gump-general mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Adam R. B. Jack" <aj...@trysybase.com>
Subject Re: HEADS UP: (was Re: Gump and SVN: Any takers for candidate project?)
Date Tue, 25 Nov 2003 14:38:35 GMT

> Just go ahead.  I still run "traditional gumps" on my machine (and
> think I'll need to update my svn client to talk to the Apache
> installation), so I can tell you.

Ok, it's in. Yup things oughtn't break to seriously.

Here is the rub, though. To try to make SVN handling similar to CVS handling
I add a 'dir' attribute onto the SVN entry in the project, so <svn
repository='apache-svn-sandbox' dir='testdir' gives:

        http://svn.apache.org/repos/test/testdir/

As I've said, I've tried avoiding making GOM changes so as not to make
traditional Gump stumble, but what about this sub-directory? Without work on
traditional Gump, all SVN projects will certainly fail (and also draw down a
lot of extra stuff from an SVN root.) Also, other attributes could be added,
and I really ought clean-up <cvsweb [the viewcvs link] on <repository since
it isn't a fit for SVN.

I like traditional Gump to exist as a comparison run, but clearly don't want
to force such work, 'cos that is counterproductive to Python Gump taking
over.

Perhaps I need to create a new complete profile (profile/gumpy.xml for want
of a better name) and only do new stuff in there. This would probably work
for new SVN projects, but what about other metadata objects (should the need
arise to tweak them). Any thoughts on all this?

regards,

Adam


Mime
View raw message