I've thought about the dependencies problem a little bit. Maybe add some kind of hook so that the user can define the dependencies of the computed field himself.
If he doesn't define them, the current behavior occurs. If he defines them, the behavior could be as follow: 1) All the dependencies are present in the input The field is updated from the input 2) Part of the dependencies are present in the input The current behavior occurs. Alternatively, try to read the missing dependencies from the DB (for an update) or from the default values (for an insert), but I'm not sure how complicated that would be. 3) None of the dependencies are present in the input The computed field is not updated. >From the above, you'd get the best of all worlds. 1) User friendliness: The user doesn't have to define dependencies if he doesn't want to and he'd get the current behavior. 2) Flexibility: The user can define dependencies in which case the computed field doesn't restrict updates that don't include the dependencies on the field. 3) Correctness: Exceptions due to mistyping variables are not caught and properly raised. On Dec 23, 9:11 am, mdipierro <mdipie...@cs.depaul.edu> wrote: > Need to think about it. > > On Dec 23, 3:51 am, cjrh <caleb.hatti...@gmail.com> wrote: > > > > > On Dec 23, 2:04 am, mdipierro <mdipie...@cs.depaul.edu> wrote: > > > > I can add a try except so that web2py does not tries to compute fields > > > when there is not enough information. The downside is that a typo will > > > result in a silent failure to compute. Perhaps we should do this > > > anyway. > > > If you're going to be doing work in the area of compute anyway, is > > there a posssibility of making another change that the current value > > of the compute-ed field can be passed to the compute method?- Hide quoted > > text - > > - Show quoted text -
