Session Start: Mon Mar 06 22:04:08 2006 Session Ident: #twiki_edinburgh [22:04] * Now talking in #twiki_edinburgh [22:04] Hello to all [22:04] hi kenneth! [22:05] cdot tells me there are some really good mexican reds, i need to try more! :) [22:06] Hellow everybody. Can we get started? [22:07] fine with me [22:07] is lynnwood joining? [22:08] I guess he would be here by now if he was; he's not in #twiki [22:09] sf has some problems with mail [22:09] possibly not all got the invitation [22:09] bitca was going to join, but last I heard she was being mugged by her cats [22:09] well, it is 7 min past [22:09] shall we start? [22:10] http://twiki.org/cgi-bin/view/Codev/EdinburghReleaseMeeting2006x03x06 [22:10] we have two items on the agenda [22:10] 1. TWiki 4.0.2 release [22:10] 2. Post-Dakar development model [22:10] any other subjects to discuss? [22:11] who is taking the minutes? [22:11] Possibly edin* focus [22:12] yea, we need closure on this [22:12] btw, shall we rename edinburg to edin? [22:12] saves typing [22:12] Good [22:13] or we can just nickname it Edin* [22:13] you can pronounce it the way the locals do; enburra [22:13] I never memorize if its burough, borogh, borough [22:14] Ed* [22:14] hell, keep it simple; Entebbe [22:14] In the end its just like Chris the Burgh [22:14] or is it Chris de Burgh, sigh [22:14] and the german word for fortress - burg [22:14] can we start? [22:15] sure [22:15] but who is taking the minutes? [22:15] Lavr: did such a superb job on the agenda, it only seems fair to let him so the minutes as well [22:16] steffen is not here today, someone else should jump in [22:16] OK. If I can do it afterwards based on log. I type too slowly to both chat and minute [22:16] ok, that is fine [22:16] * bitca has joined #twiki_edinburgh [22:16] a simultaneous edit feature would be nice for there types of meetings [22:17] if the experiment at WikiSym is anything to go by, simultaneous edit is a disaster waiting to happen [22:17] it just ended up with 10 people all taking their own minutes on the same page [22:17] ---+ TWiki 4.0.2 release [22:18] where do we stand, what needs to happen? [22:18] crawford, you have been most active [22:18] for the most part, we are waiting for 2 things [22:18] (1) merges and (2) feedback [22:18] I have already merged 70% of the fixes, but I have run out of time and the will to live [22:19] as suggested by my recent mail [22:19] the remaining merges really need Steffen [22:19] as I can't nail down what he did. [22:19] That's the "most part". The rest is Pattern Skin [22:19] and I can't comment on that. [22:20] what in develop is not going into twiki 4 patch release? [22:20] most of the checkins probably should not go in, IMHO. [22:20] they are fine for 4.1, but not for 4.0.2 [22:21] 4.0.2 should be bugfixes *only* [22:21] what about all the doc updates? [22:21] ? [22:21] what doc updates? [22:21] and what if we talk about 4.1 instead of 4.0.2? [22:21] there have been many doc commits in develop marked as for twiki4 [22:22] have there? [22:22] * CDot looks [22:22] possibly we should first talk about 2. Post-Dakar development model since there are some confusions? [22:23] I'm utterly confused myself [22:23] i am for sure [22:23] As long as PatternSkin update is uncluded! That is the important one for closing THE BUG + More. And it needs to get merged fast so we can give it proper test time. It will need a lot of testing which I am already doing now. [22:23] included .... not uncluded :-) [22:24] Nice coining of a term, howeveer [22:24] so what to include and what to unclude [22:24] so, Lavr is saying that the new pattern skin *is* a bugfix, and should go in [22:24] right? [22:25] (to 4.0.2, that is0 [22:25] i agree [22:25] Well, it's both, so yes [22:25] except that users will need some work to it [22:25] that would make it a major release [22:25] although 'major' work is prob too much [22:25] er, or a minor release (which would be 4.1) --- you're saying it's not a point release? [22:26] there are bugfixes that are, actually, *major* features [22:26] No. The admins. It is pretty safe to TOPICS. That is where the investment is. Fixing a top bar and maybe a few other tayloring is not the end of the world. The BUG is. [22:26] I assume everyone has read http://twiki.org/cgi-bin/view/Codev/ProcessForTWiki4PatchReleases? [22:26] y [22:26] y [22:26] rofl. [22:26] y [22:26] brievly [22:26] * wbniv needs to reread it from all the updates that happened while i was sleeping :) [22:26] or, rather, *blush* [22:26] * wbniv loading it up again...; [22:27] ok, lets go to agenda item 2 first [22:27] ok, then the distinction between "patch", "minor" and "major" at least has a working definition [22:27] we are talking here about what goes into a *patch* release, 4.0.2 [22:28] if everyone agrees that what Lavr has reported is a *bug* rather than a functional change, it should go to 4.0.2 [22:28] if it is a *functional change* then perhaps we should be talking about 4.1.0 instead [22:28] but that bugfix may force admins to update their skins. [22:29] Toughinski [22:29] People are copying text from TWiki into MS word to be able to read the topics. That is a show stopper problem. [22:29] the bug fix has led me to do a functional change [22:29] And, for god's sake, twiki.org should be running Dakar!! [22:29] functional change or not, it is a usability bug that needs to be fixed in 4.0.2 [22:29] But I advice to update ASAP [22:29] it's a semantic point, but I agree with CDot, we should be really talking about 4.1.0 because of the Pattern bugfix. [22:30] I don't care actually [22:30] I will go with the majority on this; but what I *don't* want to happen is this to be used as a way to open the floodgates on 4.0.2 [22:30] I think that we don't even have to vote if pattern goes or not into the next release. [22:30] Then call it 4.1. The important thing for me is to get a release without this bug i.e. with the new PatternSkin. [22:30] so we end up with *every* checkin on DEVELOp going in [22:30] Yes [22:30] which I see as a *major* mistake [22:30] To what Lavr said, that is [22:31] the functional change i see is to move the jump box and search box from the sidbar to the topbar [22:31] that requires admins to update _all_ their sidebars [22:31] if users keep their webleftbars, nothing will change for them [22:31] it just works [22:31] instead they can change the top bar [22:31] PeterThoeny: Functional change don't always mean UI changes. The css work "differently" than before. [22:31] which they prob have already [22:31] don't they have 2 boxes in the soidebar and wo in topbar? [22:32] its no big deal, small change [22:32] s/wo/2/ [22:32] I dunno. I'm using the new wonderful patternskin and the jump and search boxes moved to the top. [22:33] AFAIK, I didn't do anything to cause that. (Not that I'm unhappy) [22:33] arthur: assuming this change is done, are there any files/topics an admin needs to touch on upgrade? [22:33] the thing is for uers that have been working on reskinning [22:33] depends on what they have changed [22:33] My understanding is that the pattern skin changes can be treated as a bug and go into 4.0.2, but they need to be reworked to revert the *other* changes that have since been done on DEVELOP, yes? [22:33] The jump/search at top was one of the first changes I did to 4.0.0. It is a natural thing and much better for upgraders from Cairo. [22:33] assuming no skin mods? [22:34] PeterThoeny: who was that question aimed at? [22:34] the only thing is the webleftbar and webtopbar [22:34] arthur: assuming admin did not skin mods, do any files/topics need to be touched? [22:34] they might need to edit them [22:35] can be quite some work if they have 100+ webs [22:35] 'touched' meaning changing again after upgrading? [22:35] yes [22:35] no, change webleftbarsearch to empty and they're done [22:36] that could still be a lot of work [22:36] one topic? [22:36] ok, one place, so problem then [22:36] 100 topics, if oyu have 100 webs [22:36] Maybe if you change the top bar template to not be a table the WebTopBars they already have made will still work. That was the only problem I ran into. [22:36] TWiki.WebLeftbarSearch [22:36] ArthurClemens: ok, sorry [22:36] crawford: the search boxes in the sidebar are all includes, so fixing one include fixes all 100 topics [22:36] s/topics/webs/ [22:37] what is the best place to document this? [22:37] an upgrade topic? [22:37] in the release notes for 4.0.2 [22:37] in this case i do not see an issue with moving the boxes from the leftbar to the topbar [22:37] Yep. Release note it the right place. [22:37] a patch