[PATCH] Change pow(, 0.5) -> sqrt() in osscan[2].cc

[Kris Katterjohn <kjak () ispwest com>]

The attached patch just uses sqrt() instead of pow(,0.5) on everything
instead of just Linux in osscan.cc and osscan2.cc. Apparently (judging
by the comment in osscan.cc), sqrt() is used on Linux because pow() is
broken in some libc versions. sqrt(X) is the same as pow(X,0.5), so we
can just use sqrt() and avoid the #ifdef.

It's a diff against 4.20ALPHA6

Thanks,
Kris Katterjohn

--- x/osscan.cc 2006-08-28 22:26:10.000000000 -0500
+++ y/osscan.cc 2006-09-02 22:04:38.000000000 -0500
@@ -959,13 +959,7 @@ static FingerPrint *get_fingerprint(Targ
        /*       printf("The sequence sum is %e\n", seq_inc_sum);*/
        seq_inc_sum /= (si->responses - 1);
 
-       /* Some versions of Linux libc seem to have broken pow ... so we
-          avoid it */
-#ifdef LINUX       
        si->index = (unsigned int) (0.5 + sqrt(seq_inc_sum));
-#else
-       si->index = (unsigned int) (0.5 + pow(seq_inc_sum, 0.5));
-#endif
 
        /*       printf("The sequence index is %d\n", si->index);*/
        if (si->index < 75) {

--- x/osscan2.cc        2006-09-02 20:11:42.000000000 -0500
+++ y/osscan2.cc        2006-09-02 22:04:45.000000000 -0500
@@ -1735,11 +1735,7 @@ void HostOsScan::makeTSeqFP(HostOsScanSt
       seq_stddev /= hss->si.responses - 2;
 
       /* Next we need to take the square root of this value */
-#ifdef LINUX       
       seq_stddev = (unsigned int) (0.5 + sqrt(seq_stddev));
-#else
-      seq_stddev = (unsigned int) (0.5 + pow(seq_stddev, 0.5));
-#endif
 
       /* Finally we take a binary logarithm, multiply by 8, and round
         to get the final result */


_______________________________________________
Sent through the nmap-dev mailing list
http://cgi.insecure.org/mailman/listinfo/nmap-dev
Archived at http://SecLists.Org