[
Date Prev][
Date Next][
Thread Prev][
Thread Next][
Date Index][
Thread Index]
[
List Home]
[stellation-res] Big changes - should we go for it?
|
I've got two potentially extremely large changes that I'd like to
propose.
First: I think that we should go directly for WVCM, and drop the interim
API.
After spending the last few months working on schemas and
semantics, I've learned a whole lot about how a WVCM-ish
repository should work. And now, looking at starting to implement,
I'm realizing that there may be no need for an interim API - once
you take what I proposed for the interim API, and make the necessary
adjustments for the SCL comm layer, you wind up with something
that is semantically, if not operationally, extremely close to WVCM.
WVCM is more complicated than the interim API - but not dramatically
so. Starting with basic support only for client workspaces, I think it
will
take a bit longer to do WVCM than to do the interim API, but because
they are operationally so different, I would guess that of the code
that implements the client API level of the system, less than 20%
would be reusable for WVCM. (The operational layer is really
quite different, but the same command language can be used
underneath both.)
So I see very little point in the interim API anymore - WVCM really is
better,
and I don't think that we'll gain much by switching from the current API
to an interim API, and then switching again from an interim to WVCM. I'd
rather just do the one switch.
I'm pretty convinced that this is the right way to go. Any objections?
The second big change that I'm proposing is a change in project
structure
in the self-hosting repository. I think that it was a mistake to make
each
of the Eclipse projects into a separate Eclipse project. With the way
that
we'll be changing the repository for WVCM, the number of Eclipse
projects is going to increase significantly - and most of them will
almost
always be checked in together.
So I'd like to switch to a single Stellation project in the repository,
with each
of the Eclipse projects as a subdirectory - so in essence, our current
Eclipse workspaces will become the SCM project in Stellation.
This one affects everyone equally, so I'd like to put it to a vote -
should
we switch to a single project? Or should we stick with what we're doing
now?
-Mark
Mark Craig Chu-Carroll, IBM T.J. Watson Research Center
*** The Stellation project: Advanced SCM Research
*** http://stellation.eclipse.org
*** Work: mcc@xxxxxxxxxxxxxx/Home: markcc@xxxxxxx