[03:40:08] --- filip.navara has joined
[06:21:48] --- filip.navara has left
[12:57:45] --- kmurchison has joined
[13:39:29] --- avel has joined
[13:55:05] --- alexeymelnikov has joined
[13:58:43] --- cyrus_daboo has joined
[13:59:19] <alexeymelnikov> Hopefully, I can get the rest of people to talk about Sieve interop soon
[13:59:41] <kmurchison> OK
[14:03:05] <cyrus_daboo> Fine.
[14:09:44] --- barryleiba has joined
[14:09:48] --- Dave Cridland has joined
[14:09:59] <Dave Cridland> Hi.
[14:10:08] --- arnt has joined
[14:10:14] <barryleiba> Hola
[14:10:21] <arnt> hi
[14:10:27] --- aaronstone has joined
[14:10:35] <arnt> dobry den
[14:10:47] <avel> Hello!
[14:10:48] --- filip.navara has joined
[14:10:55] <cyrus_daboo> Hello...
[14:11:07] <cyrus_daboo> Enjoying some beer in Prague?
[14:11:13] <Dave Cridland> Sadly not.
[14:11:21] <Dave Cridland> We're doing Sieve instead. :-)
[14:11:28] <avel> !!!
[14:11:29] <avel> :-)
[14:11:33] <kmurchison> This should be a bar BOF
[14:11:34] <cyrus_daboo> That's Ok then.
[14:11:55] <cyrus_daboo> Except I am stuck at my desk!
[14:11:58] <Dave Cridland> Alexey is talking about interop testing and moving Sieve to draft, potentially.
[14:12:25] <kmurchison> I'm in my home office. The beer fridge is only 5 second away ;)
[14:12:35] <Dave Cridland> Alexey says we should create a set of Sieve scripts and associated messages.
[14:12:40] <avel> Ken beat me, mine is 30 seconds away.
[14:12:54] <Dave Cridland> Barry points out that at least two people need to create the scripts.
[14:13:21] <Dave Cridland> Alexey says he has a set of 30 or so scripts that test basic features.
[14:13:37] <Dave Cridland> Barry says that creating scripts is out of scope.
[14:14:05] <cyrus_daboo> I agree with Barry.
[14:14:08] <avel> I think it would be good to see first, how many servers and clients there are, and the capabilities of each. Even though I have a picture of this in my mind, more-or-less.
[14:14:08] <Dave Cridland> Arnt thinks he has about 100 scripts, some of which would be useful for testing.
[14:14:22] <Dave Cridland> Ahem.
[14:14:31] <cyrus_daboo> What we need is a set of input messages, a set of scripts, and a description of the results of apply those scripts to the inputs.
[14:14:36] <Dave Cridland> Some of which would be useful testing other Sieve implementations to his.
[14:16:17] <kmurchison> The Cyrus Sieve source has a harness for testing comparators that might be of some use if I can make it generic enough
[14:17:25] <Dave Cridland> (Some more to-ing anf fro-ing between Alexey and Barry about what exactly we need to test)
[14:18:09] <cyrus_daboo> We need to test that each sieve system filters the same messages in exactly the same way.
[14:19:03] <aaronstone> I have a command line tool to exercise my engine, that takes a script and checks its syntax, or takes a script and a message and reports the actions taken
[14:19:33] <Dave Cridland> I should point out Alexey is not really on Jabber.
[14:19:47] <aaronstone> i'd guess that it would be the easiest piece of code for Ken to extend, basically just swap out the callback registration to the Cyrus functions
[14:19:47] <Dave Cridland> S you need to type really loudly for him to hear.
[14:20:34] <kmurchison> Aaron: are we discussing trying to make the Cyrus 'test' program usable for everybody?
[14:21:49] <Dave Cridland> Barry says we need to test in combination as well as individually.
[14:21:57] <Dave Cridland> And we need to test every MUST and MUST NOT.
[14:22:07] <cyrus_daboo> What does "in combination" mean?
[14:22:48] --- alexeymelnikov has left
[14:23:09] --- filip.navara has left
[14:23:09] --- filip.navara has joined
[14:23:11] --- filip.navara has left
[14:23:18] <barryleiba> If the spec says that using fileinto and discard together makes <this> happen...
[14:23:31] <barryleiba> then we need to test that using fileinto and discard together makes <this> happen.
[14:24:07] <avel> Shall we also test ManageSieve? (e.g. SASL, script upload)
[14:24:19] <aaronstone> Ken: I think I just proposed porting my port of your tool back to your code base. Thus, nevermind, I think we've got everything covered ;-)
[14:24:29] <cyrus_daboo> Barry - Ok I see what you mean.
[14:24:31] <Dave Cridland> Avel: Nope. Just 3028bis base spec revision.
[14:24:37] <avel> OK.
[14:25:14] <Dave Cridland> avel: So we're testing the engines themselves.
[14:25:38] <aaronstone> Avel: have you seen the complete Sieve parser written in PHP? http://libsieve-php.sourceforge.net/
[14:25:40] <avel> Yes, I thought this interop had a broader scope. However, I _can_ test an experimental sieve parser
[14:25:41] <kmurchison> Aaron: Thx, I'm always in favor of less work for me.
[14:25:44] <Dave Cridland> The new addition to 3028bis is the unicode encoding Thing
[14:25:51] <avel> Hmmm Aaron you beat me to it by milliseconds
[14:26:15] <Dave Cridland> avel: 2,000 milliseconds, but yes.
[14:26:21] <avel> :-)
[14:27:07] --- kjetilho has joined
[14:31:10] <Dave Cridland> (Lots of muttering about if anyone implementing encoded-character extension)
[14:32:02] <kmurchison> Cyrus Sieve doesn't have it
[14:32:17] <avel> (which extension is that?!)
[14:32:26] <Dave Cridland> 3028bis section
[14:37:45] <Dave Cridland> Arnt's now arguing that we shouldn't have it.
[14:38:16] <Dave Cridland> Or rather, arguing that if the implementation uses UCS2/UTF-16 for strings, it cannot handle arbitrary octets in strings.
[14:38:53] <Dave Cridland> Barry says:
[14:39:09] <Dave Cridland> Someone needs to go through the spec and pick out each item that needs to be tested.
[14:39:56] <Dave Cridland> Barry volunteers for doing this, but we need more volunteers.
[14:40:14] <Dave Cridland> Arnt says he has done this for RFC3028, and will do some more for 3028bis.
[14:41:10] <Dave Cridland> So we then need some test scripts, and some test messages.
[14:42:05] <Dave Cridland> We also need to talk to the ADs about how strict they need the testing to be.
[14:48:51] --- barryleiba has left
[14:48:51] --- barryleiba has joined
[14:51:17] --- barryleiba has left
[14:52:40] <Dave Cridland> Kjetil and Alexey will do the scripts.
[14:53:29] <Dave Cridland> List of features (Barry, Arnt) will produce feature test list by first week of May.
[14:54:58] <Dave Cridland> Jully, discuss status in Chicago.
[14:55:41] <Dave Cridland> And thus concludes the Sieve Interop Discussion Meeting.
[14:55:55] --- Dave Cridland has left
[14:55:59] --- avel has left
[14:56:07] --- kjetilho has left
[14:56:09] --- aaronstone has left
[14:58:19] --- arnt has left
[15:00:54] --- cyrus_daboo has left
[15:02:04] --- kmurchison has left
[15:10:04] --- arnt has joined
[15:17:36] --- arnt has left