[08:40] <petski> seb128, already had some time to look at bug 276603 ?
[08:41] <seb128> petski: no, I commented after midnight went to bed and I'm just starting my work day
[08:42] <petski> okay, no rush :)
[08:42] <petski> Good morning :)
[08:43] <seb128> right, there is no rush, good morning or afternoon or evening whatever match your timezone to you ;-)
[08:51] <mvo> hey seb128! did you had a chance to use the patched dpkg in my ppa? any explosions yet from it? I would like to know if it gives you further performance improvements
[08:51] <mvo> (no rush, just curious)
[08:52] <huats> morning everyone
[08:53] <didrocks> hey huats
[08:54] <didrocks> seb128 & mvo & petski as well ;)
[08:55] <petski> morning :)
[08:58] <seb128> mvo: no, I just tried your bzr thing, what is the dpkg update supposed to speed up, the index reading?
[08:58] <seb128> lut didrocks mvo
[08:58] <mvo> seb128: the dpkg update avoids reading the available file for most operation (like unpack/configure). that saves another ~5mb of disk reads
[08:59] <seb128> mvo: ok, will install this one and let you know if it feels faster then ;-)
[08:59] <mvo> seb128: excellent
[09:00] <didrocks> mvo: do you have a Spec for those improvements (just courious, I think you parse /var/lib/dpkg/status et clean old deb?)
[09:01] <mvo> didrocks: I have no spec, but there is some explaination in the script, its pretty simple, just purging stuff that is really old
[09:01] <mvo> (purging not in the dpkg sense, but in the get-it-of-the-file sense)
[09:02] <didrocks> mvo: sorry, I didn't see today where your script was (in your ppa?)
[09:02] <mvo> didrocks: http://bazaar.launchpad.net/~mvo/%2Bjunk/compact-dpkg-status/annotate/head%3A/compact-dpkg-status
[09:02] <didrocks> mvo: thanks
[09:03] <didrocks> I will take a test tomorrow
[09:03] <mvo> didrocks: I added a system-cleaner plugin for it now too and will make it part of the dist-upgrader I think
[09:03] <mvo> the script is step 1, the dpkg patch step 2
[09:03] <mvo> together performance should be much better indeed
[09:03] <didrocks> mvo: BTW, you might have an answer I keep asking myself.
[09:04] <didrocks> why md5sums of conffiles are not embeeded in the package? It seems to be generated at install time
[09:04] <didrocks> (in the status file and not in the md5sums one)
[09:04] <didrocks> I thought the reason why was if we modify them in postinst, the change will be taken into account
[09:05] <didrocks> but apparently, this is not the case :/
[09:05] <seb128> mvo: do you have an intrepid system handy? or do you use jaunty now?
[09:05] <mvo> seb128: my laptop is still intrepid
[09:05] <didrocks> (I have a meeting now, will be back on a couple of hours)
[09:06] <seb128> mvo: when you have it booted can you test build and sponsor the change on bug #276603? the debdiff is trivial but I've no intrepid box here to do that
[09:07] <seb128> having to keep hardy, intrepid and jaunty installations is no fun, I should set up an intrepid vm or something
[09:08] <mvo> yeah
[09:08] <seb128> thanks
[09:10] <mvo> seb128: building it now
[09:12] <petski> Regarding 276603: Some users already tried a simular dpkg from my PPA with good results
[09:12] <seb128> dpkg?
[09:13] <petski> yeah, package, whatever
[09:13] <mvo> petski: oh, what ppa is that?
[09:13] <mvo> nevermind
[09:13] <huats> hello mvo and seb128
[09:13] <petski> mvo: it's described in the bug comments
[09:13] <mvo> wrong context :)
[09:14] <huats> pitti: sorry to bother you with that. Did you have some time to look at the gcalctool SRU ?
