On Sun, 2002-04-21 at 10:47, Marc wrote: > Andrew C. Oliver wrote: > > >Hi All and especially those of you who ride Kangaroos to work everyday, > > > >It occurs to me that some of the things I'm about to do are heavily > >HSSF-2.0 oriented and yet I'm not yet sure we'll be ready for 1.5. > >Everytime I think we're RealClose(tm) a Gigantic moth sized anywhere > >from little to gigantic flies into the works and gums them all up. So > >"to branch or not to branch" that is the question! > > > >Secondly, after we release 1.5, we'll undoubtedly have a 1.5.1 or > >something of the such to capture bug fixes/etc. So the question before > >you: > > > >do we create a branch for 1.5 and continue all bug fixes there while > >continuing further development on the head. > > > >I vote +1 - if Glen doesn't have the necessary CVS-foo I'll fake it. ;-) > > > >Later votes or discussions can determine the status of particular pieces > >of code. I say all formula related stuff goes in the head and not in > >1.5. > > > >-Andy > > > There are certain practices that I feel must be justified to the > satisfaction of everyone actively on a project -- deliberately violating > encapsulation is one example. Branching is another. It can be (and > usually is) a collossal headache and it is something that I will > grudgingly admit *might* be an acceptable practice after a major > release, to allow for bug fixes while new development continues. > > This does not sound like such a case. > > So, I vote -1. > > I can be persuaded by a convincing argument to retract that vote. Such > an argument should include a restatement of our 1.5 objectives and why > 1.5 is being treated as if it were a major release, and why the new work > cannot be performed while 1.5 bug fixes are applied. Also, how about an > explanation of what that new work is? >
1.5 - we've got major bugs and minor bugs and the big packaging issue that is not captured in 1.0.2 development builds should have formulas enabled... 1.5 should not. I'm queuing patches that will make major changes to formula support. Libin's patches for Named ranges break some charting unit tests (no idea why yet) and I really don't think those should go in a "production" release. 1.5 is intended to be a production release. That being the case the inevitable 1.5.1 should be a production release. Development builds will start to have things in them that are totally NOT ready for production (formula stuff especially) and 1.5.1 should not capture those changes. We're only talking about 1 branch. Not two. And only MINOR changes will happen to the 1.5 branch...bugfixes, etc. (because the dev work will continue at the HEAD)/ Thoughts? > Not trying to be a PITA (when it comes naturally, there's no need to > *try*), but this is a rather serious action we're contemplating, and I > cannot endorse it, indeed, I must oppose it, absent serious discussion. > -- http://www.superlinksoftware.com http://jakarta.apache.org/poi - port of Excel/Word/OLE 2 Compound Document format to java http://developer.java.sun.com/developer/bugParade/bugs/4487555.html - fix java generics! The avalanche has already started. It is too late for the pebbles to vote. -Ambassador Kosh
