[Vortex] [patch] partial patch for dependencies on msg ordering

Francis Brosnan Blazquez francis at aspl.es
Mon Nov 24 10:26:10 CET 2008


Hi Sam,

> The dependency on msg ordering is pretty wide-spread.
> 
> I removed assumptions about order or received messages. However, that
> causes reply to be broken:
> 
> (proc 6436): (warning) vortex-channel: Received a rpy request for
> message 102 while already waiting to reply to 0, storing and unlocking
> caller for later deliver
> 
> That reply will never get delivered,.
> 
> The unreplied msgs needs to be an ordered list, not a set, so that
> replies can be made in same order as reception.

Right. 

> I'll keep going with the patch if you are interested in it. If you'd
> like to fix this yourself, or just don't want to fix it all, I'll
> leave it.

Sure I want it fixed ;-)

I think the fix will take a bit longer than placing a list and other
minor changes. As you noted, current (wrong) assumption about msg
numbers is deep inside Vortex implementation. 

The key issue is not only providing a fix but spending some time
implementing a regression test and ensuring the fix not only makes
beepcore-c to work, but all cases in vortex-regression-client. I think
it is quite work. If you want to do it, go ahead Sam.

Cheers!

> Cheers,
> Sam
-- 
Francis Brosnan Blazquez <francis at aspl.es>
Advanced Software Production Line, S.L.



More information about the Vortex mailing list