Re: [PATCH 2/8] revert: decouple sequencer actions from builtin commands

Ramkumar Ramachandra wrote:

                                                            Once the
painful move to sequencer.c is completed, we can think about all these
things.

Honestly, moving code verbatim between files is very easy.  Repeatedly
rebasing a patch that carries out such a move would presumably be
hard, though.  But this pain is unnecessary!

Just like I haven't been reviewing the code movement, I'd be perfectly
happy to read a "patch" that says

	"And then we move the functions from the following list to
	sequencer.c.  I'll send a patch doing so once work has settled
	down in patches earlier in this series."

Now you are telling me that in the super-final future my worries are
valid, but I should forget about them today, because later in this
series there is some code movement.  That we need to get this painful
part over with.  I would be much more comforted if you said that in
the future my worries were _not_ valid, that the current design is a
good one, and that these patches are not making the program worse;
otherwise, wouldn't it be better to skip whichever are the
questionable patches and just carry out the code movement, which
doesn't depend on them?