On Jul 5, 2006, at 12:25 AM, John Sisson wrote:
Jacek Laskowski wrote:
On 7/3/06, Jason Dillon <[EMAIL PROTECTED]> wrote:
NOTE... the m2 build in trunk is already broken... this patches help
FIX MANY OF THOSE PROBLEMS!
NOTED, but... it's not broken. it has never worked so we can pretend
to call it broken. It's a small, but important point we cannot
dismiss.
Since the official build is still m1 and this will not affect the m1
build, I don't see why your point about breakage is applicable at
all.
...
When I first created the m1 build for Geronimo years ago there were
certainly a few moments of breakage due to build changes, but since
there was no commit by committee junk going on then it was easy to
just fix when things happened to get a bit askew.
The branch idea was just to make it easier to actually make
progress,
as I am move on this stuff way way faster than the lot of you can
react to emails and JIRAs which often (as this one did) need several
sets of emails to clarify.
That's the point in RTC - discussing, discussing, over and over
again.
I'm not in favour of RTC, but some of its rules are fine. It fosters
discussions we lacked. That's the main point of RTC. Isn't it funny
that you've mentioned it as an argument against RTC?
What's wrong with committing changes made in the branch back to trunk
once they've been tested? My proposal is not to wait until the
migration is done, but rather apply it in small portions, gradually.
It should work, shouldn't it? I'd greatly appreciate your comment on
it as I guess I don't see the whole picture and keep thinking the
branch might help when others have already seen it would fall short.
Can we avoid the concerns that have been aired regarding svn
merging issues when directories are reorganised by leaving the
reorganization of directories as a last phase of the m2 migration?
I would have thought that we could move further along with the
migration without reorganizing directories (AFAIK, maven should be
able to work with existing directory structures, although doing so
may incur more work). We would also need to coordinate the
reorganization of directories with the owners of other branches
from trunk, to minimize the impact on them.
I would prefer to wait to reorganize the directories until after the
work in the dead-1.2 branch is merged with trunk. I plan to go back
to this activity now. Other committers may wish to note that merging
the work in dead-1.2 should not need RTC as it is already part of a
main development line.
thanks
david jencks
John
--jason
Jacek
