Re^3: HTML::Template, pseudo trees and indention.

[dragonchild]

*claps sardonically* I was waiting for that response. You did it with H::T, not within H::T. Your node is an excellent example of twisting the problem to suit the answer. Not only is your template nearly unreadable the first 3 times one reads it, but your code is nearly impossible to follow as well. The point wasn't to see if you could code a solution that involves H::T. The point was to see if H::T could provide a solution.

And, yes, any recursive solution can be written iteratively. That doesn't mean that H::T can handle it any better.

[maverick]

/me reads question again.

...only for strict definitions of "twist". He already has the tree as an array. He already has the depth of the nodes in the tree. The only thing he's missing is the not-so-magical trick of having H::T produce html to create the changes in depth. For that you need to know how much and which way the depth changes between any two nodes and then use a tmpl_loop to generate the right number of opens / closes.

Granted the code needs comments, but impossible to follow is a stretch. jeffa figured it out sans comments in the time it took me to type this respone...can't be that bad :)

[dragonchild]

jeffa understanding something and your average programmer understanding it are two completely different things. You see, jeffa actually has (at least) half a brain and is willing to use it. :-)

I might have been a little harsh in my response. But, I feel that BUU's tree representation, while compact, is mostly unusable. Given a more useful tree representation (such as Tree or Tree::Simple), one is inevitably led to use one of the tools that already exist for reformatting objects of those classes (which I list in my original response). Why reinvent the wheel?

[maverick]

hehehee. why? simple. You said it "couldn't be done, period".

Naturally, we can debate the merits of near a infinite number of possible solutions (TIMTOWTDI after all), but that wasn't the point :)