[08:50:10] --- pguenther has become available [08:50:13] --- pguenther has left [10:31:59] --- stpeter has become available [10:32:16] --- stpeter has left [11:05:15] --- pguenther has become available [11:46:37] --- pguenther has left: Disconnected [12:07:34] --- ohm has become available [12:14:37] --- amelnikov has become available [12:30:46] --- Jim Galvin has become available [13:38:50] --- Jim Galvin has left: Replaced by new connection [13:42:05] --- Jim Galvin has become available [14:12:57] --- cyrus_daboo has become available [14:13:24] --- cyrus_daboo has left [14:36:01] --- Jim Galvin has left [15:12:18] --- cyrus_daboo has become available [15:25:12] I hear barry on mp3 [15:26:16] --- paf has become available [15:26:17] --- tonyhansen has become available [15:26:45] --- robsiemb has become available [15:26:54] --- cnewman has become available [15:27:09] --- paf has left [15:27:19] In the worst case I could scribe into jabber from the audio stream... [15:27:31] agenda is on the screen [15:27:37] --- cnewman has left [15:27:43] --- paf has become available [15:27:59] --- paf has left [15:28:16] --- cnewman has become available [15:28:18] --- lha has become available [15:28:24] --- raeburn has become available [15:28:27] --- cnewman has left [15:28:52] only diff in agenda from what was posted is that alexey's items have been moved up [15:29:08] sieve is an official WG [15:29:32] Actually audio quality is not bad [15:29:39] 3028bis: [15:30:14] changes from rfc: split refs; update refs to current versions; replace e-mail with email; 3028 errata; [15:30:43] --- robsiemb has left [15:30:55] Did 3028bis relax the side-effects, or is Phil still working on that? [15:30:55] --- robsiemb has become available [15:30:55] reject cancels implicit keep; acap references are changed to comparator draft; update grammer for legal utf-8, errors and typos; ipr boilerplate [15:31:01] 3028bis issues: [15:31:15] philip has charts about issues [15:31:36] --- robsiemb has left [15:31:50] --- robsiemb has become available [15:31:54] test [15:31:54] sync with collation draft; cr&lf are not whitespace -- legal in comments, strings or multiline literals? [15:32:38] should there be backslash escapes in quoted strings? (or out of scope?); do we update 821/822 dependencies to 2821/2822? [15:32:55] Does the collation draft say anything about the i; version and backwards compatability? [15:33:25] I think supporting multiple aliases for comparators should be no brainer [15:33:36] discussion [15:33:44] --- raeburn has left: Lost connection [15:34:48] There is more than one reason to recycle the based spec at Proposed [15:35:31] Ned: several reasons we'll have to recycle at proposed, and take the 6 month hit on moving to draft [15:35:58] I agree with Ned and that is what we were proposing - submit to IESG as proposed and then do implementation and move to draft [15:36:14] Changing 821/822 --> 2821/2822 will preclude moving to Draft in any near future, even though this is a right thing to do. [15:37:27] philip: how should cr&lf be handled? [15:37:39] these are bare cr's & lf's [15:38:09] barry: body searches may require them [15:38:54] philip: backslash escapes? spec only defines \ followed by char, saying it's that char, with no other meta characters allowed [15:39:32] would like to match against at least \0, and one implementation supports it [15:40:48] adding this definition would break current implementations [15:42:53] how do we add this? 1) change base spec to include \X; 2) extensions could be used to extend base definition; 3) leave alone, but extend variables spec to define ways of defining special characters [15:43:22] barry: does anyone have scripts that would be broken? [15:43:46] phil: none known, except for an error case [15:44:09] barry: pref is to change base spec [15:44:36] phil: needs to codify what \ extensions are definied [15:44:44] ned: he agrees [15:45:13] rob: can we do it with a capability? [15:45:23] nb: this was option 2 [15:45:52] rob: can this room speak for all implementations? [15:46:26] We could have a generic 3028bis 'require' capability that would cover all the things in the bis document that are not bacwards compatability, rather than having separate extensions for all these little things. [15:46:53] Cyrus: I agree [15:47:40] "sieve2" :-) [15:47:50] phil: will float the options on the list [15:47:52] sieverev1 ...