--- Log opened Wed Feb 04 10:06:27 2009
* jlaska in another meeting, but lurking10:06
-!- wwoods changed the topic of #fedora-meeting to: Fedora QA Meeting | init10:06
-!- nicubunu [n=nicubunu@mail.apsro.com] has left #fedora-meeting ["Fading into sunset"]10:06
wwoodsf13, poelcat, jds2001: ping10:07
wwoodsanyone else here, say hi for the log10:07
-!- tk009 [n=tk009@adsl-074-166-251-205.sip.bct.bellsouth.net] has joined #fedora-meeting10:07
* viking_ice Hello 10:08
wwoodsI know leitz was here, but he's apparently going to brb10:09
wwoodsanyway, I'll get started with a quick review of Alpha, and anyone joining late can just scroll through it10:09
-!- wwoods changed the topic of #fedora-meeting to: Fedora QA Meeting | F11Alpha status10:09
wwoodsfor reference, the tracker bug: http://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=F11Alpha10:09
buggbotBug 476773: medium, low, ---, notting@redhat.com, ASSIGNED, Fedora 11 Alpha tracker10:10
wwoodsand the installer test results: https://fedoraproject.org/wiki/QA:Fedora_11_Alpha_Install_Test_Results10:10
wwoodsAs far as I know, Fedora 11 Alpha is on its way to the mirrors10:11
wwoodsthis test cycle was a bit hectic. things kept going wrong and it kept taking a while to notice10:11
wwoodswe ended up going through *9* sets of release candidate images before we got it all together10:12
wwoodsone of the major problems was that f13 was building images in two different places - at home, and in the facility in phoenix where we do the official builds10:12
wwoodsand then there's also rawhide, which is built automatically in phoenix10:13
wwoodsjlaska and I were testing the disk images f13 made in phoenix, f13 was testing his home-built images, and clumens was testing rawhide10:13
wwoodsand we were all getting different results10:13
-!- leitz_brb is now known as leitz10:13
wwoodsthis caused a lot of confusion10:13
-!- adamw_stupidnick [n=AdamW@nat/redhat/x-1aa3d8d3927ba7e7] has joined #fedora-meeting10:13
wwoods(clumens is in a different office from jlaska and I)10:13
wwoodserr. jlaska and me. stupid english.10:14
wwoodsanyway. we're going to try to solve that problem by setting up some notification system10:14
wwoodsso when stuff gets built in phoenix it can notify people, and we can get it automatically synced to our various offices/homes10:15
wwoodsso we get notification of new builds *and* copies of the official builds10:15
wwoodsnaturally this will all be developed in the open so you can join in the fun.10:15
-!- stickster [n=pfrields@fedora/stickster] has quit [Read error: 110 (Connection timed out)]10:16
-!- stickster [n=npfrield@ip72-205-14-2.dc.dc.cox.net] has joined #fedora-meeting10:17
leitzNot sure that I understand the build process, but what prevents us from twiddling at home and building in pheonix?10:17
leitzIs there a difficulty uploading changes?10:17
wwoodsdefine "twiddling"?10:17
wwoodsleitz: it turned out that the images that f13 was building in his "lab" were slightly different from the ones built in phoenix, somehow10:17
wwoodsthe ones from PHX would crash at the timezone screen; the ones f13 and clumens were using didn't10:18
wwoodsthis was bug 48290710:18
buggbotBug https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=482907 medium, low, ---, anaconda-maint-list@redhat.com, CLOSED RAWHIDE, f11alpha anaconda hanging when moving to timezone screen10:18
wwoodseveryone's welcomed - and encouraged - to test the official bits at home, on whatever hardware you can get your hands on10:19
leitzSo he was using a pheonix build that crashed on others but not in his lab?10:19
leitzI want his lab...10:19
wwoodsno, he was using home-built images that worked on his machines10:20
wwoodsbuilt from the same package set, using the same tools10:20
wwoodsbut some minor variation in the host system performing the build caused major differences10:20
wwoodsso his worked and the "official" ones crashed10:20
wwoodsso, for the moment we can't assume that test results for home-built images are valid10:20
wwoodswe need to make sure everyone tests the same bits.10:21
leitzUntil we identify the "minor variation".10:21
wwoodseventually we might be able to set up a build script that eliminates that factor10:21
wwoodsso people can build their own identical images at home10:21
wwoodsuntil then - home-built images don't count for release testing.10:21
leitzDepends, there are a *lot* of potential factors.10:21
wwoodswhich means we need to provide bits10:22
wwoodswhich means we need a site that only rel-eng/qa is allowed to get at, because otherwise it gets overwhelmed with swarms of onlookers who aren't actually going to contribute10:22
leitzAnd we have a document that says "Here's how to grab the latest proper build and try it out?"10:22
wwoodsand then we can't get images, we can't get testing done, etc.10:23
wwoodswe have some of that - https://fedoraproject.org/wiki/QA:Fedora_11_Install_Results_Template10:23
wwoodswe have an actual test plan from F10 - https://fedoraproject.org/wiki/QA/TestPlans/Fedora10Install10:24
wwoodswhich needs to be adapted for F1110:24
wwoodsand clarified10:24
viking_iceboth of these need to be rewritten and simplified10:24
wwoodsbut, yeah, basically, the QA/rel-eng team should get notification of new candidate builds and where to get 'em10:24
leitzOkay, let me go over the process after the meeting and I'll see if it's idiot proof. And where it could use some clarity.10:25
wwoodsand we'll have a page to coordinate testing (and explanations of the test cases)10:25
* leitz is great at idiot testing...10:25
-!- lvillani [n=lvillani@kde/developer/lvillani] has left #fedora-meeting ["Leaving"]10:25
wwoodsit's pretty rough right now, there's a lot of institutional knowledge buried in here10:25
leitzyeah, I noticed.   :)10:25
-!- RodrigoPadula [n=Rodrigo@] has joined #fedora-meeting10:26
leitzThat's what I'm looking to deal with.10:26
wwoodsdefinitely in need of simplification, which is why I haven't just copied the F10 one for F1110:26
wwoodsI want to distill the F10 info into something clearer for F11.10:26
viking_iceCloser to the new look and feel10:26
wwoodsI like it!10:27
viking_iceand guys please refrain from writing your life story in the comment field10:27
leitzviking_ice, I'll send you my notes for review, okay?10:27
wwoodsheh, yeah, those tables get all wonky if one of the fields gets really big10:27
viking_icemagic of using the wiki10:27
viking_icebut as always that's what we got that's what we use for now..10:27
wwoodsso there are still a couple of outstanding issues with F11a - bug 478999 is the big one10:27
jlaskadifferent teams need to gather and display results in different formats ... which makes wiki great for playing around with different organizational approaches10:27
buggbotBug https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=478999 high, medium, ---, ajax@redhat.com, ASSIGNED, old-style bitmap fonts not provided by X server10:27
wwoodserr, wait, sorry10:28
wwoodsbug 48066710:28
buggbotBug https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=480667 medium, low, ---, pjones@redhat.com, NEW, nash unable to find dm devs by uuid or label causing boot to fail10:28
wwoodsthat one shows up as a system failing to boot at first boot10:28
wwoodsthe workaround for that is to add boot_delay=5, or change root=UUID... to root=/dev/VolGroup00/LogVol00 (or similar)10:29
pjonesin theory, hansg might have a fix for that.10:29
wwoodsright - there should be a fix Real Soon Now, but it'll probably come up when people start playing with F11a10:29
wwoodshopefully it'll be in the release notes, otherwise.. we'll find other ways, but spread the word10:29
wwoodsanything else on F11Alpha specifically?10:29
viking_icecan the boot_delay= be used as a workaround for usb external disks10:30
wwoodsviking_ice: I'm not sure what problem you're referring to10:30
-!- cschu [n=cschu@nat/redhat/x-f2b23a5bdb378d61] has joined #fedora-meeting10:30
wwoodsgot a bug id?10:30
* viking_ice diggs it up10:30
-!- cassmodiah [n=cass@fedora/cassmodiah] has quit [Read error: 104 (Connection reset by peer)]10:31
