UbuntuIRC / 2009 /06 /11 /#ubuntu-testing.txt
niansa
Initial commit
4aa5fce
=== asac_ is now known as asac
=== mathiaz_ is now known as mathiaz
[07:09] <ara> good morning!
[07:55] <persia> Has anyone else looked at the Ubuntu Studio alpha 2 builds? I can't seem to get the -rt kernel to boot post-install.
[08:03] <ara> persia: I haven't try
[08:03] <ara> tried, even
[08:03] <persia> ara, No worries. I'll bug the team. Might be that the images fail.
[14:00] <ara> jcollado, eeejay: people are already talking about mago in the list ;-)
=== thekorn_ is now known as thekorn
=== torkiano is now known as torkiano|afk
[15:41] <persia> So, I'm trying to test something in a schroot. I've used schroot -p to get access to my environment (so X works, etc.). Unfortunately, /dev doesn't contain the device I seek. Does anyone happen to have a strategy to work around this?
=== torkiano|afk is now known as torkiano
[16:52] <sbeattie> persia: did you come up with a solution? I've not used schroot, but I've bind-mounted /dev successfully into other chroots before to solve that problem.
[16:55] <persia> sbeattie, I ended up finding a way to do my test without direct device access. I'll give bind-mounting a shot, although I worry about permissions (I tend to use schroot to run things I don't trust on my local system :) )
[16:57] <sbeattie> you could bind mount just the specific device you need access to, rather than all of /dev
[16:57] <persia> Really? How does one do that?
[16:59] <stgraber> just like you would for a directory
[17:00] <sbeattie> yep; e.g. mount -o bind /dev/sdb1 /tmp/sdb1 (where you've touched /tmp/sdb1 to make it exist first)
[17:02] <persia> Oh, nifty. I'll definitely try that next time I launch, as it's minimally invasive.
[17:33] <henter> 06270627hii
=== fader is now known as fader|lunch
=== rmcbride_ is now known as rmcbride
[17:48] <slangasek> we could use some help testing kubuntu ISOs for alpha2, currently there are 0 test results reported in for any of the ISOs
=== MaWaLe is now known as Mosquitoooo
=== fader|lunch is now known as fader
[19:17] * fader does some kubuntu alternate amd64 testing
[19:17] <eeejay> hey cr3
[19:18] <cr3> eeejay: yo homie
[19:18] <eeejay> cr3, it's looking good
[19:19] <eeejay> cr3, now i am at the stage where i am dealing with the test results
[19:19] <eeejay> cr3, is the launchpad_report xml format what we use?
[19:19] <cr3> eeejay: excellent, but are you still firing report-result in your MagoCommand?
[19:20] <eeejay> cr3, no, and that is another buggie
[19:20] <cr3> eeejay: ideally, you shouldn't be concerned with the resulting report and you should only have to fire a report-result event
[19:20] <eeejay> cr3, the whole "questions" concept is very manual testing
[19:20] <cr3> I'm not sure I understand
[19:21] <eeejay> cr3, right now i added a "prompt" option to the plugin's options
[19:21] <eeejay> cr3, when it is false, the tests are added as shell plugin tests
[19:21] <cr3> eeejay: can you push your latest code so that I can have a look at that?
[19:21] <eeejay> cr3, when true: "manual"
[19:21] <eeejay> cr3, it is pushed
[19:22] <cr3> eeejay: interesting...
[19:22] <eeejay> cr3, how bad is it?
[19:23] <eeejay> cr3, with prompt being true or false, I want the test results to have the same fields, namely:
[19:23] <cr3> eeejay: it's good, I really like how you can set the plugin on the fly like that :)
[19:24] <eeejay> 1. pass/fail/skip 2. duration 3. message
[19:24] <cr3> eeejay: hm, that's reasonable but I wanted to get someone else's point of view regarding #3
[19:24] <cr3> fader: yo, got a minute?
[19:25] <fader> cr3: For you I've got three!
[19:25] <eeejay> cr3, the whole "comment" field is arbitrary for these kinds of tests. it is a nice addition, but that is not where the data should be attached
[19:25] <cr3> eeejay: by the way, correct me if I'm wrong but it doesn't seem like you need the mago_prompt plugin anymore, right?
[19:25] <eeejay> cr3, correct
[19:25] <eeejay> cr3, it is staying around - just in case :)
[19:25] <cr3> eeejay: so where should the data be attached then?
