On Mon, 2006-09-04 at 02:25 +0200, Henrik Nordstrom wrote: > mån 2006-09-04 klockan 09:21 +1000 skrev Robert Collins: > > > Right now, 3.0 is the top priority. However, I think the big step is > > saying 'using the STL is allowed' - we should always be considering any > > individual change with rigour - and doing a single large branch is going > > to be much harder than just having a series of branches which say 'here > > is an improvement X, and it happens to use this STL feature.' > > Again, I am fine with forking 3.0 pretty soon to allow for these things > to get started in an incremental manner if you prefer doing further > refactoring instead of focusing on bugfixes.
I think the best use of the time I have available is in improving the testing & development facilities for squid, rather than working on new features. So yes, refactoring is part of that. > I very much prefer this kind of transition to be done incrementally over > time instead of a big batch. Agreed. > I don't agree that forking would seriously further delay 3.0 unless the > developers wanting to see 3.0.STABLE can't focus. People wanting to work > on new features is likely to do that anyway rather than fixing bugs, and > I rather have those efforts collected incrementally (with due review) in > HEAD than floating around and bitrotting elsewhere like we have done in > the past.. I'm not claiming that forking for 3.1 as a new trunk will delay 3.0 unless everyone switches ;). What I'm really trying to say is 'lets get 3.0 out the door!'. -Rob -- GPG key available at: <http://www.robertcollins.net/keys.txt>.
signature.asc
Description: This is a digitally signed message part
