[Seaside] Wierd totally fatal bug problem

[Avi Bryant]

On Tue, 14 May 2002, Tim Rowledge wrote:

> This may or may not be anything to do with seaside, or even comanche,
> but I thought I'd see if anyone else has seen anything like this
> recently.

<snip description of image corruption and chaos>

Ok, I'm glad to hear that I'm not going mad; I've been seeing very similar
symptoms recently, but nobody else has been reporting anything.  What I
usually see is this

- sometimes after an image is saved, Comanche stops responding; it still
accepts connections but it doesn't do anything with them
- restarting Comanche on that same port doesn't do anything
- starting a new IAKom service on a different port often works, but not
always
- once this happens, that image is unreliable; it is liable to lock up,
for example, ten or so seconds after startup

I've been playing around with different VMs, different base images,
different versions of Comanche, different versions of Seaside... haven't
isolated the problem yet, although using the very latest SourceForge VM I
haven't actually run across it yet (using Ian's latest VM, however, I
definitely did).  The frustrating part is that I *know* this never
happened until a few weeks ago, but I can't imagine what's changed.  Tim,
what versions of everything are you running?

It only seems to happen when Comanche manages to auto-start the service;
whenever I start up a VM and can't connect right away, I know it'll work
once I manually get things going.  I have no idea why this might be,
however.

I'll keep trying to hunt this down, but if anybody has any clues, please
let me know.

Avi