[19:26] <cr3> fader: so, in checkbox, when a manual test has a corresponding script which can be run from the "Test" button, would it be reasonable for the output of the command to populate the comment box?
[19:26] <eeejay> cr3, result.data is fine, but the way it is martialled in launchpad_report doesn't seem right
[19:27] <fader> cr3: From a simplicity of test creation point of view, sure. From a user-experience point of view, I don't like it. :(
[19:27] <fader> IMO the comments box should be for the user and for the user alone. If I start seeing stuff pop up in there I'm going to get nervous and wonder if it's safe to edit.
[19:27] <eeejay> cr3, fader, i think we might need to introduce a new UI for this. it doesn't match the manual testing experience
[19:27] <fader> Plus the cases where people will want to edit or delete the info that gets populated there... it just feels unsafe :(
[19:28] <fader> (Sorry, eejay)
[19:28] <cr3> fader: fyi, I have changed the behavior of the radio buttons to be set to pass/fail depending on the result of running the script. I thought that was reasonable though
[19:28] * eeejay agrees
[19:28] <fader> Yeah, I saw that discussion... that's a Good Thing.
[19:28] <eeejay> cr3, akshully, that might be problematic..
[19:28] <fader> Nice little usability boost!
[19:28] <eeejay> let's say video test
[19:28] <eeejay> the gst script returns status 0
[19:29] <charlie-tca> anyone else having issues restarting the desktop cd?
[19:29] <eeejay> doesn't mean the user saw the video correctly
[19:29] <fader> eeejay: It seems to me that as long as the user can change the value to 'fail' that's not a big deal
[19:30] <cr3> eeejay: for most cases though, it seems that the behavior is at least helpful to the user, not authoritative for the user. otherwise, the test wouldn't have to be manual anyways
[19:30] <fader> We're saying "Checkbox thinks this thing worked correctly. Did it really?"
[19:30] <eeejay> fader: cr3: i would like to see an alternative show_test() method, instead of changing the behavior of this one
[19:30] <cr3> eeejay: what you're asking for is a way to select the tests which the user may or may not want to run, right?
[19:30] <eeejay> fader: cr3: the radio button change is very subtle, not really noticeable
[19:30] <eeejay> cr3, yeah. but not upfront
[19:31] <eeejay> cr3, allow the user to watch the test
[19:31] <eeejay> cr3, and then populate a UI
[19:31] <fader> I forget.... what does the radiobutton default to now after a test?
[19:31] <eeejay> cr3, which could even be read-only, except for a comment field
[19:31] <cr3> fader: it remains in skip state
[19:31] <eeejay> .. which i think is the correct behavior for manual tests
[19:32] <cr3> eeejay: wait, are you saying you want the test to: 1. run automatically; 2. populate the ui; 3. ask for feedback from the user?
[19:32] <eeejay> cr3, 3 - give the user an _option_ for additional comments
[19:33] <cr3> eeejay: as provided by the current ui, right?
[19:33] <eeejay> cr3, just like the current manual plugin
[19:33] <eeejay> cr3, could we talk on the phone?
[19:33] <fader> Heh
[19:33] <eeejay> fader: cr3: skype conference?
[19:33] <fader> eeejay: You don't realize that cr3 has a direct brain-link to IRC :)
[19:33] <cr3> eeejay: that's what I once did but then I introduced hibernate and suspend tests, imagine the user experience then: user is running a test and then clicks the "Next" button. suddenly, his machine hibernates for no apparent reason :)
[19:34] <eeejay> cr3, brilliant :)
[19:34] <cr3> eeejay: my motivation at the time is that I found it redundant for the user to have to click on the "Test" button, so I thought it was helpful to just run the test :)
[19:35] <cr3> eeejay: ideally, I would like to solve both hibernate and mago use cases in a consistent way instead of introducing all kinds of variations
[19:35] <eeejay> cr3, i agree
[19:35] <cr3> eeejay: even in the case of mago tests, I think the user might get surprised if applications suddenly start popping up without their prior and explicit consent or, at least, some action on their part
[19:36] <eeejay> cr3, the action on their part is "run test", we could make it clearer
[19:36] <cr3> that being said, lets revisit the problems with the current manual user experience
[19:36] <eeejay> "run automated UI test"
[19:36] <eeejay> dunno
[19:37] <eeejay> cr3, the current manual experience is good, imho
[19:37] <eeejay> cr3, that is why i don't want to touch it, but create a new UI that matches mago tests
[19:37] <cr3> eeejay: I would have the description of the test be explicit but keep a consistent "Test" button which can consistently have the same shortcut (Alt-t)
[19:38] <eeejay> cr3, i am just going to come up with a patch as a proof of concept
[19:38] <eeejay> cr3, but first i will eat a sandwich
[19:39] <sbeattie> iiiinteresting, ubiquity just crashed on the kubuntu live installer.
