[AG-TECH] Patched TVS

[Ivan R. Judson]

Lotta stuff going on with the server. We've been waist deep in server guts
sorting out the hanging problems, weird client/server protocol bugs, and
other stuff.

You might notice text is working again on the TVS, this is one side effect
of our work. We're pretty close to understanding protocol/state bugs, tom
and eric have put a lot of good work into hammering that into shape.

As we progress, we've taken a little more progressive approach to the TVS,
making it the "latest stable cvs", so that it has the latest bug fixes, etc.
We encountered a new bug this morning with service certs that we'll try to
get closed tomorrow that (along with the protocol/state fixes) should really
go a long way towards fixing not only 1057, but many of the other problems
users are seeing as well.

--Ivan 

> -----Original Message-----
> From: Christoph Willing [mailto:willing at itee.uq.edu.au] 
> Sent: Thursday, July 15, 2004 6:24 PM
> To: Ivan R. Judson
> Cc: 'AG-TECH'
> Subject: RE: [AG-TECH] Patched TVS
> 
> On Thu, 2004-07-15 at 20:54, Ivan R. Judson wrote:
> > > Ivan,
> > > 
> > > At 6pm Chicago time, the newly running TVS is working 
> _much_ better.
> > > 
> > > The changes seem to have fixed the "hang on room changs/exit" 
> > > bug #1057, which I thought was a AGtk 2.2 VenueClient bug. 
> > > I've just been through all the top level rooms & back 
> through main 
> > > lobby without any hangs at all. More testing to do I guess, but 
> > > looking good!
> > > 
> > > Have the patches made it to CVS yet?
> > 
> > Yes, the TVS is running the latest tested CVS code, the 
> patches are in 
> > cvs already.
> > 
> > We have another round of fairly thorough review to do on 
> client/server 
> > state machine sanity, I think it will shake out the 
> remaining common 
> > problems users are seeing. I did not anticipate the previous fixes 
> > would correct #1057, I'm not even confident I can claim 
> it's fixed, I 
> > wish I could break it reliably :-).
> >  
> 
> Ivan,
> 
> I couldn't understand the connection (if any) with #1057 either.
> Nevertheless the hangs had completely stopped when accessing 
> TVS rooms, yet immediately manifested when accessing rooms at 
> the APAG server (which is still running the AG2.1.2 server). 
> Hence the thought of a connection between the TVS patches and 
> bug #1057.
> 
> This morning (~6pm your time), #1057 is back again in full 
> force on TVS.
> Actually I noticed it late yesterday (2am your time), approx 
> 9 hours after yesterday's good 6pm result, then checked it 
> again just now.
> 
> Did anything change at the TVS since 6pm yesterday?
> 
> 
> chris
> 
> 
> -- 
> Chris Willing                     Phone: +61 7 3365 8350
> QPSF Access Grid Manager          willing at itee.uq.edu.au
> University of Queensland
> 
>