[09:14] <petski> my bad :)
[09:14] <huats> pitti: and hello btw :)
[09:17] <seb128> mvo: your patched dpkg seems to behave fine, I'm on my desktop right now though which has a newer installation than my laptop and a faster disk so I don't notice the slowness that much, I'll try on my laptop too a bit later
[09:20] <mvo> ok, thanks
[09:30] <mvo> seb128: seems to be working ok
[09:31] <seb128> mvo: thanks, so you can do the upload ;)
[09:34] <mvo> ok
[10:01] <mpt> asac, there is no standard mechanism for showing dialog/form errors that I know of
[10:02] <pochu> good morning
[10:02] <pochu> seb128: hi, is anybody doing vinagre 2.25.5?
[10:02] <seb128> pochu: hey, not that I know no, want to do it? ;-)
[10:02] <mpt> asac, Web pages actually tend to do that better than native dialogs do, because they have more room, so they lengthen the page and insert an error saying "hey, you forgot this bit" (Launchpad has primitive examples of these)
[10:02] <pochu> sure :)
[10:04] <mpt> asac, it's fairly common to disable the OK/equivalent button in a dialog until the form values are acceptable, but occasionally it's not obvious why they're unacceptable yet, so the OK button is kept available and when clicked it brings up an annoying error alert explaining the problem
[10:04] <asac> mpt: yeah. but what to do for native form dialogs?
[10:05] <mpt> asac, so as usual the answer is It Depends :-)
[10:05] <mpt> asac, what's the dialog in question?
[10:05] <asac> mpt: network-manager connection editor dialogs
[10:07] <mpt> (That really should be renamed. Text is edited. Connections are not.)
[10:07] <mpt> asac, just the password field? or other fields?
[10:08] <asac> mpt: all kind of fields
[10:08] <asac> fields or even field combinations
[10:08] <asac> e.g. you can select a few checkboxes, but never disable all
[10:09] <asac> hence it was a general question :)
[10:12] <mpt> asac, I'm sorry there isn't a general answer :-)
[10:12] <mpt> It depends on whether it's obvious enough why a set of values is unacceptable without any extra hint
[10:14] <mpt> It would be nice if there was a standard way of highlighting unacceptable form controls (e.g. flashing them pink or putting an error icon in them or something)
[10:15] <mpt> hmmmm
[10:15] <mpt> there could be a little error icon in each unacceptable field, and clicking the icon opens a balloon explaining the exact problem
[10:16] <mpt> so that you don't need to dismiss anything if you can see yourself what the problem is
[10:30] <pochu> seb128: do you know what could be causing bug 321916? I can reproduce it on Jaunty, but not in Debian (both with 2.25.5). Might it be GTK+?
[10:35] <seb128> pochu: no idea about this one
[10:36] <pochu> ok, will ask jwendell
[10:36] <pochu> seb128: bug 322247
[10:37] <seb128> pochu: thanks
[10:37] <pochu> in case you can sponsor it :)
[10:38] <pochu> heh
[10:38] <pochu> evolution's configuration dialog doesn't fit in my VM
[10:48] <seb128> right, that's a known issue for notebooks, etc
[10:57] <seb128> pochu: why do you remove the copying and install now?
[10:58] <pochu> seb128: it was done in the past, but I accidentally removed that code from debian/rules when I splitted vinagre into vinagre-common, and then undo the change
[10:58] <pochu> seb128: lintian complains, and they are not needed at all
[10:58] <seb128> ok
[10:58] <pochu> we have copyright, and you don't INSTALL as it's already installed :)
[10:58] <pochu> don't want
[11:09] <pitti> huats: frankly I don't think we should bother about this too much; intrepid SRUs already take more than appropriate time away from devs..
[11:12] <huats> pitti: ok...
