On Wed, 2017-04-19 at 14:50 -0700, Gary Gregory wrote: > On Mon, Apr 3, 2017 at 9:06 AM, Oleg Kalnichevski <[email protected]> > wrote: > > > On Sun, 2017-04-02 at 15:07 -0700, Gary Gregory wrote: > > > On Thu, Feb 16, 2017 at 6:47 AM, sebb <[email protected]> wrote: > > > > > > > ... > > > > > > > > > > > > So now yes, I can see a reason for a Timeout subclass. > > > > > > I can see in the code (e.g. > > > org.apache.hc.core5.http.HttpConnection.getSocketTimeout()) where > > > timeouts > > > with values of 0 and -1 are special with 0 = disabled and -1 = > > > undefined. > > > > > > So I could see factory methods like Timeout.disabled() and > > > Timeout.undefined() that return singletons; or just define public > > > statics. > > > > > > Then the code could call aTimeout.isDisabled()/isUndefined() > > > instead > > > of > > > comparing to magic numbers. > > > > > > Thoughts? > > > > > > Gary > > > > > > > Sounds good to me. I would also make it illegal to create Timeout > > instances with negative duration. > > > > I have a first cut of the Timeout class in svn trunk. It allows for a > -1 > timeout as a special case so I wonder if we want to hide that more. > Also > there is some awkwardness to the two isUndefined*() methods. >
I suggest #isUndefined() be dropped and null be used where appropriate. Oleg > Tweaks and comments appreciated. > > Gary > > > > Oleg > > > > > > > > ----------------------------------------------------------------- > > ---- > > To unsubscribe, e-mail: [email protected] > > For additional commands, e-mail: [email protected] > > > > > > --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
