Grzegorz Kossakowski schrieb:
Alexander Klimetschek napisaĆ(a):
Yes, this would indeed be better. The super block must not know of its
concrete children, but be able to call polymorphic matchers - true OO.
Then I can remove the cyclic dep.
Alex
BTW: When I started to use blocks and especially using the inheritance
feature, I was surprised that I manually had to include the
forward-**-to-super-block matcher at the end of the child sitemap....
I'll create issue for this one. Once again: would you like to implement it? If
so, just put a message on list.
I could easily create the patch for the warning message.
Regarding the other solution (automatically resolving the circular dep and
thus allowing it): I would now agree to skip this. A bit of research showed
that generic circular deps handling is difficult. There are solutions if one
specifically knows the two objects (see [1] and [2]). But here we are in a
piece of code that has to work generically.
Instead, it would be nice to get the real forwarding implemented soon - but
I don't know yet where to tackle the servlet-services code to get this done.
Alex
[1] http://forum.springframework.net/archive/index.php/t-94.html
[2] http://forum.springframework.org/archive/index.php/t-10186.html
--
Alexander Klimetschek
http://www.mindquarry.com