I think there's a solution to RTC and having a place for experiments
like this where nothing's known ahead - a branch. With a branch you
can do whatever you want and no RTC rules apply there. I think it
would help us all. Interested? Count me in! ;-)
I don't really consider this work experimental... gshell is
experimental, but the m2 work that I am doing is just the application
of experience with this system and other build systems for the past
years (and years).
But... lets see what others have to say.
I may create an experimental branch to see how well SVN actually
handles merges and remerges of an entire (massive) branch. And if
that ends up being relatively feasible then I would consider creating
a branch for the remaining m2 work.
Though... even with a branch, we would have to RTC to merge back to
trunk, which may take several weeks... which is not acceptable IMO.
--jason