[11:13] <seb128> pitti: not easy to say to users to wait another 3 months to get their issue fixed though, but I agree that this sru is probably not an important one
[11:22] <hggdh> seb128, perhaps we should include a warning about going back to evo 2.24 after upgrading to 2.25.x due to the folders.db issue
[11:23] <hggdh> and the dirty column warning on first run of 2.25.x
[11:25] <seb128> hggdh: there is not so many people downgrading versions but I'll talk to srag about it when he's around before uploading the new version
[11:27] <hggdh> Oh, yes, of course. But, still, a warning on upgrade would be warranted -- all evo users will get the pop-up on upgrading to 2.26
[12:05] <pochu> seb128: thanks for the upload
[12:06] <seb128> pochu: you're welcome
[13:00] <asac> mpt: sorry was on the call. i think the icon is a good thing
[13:01] <asac> mpt: i will try to discuss this with NM upstream a bit.
[13:02] <seb128> Keybuk: is bootchart supposed to work out of the box in jaunty?
[13:02] <asac> mpt: oh ... if you want to provide input on what you want to see changed in applet, it would be a good time to do so now
[13:03] <asac> mpt: we are currently defining things on TODO list and priorities. having suggestions on UI would be good thing
[13:42] <mpt> asac, I would love to help with the design, but I don't understand enough about networking (e.g. what things are mutually exclusive and what aren't, how common each option is)
[13:49] <mpt> asac, I think it needs either a couple of days of solid meetings between designers and networking experts, or alternatively, about six months of equivalent discussion on the mailing list
[13:50] <asac> mpt: ok. so you are not working on it right now?
[13:50] <asac> thats ok too ;) ... i just thought you already worked on that ;)
[13:50] <asac> mvo: can you look at the gobby document again?
[13:50] <asac> mvo: i did some major polishing i hope
[13:51] <asac> and completing stuff
[13:56] <mvo> asac: on the phone right now
[13:56] <asac> mvo: thats ok ;)
[13:57] <mpt> asac, I've done some sketches of improving just the menu
[13:58] <mpt> but if the windows can be sorted out too, that would alter what the menu looks like
[13:58] <asac> mpt: great. its not super urgent, but if we could gather all ideas and drafts for further discussion that would be great.
[13:59] <mpt> asac, will you be able to explain to me all the intricate details of "System setting" and IPv4 routes and simultaneous wired and wireless connections etc?
[14:00] <asac> mpt: i guess so ;)
[14:00] <mpt> ok
[14:00] <mpt> asac, are any of the other NM developers in Berlin or nearby?
[14:00] <asac> if there are open questions at end of day i can certainly find out
[14:01] <asac> not that i know. actually i dont know where tambeti is from. i will ask
[14:02] <mpt> ok, cool
[14:02] <asac> but i doubt he is german
[14:02] <asac> Ingo Tambet
[14:04] <mpt> Estonia, apparently
[14:28] <rickspencer3> tedg: ping
[14:29] <tedg> Morning rickspencer3
[14:29] <tedg> You're up too early!
[14:30] <rickspencer3> not for me :)
[14:30] <rickspencer3> I get up at 6am at the latest, otherwise, I miss everything :(
[14:31] <rickspencer3> tedg: I'm still confused about the power icon stuff. Are you guys going to do work to add more levels of power indication to support a larger icon set?
[14:31] <tedg> rickspencer3: I dont' think so.  I think Ken is just going to change where red begins.
[14:32] <tedg> I haven't talked with him about that directly, but that was my impression from the e-mail.
[14:32] <rickspencer3> so just change the icons for Jaunty, and add all the levels in Jaunty +1?
[14:32] <tedg> rickspencer3: Correct.
[14:32] <rickspencer3> ok
[14:33] <rickspencer3> it seems that upstream would be interested in having the extra levels, even if they don't use them all
[14:33] <tedg> rickspencer3: Uhm, yes and no.  They'd probably be interested if someone would draw the gnome icon theme icons for them :)
[15:11] <rickspencer3> tedg: okay, so I wasn't crazy
[15:11] <dobey> what levels/
[15:12] <tedg> rickspencer3: I wouldn't couple your sanity with the logic of upstreams, that'd be a bad choice :)
[15:12] <tedg> dobey: How many levels of icons are shown by GPM, it's currently 6.  There was some discussion on increasing it to 10 or more.
