DESCRIPTION

Added touch operation

Added an operation to reset the expiration time of a key. I have
tested this code thoughly and will check in the unit tests in a
later commit

Change-Id: Iee3045bdb58142dd44c9e8c723dffe7abdbcb230


COMMENTS

author: Kenny Maynard
date: 2011-06-03 23:08:06.001000000

Uploaded patch set 6.

-------------------------------------
author: Kenny Maynard
date: 2011-06-03 23:09:17.238000000

Patch Set 6: Verified

No more asyncTouch(), just touch(), but it's async.

-------------------------------------
author: Noah Dean
date: 2011-06-03 23:09:49.803000000

Patch Set 6: Looks good to me, approved

The tests are blocked on another change.  I believe they exist, but this code looks pretty good so far.

-------------------------------------
author: Jakob Lozano
date: 2011-06-03 23:11:01.373000000

Patch Set 6:

I think there's something to be said for consistency in naming the method.  I don't mind having asyncTouch() and no touch().  Right now, people familiar with the API may assume think touch() is synchronous even when it isn't.  I suggest putting it back.

-------------------------------------
author: Noah Dean
date: 2011-06-03 23:12:55.036000000

Patch Set 6:

Well, set is async now.  That's the original sin of inconsistency I was thinking about.  Given both are inconsistent, my vote would be for the smaller one.  I don't have a huge amount of care, though.

-------------------------------------
author: Jakob Lozano
date: 2011-06-03 23:14:28.873000000

Patch Set 6:

OK.  Maybe for the 2.7 release we making everything consistent.

-------------------------------------
author: Noah Dean
date: 2011-06-03 23:18:37.785000000

Patch Set 6:

That's more of a 3.0 thing -- where we actually break the API.

-------------------------------------
author: Andreas Short
date: 2011-06-03 23:19:35.896000000

Patch Set 6: I would prefer that you didn't submit this

(1 inline comment)



Line:72, src/main/java/net/spy/memcached/protocol/ascii/AsciiOperationFactory.java -> This should have a message indicating why it's unsupported.  Simply saying touch is a binary protocol only operation would be sufficient.  Would save someone a lot of hassle.

-------------------------------------
author: Noah Dean
date: 2011-06-03 23:21:36.403000000

Patch Set 6:

I think keeping it like set makes sense.  I don't think application developers generally are going to really care about waiting for the touch to finish, but it'll be possible.

Having asyncSet and asyncTouch would make sense in building out a new API.  It's a mistake, but I do think this is consistent in spirit.

Blair, do you think it's OK to go with this, or do you feel strongly that it should be asyncTouch?

-------------------------------------
author: Jakob Lozano
date: 2011-06-03 23:23:09.498000000

Patch Set 6:

I see your argument for naming it touch() to be consistent with the async set().  I don't feel strongly about the naming, though.

-------------------------------------
