sscanf() %n Bug Introduced in 20000221 Snapshot

[Mark Rahner]

Before making my initial post, I verified that the bug was still present in the 20000309
snapshot.  I've cc'ed this message to the newlib address you suggested.

-Mark

Chris Faylor wrote:

> sscanf comes from the newlib library.  Try sending email to newlib@sourceware.cygnus.com.
>
> Also, try a new snapshot.
>
> cgf
>
> On Thu, Mar 16, 2000 at 04:03:50PM -0500, Mark Rahner wrote:
> >I apologize for sending an HTML formatted message last time.  Here's the
> >plain text version...
> >
> >This message documents a problem with sscanf() %n format specification
> >which began, to the best of my knowledge, with the 20000221 snapshot.
> >This problem did not exist in the 20000211 snapshot.  I'm using
> >Windows98 and Cygwin Beta 20.1 with upgrades consisting of Mumit's
> >2.95.2 compiler and the 200000221 snapshot.
> >
> >Compiling the following using "g++ sscanfBug.C -o sscanfBug.exe"
> >
> >     #include <stdio.h>
> >     #include <iostream>
> >
> >     int main () {
> >
> >         const char *str = "0.02";
> >         double num;
> >         int len;
> >
> >         cerr << "str is <" << str << ">\n";
> >
> >         sscanf( str, "%lf%n", &num, &len );
> >
> >         cerr << "num is <" << num << ">\n";
> >         cerr << "len is <" << len << ">\n";
> >     }
> >
> >produces an executable which, when run, writes the following to standard
> >error:
> >
> >     str is <0.02>
> >     num is <0.02>
> >     len is <2>
> >
> >The correct value for len is 4.  Please note that this is not the same
> >problem discussed on this mailing list last June.  Specifically, there
> >is no space preceding %n.  I apologize in advance for not presenting you
> >with a patch; after briefly attempting to track this bug myself, I
> >realized that I wasn't getting anywhere.
> >
> >Hope this helps,
> >Mark Rahner
>
> --
> Want to unsubscribe from this list?
> Send a message to cygwin-unsubscribe@sourceware.cygnus.com