What ERB does is to take your template and turn it into pure Ruby code
that builds up a string using a variable name called (by default)
'_erbout'. When you call result(), the ruby code is eval'd in the
context of the binding passed (or TOPLEVEL_BINDING if no binding is
passed).
To do a sub-template, you need to eval the sub-template and insert
it's results mid-stream. But all the eval-ing happens at once (when
you do the outermost template), so you need to use a different output
variable name in your subtemplate, or your subtemplate will stomp on
the output of your outer template. Hence the '_erbout2' parameter to
new(). If the subtemplate called yet another subtemplate, you would
need another variable name, and so forth, for each level of nesting.
I have the same problem.
I am translating an old ASP-application (pre dot.net) that uses textual
includes to implement DRY.
When I try to use ERB for this I get scoping problems
* Variables or functions declared in the subtemplate I try to read with
erb are not 'remembered' when returning to the main view (that did the
erb call)
Note. if a variable was declared the first time in the main view it is
possible to change the value of that variable in the subtemplate and
that new value is 'remembered'.
I can't see how you can implement good DRY-solution in ruby/rails given
the above problem but it should be possible since ruby/rails is a much
newer/more modern framework than old asp.
Of course you can put the methods in an helper method but then you don't
have access to the <% %> tags and you don't have access to variables
declared in the view.
I can't see how you can implement good DRY-solution in ruby/rails given
the above problem but it should be possible since ruby/rails is a much
newer/more modern framework than old asp.
Of course you can put the methods in an helper method but then you don't
have access to the <% %> tags and you don't have access to variables
declared in the view.
As is was using ERB without Rails i made the switch to a template engine
called tenjin. There it's possible to have clean subtemplate resolution.
Yes I tested quite a bit with different renders and partials. What I
remembered was that it couldn't asked the variables declared before the
render-statement. i.e. when you start rendering the subtemplate all the
variables from the main template are forgotten.
Of course one solution is to declare the variables with @varname in the
controller but I am working with a automatic translator from asp and I
would really need to have the code in the views for a start with a
subtemplating system that works just like a text-include (like in asp
with <--#include file)
The only thing that comes to my mind is the :locals hash you can pass in a call to render :partial. But it'd probably be as much of a pain to figure out how to have your transliterator make those calls as it would be to have it put variable creation code in the controllers...
