[Grok-dev] Re: we need a MethodGrokker

[Philipp von Weitershausen]

Martijn Faassen wrote:
> [... a lot of sense ...]

I was thinking the same thing when I made the refactoring. I don't have 
much else to say, I think I agree comple.... CRASH BANG BOOM KRZZZZ 
FYOOP SLURP FIZZZ (this is the universe imploding due to this anomaly :)).

> * note that the original XMLRPCGrokker only did the check_permission for 
> the class-level permission, we do it for all permissions (also the 
> directive level permissions). You'd think that would be the right thing, 

Yes, I stumpled over this as well. I think it's the right thing.

> but perhaps there's some subtlety that the tests cover that I forgot 
> about. There's a performance argument in not doing it for permissions 
> multiple times though, but we already didn't do this for class-level 
> permissions.

Performance perschormance :) It's just a getUtility call each time. 
Doing 1000 of those takes roughly 3ms on my machine.