[ODE] Proposal: ODE "unstable"

[Russ Smith]

BTW, all of this information has been added to the wiki:

  http://ode.org/cgi-bin/wiki.pl?HowToGetYourPatchCheckedIn

russ.

--
Russ Smith
http://www.q12.org/

On Mon, 14 Mar 2005, Russ Smith wrote:

>
> hi,
>
> let me try and address some of these concerns. ODE is a slowly moving
> project right now, as all of the developers have a (i guess) lot of other
> things in their lives besides ODE :) - however, there's a few things that
> can be done to get patches into CVS faster. here's what i've done:
>
>   * created an UNSTABLE branch on CVS. this is where untested, potentially
>     dodgy code should go. the UNSTABLE branch (or parts of it) can be
>     merged to the more stable trunk periodically, after a sufficient
>     period of testing. obvious bug fixes can continue to go on the trunk.
>
>   * i you want to check in your patch, but think that the project
>     developers are being too slow, ask to get 'developer' status on
>     sourceforge (make sure to include your sourceforge user id in your
>     request). as a developer, you can check in your change yourself!
>     yes, this is more work for you, but it's less work for the project
>     admins (you might think it takes just ten seconds to check in a
>     patch, but that true only if you are checking in 100 at a time).
>
>   * when you check in your patch, remember to post a message to this list
>     saying what you've done and why.
>
>   * what about the security and code-stability issues of having too many
>     people with CVS access? well, i'm not too worried about malicious
>     people who try to trash the project - (1) the ODE community by and
>     large has been extremely well behaved compared with other open source
>     projects (pat yourselves on the back), (2) there's a CVS mailing list
>     that publishes new patches so there's plenty of visibility, and (3)
>     CVS never forgets, so it's easy enough to recover from any problems.
>     if this free-for-all turns out to be a problem (and i don't think it
>     will), we can always try something else.
>
>   * i've upgraded four developers to admin status (adam moss, jeffrey
>     smith, martin martin and nate waddoups). all project admins can
>     add new developers. so if you want to have developer status to check
>     in your patch, there should be *somebody* who can help you.
>
> on the issue of subversion, ODE won't be using it anytime soon. i agree
> that it's better than CVS, but CVS is what sourceforge provides right now,
> so changing over would be more trouble than it's worth.
>
> russ.
>
> --
> Russ Smith
> http://www.q12.org/
>
> On Mon, 14 Mar 2005, Tanguy Fautre wrote:
>
> >
> > I totally agree,
> >
> > I've posted two patches last month, one to get ODE to compile on QNX and
> > another to get OPCODE working on AMD64, and none of them ever made it to
> > the repository.
> >
> > I've been using Subversion for a year now, I never had any problem with
> > it. So I'm all for it.
> >
> >
> > Regards,
> >
> > Tanguy
> >
> >
> > Colin Bonstead wrote:
> > > It seems like there hasn't been much progress on ODE lately, despite the
> > > fact that there are lots of people using it and willing to submit
> > > patches.  Part of this seems to be because the maintainers don't want to
> > > apply any patches that aren't fully implemented, or have some bugs.
> > > Instead they go into the contrib folder, or don't ever get added.  The
> > > problem is, very few people are going to bother applying the patches
> > > from the contrib folder, so the bugs or missing features are never going
> > > to get taken care of.
> > >
> > > What I'm proposing is having an "unstable" tree too, which has all the
> > > contributions in it.  People can build that tree, fix bugs, and things
> > > can eventually get promoted to the main tree.
> > >
> > > I personally submitted a patch for trimesh preprocessing way back in
> > > June of last year.  Here's the mail:
> > >
> > > http://q12.org/pipermail/ode/2004-June/013183.html
> > >
> > > Nothing ever happened with it, probably since I only added support for
> > > it to the ccylinder collider (although any collider that doesn't have
> > > support should just behave as it always has).  I can confirm this
> > > optimization does work since we're using it in Myst V.
> > >
> > > While I'm throwing things out here, I might as well also suggest that
> > > ODE switch to Subversion instead of CVS.  We've been using Subversion
> > > here at Cyan for almost a year with no problems, and it's WAAAY better
> > > than CVS.
> > >
> > >
> > >
> > > _______________________________________________
> > > ODE mailing list
> > > ODE at q12.org
> > > http://q12.org/mailman/listinfo/ode
> > >
> >
> > _______________________________________________
> > ODE mailing list
> > ODE at q12.org
> > http://q12.org/mailman/listinfo/ode
> >
>
>
> _______________________________________________
> ODE mailing list
> ODE at q12.org
> http://q12.org/mailman/listinfo/ode
>