\ [15:48:04] discussion [15:48:23] ned: this is how imap did it, how do people feel about that? [15:48:31] chris: don't unless you really really have to [15:48:53] Very painful in IMAP, but sometimes unavoidable. [15:49:09] on to 821/822 => 2821/2822 [15:49:21] --- robsiemb has left: Disconnected [15:49:37] barry: given we're recycling anyway, this should be okay [15:50:18] --- robsiemb has become available [15:50:27] --- robsiemb has left [15:50:34] --- robsiemb has become available [15:51:43] phil will send update to the list [15:51:43] --- robsiemb has left: Disconnected [15:51:51] on to imap flags draft [15:52:20] Tony is channeling for alexey while barry describes what was sent earlier [15:52:28] --- Hollenbeck has become available [15:52:44] new draft is out; changes: spelling, typos, minor edit, implementation notes error [15:53:23] suggested changes for next revision: make 1st parm optional, so can omit the name of the var, in which case the option is performed on a default global var; if no :flags, use default global var; [15:53:38] The suggested changes were discussed between Ned, Ken and myself. This is for backward compatibility with deployed -03 (CMU implementations). [15:53:59] this brings question: what happens when keep is specified more than once [15:54:16] head nods in the room; on to the next topic [15:54:17] As Rob is in the room, I want to hear his opinion [15:54:22] refuse draft? [15:54:25] no draft yet [15:55:12] philip: at san diego lunch bof, should reject be pulled out of base spec and combined with refuse, and provide a combined semantics that lets the implementation choose the best mechanism [15:56:06] There are two separate issues: (1) pull reject out of the base spec and put it into refuse, (2) technical details on whether we have a single action or multiple. [15:56:25] philip: list feedback caused a back off of this proposal [15:56:35] I think (1) is a good thing. I am still not sure about (2), there are multiple choices there [15:56:53] We agreed in DC to pull reject out of base spec into refuse document. That hasn't been done so far because we were waiting on the -00 base spec. Now that is here the refuse authors need to get to work on that... [15:56:54] --- ajf101 has become available [15:56:55] --- Hollenbeck has left: Replaced by new connection [15:57:07] more efficient to have smtp do the bounce [15:57:13] --- dbrashear has become available [15:57:30] randy: does the end user care? or is it the implementors who care? [15:58:20] IMHO, users should not care, unless they know the difference between MDN and DSNs :-) [15:58:22] barry: not every implementation can do refuse [15:58:43] --- robsiemb has become available [15:59:18] --- paf has become available [15:59:22] tony: channeling [15:59:26] --- paf has left [15:59:40] philip: if users knew they had a refuse option, they'd choose it [16:00:04] barry: don't know if there's an intermediate relay, so can't tell anyway, [16:00:04] --- robsiemb has left: Disconnected [16:00:09] ned: only the MSA can tell [16:00:26] Agree with Barry - we need to wait for the draft. [16:00:39] --- ajf101 has left [16:01:14] --- paf has become available [16:01:20] back to flags momentarily to answer alexey's query: rob was one of the people nodding [16:01:43] 3431 bis: new draft is out -- added definiition of operators [16:01:55] not in i-d repository yet [16:01:58] Actually the draft is in the id-tracker... [16:02:00] --- Hollenbeck has become available [16:02:06] subaddress 3598bis -- new -00 is out [16:02:16] The draft was published, but CRLFs are wrong [16:02:31] added note regarding local-part processing with multiple separator characters, fixed envelope test to only use "to" address [16:02:32] Subaddress is pretty much in the same state as relational - very minimal changes. [16:03:17] philip: change was for leftmost character, or do whatever local mta does [16:03:45] --- lha has left [16:03:47] on to spamtest-bis: -00 is out [16:03:47] --- paf has left: Disconnected [16:03:56] added :percent argument [16:04:04] issues: what range to use? [16:04:47] problems: relational only defines unsigned numeric comparator, not signed; can we get away with 0 for untested? [16:05:04] philip confirmed unsigned [16:05:13] --- lha has become available [16:05:31] --- dbrashear has left: Lost connection [16:05:55] barry: jutta said: you can have 10 numbers that you don't know what they mean, versus 100 numbers that you don't know what they mean [16:06:12] lisa: have a :tested/:untested separate from numbers [16:06:23] randy: agrees [16:06:32] OK, taking Lisa's suggestion: [16:06:49] randy: tools are using 0-100 now -- don't want to cause confusion [16:06:57] spamtest :untested will return true if untested, and spamtest :percent will return 0 - 100. I like that. [16:08:08] --- Hollenbeck has left: Replaced by new connection [16:08:14] --- Hollenbeck has become available [16:08:29] --- klensin-ietf has become available [16:09:05] Yes, I am spamtest author... [16:09:48] I will scribe for now. [16:09:54] On to variables draft. [16:10:04] last call complete [16:10:09] some minor changes done [16:10:25] --- Hollenbeck has left: Replaced by new connection [16:10:26] --- Hollenbeck has become available [16:10:30] remaining issue is what to do about ${0}? [16:10:52] someone is talking but I cannot hear.... [16:11:13] Philip: return match string can be expensive [16:11:21] --- tonyhansen has left: Replaced by new connection [16:11:28] --- robsiemb has become available [16:11:58] --- tonyhansen has become available [16:12:01] Consensus is to use separate test On to variables draft: -01 Moved refs from appendix into doc; Added SET example with a comparator Change d”highest value” to “largest value” Update utf-8 ref Remaining issue: what to do about ${0}, return match string or # of matches Phil: often expensive to return match string [16:12:02] Ned: prefers reteurn the match string, because for glob style matching it is the only way. [16:12:17] on to vacation draft [16:12:20] vacation draft: [16:12:33] I'll let tony take over now. [16:12:38] Ned taking on as editor [16:12:53] outstanding issue: add :from arg, use auto: [16:13:29] ned: any suggestions on what draft should say for controls on what :from says? [16:13:37] --- randy has become available [16:13:44] dave crocker: how is diff from 2822 from? [16:13:46] ned: not [16:13:54] ned: implements may set policy [16:14:27] philip: don't vac messages be submitted via MSA? [16:14:28] phil: do we need to say anything? [16:14:43] --- Hollenbeck has left: Disconnected [16:15:04] We need to spell out the possibility of spoofing in Security Considerations. [16:15:19] dave: his concern: two choices -- point to another doc for policy, or try to say policy and be incomplete and unhelpful [16:15:42] One issue might be that the sieve engine has a different trust relationship with the submission server than indicidual users/MUAs [16:15:48] barry: is it reasonable for us to just say we follow 2822's rules [16:16:01] --- lha has left [16:16:06] so saying that the same policies apply may not be enough. [16:16:23] ned agrees w/ cyrus [16:16:53] ned: extract from and become wierd remailer service? [16:17:14] randy: thinks it would be useful to mention something in there [16:17:37] randy: yes, forging is able to be done, but worth mentioning something about MSA's trust relationship [16:17:56] ned: will add something to security considerations section [16:18:09] on to body test: -00 [16:18:27] minor changes. can we last call? [16:18:46] scott: how can -00 be only minor changes? [16:19:01] philip: minor changes from jutta's previous draft [16:19:26] jutta had 4 revs [16:19:41] wglc announcement will come out soon [16:19:49] edit header draft: also ready for WGLC [16:19:53] notify draft [16:20:02] no new draft; barry is taking it on [16:20:37] discussion on list about using associative arrays, variables interactions, URI use? [16:21:14] previous authors were a coworker of barry and someone we can't find [16:21:25] more discussion on the list [16:21:33] regex: no new draft [16:22:03] ned: had an action item to look at i18n issues [16:22:47] ned: discussed stuff with jon klensin, will send stuff to list [16:22:52] Thanks Ned! [16:23:04] Several