qwait() interface has changed, and silently fails if used the "old" way

[dellard]

If the kind parameter is set to something invalid, qwait() returns None, but doesn't raise any error or warning to let the user know that the result is bogus.

The reason why the kind parameter might be set to something invalid is that the interface has changed. The order of the kwargs has changed, so any code that specified the actual parameters positionally is now bind junk (probably qubit references) to the kind parameter. This showed up in the pyqgl2 tests, which used positional notation. I don't know if this bug appears anywhere else, but it might have been fixed everywhere else already since the change to qwait() is several months old.

[dieris]

Closed by #186?

[dellard]

Now that #186 is closed, a blatantly incorrect use of qwait() should raise an error.

I don't know if we want to spend any effort to search through the code to find any such errors (which was the original intent of this ticket) or just wait for the errors to appear. The fact that nobody noticed this problem recently suggests that this problem was already fixed everywhere except the pyqgl2 code, so maybe it's not worth worrying about.