State machine logic broken when appending

[rconstanzo]

Steps to reproduce:
Start playing back loop, then go into append mode, then try to end append mode with play (or stop).

Actual results:
The append state machine gets broken and non-responsive.

Expected results:
Play should carry on playing from wherever the playhead is. stop should define the new ‘initial loop length’ and stop the state machine.

[rconstanzo]

I didn't check this with 08c108b but as of 1209be5 this is kind of working, kind of not.

It appears to work fine if you are going from play to append and back again. But if you are in a stop-ed state and send an append message, this is half working.

Steps to reproduce:
Record a loop and stop playback. Now send append message ([karma~] should now be 'primed' for initial loop recording (appending)). Now send a play message.
(following append with record works perfectly fine)

Actual results:
[karma~] starts playback from the start of the loop, but remains in the initial loop 'append' state with no way out of this state.

Expected results:
I don't remember karma~ 1.0 behavior for this, and it's unclear from the original helpfile about this specific edge case, but one of two things should happen. Either it immediately leaves the append state and starts playback from the start of the loop OR it plays back as it does now, but allows you to then record into the loop as if you go into append mode from normal playback.

Comments:
This is kind of confusing and probably easier to talk about over Skype, but basically if you send append and then play, you are stuck there with no way out.

*also downgrading this to low as I don't really anticipate append-ing in the current project.