basho/riak_test

A return of `fail` should actually report failure

A wrapper must be used to convert a test's return to an exit code.
Otherwise the returned term is ignored and tests that return 'fail'
are treated as passing.

Add a test that always fails to make sure that the pass/fail logic is
working.

Check git exit code and fix pipe cleanup

Recently Scott was running into an issue running `verify_handoff`
where his old data was not being properly reset when running
`setup_harness`. I noticed we were using `os:cmd` which doesn't check
the exit code of the command. I modified `run_git` to use `cmd` as
well as verify the exit code is 0. This allowed Scott to catch the
real issue which turned out to be a bad path in his config.

While making this modification I noticed a bug in the pipe cleanup
code. The `file:del_dir` call is actually returning `{error, eexist}`
because there is a `bin` directory under each pipe dir which had not
yet been deleted. Rather than spend time writing a recursive delete in
Erlang I changed the code to use `cmd` and to confirm an exit of 0.

I modified `stop_all`, which is used by `setup_harness`, to also use
`cmd` and check exit codes.

Finally I make sure that `spawn_cmd` flattens the list before passing
it along as `open_port` wants a string not an iolist.

add bash/zsh completion for test names

Add option to allow tests in a directory to be explicitly skipped

Conflicts:
	src/riak_test_escript.erl

Conflict was a simple whitespace issue.

Need to activate a bucket type before it can be used. Use the function
built into `rt` for creating bucket types.

Test new MR Search Switch and "index" spec

The "bucket" is the "index".  Chalk it up to poor naming make a long
time ago.  We're only human.

Add ability to upgrade and update app cfg

Faster wait_for_cmd that preserves order

Remove the receive clause for port data.  This can cause
`wait_for_cmd` to needlessly loop through all stdout msgs before
getting to the port exit.  By just doing selective recv on the exit
msg Erlang will implicit use the "save queue" and place the data msgs
back on the mailbox in the original order of arrival.  In fact, the
previous method could rearrange the stdout ordering.

Add the ability to upgrade node and update its app.config in one go.
This is useful when testing upgrade path on Yokozuna where the old
app.config should be copied over but modifications should also be made
before restarting the node.

Add code coverage analysis to riak_test

Update verify_capabilities for Riak 1.4/1.3/1.2

This allows user to view logs from last verify run.

Add upgrade test for Riak Control

Some issues have been found when running a mixed cluster with Riak
Control enabled.  This tests the typical upgrade scenario of a
customer and verifies that Riak Control at least doesn't crash any of
the nodes while in a mixed cluster state.

See basho/riak_control#120

* My machine is too fast, changed timeout to 1ms.

* Fixed match of return from `http_query`.

* The proplists key is binary, not an atom (i.e. <<"keys">>).

While running the `verify_staged_clustering` test on my local
workstation I hit a race when force replacing dev3 with dev2.  The
call to 'cluster plan' was happening before the ring had converged and
thus failing, but the test assumed the first call always succeeds.

This patch simply mimicks what a user might do and retry the 'cluster
plan' command over and over until it succeeds.

Pull riak http client from master

Need to pull from master or else a version mismatch occurs for the
protobuffs client.

Intercepts provide the ability to easily and efficiently intercept
function calls.  Giving the ability to change the code being executed
as well as affect local and global state.