Re: Headcount
Magnus Lie Hetland <[email protected]>
| Newsgroups | gmane.comp.python.anygui.devel |
|---|---|
| Message-ID | <[email protected]> |
Samuele Pedroni <[email protected]>: [snip] > there's a > > if not self.widget: return > > at the top of all setFoo methods (except setContainer's that trigger widget > creation) Hm. That's a bit like what we had in 0.1. It's OK (the None stuff seems like a clean solution) but perhaps we could put this check in the dispatcher (i.e. in the pull/push methods of the Wrapper) instead? IOW, check for self.widget == None -- if not, start looking for an accessor. I definitely think this is a refactoring worth doing (even though I haven't really looked at your new code yet ;) [snip] > Another question is: now proxy.wrapper is kept in state and there's a > comment asking whether it should be that way. Yeah; It really shouldn't. Now that we have initial underscores back again, we could start using proxy._wrapper, of course... (Or we could exempt it from state handling like proxy.state itself...) > For now I have added a non-in-state _wrapper_set flag that is set to true when > wrapper is set to a concrete wrapper. Hm... OK. Doesn't sound like an ideal solution. > I need to check for wrapper presence in push and pull that are called by the > Attrib constructor before wrapper is set > in order to do the right thing, i.e. return without after doing nothing. I see. How about simply generalising the test if attr == 'state' to something like if self.isStateVariable(attr) or something in Attrib? Then it would be easy for Proxy to override the default Attrib behaviour (only dropping 'state' and anything beginning with '_') and make 'wrapper' a non-state variable. > Should wrapper be out of state and become _wrapper, then we need a getWrapper > or all occurences of wrapper > in the backends should be renamed _wrapper. With the suggested scheme over, it will still be a legal attribute (simply proxy.wrapper) but would not be put in proxy.state. > regards. -- Magnus Lie Hetland The Anygui Project http://hetland.org http://anygui.org ------------------------------------------------------- In remembrance www.osdn.com/911/