viking_iceBug 47882410:31
-!- balor [n=balor@] has quit [Read error: 110 (Connection timed out)]10:31
buggbotBug https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=478824 medium, low, ---, kernel-maint@redhat.com, NEW, Installing to an external usb connected HD fails...10:31
viking_icethe installation works fine10:32
wwoodsviking_ice: it sounds like it might be a similar bug10:32
wwoodstry boot_delay=5 and see if it helps10:32
wwoodspjones: got an opinion on that one?10:33
viking_icemore curios why both you and clumens claimed it work for you ?10:34
wwoodsas you can see from the test results page - we haven't tested it in F11a10:34
wwoodsbut we tested in F9 and F10 and it worked then10:34
pjoneswwoods: I love how he contradicts himself as to where the failure is...10:34
wwoodsif we had notified you about the F11a RC builds, would you have been able to test it?10:34
wwoodspjones: be fair, it's a subtle and confusing problem10:35
-!- kital [n=Joerg_Si@fedora/kital] has joined #fedora-meeting10:36
viking_icewwoods: me?10:36
viking_iceas in test it10:36
viking_icesure I have a spare disk and could free an ide one as well10:36
wwoodsviking_ice: sure - we had candidate images but we were testing everything else as hard as we could and none of us are easily equipped to test external USB at the moment10:36
pjoneswwoods: so that screenshot definitely looks like we're just not waiting long enough in that case for the disk to show up.10:37
wwoodsso it might just be another manifestation of 48066710:37
viking_iceI downloaded image a week or go from alt to test but i'm pretty sure the dvd drive in my laptop is failing10:37
pjonescould be.10:37
wwoodsokay, we'll retest and see what helps10:37
wwoodsthanks for the info!10:37
-!- nicubunu [n=nicubunu@mail.apsro.com] has joined #fedora-meeting10:38
wwoodsviking_ice: there are new images on alt - live images too. I'll try to scrounge up a USB disk to test again10:38
wwoodsbut yeah, that section of the test results page is blank10:38
-!- gratnam1 [n=gratnam1@bas1-toronto42-1279600068.dsl.bell.ca] has joined #fedora-meeting10:38
wwoodsso we need help testing, wherever we can find it.10:38
-!- nicubunu [n=nicubunu@mail.apsro.com] has left #fedora-meeting ["Fading into sunset"]10:38
viking_iceI have access to atleast 5 various types of laptop to test and play with10:39
wwoodsanyway, let's move on to the organizational-type stuff10:39
wwoodswe'll be talking about F11a and rawhide a lot, esp. with test days and such coming up10:39
* leitz wants viking_ice's lab too...10:39
viking_icein the evening..10:39
-!- wwoods changed the topic of #fedora-meeting to: Fedora QA Meeting | Wiki use10:39
wwoodsthe next topic was wiki use10:39
viking_iceThank good for people that buy vista and think their old laptop is crap ( which I gladly collect )10:39
wwoodsviking_ice: heh! nice10:40
wwoodsso we've got our own namespace on the wiki - "QA:"10:40
wwoodsthis helps keep qa-specific stuff (test plans, test results pages, test cases, per-package info, other in-depth documentation) 10:41
wwoodsseparate from the end-user oriented stuff10:41
-!- cschu [n=cschu@nat/redhat/x-f2b23a5bdb378d61] has quit ["chramst..mlask..hehehe"]10:41
wwoods*but* some of the pages we write - like the Test Days stuff - should probably still be in the main namespace, since we *want* normal users to be able to search and stumble upon it10:41
leitzDo we really want that? Or have news items posted in users areas?10:42
wwoodswell, both, honestly10:42
viking_icenah not really we just create a testday page with links to already existing test day's and the new one10:42
leitzWe already hit a snag on centralization.10:42
wwoodstest days should get mailed to the lists, we should notify the FWN guys, it should be on the wiki, etc.10:43
viking_iceso we can just as well have the $date in the QA: name space10:43
wwoodsviking_ice: right, we've got the category page for that10:43
leitzI'd suggest one place to rule them all , and what Chris Brogan calls "outposts"10:43
wwoodsso should https://fedoraproject.org/wiki/QA/Test_Days/2009-02-05 get moved to QA:Test_day_20090205?10:43
wwoodswe obviously want all test days to be in the Test Days category10:44
viking_iceI would like to have that page made of Test Case and  Passes10:44
wwoodsoh I guess we'd want to name them by the theme, really?10:44
leitzPlaces where we attract interest, for those interested.10:44
-!- RodrigoPadula [n=Rodrigo@] has quit ["Saindo"]10:44
wwoodsor by the date - either way, the important bit needs to be at the beginning10:44
wwoodsto make the category page useful10:44
* viking_ice far from being wiki expert and would like to avoid wiki masters wrath 10:45
wwoodse.g. QA:2009-02-05_Test_day or QA:ext4_test_day10:45
leitzDefine important?10:45
* viking_ice ducks 10:45
leitzWe need to sort, plus we need to minimize user time spent on non-interesting topics.10:45
jlaskaso ... hrmm, wiki ... what's the goal with the renaming?10:45
jlaskais it to improve searching in the QA namespace?10:46
leitzOr marketing?10:46
wwoodsjlaska: yeah, and to keep stuff that isn't end-user-focused from cluttering the main namespace10:46
leitzOr wwoods "both"?10:46
viking_iceclean up and serach among other things10:46
jlaskawwoods: so I can just prefix all the test day content with QA: instead of QA/10:46
wwoodse.g. normal users searching for 'anaconda' shouldn't have to weed through our 47 anaconda test cases10:46
wwoodsto find the page with anaconda options on it10:46
jlaskayeah i like the QA: namespace, that's a good idea10:47
wwoodsBUT, since test day stuff *is* arguably end-user focused10:47
wwoodsI can see that one either way10:47
jlaskais there a mandate from the wiki/infrastructure folks?10:48
leitzI'd recommend centralization with QA team personal marketing to users.10:48
wwoodsjlaska: there are some guidelines - stop using slashes, for one thing10:48
wwoodsevery time you're thinking of doing QA/[something]/[morestuff]10:49
jlaskabut that was mainly a result of everything existing in the primary namespace?10:49
wwoodsmaybe it should be more like: QA:Ext4_root_test_case10:49
-!- gratnam1 [n=gratnam1@bas1-toronto42-1279600068.dsl.bell.ca] has quit [Remote closed the connection]10:49
-!- alexxed [n=alex@dnt-gw-teamlingo.dnttm.ro] has quit [Remote closed the connection]10:49
leitzwwoods, I like it.10:49
* viking_ice yes wiki masters dont / still got a mark after the evil whip 10:50
wwoodsjust general cleanup, honestly. it's a good idea if you ask me. "Ext4 root test case" is easier to remember than "QA/TestCases/Ext4Root"10:50
-!- petreu [n=peter@fedora/petreu] has quit ["( www.nnscript.de :: NoNameScript 4.02 :: www.XLhost.de )"]10:50
jlaskafor some maybe10:50
jlaskabut this is a naming debate ... there's no right answer on that is there?10:50
wwoodsactually we could follow the wikipedia convention and do e.g. Ext4 root (test case)10:50
wwoodsthat's why we're setting guidelines10:50
wwoodsso we can pick something we like and stick with it10:50
adamw_stupidnickslashes are bad for a flat wiki because they make people think they're directories when they aren't...10:51
-!- adamw_stupidnick is now known as adamw10:51
jlaskaright ... and we've got them from our previous moin instance10:51
adamwgrr, brb10:52
-!- adamw [n=AdamW@nat/redhat/x-1aa3d8d3927ba7e7] has quit ["Ex-Chat"]10:52
jlaskaso we should move test cases and plans into the QA: namespace, that sounds like a clear goal right?10:52
wwoodsright. we're only using 'em because moin made us do it10:52
wwoodsyes, definitely10:52
jlaskatest days ... wwoods you noted they could go either way10:52
leitzIf QA were to publish a book, would the wiki be it or just the brain-storm session? That is, how "final draft" should it be and what will we present to the outside world when they ask about us?10:52
-!- adamw [n=AdamW@mandriva/channel-support/AdamW] has joined #fedora-meeting10:52
jlaskawwoods: is it worth renaming all existing content to fit into the model ... or just all content going fwd?10:53
