Re: name:variable

On Nov 27, 2007, at 4:16 PM, Sovande Ulv wrote:

> Would it be possible for name:var to return a value so it can be used
> in statements such as this:
>
> <?cs if:var:key == name:header ?>
>

Ah hmm, the problem was not that name:header does not return a value.  
The problem is that I use variable names with '-' in them and that  
'var:key'  and 'name:header' evaluates to a minus expression when key  
or header contains a '-'.

So the expression, <?cs if:var:key == name:header ?>  becomes  <?cs if: 
0== 0 ?>. The problem lies in csparse.c:parse_tokens() which scan args  
and if it finds a '-' or '+' in the arg assume (wrongly in my case)  
that it deals with a numeric expression.

I understand that '-' and '+' may be tricky in variable names, but  
this could easily be handled by using the parse->hdf tree as a symbol  
table and look up a variable there first. At least, this will handle  
hdf variables. "In place" variable set with "set:variable.." is  
another case.

The reason I use variables with '-' in the name is that it is very  
convenient to define HTTP headers as they come in, directly in a hdf  
tree like so:

Request.Header.Host = localhost:8080
Request.Header.Connection = keep-alive
Request.Header.Accept-Encoding = gzip, deflate
Request.Header.Accept-Language = en-us
Request.Header.Accept = text/xml,application/xml,application/xhtml 
+xml,text/html;q=0.9,text/plain;q=0.8,image/png,*/*;q=0.5
Request.Header.Cache-Control = max-age=0
Request.Header.User-Agent = Mozilla/5.0 (Macintosh; U; Intel Mac OS X;  
en-us) AppleWebKit/523.10.3 (KHTML, like Gecko) Version/3.0.4 Safari/ 
523.10

Request headers can then be accessed and printed like so,

  <?cs each:header = Request.Header ?>
     <li> <?cs name:header ?> = <?cs var:header ?>
  <?cs /each ?>

Though direct access to e.g Request.Header.Accept-Encoding  using  <? 
cs var: Request.Header.Accept-Encoding ?> does not work since it  
evaluates to "Request.Header.Accept - Encoding" -> 0 Though <?cs evar:  
Request.Header.Accept-Encoding ?> prints correctly -> gzip, deflate,  
but unfortunately evar cannot be used in a each-loop.

I believe that the clearsilver CGI module also would benefit from this  
as well, i.e. allow to define variables with '-' in them.  This way  
the cgi module does not have to hardcode header names it will  
recognize but can define all headers as they come in, in a hdf tree.

Would you be interested in a patch for this Brandon, or maybe you  
already have thought about it and have a workaround?

Best regards