Could this be he problem that Brian encountered that on some compilers
they use strict assignmen rules and on others they don't? This may cause
the assignment to change the value when assigning to an unsigned long.
Technically, it's supposed when assigning a long to an unsigned which is
defined as unsigned = abs(signed). Just a thought.
Andy
On Tue, 9 Oct 2012, Gerardo Ganis wrote:
> Follow-up Comment #8, bug #97585 (project xrootd):
>
>
> Hi Brian,
>
> I've also managed to try that one and found the same.
> The problem is in the function
>
> int BN_cmp(const BIGNUM *a, const BIGNUM *b)
>
> The execution enters this function with two equal numbers but for some
> reason, which I have not yet found, the byte by byte check at the end of the
> function fails:
>
> for (i=a->top-1; i>=0; i--)
> {
> t1=a->d[i];
> t2=b->d[i];
> if (t1 > t2) return(gt);
> if (t1 < t2) return(lt);
> }
>
> The problem is in the assignments t2 = b->d[i]; t2 and b->d[i] are
> unsigned long.
>
> With 1.0.1c (the version run by Ubuntu) the problem disappears. Looking at
> the release notes it is not obvious which change could have fixed it (assuming
> that it is fixed ...); also a diff between 1.0.1c and 1.0.0j does not spots
> anything obvious.
>
> I'll continue debugging ...
>
> Gerri
>
>
> _______________________________________________________
>
> Reply to this item at:
>
> <http://savannah.cern.ch/bugs/?97585>
>
> _______________________________________________
> Message sent via/by LCG Savannah
> http://savannah.cern.ch/
>
> ########################################################################
> Use REPLY-ALL to reply to list
>
> To unsubscribe from the XROOTD-DEV list, click the following link:
> https://listserv.slac.stanford.edu/cgi-bin/wa?SUBED1=XROOTD-DEV&A=1
>
########################################################################
Use REPLY-ALL to reply to list
To unsubscribe from the XROOTD-DEV list, click the following link:
https://listserv.slac.stanford.edu/cgi-bin/wa?SUBED1=XROOTD-DEV&A=1
