Re: asdbget - precision

Sean Porterfield wrote:

> Response at end...
>
> "Jason M. Felice" wrote:
>
> > On Wed, Mar 15, 2000 at 05:20:28PM -0500, Sean Porterfield wrote:
> > > Toward the end of this function, what is the purpose of "v = -v;" ?
> > > It's what is causing my precision to show as negative on some fields in
> > > my XML (I haven't yet found a pattern or reason).  Before I just comment
> > > it out, I wanted to see if it had a purpose.  It may well be indicitive
> > > of a hidden bug somewhere, of a misunderstanding due to OS400 version
> > > differences.
> > >
> > > (I'm on V4R2)
> > >
> > > /* Convert a signed decimal field to a C `int' */
> > > static int
> > > signed2int (buf, len)
> > >         Byte *buf;
> > >         int len;
> > > {
> > >   int v = 0;
> > >   while (len)
> > >     {
> > >       v *= 10;
> > >       v += buf[0] & 0x0f;
> > >       len--;
> > >       buf++;
> > >     }
> > >   if ((buf[0] & 0xf0) == 0xd0)
> > >     v = -v;
> > >   return v;
> > > }
> >
> > This converts an AS/400 signed field to an integer (we should probably be
> > converting to a long, but that doesn't matter on 32-bit architectures).  
>This
> > is IBM's zoned-decimal format, which stores one digit per byte and 
>zone-shifts
> > the last digit if the field is negative.
> >
> > Basically, ebcdic digits zero through nine are X'F0' - X'F9', but the last
> > digit is converted to X'D0' - X'D9' for a negative number.
> >
> > However, looking over it for the fifth time, the function is in error.  Try
> > this instead:
> >
> > /* Convert a signed decimal field to a C `int' */
> > static int
> > signed2int (buf, len)
> >         Byte *buf;
> >         int len;
> > {
> >   int v = 0;
> >   while (len)
> >     {
> >       v *= 10;
> >       v += buf[0] & 0x0f;
> >       len--;
> >       if (--len == 0 && (buf[0] & 0xf0) == 0xd0)
> >         v = -v;
> >       buf++;
> >     }
> >   return v;
> > }
> >
> > It seems we were checking the zone of a random garbage character (the first
> > character after the end of the field).  I've just fixed this in CVS.
> >
> > -Jay 'Eraserhead' Felice
> >
>
> CVS code doesn't have the len--; line in it.  When I keyed exactly as you 
>have it
> here, I got a seg fault/core dump when it was "RETR QTEMP/FIELDS"
>
> Using the code from CVS does not cause the crash.  I haven't checked for any 
>other
> problems yet.

Well, it doesn't crash that way, but it does still show negative numbers on my
precision in some fields.

argh!


+---
| This is the LINUX5250 Mailing List!
| To submit a new message, send your mail to LINUX5250@midrange.com.
| To subscribe to this list send email to LINUX5250-SUB@midrange.com.
| To unsubscribe from this list send email to LINUX5250-UNSUB@midrange.com.
| Questions should be directed to the list owner/operator: david@midrange.com
+---