Hi brent

I really appreciated your effort to help us the KCP people, but if this 
complicates your code and make your browser (that is starting to be 
central for the harvesting process) fragile, just discard this 
functionality, we will find another way to deal with that.

Stef


On Friday, August 8, 2003, at 05:04 AM, Brent Vukmer wrote:

> Er.. Hi, Tim.  The KCP Archive Viewer was an *experiment* in loading 
> changesets from the KCP Swiki page.  I don't think I made the 
> experimental nature clear; sorry my lab experiment blew up in your 
> face :)
>
> My bad for not testing more carefully, and for not making clear that 
> the "Swiki Archive viewing" functionality was very immature.
>
> Since the KCP Viewer is just loading changeset files, not mail-message 
> files,
> it makes sense that the timeSent could be nil...
>
> Regarding the KCP Viewer's failure to load a bunch of posts; that's 
> because the KCP folks decided at some point to upload their changesets 
> to a different URL.  I didn't realize this, and the BFAV depends on 
> having one well-known URL
> for the archive it's viewing, so...
>
> The BFAV originally modelled the current harvesting process very 
> closely.  Trying to extend it to also model the project-Swiki-page 
> process makes the BFAV packages more complicated, increases the amount 
> of QA I Should Be Doing, etc. Frankly I just don't know if it's 
> needed.  I may just remove the functionality.  Is anybody else using 
> the KCP Viewer?  Any votes for keeping it?
>
> Regarding the BFAV regularly failing to load posts from the 
> swiki.gsug.org/sqfixes URL -- I know that this is an issue; it seems 
> to be mainly due to dropped connections on the server's end.  I'm 
> working with Ken Causey and other folks to set up Bug Fixes Archive 
> II, which will use Apache to serve static files, and which should be 
> much better at handling lots of connections.
> <winmail.dat>