On Tuesday 02 April 2002 11:01, Geoffrey Talvola wrote:
> Kendall Clark wrote:
> > On Tue, Apr 02, 2002 at 01:15:23PM -0500, Geoffrey Talvola wrote:
> > > Actually, I don't mind if other major work goes into
> >
> > Webware before 1.0,
> >
> > > like Jay's multi-application version of WebKit or some of
> > > Tavis's experimental work.  I just want to make sure we're
> > > making
> >
> > progress toward a
> >
> > > release that we can call 1.0.
> >
> > Hey, that's pretty funny, Geoff.
> >
> > That's *exactly* what I wanted to say, i.e., I do hope that
> > some of the
> > experimental stuff makes it in before 1.0 (well, I don't care
> > about release
> > nubmers, per se; I just would like to see more of that stuff
> > migrate more
> > regularly). A good deal of the Tavis and Jay stuff strikes me
> > as *very*
> > cool, including the dependency stuff, ideas for a CacheKit,
> > some of the url
> > traversal changes, etc.
> >
> > I just figured Chuck was prolly dead set against it. :>
> >
> > As an outsider, there does sometimes seem to be some kind of
> > slightly odd
> > tension between the experimental stuff and the 'mainstream'
> > Webware -- at
> > least, a tension among some of the various proponents about
> > what goes in and
> > what doesn't.
> >
> > Anyway, thanks for making this point publicly.
>
> I'm speaking for myself only, but the biggest problem with
> incorporating Tavis's changes, and the reason it hasn't even
> started, is that it is a _rewrite_ of Webware where many parts have
> been rewritten in a major way. We don't necessarily want _all_ of
> the changes, but in the current form it's impossible to just grab
> some of the changes without MAJOR effort to understand it.  It's a
> big effort that I'm not sure any of the Webware developers really
> wants to undertake.

That's a fair assessment, although I'd use the term _refactoring_ 
rather than _rewrite_.  For my perspective, many of the changes 
cannot be decoupled. They are heavily interrelated. Trying to 
implement this stuff on a piece-by-piece basis will render the cvs 
codebase unusable for days at a time. Using a separate devel-branch 
makes much more sense than patches.  I've politely asked for cvs 
access several times to be able to work on a branch, but never got 
any response.  That's why this stuff is in a separate repository at 
sourceforge.

Geoff's right, splitting them out into individual patches is a HUGE 
amount of work, even for someone who knows the code inside out.  It 
would be much less work to just bring the WebwareExpRefactoring 
package into a state that is acceptable to the rest of the webware 
developers than vice-versa.  Technically at least; politically that's 
another issue.

I've been writing patches for the minor changes that can be decoupled 
: e.g. the security fixes and the email error handling. Other stuff 
that I've been working on I've submitted in PEP form to the mailing 
list to stimulate discussion and interest. Unfortunately, nothing 
much has come of those postings.

> If we could decouple the changes from each other, and have a
> separate discussion of the merits of each proposed change, then
> phase in the new features gradually as patches, then I could see a
> path for migrating Webware towards what Tavis has done.  It'd
> probably have to be Tavis who does most of the work to create those
> patches.

Honestly, I'm not interested in doing all the work myself :)  I've 
written and tested these refactorings myself.  There's months of work 
in there.  Now it's up to other people to at least look at the code 
and comment on it.  Then we can decide how to proceed.

Tavis

_______________________________________________
Webware-discuss mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/webware-discuss

Reply via email to