|
Boost : |
From: Gennadiy Rozental (rogeeff_at_[hidden])
Date: 2008-07-18 23:14:39
Emil Dotchevski <emil <at> revergestudios.com> writes:
> If the motivation for branching is to prevent interim commits of work
> in progress, then I suggest stating *that* as requirement.
>
> OTOH, anything that hasn't been tested should be considered work in
> progress, and to test a change it has to be committed to Trunk -- not
> only because Trunk is the thing that gets tested, but also because
> many changes can produce ripple effects throughout Boost. Add to that
> the testing lag, and it's a given: Trunk is potentially not very
> stable.
For what it worth, I concur. Until we have an ability to run tests against
branch, trunk *is* a development area and I do not see an alternative.
Gennadiy
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk