On Sun, 2004-04-25 at 12:06, Tels wrote:

*snip accurate description*

> The current state is that dave takes chromatic's[0] works, fiddles with it a 
> bit, releases it as his own, then leaves chromatic to "backport" his 
> changes. In the meanwhile, chromatic works on *his* branch and adds things 
> that are not yet in Dave's version.

Right.  When I started the branch in Subversion, it was the only
branch.  No one had access to Dave's code in CVS, but Tels had some
patches that I incorporated.

Then Dave returned and gathered code he had in CVS and added some new
features from the Subversion branch, releasing that as 2.0.x.

> As I wrote: a mess which can partly be explained that both sides want to (or 
> can) work only on *their* branch. I hope that this can be solved somehow.

It's not quite that bad.  There's no svn client that Dave can use on his
Zaurus, so he has to pull changes from the svn-commit-modules-sdl at
perl.org list.  I don't have access to Dave's CVS, so I have to backport
patches from his release.

I've kept the version number in the snapshots as they were when I
started, because there's been no official release.  Effectively, it's
the same version as on the CPAN without the simplified objects.  (There
should be no API differences, though there may be some differences in
constant handling.)

My goal is still to make the build work on Win32 before applying that
patch.  I really don't want to debug failing tests due to two things at
once.

-- c

Reply via email to