Fixed overflow detection in argument to long convertion function in general, and in expire/ttl pairs specifically, addressing issue 54

support for write operations against expiring keys, by master-controlled expiring in replication and AOF synthesizing DEL operations

Return an error when the resulting value is not a number (NaN). Fix
ZUNIONSTORE/ZINTERSTORE to clean up when a weight argument is not a
double value.

When +inf and -inf are added, the result is NaN. We don't want NaN
scores in a sorted set, so agreed on the result of this operation being
zero.

In order to trigger the bug what's needed is to call ZINTERSTORE
resulting into an empty set created, bug against a key that already
existed. The command was not propagated, so the replica ended with the
key that the master removed. Sequence of command to reproduce:

redis-cli hset 446 34 905
redis-cli hset 446 393 911
redis-cli zadd 966 0.085412045980529885 652
redis-cli zadd 645 0.25081839284432045 280
redis-cli zinterstore 446 2 966 645

use the function deprecated attribute if compiling with GCC to get warnings for malloc/free usages. We always want to use our zmalloc/zfree versions for memory usage tracking

calloc is more effecient than malloc+memset when the system uses mmap to
allocate memory. mmap always returns zeroed memory so the memset can be
avoided.  The threshold to use mmap is 16k in osx libc and 128k in bsd
libc and glibc. The kernel can lazily allocate the pages, this reduces
memory usage when we have a page table or hash table that is mostly
empty.

This change is most visible when you start a new redis instance with vm
enabled.  You'll see no increased memory usage no matter how big your
page table is.

zmalloc calls abort() so _dictPanic will never be called.

other shared objects where created in the I/O thread in createStringObjectFromLongLong. Fixed as well.

don't use object sharing inside I/O threads, as a fix for a well known instability of VM introduced with the new object sharing code

Fixed a crash loading the AOF file containing MULTI/EXEC, a result of WATCH implementation. Test needed...

WATCH will now consider touched keys target of EXPIRE command after the WATCH is performed, but not before

Fixed MONITOR output for consistency: now integer encoded values are also formatted like this: "3932"