= ReviewerMeeting20081126 = == summary == * al-maisan mentored by gmb * barry to move OCR to another day * beuno wants more graphs and data collection on new features * `from canonical.launchpad.interfaces import branch` is okay * Top merge proposal bugs: * 245257 - 5 (diffs/bundles) * 202000 - 3 (complete diffs) * 300462 - 2 (better follow up diffs) * 297547 - 2 (CC someone via email) * 297669 - 2 (Approve a mp via email) * 295156 - 1 (one mail on merge proposals) * 297640 - 1 (quote comments in web ui) * 301836 - 1 (dashboard - difficult to figure out what to do) * 264905 - 1 (access via launchpadlib) * 302792 - 1 (do not strip diffs from emails to m-ps) (note: there was some additional post-meeting feedback) == log == {{{ Nov 26 10:00:37 #startmeeting Nov 26 10:00:38 Meeting started at 09:00. The chair is barry. Nov 26 10:00:38 Commands Available: [TOPIC], [IDEA], [ACTION], [AGREED], [LINK], [VOTE] Nov 26 10:00:43 * bac (n=bac@canonical/launchpad/bac) has joined #launchpad-meeting Nov 26 10:00:51 hello everyone and welcome to this week's ameu reviewers meeting. who's here today? Nov 26 10:00:54 me Nov 26 10:00:55 me Nov 26 10:00:56 me Nov 26 10:00:56 me Nov 26 10:01:00 me Nov 26 10:01:07 me Nov 26 10:01:34 * flacoste (n=francis@canonical/launchpad/flacoste) has joined #launchpad-meeting Nov 26 10:01:43 me Nov 26 10:01:47 moo Nov 26 10:01:48 me Nov 26 10:01:53 * abentley1 is now known as abentley Nov 26 10:02:18 * mars was thinking of 'moo', but chicken'd out Nov 26 10:02:31 cluck cluck Nov 26 10:02:47 [TOPIC] agenda Nov 26 10:02:48 New Topic: agenda Nov 26 10:02:58 * Roll call Nov 26 10:02:58 * Getting graphs in place before rolling out features (way before) - beuno Nov 26 10:02:58 * Merge proposal bugs Nov 26 10:02:58 * 245257 - 5 (diffs/bundles) Nov 26 10:02:58 * 202000 - 3 (complete diffs) Nov 26 10:02:58 * 300462 - 2 (better follow up diffs) Nov 26 10:02:58 * 297547 - 2 (CC someone via email) Nov 26 10:02:58 * 295156 - 1 (one mail on merge proposals) Nov 26 10:02:58 * 297640 - 1 (quote comments in web ui) Nov 26 10:02:58 * 297669 - 1 (Approve a mp via email) Nov 26 10:02:58 * 301836 - 1 (dashboard - difficult to figure out what to do) Nov 26 10:02:58 * 264905 - 1 (access via launchpadlib) Nov 26 10:02:58 * If there's time, the old boring script Nov 26 10:02:58 * Next meeting Nov 26 10:02:58 * Action items Nov 26 10:02:58 * Queue status Nov 26 10:02:58 * Mentoring update Nov 26 10:02:58 * al-maisan looking for a mentor Nov 26 10:03:14 [TOPIC] * Getting graphs in place before rolling out features (way before) - beuno Nov 26 10:03:19 me Nov 26 10:03:25 beuno: the floor is yours Nov 26 10:03:27 New Topic: * Getting graphs in place before rolling out features (way before) - beuno Nov 26 10:03:43 hello hello Nov 26 10:03:57 so, I sent out an email to the list about this Nov 26 10:04:07 and since everyone seemed on board with the idea Nov 26 10:04:15 I'd like to formalize it somehow for the reviews Nov 26 10:04:48 the main idea is to have enough information beforehand being collected, to be able to track the impact of the feature/change Nov 26 10:04:52 me (late) Nov 26 10:05:01 me too Nov 26 10:05:26 beuno: I don't really get it. What features do we graph, and how do we know what we care about? Nov 26 10:06:45 I wonder more about how much overhead this introduces to our development process Nov 26 10:06:59 beuno: like abentley i'm a little confused and would benefit from a fleshed-out example. Nov 26 10:07:16 so, if you're going to add diffs to merge proposals, make sure we have a graph on how many merge proposals we have per day Nov 26 10:07:27 maybe how many comments per merge proposal Nov 26 10:07:37 and that we we can measure what kind of impact the feature had Nov 