> >  > >       * 155-158: I'm a bit confused by this.  Why not just put the
 > >  > >         contract_latest(&ct) call directly in the `default' arm of
 > >  > >         the swtich statement?
 > >  > 
 > >  > I'm just reducing code duplication.  The contract_latest needs to be
 > >  > done before ct_tmpl_clear and close, but those two things need to be
 > >  > done for the -1 (failure) and 0 (child) cases as well.
 > > 
 > > It was unclear to me why contract_latest() needed to be called before
 > > those two things while contract_abandon_id() could be done afterwards --
 > > but I didn't dig into the libcontract API details.
 > 
 > We're all building those grass runways on the island.

Does that mean the current structure is necessary?

-- 
meem

Reply via email to