At 12:14 AM 9/6/00 -0400, Bradley M. Kuhn wrote:
>Dan Sugalski wrote:
> > The decisions should be based on technical merit and general availability.
>I would include "available under a free software license" as part of the
>definition of "general availability".

You would, but in this case I don't. We can give anyone we want a client 
and get them access to the repository.

>For example, perforce doesn't ship standard with operating system
>distributions such as Red Hat, Debian, and FreeBSD, because it isn't under a
>free software license.  However, CVS does ship with those systems.

So? CVS is still inferior, and there are a *lot* of other operating systems 
in the world. Having to go fetch one program to check things in really 
isn't a hardship.

>Also, what if people want to learn how the source system works on their own,
>and experiment with it?

They can go do it on a test repository they set up on a scratch machine. 
I'd rather people not play with the p6 repository honestly--we're going to 
have enough problems of our own.

>With only 100-user license, we are pretty tight as
>to who can do that.  There are probably more than 100 people on the various
>perl6 mailing lists who'd want to at least experiment with the system on
>their own.

Anyone who wants it can get a 2-user license to perforce to set up and run 
wherever, which is just dandy. (And I don't want the perl repository being 
a test anyway) Ask can always get more licenses if he needs, so the 99-user 
current limit (I's like one of the licenses dedicated to anonymous 
read-only access) isn't a big deal.

>On another point, developers typically want to run checked-out versions of
>the software if they are actively participating.  I have always been sad I
>was unable to do this with perl5, even though I wasn't an active developer.

Huh? You've always been able to. The development tree (or one version of 
it) has always been available from activestate, and a perforce ID or 
anonymous access has been available for the asking.

>I would hate to see perl6 development go the same route.  The code would
>remain much more public and accessible if it were available via anonymous
>CVS.  Isn't that something we want?

There were plans to set up a read-only CVS mirror last I knew, and I don't 
think that's changed. The perl 5 tree's available via FTP and rsync, and I 
don't see any reason to change that for perl 6. And, honestly, if the effort

>One might argue that we can use a "snapshot" system can solve this problem.
>In my experience, The "snapshot" system can be quite inconvenient, for both
>those submitting patches and those accepting them.  It just doesn't beat
>anonymous CVS access.

That's fine. We set up anonymous perforce access, or people can use the CVS 

Either way it doesn't matter, since I feel strongly that we should be 
accommodating people working from a point-release version *anyway*. 
Compared to that, snapshots or anonymous repository access is easy going.

>Finally, most free software and open source projects have standardized on
>CVS.  Do we really have a compelling reason to go against the standard?

Perl 5 uses perforce, and perforce is a lot nicer. I don't much care what 
the other projects use--that's largely irrelevanyt, and in many ways a 
misleading argument. (It's the same sort of argument that pushes MySQL as a 
database since "most folks are using it", regarless of its inadequacy for a 

Perforce is much nicer for the people who need to manage the repository, 
and that's the important thing. We can get easy-enough access to the 
repository for the casual hacker and that's fine.


--------------------------------------"it's like this"-------------------
Dan Sugalski                          even samurai
[EMAIL PROTECTED]                         have teddy bears and even
                                      teddy bears get drunk

Reply via email to