From: Makoto Kuwata <kwa@...>
Date: 2004-03-20T06:30:46+09:00
Subject: Re: [ANN] Kwartz (2004-03-19) released

Andreas,

Thanks for your suggestions.

Andreas wrote:
>
>1)
><span id="if:logged_in">
>#{username}#
></span>
>
>results in
>
>if logged_in then
>  print username, "\n"
>end
>
>But according to
>http://www.kuwata-lab.com/webtech/kwartz/users-guide.en.03.html#directive-if
>the result should look like this:
>
>if logged_in then
>  print "<span id=\"if:logged_in\">\n"
>  print username, "\n"
>  print "</span>\n"
>end


Tag <span></span> or <div></div> are deleted if they have only directives.
If you don't want to delete them, use <SPAN></SPAN> or <DIV><DIV>,
or add other attributes like <span id="if:logged_in" class="foo"></span>.

See
http://www.kuwata-lab.com/webtech/kwartz/users-guide.en.03.html#directive-notes


>
>2)
>IMO something like this would be much cleaner than eval()'ing the code
>generated by kwarts, because the ruby file has to be parsed only once,
>and it isn't always wanted to send the output to STDOUT.
>
>template = Kwartz.new(File.read('template.html'))
>string = template.expand(data)


You should use ERB if you want to get result as string.
   ------------------
   s = File.open('file.eruby') { |f| f.read() }
   require 'erb'
   erb = ERB.new(s)       # or ERB.new(s, $SAFE, '%')
   str = erb.result(binding())
   ------------------

And executing the output script is enough fast.
See 
  http://www.io.com/~egabriel/bench/results.html
The benchmark shows that eruby (= ERuby::import()) is faster than
any other template system in Ruby.
(ERB works almost as fast as eruby.)
It is one of the reasons why I adopted the current way in Kwartz.

If you want to get more speed, try erbscan.
http://raa.ruby-lang.org/list.rhtml?name=erbscan
I have not tried it yet, but I heared that erbscan makes ERB very very fast.


>
>3)
>ruby output could also support sanitizing (CGI::escapeHTML).
>

Yes, you are right.
I do that in the future release.


--
regards,
kwa