26 10:07:44 is it making users use the feature more? Nov 26 10:07:49 are they commenting less? Nov 26 10:08:08 basically enough information to be able to interpret them Nov 26 10:08:17 beuno: i have no idea how to create graphs. i know how to collect data i might be interested in in my db tables tho Nov 26 10:08:46 beuno: are we basically talking about making sure we're collecting interesting usage data for our new features, or is there more to that? Nov 26 10:08:51 abentley, a good example of a back-fill graph would the the 'Review type' feature. Do people ever set the review type? How often? Nov 26 10:09:00 barry: i think it's adding cricket graph Nov 26 10:09:01 barry, it's just opening an RT ticket with the sql query to be ran every day Nov 26 10:09:25 * kiko-afk (n=kiko@canonical/launchpad/kiko) has joined #launchpad-meeting Nov 26 10:09:27 hey there Nov 26 10:09:32 can you guys ask martin through email Nov 26 10:09:35 kiko-afk, you're here! Nov 26 10:09:35 or next century Nov 26 10:09:38 kthxbye Nov 26 10:09:39 * kiko-afk (n=kiko@canonical/launchpad/kiko) has left #launchpad-meeting ("Left to bother somebody else") Nov 26 10:09:45 * beuno closes laptop Nov 26 10:10:14 okay! i guess we'll continue that discussion on the ml Nov 26 10:10:17 now *that* was a drive-by Nov 26 10:10:19 From now on let it be known: Nov 26 10:10:24 kiko is beuno's mother Nov 26 10:10:29 Ha! Nov 26 10:10:46 now that they're gone, it's time to party! Nov 26 10:10:53 [TOPIC] * Merge proposal bugs Nov 26 10:10:54 New Topic: * Merge proposal bugs Nov 26 10:11:14 thanks everyone who sent me their feedback. here are the totals so far: Nov 26 10:11:22 * 245257 - 5 (diffs/bundles) Nov 26 10:11:22 * 202000 - 3 (complete diffs) Nov 26 10:11:22 * 300462 - 2 (better follow up diffs) Nov 26 10:11:22 * 297547 - 2 (CC someone via email) Nov 26 10:11:22 * 295156 - 1 (one mail on merge proposals) Nov 26 10:11:22 * 297640 - 1 (quote comments in web ui) Nov 26 10:11:22 * 297669 - 1 (Approve a mp via email) Nov 26 10:11:22 * 301836 - 1 (dashboard - difficult to figure out what to do) Nov 26 10:11:22 * 264905 - 1 (access via launchpadlib) Nov 26 10:11:23 beuno: I am skepical that we'll know the right things to graph before we land the feature. I am skeptical that it will be physically possible to create the graph before we land the feature. Nov 26 10:11:40 the first number is the bug number, second is the # of people who named that bug, followed by a brief summary Nov 26 10:11:41 * jtv has quit (Read error: 110 (Connection timed out)) Nov 26 10:11:42 bug301836 += 1 Nov 26 10:12:04 bac: i guess make that 2 :) Nov 26 10:12:20 it's clear to me that 245257 is our top priority Nov 26 10:12:30 202000 and 300462 are closely related Nov 26 10:12:31 * al-maisan (n=al-maisa@p5087B706.dip.t-dialin.net) has joined #launchpad-meeting Nov 26 10:12:42 my idea about 297669 is similar to what 301836 will provide (i.e. if we finally start using the merge-proposal status and set it as "Approved" instead of just voting, we'd be able to see what is there that still needs to be reviewed) Nov 26 10:12:52 does anybody have any addition comments, and then i'd like to get abentley and rockstar 's feedback Nov 26 10:13:36 barry, I know MOST of those are being worked on in some form or another. Nov 26 10:14:15 rockstar: ok, so they'll be done by next week? ;) Nov 26 10:14:19 * danilos runs and hides Nov 26 10:14:34 danilos, well, for some of them, yes. Nov 26 10:14:41 rockstar: i'm off for thanksgiving weekend, so if you could land those my monday, that would be great kthxbye Nov 26 10:14:47 :) Nov 26 10:14:59 barry, but I need turkey too! Nov 26 10:15:07 abentley: clearly 245257 is our #1 request. any word on that? Nov 26 10:15:24 