dccl-dev team mailing list archive

[Nathan Knotts <nknotts@xxxxxxxxx>]

Nathan Knotts has proposed merging lp:~nknotts/dccl/fill_n into lp:dccl.

Requested reviews:
  DCCL Developers (dccl-dev)

For more details, see:
https://code.launchpad.net/~nknotts/dccl/fill_n/+merge/251376

In http://bazaar.launchpad.net/~dccl-dev/dccl/3.0/revision/314, I used std::memset to zero-initialize the byte string. This compiled and worked for me on OS X/clang 3.5. I began testing on Ubuntu 14.04/Centos7 (gcc 4.8) and the std::memset was not found because of a missing <cstring> include. Instead of adding the <cstring> include, I switched to using std::fill_n because <algorithm> is already included and std::fill_n is a slightly higher level call. Also, this stack overflow post, http://stackoverflow.com/questions/1373369/which-is-faster-preferred-memset-or-for-loop-to-zero-out-an-array-of-doubles, seems to prefer std::fill as it is more modern/C++ then std::memset.
-- 
Your team DCCL Developers is requested to review the proposed merge of lp:~nknotts/dccl/fill_n into lp:dccl.

=== modified file 'src/bitset.h'
--- src/bitset.h	2015-01-26 14:51:34 +0000
+++ src/bitset.h	2015-02-28 23:01:29 +0000
@@ -333,7 +333,7 @@
             assert( max_len >= len );
 
             // initialize buffer to all zeroes
-            std::memset(buf, 0, len);
+            std::fill_n(buf, len, 0);
 
             for(size_type i = 0, n = this->size(); i < n; ++i)
                 buf[i/8] |= static_cast<char>((*this)[i] << (i%8));