Comments on: Pre- and postconditions for CFML functions? http://riait.co.uk/2008/06/25/pre-and-postconditions-for-cfml-functions/ just another ria (sort of) weblog... Thu, 04 Dec 2008 01:17:26 +0000 http://wordpress.org/?v=2.6 By: radekg http://riait.co.uk/2008/06/25/pre-and-postconditions-for-cfml-functions/#comment-55 radekg Thu, 26 Jun 2008 09:32:55 +0000 http://riait.co.uk/?p=22#comment-55 @todd: it really depends on conditions developer would like to check. the idea is pre- and postcondition attributes should allow any valid CFML expression. it is basically moving test expressions from inside function to well defined attributes but the advantages are what I described. developer does not need see CFML code to know what required/returned values are. For structures it may be for example: postcondition="structKeyExists(result,'requiredKey') and len(result.reuquiredKey) gt 2" for array it could be: postcondition="arrayLen(result) gt 0" for query: postcondition="result.recordCount gt 0" and so on. @todd: it really depends on conditions developer would like to check. the idea is pre- and postcondition attributes should allow any valid CFML expression. it is basically moving test expressions from inside function to well defined attributes but the advantages are what I described. developer does not need see CFML code to know what required/returned values are.
For structures it may be for example:

postcondition=”structKeyExists(result,’requiredKey’) and len(result.reuquiredKey) gt 2″

for array it could be:

postcondition=”arrayLen(result) gt 0″

for query:

postcondition=”result.recordCount gt 0″

and so on.

]]> By: todd sharp http://riait.co.uk/2008/06/25/pre-and-postconditions-for-cfml-functions/#comment-54 todd sharp Thu, 26 Jun 2008 00:09:32 +0000 http://riait.co.uk/?p=22#comment-54 It is an interesting proposition, but do you think it is only useful for simple use cases like strings/integers, etc? How could it be used for objects/arrays/structs/queries? Would you expect an exception to be thrown if it does not meet the conditions? It is an interesting proposition, but do you think it is only useful for simple use cases like strings/integers, etc? How could it be used for objects/arrays/structs/queries? Would you expect an exception to be thrown if it does not meet the conditions?

]]>