of us took at action item, but my impression was that we got it sorted ot much more easily than I expected. [16:23:22] mime part tests: no new draft, ned has issues [16:23:40] The plan for mime test was for Tony and I to merge our drafts - we have not done so to date :-( [16:24:40] ned: is a fan of modifying header test, more than looping [16:24:55] in particular, would like to use :exists on inner headers [16:25:14] mime-exists, or :mime [16:25:24] tony: turning to mike [16:25:56] --- klensin-ietf has left [16:26:47] tony: short description of loop [16:26:57] ned: almost turing complete? possibly too complex [16:27:31] philip: does any of this belongs in sieve? [16:28:05] dave: was pushed over the edge? [16:28:36] --- pguenther has become available [16:28:47] --- levigner has become available [16:29:04] dave: most real world filters in the masses are really simple things [16:29:44] --- masinter has become available [16:29:48] tony: other place loop works well is in combination with replace-body-part [16:30:38] ned: sees two applications of sieves: 1) GUI that generates simple sieves, 2) system wide sieves [16:31:11] ned: #2 people want everything, but are also worried about performance [16:31:41] --- randy has left: Disconnected. [16:32:06] barry: if have :mime, how do you distinguish between inner mime From test and outer mime From test [16:33:03] brent welsh: prefers an escape into a more powerful language [16:33:40] barry: that concept was pushed back on before [16:34:11] dave: core set of requirements that are simple, has sieve crossed the line [16:34:32] ned: none of these things are in the base, but extensions [16:35:30] some extensions will never be implemented in some implements, or have site policies to disable [16:37:25] My audio is dropping out... :-( [16:37:33] tony: gave some context behind [16:38:02] options will be taken to the list [16:38:18] milestones! [16:38:25] some are done, some are late [16:38:41] vacation, 3028bis, variables are all late [16:40:04] phil: base speci will be out in a week, but will then be out of contact for a bit, so responses to issues will be delayed [16:40:09] We are missing a milestone for the next vacation draft submission. When can Ned do that? [16:40:44] ned: this week [16:41:12] lots more milestones moved out ot march, april may and june [16:41:21] several are going to wglc [16:41:28] --- klensin-ietf has become available [16:42:02] Milestones: we've got a lot going on in Mar/Apr - is it too much? [16:42:17] Should we spread things out a little bit more? [16:43:01] barry: most is wglc's and things going to iesg [16:43:07] Cyrus: Items like "sending a document to IESG" is mostly work for you, me and authors. [16:43:19] philip: spread out among lots of people [16:43:32] --- amarine has become available [16:43:35] --- amarine has left [16:43:36] barry: be aggressive [16:44:09] barry: ? to scott; would he prefer a relaxed date that we beat? scott: yes. [16:44:30] I want to make sure we do get adequate review of drafts in last call. I suspect we need to push the milestones out past Paris... [16:44:32] philip: changes due to recycle to proposed will require more work [16:45:07] barry: cyrus & alexey need to shuffle the milestones [16:45:17] I think we've already tried agressive milestones, let's not do that again. [16:45:20] Alexey and I will work on milestones. [16:45:32] we're done [16:46:57] See you [16:47:52] --- robsiemb has left: Disconnected [16:48:10] --- amelnikov has left [16:48:13] Thanks, folks, sorry I wasn't there... [16:49:48] ta ta [16:49:56] --- tonyhansen has left [16:55:41] --- cyrus_daboo has left [16:58:16] --- klensin-ietf has left [17:13:44] --- levigner has left: Disconnected [17:24:54] --- levigner has become available [18:08:21] --- ohm has left [18:21:48] --- pguenther has left: Disconnected [19:29:52] --- levigner has left: Replaced by new connection [19:29:53] --- levigner has become available [19:44:53] --- levigner has left: Replaced by new connection [19:44:53] --- levigner has become available [19:44:53] --- levigner has left [19:48:56] --- levigner has become available [19:49:52] --- levigner has left [20:53:43] --- masinter has left: Replaced by new connection