[19:39] * fader wishes he could click a "give sandwich" button.
[19:39] <cr3> eeejay: I really don't see a problem with the description of the test saying something like: This will test gedit functionality. Click on the Test button to run the tests automatically.
[19:39] <cr3> fader: that would have to be run by the dbus backend so that you could also run: sudo give sandwich
[19:40] <fader> cr3: Wow, I should have seen that coming.
[19:40] <cr3> fader: that was pretty predictable indeed, I will try harder next time
[19:40] <fader> :)
[19:42] <fader> http://mirror.servut.us/kuvat/push%20button,%20receive%20bacon.jpg
[19:44] <cr3> fader: the problem is that the sign is probably for a dryer in the bathroom... even I can't come up with something to do with bacon in the bathroom
[19:44] * cr3 is lacking imagination :(
[19:44] <fader> cr3: "He doesn't know how to use the three shells!"
[19:45] <cr3> what the heck am I talking about, bacon is always good wherever you are
[19:45] <cr3> fader: very good point, I spent many a sleepless nights pondering those darn three shells
=== pace_t_zulu_ is now known as pace_t_zulu
[21:20] <davmor2> fader: how are you dude?
[21:20] <fader> davmor2: brilliant and stunningly good-looking. You?
[21:21] * davmor2 picks myself up off the floor.
[21:21] <davmor2> Dude you really need to clean that mirror from time to time ;)
[21:22] <davmor2> Fine thanks. Racing around like a headless chicken but other than that okay :)
[21:23] <fader> Heh... it keeps you busy. Off the streets and out of trouble, right?
[21:26] <davmor2> fader: I know but I could do with being paid too :)
[21:27] * charlie-tca thinks some people want it all ;-)
[21:28] <davmor2> charlie-tca: eating helps you know it's true :)
[21:28] <charlie-tca> Once in a while, anyway
[21:28] <charlie-tca> I haven't eaten in two days, again
[21:29] <charlie-tca> I am thinking I need to, though
[21:30] <fader> I have a great idea for a company. Busy people can pay me to eat for them. (Naturally I'd want high-end clients who are just too busy to eat their five-star meals.)
[21:31] <charlie-tca> heh, just might work
[21:33] * davmor2 no good looking at me then :D
[21:33] <cr3> fader: you could also offer to go to the bathroom for them as an added bonus, no charge
[21:34] <charlie-tca> ohh, I would pay for that!
[21:34] <fader> cr3: Nah, I'm going to focus on my core competency here. I'll leave the add-ons for third parties.
[21:34] * fader mutters something about leveraging synergies to grow the brand...
[21:34] <cr3> fader: you're missing out on two potential customers in this channel: charlie-tca and myself
[21:35] <fader> cr3: I've seen what you eat. You couldn't pay me enough to go to the bathroom for you.
[21:35] <davmor2> cr3: he did say 5 star meals though dude ;)
[21:35] <fader> ;)
[21:36] <cr3> yeah, meow mix probably doesn't qualify as "5 start"
[21:36] <cr3> err, "star" even
[21:37] <davmor2> cr3: It might do if your a cat :D
[21:37] <fader> I'm definitely not going to the bathroom for cats.
[21:38] <charlie-tca> Seems cleaner than dogs
[21:53] <fader> Hey, does the kubuntu oem install normally have a desktop icon for 'prepare for shipping' the way the ubuntu (gnome) oem install does?
[21:53] <davmor2> fader: yes
[21:54] <fader> davmor2: Thanks, that's what I thought.
[21:54] <davmor2> it's normally in the desktop folder
[21:54] * fader is off to Launchpad.
[21:54] <davmor2> fader: first check and see if it's in the menu
[21:55] <fader> davmor2: Yeah, I didn't see it there unfortunately.
