On Tue, Jul 30, 2013 at 9:15 AM, Erin Noe-Payne <[email protected]>wrote:

> On Tue, Jul 30, 2013 at 10:07 AM, Gornstein, Daniel S.
> <[email protected]> wrote:
> >
> >
> > -----Original Message-----
> > From: Erin Noe-Payne [mailto:[email protected]]
> > Sent: Tuesday, July 30, 2013 9:57 AM
> > To: [email protected]
> > Subject: Re: Service Layer Questions
> >
> > On Tue, Jul 30, 2013 at 6:50 AM, Gornstein, Daniel S.
> > <[email protected]> wrote:
> >> All,
> >>
> >> I am currently working on RAVE-1010 (
> https://issues.apache.org/jira/browse/RAVE-1010) updating the service
> layers to support crud operations from the rest API.
> >>
> >> I was planning to add getAll, getCountAll, and getLimitedList to the
> services with respective repositories who also implement those methods.
> >>
> >> Right now some of the services (WidgetService) return SearchResult
> objects, while others (CategoryService) return List<Category>.
> >>
> >> Should I be switching all services to return SearchResult objects?
> >
> > The rest api will probably want SearchResults objects, to fill in the
> > metadata. Possibly refactor to rename the any conflicting methods that
> > do not return a SearchResults instance?
> >
> > So does this mean if there are already getAll methods for services which
> return a list, keep them but with a different name and make a separate
> getAll which returns a SearchResult?
> >
> That is what I'm suggesting for minimum impact. If the current getAll
> methods are not used widely you could just change the method signature
> to return a SearchResults and then propagate that through. I don't
> really have a strong opinion.
>

They definitely should all be returning the same object.  Search Results is
fine so long as it encompasses the functionality we want to have for return
results.


>
> >>
> >> Also I noticed on some of the methods in interfaces for the services
> have spring annotations of @PostAuthorize or @PostFilter.
> >>
> >> I was wondering what the appropriate use cases are as they are not on
> every service interface, nor on every method in the interface.
> >>
> >> Thanks,
> >> Dan Gornstein
>

Reply via email to