barry: 300462 +1 (i didn't vote on my bug) Nov 26 10:15:31 barry: The general case will take a while, because we need a job processing system to accomplish it. Nov 26 10:15:53 rockstar: http://markmail.org/message/bve73p3qumsij3he Nov 26 10:16:14 flacoste: thanks, i'll up the tally on that one Nov 26 10:16:23 barry: And that was blocked by sabdfl until quite recently Nov 26 10:16:46 abentley: ack Nov 26 10:17:07 barry: I am currently working on support for merge directives. That will be able to use the diff from the merge directive. Nov 26 10:17:20 abentley: that. will. rock. Nov 26 10:18:05 abentley, rockstar do you think the team will find this list helpful? Nov 26 10:18:17 barry, I know I do. Nov 26 10:18:18 barry: Not really. We know all this stuff. Nov 26 10:18:55 The bugs at the top I knew are important, but the ones farther down are helpful in figuring out which bugs tagged code-review are most important. Nov 26 10:19:41 Because there are probably 30 bugs, and it's good to know which 10 of those are most important to the rest of the lp team. Nov 26 10:19:48 barry: Code review by email was intended to support attachments. We'll need to investigate why they don't. Nov 26 10:19:55 cool, what say i update the tallies after this meeting and send it on to the list. maybe someone on the team can provide a feedback on a rough alignment of these with your own priorities? Nov 26 10:20:21 * salgado is now known as salgado-lunch Nov 26 10:21:00 abentley: thx Nov 26 10:21:22 BjornT: what do you think? Nov 26 10:22:11 barry: BjornT's all sprinty with beuno, kiko and intellectronica. Nov 26 10:22:22 So his responses might be delayed somewhat. Nov 26 10:22:28 thumper is working on 295156 Nov 26 10:22:48 abentley: do you know they don't, i assumed they didn't but didn't try it out Nov 26 10:22:53 gmb: k Nov 26 10:22:54 297669 seems reasonable and easy Nov 26 10:23:13 thumper is also working on 301836 Nov 26 10:23:25 rockstar is working on 264905 Nov 26 10:23:27 abentley: are the email commands documented in the help wiki? Nov 26 10:23:34 barry: yes. Nov 26 10:23:36 * rockstar nods Nov 26 10:23:40 abentley: great! Nov 26 10:24:32 anything else on this topic? Nov 26 10:25:02 300462 will take even longer than a general solution for 202000, because sabdfl doesn't want us to do this for every project on launchpad. Nov 26 10:25:44 abentley: why not? Nov 26 10:25:46 Instead, there will be a script you can run in an EC2 instance. Nov 26 10:25:53 barry: Resource consumption. Nov 26 10:26:05 ah Nov 26 10:26:51 abentley: for the record, if attachment are handled, i consider 300462 as not really worth it Nov 26 10:27:19 abentley: developer are used to and can attach the diff to the follow-up email, i though attachments weren't supported Nov 26 10:27:38 297547 is impossible, because sabdfl doesn't want to allow subscriptions to code reviews. Nov 26 10:27:40 and thumper didn't told me they were when I discuss this aspect with him and he told me to file a general bug about that aspect Nov 26 10:28:31 flacoste: We intended attachments to work. They don't. We haven't had a chance to find out why. Nov 26 10:29:05 abentley: then fixing attachment is enough for me to consider that bug complete Nov 26 10:29:06 I only verified it yesteday. Nov 26 10:29:11 flacoste: Cool Nov 26 10:29:56 thanks everyone. i appreciate the discussion. i think it will make mp's really excellent to use. moving on... Nov 26 10:30:14 i'm going to skip around a little bit... Nov 26 10:30:17 [TOPIC] * Mentoring update Nov 26 10:30:17 New Topic: * Mentoring update Nov 26 10:30:23 * al-maisan looking for a mentor Nov 26 10:30:31 have we rounded up a mentor for al-maisan yet? Nov 26 10:30:35 barry: yes Nov 26 10:30:40 al-maisan: