OK. On a Rails note, this seems to suggest that attributes are checked from their accessors before the save, whereas I had assumed my some_attribute accessor would not be called at all during validation...
(Thank you David, this is one of the better groups I've lately monitored, and a fine complement to your book...) Thanks, Lille On Jul 26, 1:06 pm, David Chelimsky <dchelim...@gmail.com> wrote: > On Jul 26, 2010, at 11:34 AM, Lille wrote: > > > Hi, > > > class B < ActiveRecord::Base > > belongs_to :a > > validates_numericality_of :some_attribute, :greater_than => > > 0, :allow_blank=>true > > > def some_attribute > > a.numeric_attr * read_attribute(:some_attribute) > > end > > ... > > > If I create an instance of my class B in rspec with factory_girl w.out > > a parent (but not required!) > > It may not be a business requirement, but the some_attribute method above > requires it's existence in order to run without error. > > Try: > > def some_attribute > return nil unless a > a.numeric_attr * read_attribute(:some_attribute) > end > > HTH, > David > > > instance of class A, I get the following > > error: "undefined method `numeric_attr' for nil:NilClass". > > > Why? Is this an apparent idiosyncrasy of using rspec and factory_girl > > together? > > Thanks, > > > Lille > > _______________________________________________ > rspec-users mailing list > rspec-us...@rubyforge.orghttp://rubyforge.org/mailman/listinfo/rspec-users _______________________________________________ rspec-users mailing list rspec-users@rubyforge.org http://rubyforge.org/mailman/listinfo/rspec-users
