So, I'm going to adopt the following increment on this little effort, and see if anyone objects.
My view is that CTR is plenty good for plumbing like this; it is not radical, it is incrementally adjustable, and, anyway, I can always roll it back if someone is disturbed. In my idea of a perfect world these days, we'd go to git, and any time anyone was in an RTC mood, they could push a branch. Meanwhile, technically speaking, the change doesn't work without more changes. There's no point in obtaining the document if the only field in it is the idfield. So some more adjusting is needed to allow control of the fields. Since the class already has what I'd consider to be an alarmingly long constructor arg list, I plan to split out the doc retrieval into a method, make some fields protected, and expect to subclass in my own code. If there are any other low-hanging opportunities to support adaptation-via-subclass, I'll get them. That will leave the Iterable subclass orphaned for my purposes, but I don't care. The Guava people make a case for wrapping an iterator in an immutable collection to obtain an Iterable. On Wed, Mar 20, 2013 at 10:07 PM, Ted Dunning <[email protected]> wrote: > Looks like a great idea. > > We are very weak RTC. Some things are pretty obviously good ideas and low > risk so we wind up doing something like CTR. > > On Wed, Mar 20, 2013 at 8:47 PM, Benson Margulies <[email protected] > >wrote: > > > Anyone have any objections? > > > > Are we still formally RTC? > > >
