On 10/5/07, Thibaut Barrère <[EMAIL PROTECTED]> wrote: > Hi, > > it seems that when a RowProcessor used as a before_write returns nil, the > row is removed from the pipeline, but when the same processor is used as an > after_read, the nil row remains in the pipeline (and the next processor will > most likely raise a nil exception at soon as it interacts with the row). > > The difference seems to lie in engine.rb (process_control): > > after_read: rows = processed_rows.flatten > before_write: rows = processed_rows.flatten.compact > > Is there a specific reason for not having compact called as well on > after_read ?
I assumed that you would want to let your row go through your transform pipeline before compacting it to remove nils. > My feeling is that it may be worth DRYing process_control - for the moment I > didn't analyse further to evaluate the cost and possible side-effects on > existing installations though... It is probably worth it, and I have tried to refactor some of it, however I have not completely refactored it. Ideally I think that the process_control would basically be a bunch of calls to internal methods to express each of the processing steps. V/r Anthony -- Cell: 321 473-4966 Current Location: Melbourne, FL _______________________________________________ Activewarehouse-discuss mailing list Activewarehouse-discuss@rubyforge.org http://rubyforge.org/mailman/listinfo/activewarehouse-discuss