viking_icemove all I say10:53
wwoodsgeneral QA info should go in the main namespace - the main QA page for instance10:53
leitzHow much is all?10:53
wwoodsthe wiki gurus want us to move it10:53
viking_icewhich reminds me should I move the QA page from draft to front?10:53
jlaskaleitz: dunno ... maybe around 60-80 pages10:53
wwoodsthe old page name will be a redirect to the new one, so QA/blah/foo will still get you the right page10:53
wwoodsviking_ice: please do10:54
leitzWould you want to toss the "Quality Engineering (QE)" question into the morass?10:54
jlaskamoin had a nice feature where you could provide an html <form> with a textbox for the test case name, and button to create the page10:54
jlaskawhich made it easy to ensure pages landed in the right namespace etc...10:54
jlaskais there something similar we can do in mediawiki ... while we are using it for test case/plan mgmt10:55
wwoodsjlaska: not that I know of.10:55
wwoodsIIRC there's an API call that lets you create a new page with prefilled content10:56
wwoodsbut I dunno if there's a wiki macro for that or some other magic10:56
wwoodsyou'd have to ask the wiki gurus10:56
jlaskayeah ... it's tough to have the guidelines, but we don't have a way to make it "just work" as people create new content10:56
jlaskabut that's perhaps more a "nice to have"10:56
viking_iceShould I put QA/QC to QA:SQC ?10:57
wwoodsviking_ice: I'd say that general info like that - what QA is, what stuff it does, etc10:57
wwoodsshould stay in the main namespace10:57
wwoodsopinions on naming for test plans / test cases?10:58
viking_iceso when you say main namespace you mean / or :  ( just to get it right )10:58
wwoodsmain - no :10:59
wwoodsto make the wiki gurus happy you should give the pages "natural language" names10:59
jlaskawwoods: my $0.02 is that as long as it's in the namespace : ... we can call it whatever?10:59
wwoodslike "Software Quality Control" or "SQC"10:59
wwoodsavoid slashes.10:59
* viking_ice still confused.. 10:59
leitzAvoid acronym-only, unless it's real long.10:59
viking_iceNo slashes11:00
viking_iceand it should be then the final QA:Software_Quality_Control <--11:00
wwoodsprobably just Software_Quality_Control11:00
wwoodssince it's general info that non-QA-team people would find useful11:01
wwoodsstuff that's only useful/needed by QA team people should go under QA:11:01
viking_iceToday's QA ==  SQC11:01
-!- JSchmitt [n=s4504kr@p4FDD1B0F.dip0.t-ipconnect.de] has joined #fedora-meeting11:02
viking_iceSo all info for testers/triagers are in the SQC11:02
wwoodsas for test plans and test cases in QA: - I guess all test plans should be in the "Test Plans" category, and similarly "Test Cases"11:02
wwoodsbut I think that's pretty obvious11:03
wwoodsso really we need some templates/examples for test cases and plans11:03
wwoodsand to talk to wiki guys about ways to prefill a new page with a template11:03
-!- kital [n=Joerg_Si@fedora/kital] has quit [Connection timed out]11:03
jlaskawwoods: yeah, I began migrating all the test cases into additional [sub]categories last week ... https://fedoraproject.org/wiki/Category:Test_Cases11:04
jlaskaI've got a mediawiki template for a test case (https://fedoraproject.org/wiki/Template:QA/Test_Case)11:04
viking_iceI want test cases a have direction for beginners testers or some how have label of experience skill level if that test case is presuming some knowledge before proceeding11:04
jlaskawe can provide some basic blue prints, but there's nothing stopping from having images, or additional sections etc...11:05
wwoodsyeah, a skill level rating isn't a bad idea. categories for beginner, intermediate, expert..11:05
leitzviking_ice, one of my goals is to expose the skills and talents needed for tasks, and to make sure the tools required are documentd.11:06
viking_icewwoods: goes with the training class11:06
viking_icethat will be created11:06
viking_icesometime in future11:06
viking_iceTrain testers11:06
leitzBeginner is a relative term. if you have to know how to use bugzilla, for example, that gives a concrete measure of what is required.11:07
-!- daumas [n=mcronenw@adsl-65-67-98-193.dsl.rcsntx.swbell.net] has joined #fedora-meeting11:07
* leitz is still typing slow...11:07
viking_iceleitz: Beginner == livecd testing11:07
-!- pingou_ [n=Pingou@fedora/pingou] has quit [Remote closed the connection]11:07
wwoodsyeah, we'll want to have some rough outlines of what we mean by those things11:07
viking_icewhich can be any joe of the net.11:07
wwoodsthis actually kinda segues nicely into the final topic - QA team guidelines11:08
-!- wwoods changed the topic of #fedora-meeting to: Fedora QA | team guidelines11:08
viking_iceand it's safe for him to do testing without destroying his data11:08
-!- fab__ [n=bellet@bellet.info] has joined #fedora-meeting11:08
* leitz can destroy data any time, day or night...11:08
viking_iceerase != destroy11:08
* wwoods is an expert data destroyer11:09
-!- che [n=rkastl@redhat/che] has quit [Remote closed the connection]11:09
-!- fab__ is now known as fab11:09
leitzFirst Team Guidline; adamw buys bbq tomorrow.11:10
adamwi would like to register my opposition to this in the strongest terms available to a canadian!11:10
wwoodsso, you're politely declining11:11
leitzThey probably have beer...11:11
pingoubeerS ?11:11
wwoodssuddenly the entire channel wakes up11:11
adamwyeah - new euphemism: Canadian barbeque11:11
adamwmeaning, beer drinking11:11
leitzYeah, but the more meat in your gullet the better you can handle the fermentation.11:12
leitzSounds like adamw's in.  :)11:12
-!- KageSenshi [n=kagesens@fedora/KageSenshi] has quit [Remote closed the connection]11:12
adamwshould probably be noted at this point that i'm vegetarian :)11:12
leitzBeef is only refromed grass.11:13
wwoodswell, so am I, but I've got good news - we've declared bbq to be a type of fruit for the next 24 hours11:13
viking_iceis the Canadian beer good or is it equally crappy to the American one11:13
adamwwe have that power?!11:13
* wwoods is not a very good vegetarian11:13
adamwviking_ice: canadian mass-market beer like Molson Canadian is almost as bad as U.S. mass-market beer11:13
viking_iceadamw: you dont eat meat your doomed11:14
adamwbut like in the U.S. there's small local producers who make really good stuff11:14
-!- KageSenshi [n=kagesens@fedora/KageSenshi] has joined #fedora-meeting11:14
viking_icewhat you define as a good beer the local British ones are tasty and good11:14
adamwi'm originally from england. so I am qualified to have an opinion :P11:15
* pingou is offering some good belgium beers ;)11:15
wwoodsobviously, QA team membership involves an appreciation of good beer and/or liquor11:15
wwoodsif at all possible, you should send one fine bottle to a QA team sponsor11:15
viking_iceculture exchange..11:15
* leitz can only stomache actual german beer.11:15
wwoodsbecause as adamw pointed out: nobody should have to deal with bugzilla sober11:15
leitzThe Blokes don't have beer...11:15
adamwi DID bring a bottle of 2001 ontario wine. but, sadly, vincent and I drank it all already. :P11:15
leitzwarm urine, yes..beer...not really.11:16
adamwwwoods: oh yeah. do the bugzilla folks have beers in their area's company fridge?11:16
pingouidea: 1 free beer per 5 solved bugs ?11:16
adamwit would have to be the other way around11:16
adamwyou'd have to pay the beer up front11:16
viking_icewhen feeding bugzilla with report coins beer should pop out11:17
wwoodsbeer bug bounties!11:17
wwoodssubmit a fix for this bug and you get the following bottles:11:17
daumassounds great. file a bug on it.11:17
* leitz can see a lot more bugs maing it into the system...11:17
viking_iceif written in x language the more low level language the more expensive beer11:18
leitzyeah, java gives you british beer.11:18
daumaswhat would asm get you?11:18
viking_icejava gives you something like Budweiser hahah11:18
wwoodstrappist ale, hand-delivered by the monks11:18
leitzsame same11:19
pingouthe amount of alcool on the beer is according to the rank of the bug (low, medium...)11:19
