The latter is "sliding expiration". For the former one doesn't "absolute expiration" fit well the case? Il giorno 26/dic/2012 17:46, "Christoph Engelbert" <[email protected]> ha scritto:
> Someone has a good idea what to call an absolute and a relative > expiration whereas the last one is relative to the last access? > > Am 26.12.2012 16:04, schrieb Jeff MAURY: > > +1 > > > > > > On Wed, Dec 26, 2012 at 3:18 PM, Christoph Engelbert > > <[email protected]>wrote: > > > >> At least both values are checked the same way: > >> public boolean apply( Pointer<V> input ) > >> { > >> return !input.isFree() && input.isExpired(); > >> } > >> > >> > >> And I guess "Pointer::isExpired" is implemented in the wrong way to > >> get both into account: > >> public boolean isExpired() > >> { > >> if ( expires > 0 || expiresIn > 0 ) > >> { > >> return ( expiresIn + created < currentTimeMillis() ); > >> } > >> return false; > >> } > >> > >> > >> The general idea sound good to me but maybe we could find better > >> names :-) They're not really selfspeaking from my understanding. > >> > >> Am 26.12.2012 15:13, schrieb Raffaele P. Guidi: > >>> The idea was: expiresIn 3 minutes (a time lapse) vs expires tomorrow > >> 08:00 > >>> (an absolute value). Not sure it actually makes sense anymore. > >>> > >>> Ciao, > >>> R > >>> Il giorno 26/dic/2012 14:50, "Christoph Engelbert" < > [email protected] > >>> > >>> ha scritto: > >>> > >>>> Hey guys, > >>>> > >>>> I'm started documenting some of the missing interfaces / methods but > >>>> I stuck at all that fuzzy kinds of "expiresIn" and "expires". The > >>>> only thing about the last one is "-1" or > >>>> "AbstractMemoryManager::NEVER_EXPIRES" which is 0. So it seems that > >>>> keys never will expire at all. Did I missed something do we need > >>>> that second "expires"? > >>>> > >>>> Chris > >>>> > >>>> > >> > > > >
