[Zope] ZCVSFolder woes

[Rob Miller]

Hi,

Is anyone out there using ZCVSFolder to help manage Zope development? 
I've played with it a bit, and I'd love to start using it on the project 
that I'm currently managing, but I'm finding it problematic...

What I've found is that I can add CVS Folder objects with no trouble, 
and that, at first, I can enter the CVS Folder objects and perform CVS 
operations as expected.  If I wait a while, though, and then try to 
browse into the CVS Folder using the management interface again, I find 
that my http requests never get a response.  I'll click on the "CVS" 
object, the browser shows me that it's working, but I never get into the 
CVS Folder...

If I log into the server and run ./stop from the correct directory, I'll 
find that, for every time a CVS Folder access failed, one of the Zope 
python processes didn't die.  That is, if I tried unsuccessfully 3 times 
to browse into the CVS Folder, then there will be 3 python processes 
left over that should have been killed by the ./stop command.  If I kill 
these processes by hand, then I can restart the server, and the CVS 
Folders will work as expected again... for a little while.  Also note 
that I can stop Zope from responding altogether if I allow the number of 
failed CVS Folder accesses to equal the NUMBER_OF_THREADS that Zope is 
allowed to spawn.

I'm running Zope-2.4.2 on FreeBSD-4.4, and I'm using ZCVSFolder-0.2.5. 
Has anyone else experienced this?  Any ideas on how to make it stop 
happening?  I like the ZCVSFolder interface, and being able to use it 
would GREATLY facilitate our development process.

Thanks!

-r