wwoodsheh ANYWAY11:19
pingouGrimbergen, leffe...11:19
viking_icenah to much alcahol in beer ruins it11:19
adamwpython gets you coors light11:19
pingouviking_ice, tried a casteel ?11:19
wwoodsso we keep talking about "qa team members" and voting on proposals and things, but we don't have any actual guidelines for team membership or who can vote11:20
leitzTeam Guidelines, Take 2.11:20
wwoodsso here's my super-quick proposal:11:20
-!- Gaaruto [n=Gaaruto@fedora/Gaaruto] has joined #fedora-meeting11:20
wwoods1) like all the other groups in Fedora, there's members and sponsors (and administrators)11:20
wwoods2) if something needs a vote from the QA team, sponsors + admins each get 1 vote11:21
leitz2.a) Will need guideline on what needs that sort of vote, and what is hashed out in the community.11:21
wwoods3) you get to be a sponsor by doing "a bunch of good QA work" - at least a little each of: writing wiki pages (test plans, test cases, test day planning), release testing, bodhi karma-ing, etc11:22
viking_icefirst let's establish proper QC lead11:22
viking_icebefore deciding on how they vote11:22
wwoodsyeah, I think most decisions can be made by consensus of the group, but stuff that becomes hard to decide might be put to a vote11:22
* f13 peeks in11:23
wwoods3) (continued) once you've done a bunch of good stuff you can apply to the sponsors, and they'll vote on you. This is approximately how FESCo decides on new packaging sponsors.11:23
viking_iceSo If I would suggest ssh to be disable by default on the DVD that would be voted on QA and that result would enforced upstream11:24
wwoodsexcept.. they have a 9-member board11:24
wwoodsviking_ice: no, that's not a QA decision11:24
wwoodsthat's FESCo11:24
wwoodsthey're the engineering steering committee, they make those sorts of technical decisions11:24
leitz3) Suggestion: Sponsors tag members that are ready to be elevated. Eliminates voicing success except by success.11:25
-!- openpercept [n=openperc@unaffiliated/openpercept] has joined #fedora-meeting11:25
viking_icethat actually would be a QA vote as in ( SQA not SQC )11:25
wwoodsour charter is soley testing and quality assurance. we don't determine the contents of the product, except by advising on its testability11:25
-!- hanthana [n=hanthana@] has joined #fedora-meeting11:25
viking_iceagain disagreeing here11:25
poelcatlink to the charter?11:26
viking_icethe current qa should be considered SQC11:26
wwoodspoelcat: unofficial11:26
leitzAnd escalation process, as we can't suborn FESCo without at least getting them involved.11:26
viking_iceQA would have member from fesco11:26
wwoodsthat's the sumamtion of discussions with max/paul, FESCo, etc11:26
wwoodserr summation11:27
wwoodsperhaps that needs to be made official11:27
wwoodsthe current makeup of the group is: 5-member board, no official team membership11:27
* leitz eyes our FESCo rep...11:27
wwoodsmyself (lead), jlaska, f13, poelstra, jds2001. this was decided.. basically by fiat.11:27
viking_iceah so this is now currently supposed to be the board11:28
* poelcat doesn't consider myself a board member11:28
jwbdisable ssh by default on what DVD?11:28
f13I'm here because my job depends on things working, and the best way that I can ensure that is by being a part of making them work.11:28
viking_iceit's not like that in the live spin11:28
jwbit's not like what?11:29
jwbthe live spins are CDs11:29
jwband why would we disable ssh by default?11:29
viking_iceyup and with default install of an live cd ssh is not enabled by default11:29
viking_icesecurity risk11:29
viking_iceor atleast deny root access11:29
f13the only reason its disabled on Live is because there are no passwords for the users11:29
wwoodsyou guys, that is a tangent11:29
wwoodsand it is definitely a FESCo matter11:29
jwbyeah, sorry11:29
f13regular installs require you to set a password.11:30
leitzFESCo wiki page.11:30
* bpepple pokes his head in.11:30
f13... but yes, FESCo && OpenSSH maintainer.11:30
viking_iceand a regular novice user does not know that his ssh is enabled by default11:30
viking_icewhile the more experienced one has the knowledge to enable it11:30
f13viking_ice: bring it up elsewhere.  Not a QA matter11:30
f13testing that the default setting is accurate, that's a QA matter11:31
viking_icedepends of the definition of QA11:31
leitzLet's talk about QA at FESCo Friday, okay?11:31
leitzAssuming FESCo page is correct.11:31
f13QA is there to test the product definitions as provided by development and management.11:32
wwoodsI'm telling you, the QA team's purpose is to ensure the quality of the distribution (and the tools used to make it) by testing and other means11:32
f13if you want to change how the product is defined, you operate in those circles, not QA11:32
viking_iceqa in it's current form == SQC not SQA11:32
bpeppleleitz: shoot an e-mail to jds2001 or open up an item on FESCo's trac to add it to Friday's agenda.11:32
wwoodswe don't define the product, except in that we advise the other teams on what is reasonable, testable, etc.11:33
leitzTeam Guidelines, 4) Potential team members get mentored.11:33
viking_iceand the QC leads to be formed by the community and for the community not some appointed persons11:33
leitzbpepple, will do.11:33
viking_icef13: http://sqa.net11:34
wwoodsviking_ice: right - I want community members as "qc leads" (team sponsors)11:34
wwoodsand, yes, new team members should be mentored, either by other members or the sponsors11:34
* f13 has worked in multiple "QA" departments, noneofwhich determined how the product was designed.11:34
leitzviking_ice, make sure we have the questions ready to go fro Friday.11:34
leitzSponsor, not necessarily. The team is pretty good.11:35
leitzSponsors are just senior team members, right?11:35
-!- Nirmal [n=npathak@nat/redhat-in/x-7bec3536b5de53ea] has joined #fedora-meeting11:35
wwoodsessentially, yes - they're the team members that we trust enough to be able to bless new team members11:35
viking_iceQC can have differrent teams like testers and triagers and the testers team can for example have sub teams as well automation gui network etc11:36
wwoodsand to vote on things requiring it11:36
leitzSo team members get mentors, maybe a sponsor just makes sure that happens in a reasonable time?11:36
leitzOr a leader in the team, whether sponsor or not...11:37
wwoodsso. do we need a board of some static size to oversee sponsors, or are the sponsors themselves sufficient to run QA and make decisions?11:37
-!- JSchmitt [n=s4504kr@p4FDD1B0F.dip0.t-ipconnect.de] has quit [Remote closed the connection]11:37
viking_iceThe fedora QA board needs to be made up of individuals from each SIG11:37
f13only if they are interested11:38
leitzAgain, while avoiding the issue of what QA does, the scope of Sponsor voting vice community upheaval needs to be clearly defined.11:38
wwoodswell, I agree that we want a QA rep for each SIG11:38
leitzLet's gain wise counsel and avoid death by committee...11:38
wwoodsyeah I'm trying to keep this as simple as possible, keeping requirements to a minimum11:39
wwoodsbut we need to define: a) what it takes to become a QA team member,11:39
viking_iceI say first we need to come up with an QC offical board lead made from the community for the community11:39
wwoodsb) what it takes to be a sponsor, and11:39
f13the more we talk about process and governance and voting and proper representation, the more I want to just stop contributing to QA11:39
wwoodsc) what is expected of those roles + what is provided to those people11:39
viking_icebecoming a QC members should just be enough to assign to either tester or triage group11:40
viking_icefrom my perspective11:40
-!- DemonJester [n=DemonJes@fedora/DemonJester] has joined #fedora-meeting11:41
wwoodsso wait, there's a differnet FAS group that gives you bug triager permissions in bugzilla, right?11:41
wwoodsthat's all handled by bugzappers?11:41
viking_iceand how much you want to participate beyond testing/triage should is just up to that persons time and will11:41
