
> > > kGk4bBxoFTk2hpRD3TiakLM=
> > > =LIb4
> > > -----END PGP SIGNATURE-----
> > >
> > >
> > >
> > >       
>
>
>
>