Templating Engine: avioding bottle neck.

[Craige]

Technically, this thread could go in the PHP help forum, but I am not asking for help with the language specifically, but rather a code design. Thus, I have posted it here.

So my problem is in a potential design for a website templating engine I'm working on for a new site. I was wanting to do a recursive variable replacement function to the class, however my problem is with how to design it. I wanted to do a recursive function, but I'm not sure how much this is going to bottle neck the script on me.

The template file which includes a recursive replace might look like so:

code:

<!--Block: LIST--> <div style="whatever"><!-- < Var > --></div> <div style="whatever"><!-- < Var2 > --></div> <div style="whatever"><!-- < Var3 > --></div> <!--End Block: LIST-->

And what the templating engine would do is take a copy of the entire block, and use it as the template to replace variables in, x number of times. Each time, appending the result block to a static temp variable.

The recursive function would work something like so:

code:

if first iteration   get count of block variables: store in static var   get copy of block from temalte and save in static var   set static var to represent the iteration function level if ! first iteration, > last iteration   replace vars in copy of block by calling self with                                          loop num set to 1 for each var   append new block to temp static var   call self with loop num set to num - 1 if last iteration   I havn't worked it out yet, but it would finalize the iterations, and return true.   Add new set of blocks to template

I fear this will leave me with a huge bottle neck, which I want to avoid. I haven't worked out how all the code will work yet, but I think you can see what I want to do.

Ideas? Opinions?

[wtd]

I'm curious as to whether this is an academic exercise. Templating engines have been done so many times, and there are so many good ones out there that I fail to see why you wouldn't just learn one of those.

Of course, there's nothing wrong with it if it is an academic exercise. I'm just curious.

[Craige]

Nah. I've never even taken an academic course on programming. This is for personal use.

And the reason I don't use a prebuilt templating engine is simply for two reasons:
1) I enjoy learning new thins like this. Every experience helps you out later in life.

2) I generally don't use any prebuilt libraries for my personal, closed source sites. I don't have any problem using them in Open Source things I may do, but if nobody else is going to see and/or modify it (outside of a small team of coders), I prefer the code to be all original. It just means I know it in and out, and can customize it easily.

[wtd]

For simplicity, I would suggest having your templating code look absolutely nothing like the HTML you plan to embed it in.

Consider Perl's Template Toolkit for instance, that uses brackets instead of less-than and greater-than symbols.

[Craige]

wtd wrote:

For simplicity, I would suggest having your templating code look absolutely nothing like the HTML you plan to embed it in.

Consider Perl's Template Toolkit for instance, that uses brackets instead of less-than and greater-than symbols.

Actually, I'm using those due to the fact they are HTML comments. If I don't use a variable for some reason, it gets output as a comment. No big deal. Though in the programming scope of it, I should do something with every variable regardless of whether I want to display something or not.