Now I am able to commit my changes

[Brian T Rice]

On Fri, 9 May 2003, Paul Dufresne wrote:

> Well, I have added a small comment in Bag.slate, just to test if I was
> REALLY (I tried with -n -t, but wanted to see it for real) able to
> commit changes. And then I covered my traces by removing it with admin.
> :-) I was a bit surprised to see that I add admin access.
>
> I am not very at ease anyway with CVS. I did read a bit about branches,
> and maybe I would be more comfortable to add stuff in my own branch.
> Does it seems like a good idea? For minor bugs I am quite sure what I am
> doing, I think I would change directly on Main branch.

Minor bug-fixes should definitely be made on the main trunk.

> Some opinions on this?

It depends on the nature of the work you want to do. What makes this a bit
sensitive is that the source code right now is pretty much divided between
Lee and myself; he's responsible for the run-time and compiler, etc., and
I'm responsible for user-land. We often work on each other's code in minor
ways, but we generally defer to each other on the design decisions.

Basically, I'd just like to discuss it before you make a branch. If it
involves some different kind of design of the existing code, it'd be
important for us to know what's going on ahead of time. If it's just an
extension, that's probably fine to put into the trunk; just keep in mind
that Lee and I do have opinions about what fits in to the overall design
philosophy and what doesn't. We're not mean, but this is a unique project
that is already a branch of the Smalltalk family tree, and it didn't
happen in a haphazard manner. :)

If it's an entirely new library, at least let us know since I can point
out probably 2 or 3 libraries for every set of functionality, along with
some recommendations as to what would be most worth porting and how to
port it to PMD properly. The dimensioned units library for example was an
interesting example of porting that was not obvious in how to proceed.

Oh yes, and also remember that we're also looking hard at what various
Lisp systems provide when we look at a library's design, so if you're not
looking there we would probably have recommendations just based on that.

-- 
Brian T. Rice
LOGOS Research and Development
http://tunes.org/~water/