[tor-commits] [tor/master] Coverity dislikes (double) (int/int).

[nickm@xxxxxxxxxxxxxx]

commit 4c90cdc0e7850a1a97ce064fc1496e6dc258fc0a
Author: Nick Mathewson <nickm@xxxxxxxxxxxxxx>
Date:   Mon Jun 13 11:25:19 2016 -0400

    Coverity dislikes (double) (int/int).
    
    When you divide an int by an int and get a fraction and _then_ cast
    to double, coverity assumes that you meant to cast to a double
    first.
    
    In my fix for -Wfloat-conversion in 493499a3399f8a8532b4b2a80006, I
    did something like this that coverity didn't like.
    
    Instead, I'm taking another approach here.
    
    Fixes CID 1232089, I hope.
---
 src/or/routerlist.c | 10 ++++------
 1 file changed, 4 insertions(+), 6 deletions(-)

diff --git a/src/or/routerlist.c b/src/or/routerlist.c
index 46492c5..aaa8fad 100644
--- a/src/or/routerlist.c
+++ b/src/or/routerlist.c
@@ -2176,22 +2176,20 @@ scale_array_elements_to_u64(uint64_t *entries_out, const double *entries_in,
   double total = 0.0;
   double scale_factor = 0.0;
   int i;
-  /* big, but far away from overflowing an int64_t */
-#define SCALE_TO_U64_MAX ((double) (INT64_MAX / 4))
 
   for (i = 0; i < n_entries; ++i)
     total += entries_in[i];
 
-  if (total > 0.0)
-    scale_factor = SCALE_TO_U64_MAX / total;
+  if (total > 0.0) {
+    scale_factor = ((double)INT64_MAX) / total;
+    scale_factor /= 4.0; /* make sure we're very far away from overflowing */
+  }
 
   for (i = 0; i < n_entries; ++i)
     entries_out[i] = tor_llround(entries_in[i] * scale_factor);
 
   if (total_out)
     *total_out = (uint64_t) total;
-
-#undef SCALE_TO_U64_MAX
 }
 
 /** Pick a random element of <b>n_entries</b>-element array <b>entries</b>,

_______________________________________________
tor-commits mailing list
tor-commits@xxxxxxxxxxxxxxxxxxxx
https://lists.torproject.org/cgi-bin/mailman/listinfo/tor-commits