We do all development on branches. The
trunk represents the code that is currently in production. All doc
I read is based on development going on in the trunk, but I don't see how
that can work when there are always 2 or 3 releases in progress at any
one time.
Each branch is merged into the following
branch on a regular basis.
So for example, we have an April patch
branch, a June branch, and an August branch in progress. At certain
points we merge April patch branch into June branch and June branch into
August branch.
When we create the October branch, it
is created from the trunk and then August branch will be merged into October
branch.
So here is my question ...
When I create October branch I know
that it should look exactly like August. So if I created October
branch from August branch instead of the trunk I would eliminate the August
to October merge. But then that implies that December branch would
be created from October, etc ... and that does not seem right.
So should I stick with the original
strategy and always branch off of the trunk ?
Thanks,
Tim Velten
IBM Global Services
859 635 1785
address@hidden