Add utility riak_array

[hazen]

• Also run integration tests only when command-line argument is specified
• Used as a key datatype in upcoming connection pool implementation

[bookshelfdave]

going to Valgrind this in BOS

[hazen]

My test with valgrind on riak_c_cunit doing ONLY the riak_array suite shows leaks only in libevent:

==18105==
==18105== HEAP SUMMARY:
==18105==     in use at exit: 216 bytes in 6 blocks
==18105==   total heap usage: 96 allocs, 90 frees, 17,916 bytes allocated
==18105==
==18105== 32 bytes in 1 blocks are still reachable in loss record 1 of 6
==18105==    at 0x4C2C857: malloc (vg_replace_malloc.c:291)
==18105==    by 0x54A6E79: debug_lock_alloc (evthread.c:161)
==18105==    by 0x54A6A42: event_global_setup_locks_ (event.c:2882)
==18105==    by 0x409DCE: main (registry.c:56)
==18105==
==18105== 32 bytes in 1 blocks are still reachable in loss record 2 of 6
==18105==    at 0x4C2C857: malloc (vg_replace_malloc.c:291)
==18105==    by 0x54A6E79: debug_lock_alloc (evthread.c:161)
==18105==    by 0x54B87A3: evsig_global_setup_locks_ (signal.c:442)
==18105==    by 0x54A6A55: event_global_setup_locks_ (event.c:2884)
==18105==    by 0x409DCE: main (registry.c:56)
==18105==
==18105== 32 bytes in 1 blocks are still reachable in loss record 3 of 6
==18105==    at 0x4C2C857: malloc (vg_replace_malloc.c:291)
==18105==    by 0x54A6E79: debug_lock_alloc (evthread.c:161)
==18105==    by 0x54B6953: evutil_secure_rng_global_setup_locks_ (evutil_rand.c:121)
==18105==    by 0x54A6A67: event_global_setup_locks_ (event.c:2886)
==18105==    by 0x409DCE: main (registry.c:56)
==18105==
==18105== 40 bytes in 1 blocks are still reachable in loss record 4 of 6
==18105==    at 0x4C2C857: malloc (vg_replace_malloc.c:291)
==18105==    by 0x5291DB9: evthread_posix_lock_alloc (evthread_pthread.c:46)
==18105==    by 0x54A7527: evthread_setup_global_lock_ (evthread.c:358)
==18105==    by 0x54A6A42: event_global_setup_locks_ (event.c:2882)
==18105==    by 0x5291F5F: evthread_use_pthreads (evthread_pthread.c:185)
==18105==    by 0x409DD3: main (registry.c:57)
==18105==
==18105== 40 bytes in 1 blocks are still reachable in loss record 5 of 6
==18105==    at 0x4C2C857: malloc (vg_replace_malloc.c:291)
==18105==    by 0x5291DB9: evthread_posix_lock_alloc (evthread_pthread.c:46)
==18105==    by 0x54A7527: evthread_setup_global_lock_ (evthread.c:358)
==18105==    by 0x54B87A3: evsig_global_setup_locks_ (signal.c:442)
==18105==    by 0x54A6A55: event_global_setup_locks_ (event.c:2884)
==18105==    by 0x5291F5F: evthread_use_pthreads (evthread_pthread.c:185)
==18105==    by 0x409DD3: main (registry.c:57)
==18105==
==18105== 40 bytes in 1 blocks are still reachable in loss record 6 of 6
==18105==    at 0x4C2C857: malloc (vg_replace_malloc.c:291)
==18105==    by 0x5291DB9: evthread_posix_lock_alloc (evthread_pthread.c:46)
==18105==    by 0x54A7527: evthread_setup_global_lock_ (evthread.c:358)
==18105==    by 0x54B6953: evutil_secure_rng_global_setup_locks_ (evutil_rand.c:121)
==18105==    by 0x54A6A67: event_global_setup_locks_ (event.c:2886)
==18105==    by 0x5291F5F: evthread_use_pthreads (evthread_pthread.c:185)
==18105==    by 0x409DD3: main (registry.c:57)
==18105==
==18105== LEAK SUMMARY:
==18105==    definitely lost: 0 bytes in 0 blocks
==18105==    indirectly lost: 0 bytes in 0 blocks
==18105==      possibly lost: 0 bytes in 0 blocks
==18105==    still reachable: 216 bytes in 6 blocks
==18105==         suppressed: 0 bytes in 0 blocks
==18105==
==18105== For counts of detected and suppressed errors, rerun with: -v
==18105== ERROR SUMMARY: 0 errors from 0 contexts (suppressed: 2 from 2)

I've looked in the source and the documentation, but I don't see any obvioius "clean up global locks" function.

If I comment out ALL libevent from the unit test:

==18536==
==18536== HEAP SUMMARY:
==18536==     in use at exit: 0 bytes in 0 blocks
==18536==   total heap usage: 90 allocs, 90 frees, 17,700 bytes allocated
==18536==
==18536== All heap blocks were freed -- no leaks are possible
==18536==
==18536== For counts of detected and suppressed errors, rerun with: -v
==18536== ERROR SUMMARY: 0 errors from 0 contexts (suppressed: 2 from 2)

[bookshelfdave]

looks good

[bookshelfdave]

bah, feel free to ignore my doc return type comments. Like you say, they'll probably just get out of sync with the code.

============================================================================
Testsuite summary for riak_c_client 0.5
============================================================================
# TOTAL: 1
# PASS:  1
# SKIP:  0
# XFAIL: 0
# FAIL:  0
# XPASS: 0
# ERROR: 0
============================================================================

[bookshelfdave]

:+1: