Re: bug in svn version

On Thu, 20 Apr 2006 18:55:00 +0200, "Michele Petrazzo"
<[email protected]> said:
> A question about the "old", I think, spyce code: I see into the code
> that you are using the "old" class definition type, not the:
> 
> class foo(object): pass
> but
> class foo: pass
> 
> so we can't use super...
> 
> This, if done, will break the 2.2 compatibility, but leave us to user
> the "super" operator. Do you think that into the next release you'll
> change this?

I don't think "super" is a sufficient reason to break 2.2 compatibility.

But maybe a patch that added

try;
  __metaclass__ = type
except NameError:
  pass

to each module would be okay.  Still kind of ugly, though.

> Other question:
> I have into my config file the:
> 
> check_modules_and_restart = True
> check_modules_no_restart = True
> concurrency = 'thread'
> 
> but sometimes (more or less always), when I change some code into a
> module, I see into the logs:
> 
> # (Detected changed modules ['py.db'].  Restarting...
> 
> (If you don't want Spyce to do this, turn off check_modules_and_restart
> in your spyceconf file.)
> (28202, 768)
> 
> and spyce didn't restart...
> This don't seem to happen on:
> concurrency = 'fork'

Doesn't really surprise me.

I should really probably just take "fork" out; threading is a better
model for Spyce anyway, and modern Linux supports threading just fine.
(Historically, Linux threads kinda sucked vs Windows or Solaris.)

What's your reasoning behind using the forking server?

-Jonathan
-- 
C++ is history repeated as tragedy. Java is history repeated as farce.  --Scott McKay



-------------------------------------------------------
Using Tomcat but need to do more? Need to support web services, security?
Get stuff done quickly with pre-integrated technology to make your job easier
Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642