leitzf13, jlaska, I agree to a great extent. However, there's a great way to work this that I'm not sure we've yet found.11:41
-!- JSchmitt [n=s4504kr@p4FDD1B0F.dip0.t-ipconnect.de] has joined #fedora-meeting11:41
daumasend-user question: if i only have enough time to spend a few hours a week downloading and testing ~10 packages from updates-testing and putting karma on them, do i need to worry about any of this or should I just continue what I'm doing?11:41
wwoodsjds2001: is that true?11:41
f13daumas: you can just continue what you're doing11:42
-!- valente [n=vali@] has quit ["Leaving."]11:42
wwoodsdaumas: you don't *need* to worry about it11:42
f13daumas: and it will continue to be very valuable, and we'll continue to tahnk you11:42
daumasf13, wwoods: ok, thanks11:42
wwoodsbut one of the things I'm proposing that QA team members get is: recognition in bodhi - a little "QA" badge next to your username11:42
* viking_ice notes it's more than a month since I filed a ticket first in infra then in bodhi about remind testers to use bodhi and vote 11:43
viking_icein updates-testing report11:43
leitzdaumas--though if you can send us/me notes on the process and docs that bug you, I'd appreciate it.11:43
wwoodsso that rel-eng / maintainers know that this feedback comes from a recognized QA guy, who definitely knows how to test stuff and knows what they're talking about and should be taken seriously11:43
wwoodswould that be useful?11:43
viking_icewwoods: wont work unless maintainers actually push stuff to updates-testing11:43
wwoodsor should we not even bother? so far we get plenty of people using bodhi and testing stuff without needing any of this bureaucracy11:44
f13viking_ice: tickets are neat, but when there is a single bodhi developer, he has to set his priorities.  Filing tickets doesn't automatically mean code will be generated.11:44
-!- alexxed [n=alex@dyn-] has joined #fedora-meeting11:44
leitzwwoods, I'd like to bring in some actual industry standard/ certification stuff11:44
leitzSo that our QA follows some path that's proven and valid outside Fedora.11:45
wwoodsleitz: that's a rather large can o' worms - can you send a proposal to fedora-test-list for review?11:45
viking_iceback to SQA11:45
jds2001wwoods: is what true?11:45
leitzwwoods, yup. It will be on my page soon.11:46
-!- J5 [n=quintice@nat/redhat/x-f71c8f9d3cfd209a] has joined #fedora-meeting11:46
jds2001wwoods: yes, fedorabugs gives you editbugs and a few other things in bugzilla.11:46
wwoodsjds2001: bugzilla perms are controlled by a FAS group, membership in that group is overseen by bugzappers11:46
jds2001fsvo overseen :)11:46
wwoodsand therefore the QA team doesn't need to worry about oversight/sponsorship/etc. for bug triagers11:46
viking_icewwoods: we should not need any special badges over any other QC team member (tester)11:47
wwoodsjds2001: ok! that simplifies things then. thanks.11:47
jds2001but most often, a QA dude is gonna need the same permissions.11:47
wwoodsviking_ice: right, but should QA team members (testers) get recognition above non-qa-team-members?11:47
leitzwwoods, would that tie into the voting/karma discussion of some time back?11:48
wwoodsjds2001: true, I suppose, but now we have complicated overlap where we need to figure out who deserves the 'fedorabugs' permissions11:48
wwoodsleitz: voting/karma on what?11:49
leitzThat is, should a QA established (somehow) tester be more able to affect application promotion?11:49
-!- adamw [n=AdamW@mandriva/channel-support/AdamW] has left #fedora-meeting ["Ex-Chat"]11:49
jds2001so the filter i use is cla_done and thats it.11:49
-!- adamw [n=AdamW@mandriva/channel-support/AdamW] has joined #fedora-meeting11:49
wwoodswe have a lot of voting/karma discussions..11:49
viking_icefrom my perspective QC member ( tester ) is QC member all on equal ground ( except for the board of the QC )11:49
jds2001I thought we decided that at fudcon...like the first one i was at in RDU.11:49
leitzwwoods, okay, sorry. I'll dig up the one I'm thinking about and share it later.11:49
wwoodsjds2001: pretty much, yeah, since anyone who misbehaves can just be dropped11:49
wwoodsso yeah, qa membership can imply fedorabugs11:50
viking_iceas I see it with the fedora-qa trac instance I want everybody to be able to assign a case if he has the time skills and the means to do so11:50
wwoodsso that's another thing you get11:50
jds2001that takes a minor fas change.11:50
jds2001i can submit a patch for that.11:50
viking_icetesters privileges and triage privileges should be the same because we all possess the skills ( or should actually ) to do each others work11:51
wwoodsviking_ice: right, but is there a useful distinction between 'qa team member' and 'non-member'11:51
wwoodsand what do you have to do to prove that you can be trusted as part of the qa team?11:52
viking_icewwoods: what trust are you looking for11:52
adamwwwoods: i wouldn't set the bar too high on that, as a qa person cannot fundamentally break anything important on his own11:52
adamwso there's no reason to set a very high entry barrier11:53
viking_icewhat expectation do you have to a QA member11:53
-!- biertie [n=bert@] has quit [Read error: 110 (Connection timed out)]11:53
jds2001adamw: agreed.11:53
wwoodsviking_ice: that's what I'm asking you11:54
-!- spoleeba [n=one@fedora/Jef] has joined #fedora-meeting11:54
-!- hanthana [n=hanthana@] has quit [Remote closed the connection]11:54
viking_iceI want to get as many people involved if the QC members some mistake than it's our own fault not to provide proper documentation or guideline11:54
wwoodswe have some *proposed* ideas for perks for QA team members - access to release-candidate bits, triager permissions in bugzilla, badge on bodhi comments, etc.11:55
wwoodsand, of course, beer11:55
* jds2001 <3 beer :)11:55
wwoodsin exchange, we hope for assistance testing those bits, triaging those bugs, and testing those updates in bodhi11:55
viking_icefirst this bodhi badge idea does not work11:56
-!- ankit [n=ankit@] has quit ["Off for the day !!!"]11:56
adamwwwoods: from the bugzilla perspective, what I did for MDV bug triaging squad was a 'license test' - I picked out ten un-triaged bugs and asked them to give a report (on the mailing list) on what triage actions they'd take for each...11:56
jds2001viking_ice: why not?11:56
wwoodsso the question is: is it worth providing benefits to a limited group, just to entice people to participate in the group more?11:56
adamwi'd pick the bugs to give a range of common triage gotchas, so they'd probably get a few wrong and others could explain11:56
jds2001viking_ice: no code? That cna be fixed, learn python :D11:56
wwoodsor should we just drop the entire idea?11:56
viking_iceIt's up to the reporter to state if the bug he filed has been fixed or not and we can not have more value if than any other tester for that matter if our setup works but not his11:56
daumasminimum things i'd appreciate 1) fedora QA expectations 2) guidelines for bodhi: karma and comments 3) PackageKit integration with rollback support, or yum. i don't believe adding a second FESco to fedora will help.11:57
viking_iceso it gets a pass just because I have a badge..11:57
wwoodsdaumas: I'm inclined to agree11:57
wwoodsand viking_ice has a good point as well11:57
wwoodsso if there's no need for team membership, there's no need for sponsors per se11:58
adamwwwoods: i'm not sure it's right to look at it from a 'doggy treats' perspective11:58
wwoodsmentoring and documentation is still valuable, obviously11:58
jds2001the first one is unwieldy as is, we get flamed for making "arbitrary" decisions :)11:58
wwoods(and better tools to help testers do their testing)11:58
viking_iceall testers need to have access to the same bits and that is the same spins and preferably testers mirror11:58
adamwwwoods: i'd like to look at it like this: are there any types of access to any system that we can't reasonably just grant to anyone with an account. if so, what are the minimum sensible barriers to gain access to those rights. if not, grant them to all.11:58
