[Bese-devel] Some issues

[Attila Lendvai]

> But the welcome-component instance is not available and at the same time the
> 'items' initform is required. Besides, I have to indicate that the menu slot
> is a component, so I have to instance something there. Am I correct?


you can add an initialize-instance :after method to do the same. not
too nice, but you can't access the instance when initforms are
executed.


> So, when this happens (and this is happenning quite often), I do:
> (defcomponent welcome-component (simple-window-component)
>   ((menu :accessor welcome-component.menu :component
>      (standard-menu :title "" :items nil))              <--- I DO THIS JUST
> TO SAY I HAVE A COMPONENT IN MENU, maybe there's a better way
>    )


add (myslot :component t) to the slot definition.


> 2) Iterate and macrolet problems (binding issues)
>
> Sometimes some symbols seem to  be unbound when I use iterate or macrolet.
> In general, I can fix that using lambda, but maybe you know why that happens
> and how to fix it.


iterate does not add the variable it defines to the macroexpand
environment. there's no easy way to fix this, iterate should use a
compatibility layer above environments and update it.

as of macrolet, it may be a deficiency of arnesi's code walker, but i
seem to remember fixes recently. are you using the latest from darcs
(the boxset does that)?


> (:render ()
>         (<:p (<:as-html (menu.title standard-menu)))
>         (<:ul
>         (iter
>           (for (item-name .  item-func) in (menu.items standard-menu))
>              (<ucw:a :action (funcall item-func) (<:as-html item-name))))))
>
> because item-func does not get bound when referring it from the <ucw:a
> action.


you can workaround that by using a let outside the iter and setf-ing the var.

also note that iter does not create a new binding (so all you action
lambdas would capture the values from the last iteration if it
worked...

-- 
 attila