Test if runtime variable is defined

[bugzilla@…]

I wish it were possible to have an EXPRESSION to test if a variable is defined.

I would like to be able to test this:

EXPRESSION (defined '%key%')

or

EXPRESSION ('%key%' != NULL)

The problem is that %key% is not replaced in case no 'key' is passed through
query_string, and you get a literal '%key%' string in your expression instead of
a null value.


I found a workaround for this:

EXPRESSION (('%key%' lt '%key$' or '%key%' gt '%key&') and '[FIELD]' ne '%key%')

You can't test if %key% is defined, but you can test if %key% is equal to
literal '%key%'. This is my dirty way to test is a variable is defined.

I think the best approach is to replace all %vars% on mapfile with empty strings
in case they were not defined. This is the default behaviour of Perl's template
module, for example.

Could you please consider implementing a way to test if a variable is defined?

[bugzilla@…]

*** Bug 1830 has been marked as a duplicate of this bug. ***

[sdlime]

This is impossible to do as you envision because MapServer doesn't know what 
runtime variables to expect. However, you do, so why wouldn't this work.

Let's say you have a runtime variable called foo, then why not do something 
like:

  EXPRESSION (!('%foo%' =~ /foo/))

If the substitution doesn't take place, that is, foo isn't set then the regular 
expression ('%foo%' =~ /foo/) would evaluate to true.

Assuming that works (and it should) then I see no reason to pursue a more 
elegant solution, so I'll mark this as INVALID for now and we can always re-
open later.

Steve

[bugzilla@…]

Hi Steve,

your expression is more elegant (and slower) than mine.
You are right. It works, so this is not a high priority enhancement.

But this is not a nice solution, and not intuitive at all. Users must have a
happy idea in order to imagine this kind of solution. There should be, at least,
some referece to this trick on mapserver documentation.

[sdlime]

The CGI variables never make it down to the expression engine so I'm afraid 
this is the best we'll do for the moment. An expanded set of string operators 
(e.g. a strstr wrapper for substrings) might help with the speed issue.

As for documentation- the website will accept documentation from any and all 
users so here might be a good opportunity to create a short howto or snippet!

Steve

[sdlime]

FYI I did add an FAQ to the website. Feel free to add to it!

Steve