leitzdaumas--I'm working on your 1) and 2).11:58
viking_icewe can not effectively test on out of sync mirror or on outdated spins11:59
wwoodsthe only one that's a legitimate problem is access to the RC bits11:59
adamwi don't think anyone's gonna say "i want to become a qa team member in order to get access to pre-release bits!" it should be more a case of 'what should you need to be able to do before you get access to RC bits'11:59
leitzRC bits == bandwidth.11:59
viking_icebut we should restrict access to QC membership to those bits11:59
adamwthe RC bits aren't a motivator to get people involved, they're just...something you need access to to do the job11:59
EvilBobif they want pre-release bits they just have to become a mirror12:00
f13EvilBob: whic is a pretty significant expense.12:00
daumasor release a torrent..12:00
f13EvilBob: 1TB of online disk space and a ton of bandwidth.12:00
wwoodswe have limited bandwidth to the staging area so we can't just open it up to the world (we tried, it got swarmed with non-testers, none of the testers could get bits quick enough)12:00
jds2001well we dont stage a lot of our RC's to mirrors.12:00
EvilBobf13: not for a private mirror12:00
f13and what jds2001 says12:00
f13EvilBob: mirrors can't get to the RC bits until the bits move out of RC stage and into GOLD stage12:01
adamwwwoods: right. so then we simply need to define some kind of minimum level of competence or commitment or however we want to look at it in order to get access to those.12:01
adamwand a system for managing that that is consistent and understandable and doesn't depend on a single person, but has the minimum possible level of annoying bureaucracy to it12:01
EvilBobf13: ok I see the difference, thanks12:01
viking_icethat minimal level from a QC perspective would be actively report in testing or participate in a triage .12:02
viking_icethe live cd's need to be available to all12:02
wwoodsadamw: right - so, we've got a 'qa' group in the Fedora Account System, and (like all FAS groups) it defines administrators, sponsors, and users12:03
wwoodssponsors can add people to the group, admins can make other people sponsors12:03
adamwviking_ice: does it matter if the person actually has a clue (i.e. files *useful* reports and triages *properly*) or would you rather just take anyone with the enthusiasm and give them a clue as we go along? :)12:03
viking_iceI would accept anyone at this stage because we need all the help we can get in both teams and train/guide them along12:03
wwoodsso here's a proposal: restrict access to the staging area to rel-eng/qa. if you want to be in the qa group, just ask a sponsor.12:04
viking_iceteams being triage and testing12:04
wwoodsevery time there's release candidate bits, we'll mail the qa group12:04
leitzHow much time is spent managing and leading the new folks?12:04
adamwi'd say that's reasonable but if you take that route it's important to be careful that there are people with the commitment and knowledge to train the newbies12:04
wwoodsif, say, three months go by and you don't post any test results 12:04
viking_iceask the appointed leader none12:04
wwoodsyou get dropped from the group12:04
leitzSomeone loses triage time helping someone else learn to triage.12:04
wwoodsalso, I think we may be colliding with FESCo's meeting time12:05
jeremywwoods: fesco meets on fridays now12:05
leitzIsn't theirs Friday?12:05
wwoodsthey should change http://fedoraproject.org/wiki/Communicate/FedoraMeetingChannel then12:05
-!- alexxed [n=alex@dyn-] has quit [Read error: 110 (Connection timed out)]12:06
f13wwoods: I agree to that proposal.12:06
adamwwwoods: that all sounds good to me12:07
viking_iceso are we back to 15:00 because the need of one out weighs the need of many or?12:07
leitzWe agree that we need some guidelines.12:07
wwoodsviking_ice: because that's what was on the schedule12:07
adamwwwoods: and we don't really need a complex system for who's a sponsor and who's an admin at the size we're at now, right? does anyone here have a problem with just handling that in a kinda informal way?12:07
jlaskanone here12:07
wwoodsis there a better time? I know we have more people on the west coast now, so later might be better12:07
poelcatadamw: no12:07
leitzadamw--some informality kills.12:07
* poelcat isn't clear what current problem is in need of solving12:08
leitzI've gotten confused and frustrated...well, even more confused than normal.12:08
viking_icewe should have them as open as possible and adjust them if the need arises12:08
viking_iceis my perspective on the requirements to participate in QC12:08
wwoodsthe most immediate problem is: controlling access to the RC staging area (and notification of new RC bits)12:08
wwoodsbecause that was a problem with f11a12:08
poelcati thought we've always said people can just build them themselves locally and that it doesn't matter?12:09
viking_icerestrict them to QC group no more no less12:09
wwoodspoelcat: turns out it does12:09
wwoodsf13's home-built images behaved differently from the phx-built images that jlaska and I were using12:09
adamwleitz: i agree, but the danger as i saw the previous discussion is that there would be a big heavy arcane system which would *also* kill...12:09
wwoodsrawhide also behaved slightly differently, which is what clumens was testing12:09
-!- warren [n=warren@redhat/wombat/warren] has quit ["Leaving"]12:09
f13I'm still not sure, but I /could/ have had a different package set due to some rsync failures.12:10
f13I was trying to sync the freeze content and may have failed.12:10
viking_iceSo the restriction are bound to be in the QC group and if you behave badly you get dropped out of the group and hence no longer have access to the broken bits?12:11
* poelcat has always thought it was QA101 to test the actual bits you're shipping even if it is "hard"12:11
wwoodseven if that wasn't a problem, we still need to notify people when we have a frozen package list so they can *build* images12:11
wwoodswhich reduces to the same problem - who do we notify, and how do we make sure they get the right bits? well then we need to ensure they all have access to a guaranteed up-to-date mirror..12:12
f13poelcat: yes, but sometimes "hard" becomes "impossible" with our schedules12:12
viking_icecan you restrict access to mirrors ?12:12
f13mirrors never see these bits12:13
f13we churn them far too frequently for them to be of any use.12:13
wwoodsQA100 is: some testing on near-final bits is better than no testing on final bits12:13
poelcatf13: then we innovate and come up with something new :)  Isn't that our way?12:13
f13wwoods: ++12:13
f13poelcat: the inovate and something new was supposed to be "composing in mock == same results"12:13
f13much easier with a blocking freeze and everybody is working from the same rawhide12:14
adamwok, so if it's basically agreed that we restrict rc bits access to qa team members,that's great, and it just becomes about...what does it require to be a qa team member, who controls that process (i.e. who are the sponsors and admins and how is that decided)12:14
viking_iceSo are we going to waste testers time on outdated bits?12:14
wwoodsi.e. much easier for beta 12:14
wwoodsviking_ice: no, we're trying to figure out how to give testers access to up-to-date bits12:14
f13also, /functional/ testing of the software certainly doesn't have to come from the target isos.  The apache package on the iso made in phx is the same apache package in the rawhide tree12:14
wwoodswithout giving everyone in the world the opportunity to swarm that server and keep the testers from getting their work done12:14
f13the only concern really for testing the exact same bits is the installer testing.12:14
* leitz had to work with the auto glass install guy.12:15
viking_icewwoods: yes and that access should be restricted to fas account and QC membership12:15
leitzIsn't QA person not really equal to RC bit tester?12:15
viking_iceno more no less12:15
jlaskawe've got a few topics going at once ... should we record these as bullets for a future meeting?12:15