excellent, who? Nov 26 10:30:46 gmb volunteered Nov 26 10:30:48 :) Nov 26 10:31:00 fantastic! al-maisan welcome aboard. gmb thanks! Nov 26 10:31:09 thanks! Nov 26 10:31:13 * bigjools winks at gmb and hands over the fiver Nov 26 10:31:22 [TOPIC] * Action items Nov 26 10:31:23 New Topic: * Action items Nov 26 10:31:24 Heh. Nov 26 10:31:27 * barry to look into techniques for eliminating back-patching of schema types (avoiding circular imports) Nov 26 10:31:46 i have done nothing on this yet, but i plan on looking at it over the break. i think it will be easy Nov 26 10:32:13 [TOPIC] anything else? Nov 26 10:32:14 New Topic: anything else? Nov 26 10:32:22 does anybody have anything not on the agenda today? Nov 26 10:32:30 Yes. When am I a reviewer? Nov 26 10:32:37 OCR, I mean? Nov 26 10:32:39 :) Nov 26 10:33:20 abentley: ah. what would you prefer? i'm very happy to say we have great coverage for eu/am Nov 26 10:33:48 abentley: i'm also happy to move if you prefer mondays Nov 26 10:33:57 abentley, you get the Monday AsiaPac shift! :) Nov 26 10:34:16 but i think we'll be doubling up no matter what we do, unless some of us move to au/nz :) Nov 26 10:34:31 doubling up is not a bad thing! Nov 26 10:34:52 I think Mondays are best for me. But if al-maisan and gmb are doing mondays, that doesn't make sense, does it? Nov 26 10:35:10 abentley: Well, you start fairly late in our day. Nov 26 10:35:13 Fridays might also work. Nov 26 10:35:14 hmm .. different time zones Nov 26 10:35:18 So it's not *that* mcuh of an overlap. Nov 26 10:35:22 yep Nov 26 10:35:25 abentley: What's your timezone? Nov 26 10:35:28 On Monday, I haven't spoken to thumper for 2 days. Nov 26 10:35:31 gmb, same as mine Nov 26 10:35:36 ... Nov 26 10:35:38 abentley: i'm also happy to move to another day if you really want mondays Nov 26 10:35:45 mars: Which one is that? Nov 26 10:35:49 * gmb has the dumb Nov 26 10:35:53 gmb, EST Nov 26 10:35:57 Ah, okay. Nov 26 10:36:35 So, abentley is al-maisan -6, gmb -5. That's not a terrible overlap. Nov 26 10:37:12 abentley: your call. pick a slot Nov 26 10:37:14 Okay. I guess I didn't realize you were going off shift soon after I arrived. Nov 26 10:37:31 barry: Okay, Mondays please. Nov 26 10:37:51 abentley: it's yours. i'll find a time to move to Nov 26 10:37:59 barry: ta Nov 26 10:38:11 cool. any other topics for today? Nov 26 10:38:38 5 Nov 26 10:38:48 barry: Is it true that we're no longer adding entries to lib/canonical/launchpad/interfaces ? Nov 26 10:39:07 abentley: more specific imports are preferred, yes Nov 26 10:39:20 i.e. you should do "from canonical.launchpad.interfaces.branch import Branch"? Nov 26 10:39:38 abentley: correct Nov 26 10:40:05 Okay. What about "from canonical.launchpad.interfaces import branch"? Nov 26 10:40:36 abentley: i have no problem with that if it helps avoid name conflicts Nov 26 10:40:52 i.e. implements(branch.IBranch) Nov 26 10:41:06 barry: it helps reduce the number of imports tremendously. Nov 26 10:41:39 abentley: barry: That is common practice in Zope code Nov 26 10:42:28 sinzui: Also in Bazaar, though both forms are accepted. Nov 26 10:42:53 i think i'd like to see what that looks like if it's used more frequently. i like the shorter names and don't mind the longer imports, but if you want to start using that style, let's see how it looks Nov 26 10:43:13 Okay. Nov 26 10:43:48 anything else with our last 2 minutes? Nov 26 10:43:54 barry: It may be a solution to the insanely long soyuz names Nov 26 10:44:18 sinzui: that in itself would be enough! :) Nov 26 10:44:59 leave our long names alone! Nov 26 10:45:29 and with that, we're done! :) Nov 26 10:45:30 It's not the size that counts, it's what you do with it. Nov 26 10:45:34 #endmeeting }}}