[21:55] <fader> There's also nothing in ~oem or ~oem/Desktop
[21:55] <fader> Er, there are things there but not oem-config related :)
[22:03] <cr3> pip pip, cheerio
[22:07] <persia> Anyone bored and feel like testing UNR?
[22:09] <fader> BRB, got to switch network connections here
[22:11] <fader> What did I miss?
[22:17] <davmor2> fader: just to double check you had installed the oem part right ;)
[22:18] <fader> Hehe yeah
[22:18] <fader> I had the same thought but was reassured when I realized I was logged in as 'oem' and it said 'oem temporary user' (or whatever it says) as my username
[22:19] <davmor2> i'm just confirming it for you :)
[22:21] * fader hugs davmor2.
[22:22] <davmor2> persia: just finished dling it
[22:22] <persia> davmor2, Then you've beaten me :)
[22:22] <persia> Apparently all the regular testers took a break for this alpha, which leaves the rest of us :)
[22:22] <davmor2> 20meg BB
[22:23] <davmor2> persia: no it's just I've been busy ;)
[22:24] <persia> Oh. I thought plars and StevenK usually did the UNR testing. Sorry to exclude you.
[22:32] <davmor2> Oh yes
[22:32] <davmor2> persia: I just throw my hat in the ring from time to time :)
[22:35] <davmor2> fader: can you try a resize on kubuntu against an existing install please
[22:35] <davmor2> with the existing having ext4 for the fs
[22:35] <fader> davmor2: Roger wilco... I'm just finishing up an install now so you have good timing
[22:41] <fader> davmor2: You don't mind if it's an alternate install rather than desktop do you?
[22:42] <davmor2> My issue is on live but it might be good to see if it affects alt too
[22:43] <fader> Huh, I don't seem to have the option.
[22:44] <fader> Maybe I munged something in my last install... let me play with this a bit.
[22:46] * sbeattie tries to reproduce as well.
[22:56] <sbeattie> davmor2: worked here.
[22:56] <davmor2> meh just me on hw then?
[23:12] <fader> Okay, resize has shown up for me this time with only a single partition. Does it not pop up if you have too complex a partition layout for it to handle?
[23:13] <fader> My internal clock is set to EAT... back in a bit
[23:46] <eeejay> hey fader
[23:47] <eeejay> fader, have you ever edited the checkbox glade file?
[23:48] <fader> eeejay: I don't believe so... why?
[23:48] <eeejay> fader, because there are custom widgets in it
[23:48] <fader> eeejay: cr3 might snap off my fingers if I started poking around at the internals of it
[23:48] <eeejay> fader: and i think you need a glade catalog file to compliment it
[23:48] <fader> eeejay: Yeah, I can't help you with that then :(
[23:49] <eeejay> fader: darns!
[23:49] <davmor2> fader: I got prepare for shipping in my kde
[23:49] <fader> davmor2: Is this with the alt CD?
[23:49] <eeejay> davmor2: where will you ship kde to?
[23:49] <fader> Hehe
[23:50] <davmor2> eeejay: endusers
[23:50] <davmor2> fader: no live
[23:50] <eeejay> davmor2: typical
[23:50] <fader> davmor2: Hmm, I'll try it again with the alt CD next
[23:51] <davmor2> eeejay: who else would you like kde to be shipped to ;)
[23:52] <eeejay> davmor2: the world's hungry
[23:52] <davmor2> eeejay: wouldn't they still be hungry
[23:52] <davmor2> I'd of thought food would of been better :)
[23:52] <fader> o_O
[23:53] <eeejay> i think they could use all the help they could get
[23:56] <slangasek> fader: I hear there may be something missing from the kubuntu alternate OEM stuff?
[23:56] <fader> slangasek: Yeah, I didn't get the 'prepare for shipping' icon
[23:56] <fader> I filed a bug but I will retest momentarily... I can't be sure it wasn't an ID10-T error.
[23:57] <slangasek> fader: does that mean the OEM install process fails? I'm not familiar with that icon
[23:57] <fader> slangasek: There was no way to finalize the install and prep it for the end-user to receive. It kept the 'oem' user and didn't present the first boot stuff that the user is supposed to get at first boot.
[23:58] <fader> First boot first boot
[23:58] <slangasek> fader: ok
[23:58] <fader> slangasek: I will retest and see if it was the crack I had for lunch :)
[23:58] <slangasek> I guess I'm inclined to pass on documenting this in the errata anyway, supposing that OEM mode is not a significant use case for our early-alpha testers