Rename transactionv2 to transaction

[kshlm]

This commit does 3 things,

1. Move required types and functions from the old transaction package to the transactionv2 package.
2. Remove the old transaction package.
3. Rename the transactionv2 package to transaction

[Madhu-1]

21:05:04 --- FAIL: TestBitrot (18.01s) 21:05:04 --- FAIL: TestBitrot/Replica-volume (1.10s) 21:05:04 require.go:765: 21:05:04 Error Trace: bitrot_test.go:138 21:05:04 bitrot_test.go:69 21:05:04 utils_test.go:33 21:05:04 Error: Expected nil, but got: &errors.errorString{s:"strconv.Atoi: parsing "": invalid syntax"} 21:05:04 Test: TestBitrot/Replica-volume 21:05:04 --- FAIL: TestBitrot/Dist-volume (0.76s) 21:05:04 require.go:765: 21:05:04 Error Trace: bitrot_test.go:138 21:05:04 bitrot_test.go:110 21:05:04 utils_test.go:33 21:05:04 Error: Expected nil, but got: &errors.errorString{s:"dial unix /tmp/gd2_func_test/TestBitrot/w2/run/gluster/scrub-9b92170c67980722.socket: connect: connection refused"} 21:05:04 Test: TestBitrot/Dist-volume 21:05:04 === RUN TestBrickMux 21:05:14 --- FAIL: TestBrickMux (9.94s) 21:05:14 require.go:765: 21:05:14 Error Trace: brickmux_test.go:122 21:05:14 Error: Expected nil, but got: &errors.errorString{s:"os: process not initialized"} 21:05:14 Test: TestBrickMux 21:05:14 Messages: failed to kill bricks: os: process not initialized

@kshlm CI failure

[kshlm]

3 tests keep failing.

1. TestBitrot/Replica-volume
2. TestBitrot/Dist-volume
3. TestSnapshot/StatusAndForceActivate

TestBitrot/Replica-volume fails because the RPC done to the scrub daemon to fetch status, returns an empty dict. Even though the RPC is successful. I'm not sure why this happens yet.

TestBitrot/Dist-volume fails as the scrub daemon isn't running. GD2 isn't able to connect to the daemon to send the status request.

These two tests switch the failures depending on the order they run. The first test always fails because of empty dict, and the second fails because of failure to connect.

The Snapshot test fetches a list of snapshots, and the list returned is empty. This causes a out-of-index panic later when the first item in the list is accessed. This happens even though the previous tests are successful and snapshots should be present.

What is surprising to me is that all of these pass without this PR. This PR essentially doesn't change anything about the execution of transaction steps. The steps should still continue to execute as before and should work. I need to look into this more to get to the cause.

[atinmu]

retest this please