Activity (RSS)

Solve this problem:

 $[page = page_for_html_block(r)]$page.name at $r.name.

perhaps like this

$page.name at $r.name.

Closing this, it works.

Same problem, but a better example of what I have to work out:

$@annotation.nil? borks

Fixed i 0.1.7

instead of outputting @foo it outputs

expected "\"hello world\""
     got ");@output_buffer.concat((@foo).to_s);@output_buffer.concat("

where @foo = "hello wo...

Consider dropping the prefixed 'r' in the attribute expression. A colon is enough to introduce the syntax and it resembles a ruby symbol.

Pretty obvious - for the single $ you need to scan for the end of the word using a number of delimiters - $, punctuation marks, etc..

For example, a string containing (whitespace) is escaped into 
 to render the original characters in the html document. HAML substitutes a whitespace. I s...

Probably create a separate gem - renshi-sinatra - which monkey patches Sinatra::Base::Templates

Opening this again as it will be helpful in other frameworks which don't have rails helpers.

The current rdoc sucks. Write some up and install it at - http://renshi.rubyforge.org .

Yes, it looks reasonable now.

DOne.

Not needed in the ruby world (a helper can be just as effective).

Fixed.

$$ is escaped renshi for '$'

Renshi should ignore $().

The rails integration broke without my noticing it.

When transforming lines of code in $[] lines of whitespace are left. Destroy them.

When looping in 'while' and 'for' attr expressions, the next iteration output...

Done - < and > and & are encoded (see parser). No need for the others - '. ".

Perhaps no longer needed (the big compile method was moved to Parser).

Done.

For example,

Done.

Actually, jquery just uses $() (I think), so this should be fine.

Ouch - jquery uses ${}. Crap. Might have to change what renshi looks for.

Inserting a string such as 1 < 2 using before on an XML node creates an error.

This probably applies to all XML entities. Suggested workaround is to encode them...

At the moment I add a lot of helper methods on the nokogori node via a monkey patch.

The number is growing. It'd be better to hide them in a helper object to av...

Renshi doesn't take this approach.

I've decided to develop directives following a Ruby programmer's natural expectations, rather than mirroring Genshi.

No need - Rails has a built in mechanism for this (courtesy of compilable templates).

A syntax error or the like should render the line containing the error, or the line in the vicinity. Define a strategy for this.

Done for statements and conditional expressions.

See http://genshi.edgewall.org/wiki/Documentation/xml-templates.html

We're going to try avoiding processing order

The problem with doing this, though, is that it means writing the template parser specifically for the Rails platform. I wonder how haml handles Sinatra integra...

For example, a haml document

ok
= @foo
done

- if @foo
  = Time.now
- else
  there's no time like the present

compiles to this str, which is reflected as a me...

A good templating system needs to be compilable. The current prototype compiles the template live each time, it should be possible to improve this early on.

Thi...

genshi's py:def allows you to define blocks of template code that can be reused. renshi should be able to do this too:

See http://genshi.edgewall.org/wiki/Docu...

ie. they should have a structure not unlike the one in genshi:

  xmlns:py="http://genshi.edgewall.org/"
  lang="en">

... stuff goes here ...

genshi doesn't have else, but no reason renshi can't have it.