should have release notes, same as any other release [22:38] I am still concerned about the amount of work involved [22:38] on pattern? [22:38] Not much choice, however [22:38] in mergeing Arthur's work from DEVELOP to TWiki04x00x00 [22:38] just all files [22:39] + some doc topics [22:39] what about the change to templates? Leading newlines? [22:39] + locale [22:39] is there no impact from that? [22:39] what is that? [22:39] ok, probably no impact then :-) [22:39] well, there is an impact for skin writers [22:39] Pattern is self-contained in twikiplugins, isn't it? [22:40] they all need to fix their skins [22:40] just merge all changed submitted by me :-) [22:40] PeterThoeny: yes, there is; and the change should *not* go in 4.0.2 [22:40] agreed [22:40] but if Arthur has coded *assuming* the change [22:40] then his new templates *may not work* in TWiki400 [22:40] that is why testing is needed [22:40] There was an important change CC did that required all templates to be updated. It must be merged in as well. [22:40] Lavr: that is what we have just been discussing [22:41] and it absolutely *must not* go in 4.0.2 [22:41] because it would impact any 3rd party skins wot have been wrote [22:41] Then Arthur need to redo all templates again. And the in 4.1 we can so it all over again. [22:41] many companies write their own skin for internal use [22:41] correct [22:41] sorry, holiday :-) [22:41] Lavr: that is a worst-case [22:42] OK. So it both must and must not be included [22:42] 'm off [22:42] personally I think ArthurClemens's changes will "just work" [22:42] but we have to test to see [22:42] and testing is my #1 concern right now [22:42] still some locale texts need to be updated [22:42] ok, we seem to agree, take pattern into twiki4 branch & test [22:42] (still need to) [22:42] we need a running testing site for TWiki 4 [22:42] yes [22:43] be it twiki.org or twiki4.twiki.org or whatever suit us best [22:43] do not take leading space change into twiki4 [22:43] That doesn't take 3 minutes to load a page [22:43] have a server left? [22:43] not me [22:43] nope [22:43] we need a short term solution [22:44] install on develop or twiki.org [22:44] I will for sure run my Motion TWiki on it. And I will run updated SVN from my alternative machine. [22:44] that would be sufficient [22:44] is sven here? [22:44] I still want a testbed for DEVELOP [22:44] sven_: yoo hoo! [22:44] guess not [22:45] anyway, I do *not* want develop moved off of DEVELOP, personally [22:45] and that server isn't up to running 2 [22:45] me neither [22:45] it's barely up to running 1 [22:45] sven might be volunteering to install twiki4 branch on develop.twiki.org (with automated update as in develop) [22:45] that server can't support 2 [22:45] god, but it's sooooo slow. [22:45] One can barely look at a bug report [22:46] we can't stress twiki.org with additional load [22:46] I can offer my shared hosting on Dreamhost, but its not too fast [22:46] plus i do not want untested code on twiki.org [22:47] PeterThoeny: understandable; but that server is fast enough to run 2 twikis [22:47] however it doesn't have subversion, which would be a PITA [22:47] arthur: that shounds good as an interim solution [22:47] What does it take to run auto update? Just svn update; perl pseudo-install? [22:47] right [22:47] y [22:47] I might need some help on that [22:48] Which reminds me: could we *please* have -link be the default with pseudo-install? [22:48] And a cron to run it every 30 minutes? [22:48] (auto update) [22:48] bitca: raise an Item [22:48] who is volunteering to help arthur set up the system? [22:49] * bitca hides under her (virtual) desk [22:49] me, i suppose [22:49] * CDot hides under bitca [22:49] There *are* 4 of you, after all, will [22:49] i do a lot of twiki on dreamhost [22:49] * bitca likes the idea of CDot hiding under her. [22:50] wbniv: all we want is a simple test platform [22:50] nothing clever [22:50] apachelogin + htpasswd preferably [22:50] ew, not TemplateLogin? [22:51] bletch [22:51] Well. I have my merlin.lavrsen.dk which is a different ADSL line than my www.lavrsen.dk and different machine. So I will also try and setup something. [22:51] no, because of bitca's valid demand that logoff delecte session cookies [22:51] must be tested on apachelogin [22:51] k [22:51] s/demand/forceful request/ [22:52] Lavr: if you can set up with templatelogin, that would be another useful test [22:52] or vice-versa [22:52] I run Apache now. So that would be EASY for me. [22:52] cool [22:52] good, because i'd rather do TemplateLogin :) [22:52] wbniv: :-) [22:52] all agreed [22:52] I will help ArthurClemens with the merge, as far as I can [22:53] but I can't do all Steffens stuff as well [22:53] so two test machines for us to try? arthur's and kenneth's? [22:53] can someone else pick up Steffen's merges please? [22:53] I gave up on release and am running svn only [22:53] I marked them all "Waiting for Merge" on Bugs web [22:53] bitca: this *is* svn; just a differen branch [22:54] Too confusing for me [22:54] can someone else pick up Steffen's merges please? [22:54] Oh, isn't sun lending machines for free for 60 days or some such? [22:54] can someone else pick up Steffen's merges please? [22:54] Peter? [22:54] Keep trying, CDot ;) [22:55] Soronthar? [22:55] sorry, i am already overcommited [22:55] so are we all, Peter [22:55] * bitca doesn't even know what that *means* [22:55] what would be the deadline for the merges? [22:55] ASAP [22:55] which means yesterday? [22:55] well, lets count backwards [22:56] how does a merge take place? [22:56] what release date is our goal? [22:56] I wrote a detailed description on Codev, Arthur [22:56] ok [22:56] and I will hand-hold you through the process if you want [22:56] * Soronthar goes to check Steffens commits [22:57] arthur: http://twiki.org/cgi-bin/view/Codev/MergingInSubversion [22:57] PeterThoeny: I don;t want to count backwards, because people just take that as a reason to leave everything to the last minute [22:57] or drop from their hands [22:57] I would rather get things done as soon as humanly possible, so we can maximise "settling time" [22:57] but asap is not smaart [22:57] OK, I'm definitely not the right person to do a merge [22:58] agreed; but that is why Soronthar should commit a date [22:58] instead of having one imposed on him [22:58] so, if Raf can't complete the merge before - say - end of the month, then 4.0.2 doesn't go until the end of the month [22:59] etc [22:59] omg. that's 3 week away. [22:59] yep [22:59] that is fa rout [22:59] What are users supposed to do in the meantime? [22:59] not when your tax year ends on April 6, it isn't :-( [22:59] As in: poor Lavr's users [23:00] We need 2 weeks of testing to get all the bugs out of Pattern. There will be things we find. [23:00] agreed; but we are talking about Steffens fixes, not pattern [23:00] with Arthur's cooperation, we could get pattern done tomorrow [23:00] but Lavr 's saying there's going to test *testing time*, right? [23:01] which means, [23:01] there should be a field in bugs with the "who fixed this" information. [23:01] Soronthar: there is; its called "Waiting for Merge" ;-) [23:01] could set a target release date for 4.0.2 for 2 weeks from now [23:01] is there a way to see all *my* changes from the last weeks? [23:01] ArthurClemens: yes [23:01] not "where". "who" :) [23:01] Soronthar: it;'s the same thing; everything that Steffen fixed hasn't been merged yet [23:02] ah, ok. [23:02] I have actually already implemented many of Steffens fixes both at the Motion TWiki and the Motorola twiki. They were many small bug fixes that I took the chance and installed. [23:02] except the things I was able to merge easily [23:02] So where is Steffen anyway? [23:02] Lavr: I think I already merged all the simple ones [23:03] When you merge, note that the new PatternSkin closes some other bugs. And enhancements. I have listed them all in todays minutes. [23:03] back to timeline, can we set a goal to finish merge? [23:03] probably align that with finish test machine setup? [23:04] i'll tell you when develop finish loading. [23:04] I will have test machine setup probably tomorrow late evening. [23:04] i can't guarantee anything before sunday [23:05] (just checking develop.twiki.org, again three runaway processes) [23:05] (doesn't mean it won't happen before sunday, just that i can't commit to it happening before sunday) [23:05] 12 Waiting for Merge... some of them look quite "simple". To err on the sure side, a week from now. [23:05] I can do the setup tomorrow [23:05] but need some script help to update from svn [23:05] top: [23:05] PID USER PRI NI SIZE RSS SHARE STAT %CPU %MEM TIME COMMAND [23:05] 23261 develop 20 0 9292 1472 1376 R 17.7 0.2 147:41 oops [23:05] 2637 develop 20 0 9292 1472 1376 R 16.7 0.2 101:34 oops [23:05] 25666 develop 14 0 12020 11M 1392 R 16.7 2.3 0:04 view [23:05] 14994 develop 18 0 9292 1472 1376 R 16.3 0.2 96:58 oops [23:05] 19011 develop 20 0 287M 274M 1588 R 16.1 55.5 7:07 search [23:05] cronjobs no prob [23:05] Soronthar: I see 10 waiting for merge [23:05] ArthurClemens: catch me on #twiki if/when you need script help [23:06] ok [23:06] must be marked "waiting for merge" and "patch" [23:06] i killed the three processes, should be faster now [23:06] ok. [23:07] well.. that give me more "free" time this sunday :) [23:07] ok, goal is monday 13 mar [23:08] testing goal one week or two weeks? [23:08] PeterThoeny: please note that 1610 will *not* be addressed unless we can get enough detailt to reproduce it. Can you commit to providing the detail? [23:08] i will look into it again [23:08] PeterThoeny: 2 weeks [23:08] could set a target release date for 4.0.2 for 2 weeks from now [23:08] We need to set the translator to work [23:09] target release date two weeks from now with one week for setup and merge leaves one week for testing [23:09] is this ok? [23:09] Two weeks from Pattern skin merge. [23:09] the end of the second week I am skiiing [23:09] it's essentially 2 weeks testing tho, right, because the test machines are going to be setup tomorro [23:09] w [23:10] ooooh, where are you going skiing??? [23:10] well no [23:10] ArthurClemens: skiing? In Netherlands? Sounds like hard work >:-) [23:10] the merging needs to be done first [23:10] austria [23:10] better, no [23:10] ? [23:10] yep [23:11] ok, target release date mon, 20 mar [23:11] 1) set up test machine, 2) merge across, 3) test ? [23:11] I would like to request a release meeting next monday [23:11] good idea [23:11] with what agenda? [23:11] I am rather dissappointed by progress since the last release meeting [23:12] and want to stay on top of things [23:12] 1) Merge process status. [23:12] 2) Testing status [23:12] and add "doc status" (release notes etc) [23:13] ok, marked for next weeks meeting [23:13] who is merging the doc changes? [23:13] I think I already merged the changes I could see in SVN [23:13] what other changes were there? [23:14] * SvenDowideit has joined #twiki_edinburgh [23:14] locale [23:14] i marked all doc changes i did for 4.0.2 in develop as Item1611 [23:15] if they are simple changes, the the merge should be painless [23:15] brb [23:15] ok, anything else for 4.0.2? [23:16] 65 min passed [23:16] 15 min left [23:16] We finish at 5:30? [23:16] I am OK with what was covered for 4.0.2. [23:16] shall we go to 2. Post-Dakar development model [23:16] argh; 1611 is marked as "normal", and "new", so I would not have merged it [23:16] bitca: well, that's the theory, tho i'm not sure if it's ever happened yet ;-) [23:17] it has to be "urgent" or "requirement" to be considered for 4.0.2 [23:17] generally "requirement" [23:17] Oh dear. I just got an error in Bugs [23:17] not sure if this is good then [23:18] because there are many smaller enhancements that are not urgent and not required that should go into 4.0.2 [23:18] s/enhancements/fixes/ [23:18] If bugs are simple spellings, docs errors why not get them out of the way? [23:18] agreed; and they have *not* gone in, because 4.0.2 is a *bugfix* release [23:18] not an excuse for smaller enhancements to go in [23:18] every change increases risk [23:19] but a spelling correction is not for a major release, is it? [23:19] or makes it harder for users to upgrade [23:19] crawford: 1611 are mainly doc fixes that _should_ go in, but they are not bugs per se [23:19] a spelling correction can perfeclty validaly be marked "urgent" [23:19] ok [23:19] By those of us who know how to spell, that iss [23:20] if this is the rule i will mark more things as urgent [23:20] this makes future PatternSkin work a bit problematic [23:20] those are the criteria for gating releases [23:20] We all know how cranky I am about the state of the docs [23:20] just makred 1611 as urgent [23:20] as stated on the bug form [23:20] It seems bit "unnatural" to mark a spelling error urgent but if that is what it takes that is what we will do. [23:20]