leitzFor example, I'm not looking to do any RC testing soon so even though I might be helping with QA, do I need RC Bit access?12:16
f13leitz: just because you have them, doesn't mean you'd use them.12:16
wwoodsleitz: AFAICT, no, you don't12:16
f13leitz: but those that would use them, do need to have them.12:16
f13jlaska: yeah, I'm getting confused as to what we're currently talking about.12:17
wwoodsI think we've gotten it reduced to this:12:18
wwoods1) to facilitate RC testing, we should have a private mirror (or restrict access to alt) for QA12:18
-!- mintos [n=mvaliyav@nat/redhat-in/x-8a12d774d1b272de] has quit ["Leaving"]12:19
wwoods2) anyone who wants to help with RC testing can have access, but they may be dropped if inactive (esp. if the machine gets overwhelmed by downloaders who don't report any results)12:19
wwoodsdoes that sound about right?12:20
viking_iceyup aggreed +112:20
leitz2) Perhaps designate subgroup of QA doe RC testing?12:20
jds2001right, eve nthough i have crappy internet at home, getting to alt around rc time is slow as can be.12:20
wwoodsalso, we have an informal agreement that QA doesn't need a board, voting policy, or other bureaucracy.12:20
wwoodswhat it *does* need is better guidelines and documents for testers12:21
wwoodsand better tools for testing12:21
viking_iceso wwoods your make your self unemployed12:21
jds2001indeed, no board, good documents, good tools.12:21
f13leitz: I don't necessarily think we need to subgroup at this point.  If there is ever a problem, then maybe.12:21
jds2001that is my job at $DAYJOB as well.  I've been largely successful :D12:21
viking_icewwoods: how can I contact maintainers and behalf of QA regarding QA matters12:22
leitzNo board, but not sure we have the right structure yet though.12:22
f13viking_ice: not really.  wwoods' day job is to ensure that Fedora QA gets done.  How it gets done doesn't matter, just that it gets done.12:22
viking_iceWe need an offical list of QA members12:22
f13viking_ice: you just say "hey, I"m working on QA and..."12:23
f13viking_ice: why?  For what purpose?12:23
f13I'm not disagreeing necessarily, but I want to know what purpose that would serve12:23
leitzf13, a sub-group informally.12:23
f13leitz: that'll be informally formed by the people actually doing the testing and recording results (:12:24
viking_iceFirst of all because this whole appointed thing that has been done so far has failed to establish a community  and an alternative approach needs to be done to build up that community again12:24
leitzf13, as long as there's enough structure for a potential to know who to talk to and how to join, I'm happy.12:24
wwoodsin all seriousness, feel free to start your own QA group12:24
wwoodsif you want to do something more formal that's fine12:25
-!- JSchmitt [n=s4504kr@p4FDD1B0F.dip0.t-ipconnect.de] has quit [Remote closed the connection]12:25
viking_icef13: should we approach maintainers for test cases or should we *hint* them to come to use us?12:25
wwoodsif we've got KDE and GNOME surely we can accomodate multiple QA groups12:26
f13viking_ice: I don't see how bring an "official" member of the QA team or not will help with maintainer test cases.12:26
wwoodsI just don't see how it's helpful to make a group that's *harder* to manage12:26
f13they're either going to work with you or not, and being a "member" of QA isn't going to help that, nor will it force anything.12:26
f13viking_ice: I'm more inclined to attribute the lack of a large QA group to the fact that QA work is hard and not fun, particularly when not being paid for it.12:27
wwoodsI like to think that spending more time on improving the tools to make it less hard12:28
wwoodswould be the best way to get more people to participate12:28
wwoodsbut then, I'm a tools guy12:28
* wwoods has a hammer, sees a lot of nails12:28
f13and I'm just a tool12:28
viking_iceNo we had much more activity in six and seven when the testers where actually communicated to12:28
-!- themayor [n=jack@dsl081-200-018.nyc2.dsl.speakeasy.net] has quit ["bye bye"]12:29
viking_iceit died when the communication stop coming from the head12:29
adamwi think viking's perspective is the more useful one...f13 may have a point but there's nothing we can do about that, really12:29
f13I'm not sure I agree with that at all.12:29
adamwit's hard to change the *nature* of qa work12:29
wwoodsin six and seven we still produced all of the bits internally at RH, there were no spins12:30
f13I think it died in 7 when we more than doubled the size of "Fedora" and reduced the tools we had to do testing by moving things outside of Red Hat.12:30
viking_iceWe need to establish better communication to the test-list we need to redirect test topics of devel to the test list ( as used to be done )12:30
wwoodsand the distro consisted of 3000-4000 packages instead of 1000012:30
viking_icefirst you can not carry all the qa work by your self it's impossible you need the community help12:31
viking_icethere has not been sent any word on what is being considered in automation12:31
f13the community is helping, just look at the flood of incoming bugs.12:31
jwbi've seen very little test related email on fedora-devel lately12:31
-!- Renault [n=chatzill@AToulon-151-1-32-199.w83-197.abo.wanadoo.fr] has joined #fedora-meeting12:32
f13viking_ice: not any word, except for the posted video from the FUDCon session on it?12:32
jwbarguably that same content could be summarized in an email12:32
viking_icewas there a link sent to the test list about that video (12:32
f13they were announced via fedora-announce-list12:33
f13which any engaged tester should be on12:33
leitzTeam Guideline 6) #fedora-meeting gets released back to the community after we've been hashing and thrashing for 90+ minutes...12:33
viking_icejwb: it's not a question of quantity to the devel list more about it being redirect to test list12:33
viking_iceWe need more docs we need more test cases ( we are working on that )12:33
jwbif there is nothing to redirect, i fail to see how more focus on redirecting it is going to help anything12:34
-!- openpercept [n=openperc@unaffiliated/openpercept] has quit ["Leaving."]12:34
f13especially because fedora-test-list has become "Help me with this problem I'm having"12:35
f13eg end user support12:35
sticksterf13: Only for Rawhide -- or at least, that's how it *should* be used.12:35
viking_icef13: no there's very little if some ( there is one confused tester )12:35
f13viking_ice: I beg to differ.  Every day it's more of the same.12:36
wwoodsokay, so we're agreed: more info to fedora-test-list about what's going on and what to test12:36
viking_iceusually reports in vague12:36
wwoodsso people will stop spinning their wheels12:36
wwoodsand repeating the same vague test reports12:36
f13stickster: I'm not quite sure I agree with that.  If we really want it to be a list to discuss QA/testing matters, it can't be an end user support forum.12:36
viking_iceif it's for rawhide it can12:36
sticksterf13: People running Rawhide *are* testing.12:36
wwoodsyeah rawhide users aren't end-users12:36
f13stickster: yes and no.12:37
wwoodsso rawhide support is a form of testing12:37
wwoodsbut "still doesn't work" and "can't play mp3s" are not helpful12:37
sticksterBut we should be turning away questions on, say, some question on F1012:37
f13stickster: "HALP MY X IS BROKEN"  is different from "How can we test X?"12:37
wwoodsso some direction needs to be given on how to test rawhide12:37
viking_iceagain back to lack of doc12:37
sticksterwwoods: That much is true -- the moderators/leaders on fedora-test-list should have a clear place to point which indicates "Here is what we will help you with on this list viz. Rawhide."12:37
jwbpeople don't read docs12:37
jwb(no offense stickster)12:38
leitzjwb, I do.12:38
jwbyou are not "people"12:38
f13You can lead a user to docs, but you can't beat him/her about the head with them.12:38
* leitz wonders how many ways that can be taken... :)12:38
stickster:-)  But having the doc is sometimes the more important factor.12:38
viking_icejwb: are you saying that all the effort that jlaska is coming up with his virtual lab setup are then in vain12:38
jwbleitz, i mean no offense.  take that as a compliment :)12:38
leitzhehe...no problem...12:39
sticksterIf we don't document something, it doesn't exist. That's our fault as community leaders. If we do and someone doesn't read it, we can point them to it. If they refuse to look, we know we can spend our time on other pursuits.12:39
viking_icemean I was going to train testers I would certainty point to the final out from that12:39
jwbviking_ice, not at all.  i'm saying that just because we have more docs available, that does not mean we will get less emails about random problems12:39
jwbit simply means we'll have more opportunity to point people to them and say RTFM12:39
sticksterWhere RTFM is *actually useful*, that's valid. As a general answer, it's not -- in specific cases it is, as long as it's done in a helpful way.12:40
jwbi was being blunt to prove a point :)12:40
wwoodsso, yeah, we need to wtfm12:40
wwoodsso we can tell test-list to rtfm12:41
-!- mcepl [n=mcepl@49-117-207-85.strcechy.adsl-llu.static.bluetone.cz] has left #fedora-meeting []12:41
wwoodsin those circumstances where tfm would actually help12:41
wwoodsdocs good. mentoring good.12:41
sticksterWill someone take the action of writing that short page, explaining what kind of audience and questions fedora-test-list is for, and how a subscriber can write a worthwhile question/post?12:42
* stickster just making a suggestion for next action12:42
jwbit should be on the mailman page for the list itself12:42
viking_icestickster: could you create a ticket in fedora-qa trac instance for that12:42
wwoodsnobody ever reads that, but it should be there12:42
wwoods"fedora-test-list -- For testers of Fedora Core development releases"12:42
jwbthanks for reiterating my point wwoods ;)12:43
wwoodsokay. we are now at an epic 2 hours, 45 minutes12:44
wwoodsfor a one hour meeting12:44
wwoodshelp me summarize.12:44
wwoods* F11 alpha is on its way out. As always, there are some bugs, but it's mostly OK.12:45
wwoodsBiggest problem: bug 48066712:45
buggbotBug https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=480667 medium, low, ---, pjones@redhat.com, NEW, nash unable to find dm devs by uuid or label causing boot to fail12:45
wwoodsboot_delay=5 might help if your system won't boot, or do 'root=/dev/[volgroup]/[logvol]'12:45
sticksterviking_ice: wwoods: https://fedorahosted.org/fedora-qa/ticket/1012:46
f13boot_delay=5~30 PLUS removing 'quiet'12:46
wwoods* We had some problems during testing because of mismatched test images. From now on, we want to make sure testers all use the same images. More on this later.12:46
* leitz goes to get ready for phone interview...12:46
-!- leitz is now known as leitz_afk12:46
wwoods* The wiki QA: namespace is to be used for QA stuff that is *primarily* written for testers - test plans, cases, debugging tips, etc.12:47
wwoods* General interest documentation - what QA is, how it works, what we do, how to help - goes in the normal wiki.12:47
wwoods* Wiki gurus would like to remind you to use natural-language names for your pages, avoiding slashes12:48
-!- stickster is now known as stickster_food12:48
wwoods* We'll move the existing test plans/cases to QA:12:48
-!- JSchmitt [n=s4504kr@p4FDD1B0F.dip0.t-ipconnect.de] has joined #fedora-meeting12:48
-!- lcafiero [n=larry@dsl-63-249-115-153.cruzio.com] has left #fedora-meeting ["Leaving"]12:49
wwoods* The QA team has decided that it doesn't currently need a board, or voting procedures, or other formal bureaucracy12:49
wwoods* What we *do* need are more useful instructions and tools for testers.12:49
wwoods* These instructions should be clearly communicated to fedora-test-list12:49
wwoods* The one thing that we *do* need formal organization for is: allowing access to the staging area for release candidate images.12:50
wwoods* Soon we'll set up a private mirror for QA team members. Once that happens, we'll announce it and tell you how to request access.12:50
wwoodsoh, this should have been a bit earlier:12:51
wwoods* The QA team's purpose is to ensure the quality of the software produced by the Fedora project, through testing and other appropriate methods.12:52
-!- k0k [n=k0k@fedora/k0k] has joined #fedora-meeting12:52
-!- sdziallas [n=sebastia@p57A2E336.dip.t-dialin.net] has joined #fedora-meeting12:52
-!- EvilBob [n=EvilBob@fedora/bobjensen] has quit [Read error: 104 (Connection reset by peer)]12:52
viking_iceso QA == SQC12:53
viking_iceor QC12:53
wwoods** Note that this does not involve deciding the composition of the product (that's FESCo's job), except insofar as we advise other teams on whether their plans are reasonable, testable, etc.12:53
wwoodsuse whatever letters you want, we're still the QA team because that's the common parlance12:53
wwoodshave I missed anything?12:53
-!- EvilBob [n=EvilBob@fedora/bobjensen] has joined #Fedora-Meeting12:53
leitz_afkBBQ tomorrow.12:54
-!- kulll [n=kulll@] has quit [Connection timed out]12:54
-!- warren_ [n=warren@c-66-30-2-120.hsd1.ma.comcast.net] has joined #fedora-meeting12:55
daumaslooks good. probably want to include something on fedoraproject.org/en/join-fedora ??12:55
-!- fcami [n=fcami@wmh38-1-82-225-54-107.fbx.proxad.net] has joined #fedora-meeting12:56
viking_iceYeah we need to fix that12:56
wwoodslet me clarify about the mirror: it'll be restricted to members of *a certain group in the fedora account system*, but that just means that you have access to the RC bits. being a "QA team member" is distinct from that, and is largely informal.12:56
viking_icenot the fedoraproject.org/en/join-fedora but the page that comes after you click an iconn12:56
-!- warren_ is now known as warren12:56
daumasyes. and with that i'm out to lunch.12:56
leitz_afkdaumas, are you on the fedora-test-list list?12:56
wwoodsah yes:12:56
wwoods* qa team members have come to the consensus that beer is good12:57
f13fwiw, I just saw a nash patch that may fix the root mounting bug12:57
wwoodsyeah, pjones said hansg might have a fix for that12:57
viking_iceas long it's not American one will give Canadian a try when I get my hands on it12:57
* nirik would like to mention to all the QA folk that we are doing some IRC Classroom sessions this coming weekend. If anyone would like to teach a class on QA related topics, they would be most welcome to.12:58
wwoodsokay! I'm saving the log and will upload it to http://wwoods.fedorapeople.org/fedora-qa/12:59
wwoodsI'll send a summary to fedora-test-list this afternoon12:59
wwoodsthanks for your time, folks13:00
viking_iceThanks for the holding this meeting13:00
wwoodsoh! should we change the meeting time for next week?13:00
wwoods(since we don't have to worry about FESCo and I know we have a bunch of westcoasters who might appreciate something a bit later than 7am)13:00
viking_iceshould we settle on one time to always be that time?13:01
* viking_ice needs UTC 13:01
wwoodstechnically the meeting is listed as 1500UTC - 10am EST13:01
f13aah, I had it in my phone for 1600 UTC13:01
viking_iceI'm usually up 36 hours so let's try to keep somewhat in that time frame13:01
wwoodsit's 1500UTC, which is 11am EDT13:02
wwoodsbut now we're in EST so...13:02
f13I'm OK with 1600 UTC13:02
f13but no earlier13:02
viking_icelike it had been for this meeting13:03
viking_iceany one who cannot attend the QA meeting @ 16:00 UTC13:03
viking_icegoing once....13:03
wwoodsfine with me13:04
viking_icegoing twice..13:04
f131600 UTC, not to be changed when DST hits13:04
f13wwoods: you still take your lunch at 1pm anyway right?13:05
wwoodsSold! I'll move the meeting time on the wiki page - 1600UTC, regardless of US DST13:05
* viking_ice unaffected by "daylight savings" 13:05
wwoodsI take my lunch whenever I can get it.13:05
wwoodsbut I am a complete zombie before 1400UTC.13:05
-!- lfoppiano [n=lfoppian@fedora/lfoppiano] has quit ["Ex-Chat"]13:05
wwoodsthanks again, everyone13:06
--- Log closed Wed Feb 04 13:06:32 2009

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!