[15:12] <dobey> it's not like there's much difference you can show in changes for a 16x16 icon :)
[15:13] <tedg> dobey: The issues are more the < 20 icons -- where they could become "more red" even if the line isn't as descriptive.
[15:13] <dobey> tedg: yeah, ken mentioned before about possibly adding more levels to the spec, becuase the kde guys want to show more, though i don't really know if it makes sense
[15:15] <dobey> the gpm icons aren't in gnome-icon-theme anyway, they're shipped in g-p-m and installed in the private icon theme directory, as per the themeable app specific icons instructions on live.gnome.org :)
[15:15] <dobey> though i guess the human theme has them in the theme of course
[15:15] <tedg> Yes, but the theme can override them.
[15:16] <tedg> It's one of those cases where everyone is using those names but they aren't in the icon spec for some unknown reason :)
[15:16] <dobey> well, there are lots of reasons that the majority of those icons aren't in the spec
[15:17] <asac> mvo: on call still ;)?
[15:18] <dobey> one being of course that having lots of copies of the same icon with some small difference for every different type of device that has a battery, is awful
[15:18] <mvo> asac: no, reading now
[15:18] <asac> cool
[15:18] <dobey> and charging vs. discharging is also kind of annoying
[15:19] <tedg> dobey: Then put something better in the spec?  :)   High quality always wins.
[15:19] <dobey> and naming icons foo-000 is awful
[15:20] <dobey> and the people who want such icons in the spec, have generally been not helpful, unwilling to discuss the issues, and just demanding they go in as is
[15:20] <dobey> which of course, just means i will ignore them :)
[15:20] <dobey> i'm all for getting sane battery status icon names in the spec
[15:21] <dobey> but it seems nobody wants to actually discuss the issues and come to some sane compromise for it
[15:21] <mvo> asac: very nice!
[15:21] <asac> mvo: i removed the differentiation of free/non-free
[15:21] <asac> mvo: now its just: * covers important enduser use-case for which a ThirdParty Repository is essential.
[15:22] <mvo> ok
[15:22] <mvo> i noticed
[15:22] <mvo> I think that is good
[15:22] <asac> i tried to give rationals ... not really detailed, but allows us to remember its purpose
[15:22] <asac> i think i will add the same for the other requirement points like package
[15:23]  * mvo nods
[15:23] <asac> mvo: also i made ~ubuntu-archive optional in the process. the approval is done by third party review team
[15:24] <asac> not sure if that would have impact on -archive motivation ... but i guess not
[15:25] <asac> mvo: i think i will leave the bits about technical board out of it ... its the top most instant for everything in ubuntu anyway.
[15:25] <asac> unless you say it should be in there
[15:26] <asac> mvo: in a similar fashion i think its ok not to give explicit instructions how to complain about decisions of -tpr team
[15:26] <tedg> dobey: Propose a sane compromise to XDG and start from there.
[15:26] <asac> mvo: as the TB is always a fallback i guess
[15:26] <tedg> dobey: There is a lot of information there that people want to express in icons.  It'll be difficult without a whole cadre of icons.
[15:27] <dobey> tedg: i don't know what all information people want to express in icons there. i know what battery status is, and i know what the innumerable amount of devices i have, and/or have used, do to display such status to the user
[15:28] <dobey> most of them have 4-5 levels of status
[15:28] <mvo> asac: souns good, also I would like to see ubuntu-archive doing the review, they have some experience with this
[15:28] <dobey> though the 5th tends to be "dead, and the device doesn't turn on"
[15:29] <tedg> dobey: The reason you need that though is that sometimes devices lie about how dead they are :(
[15:30] <dobey> yes, i know
[15:31] <dobey> i have had plenty occasions where a laptop of mine had "30+ minutes" left, and then about 30 seconds later was all like "You have 30 seconds to plug in the AC."
[15:32] <asac> mvo: right. but i think that step shouldnt be mandatory. you probably can also sign off packages that are not too complex. especially if its "just" applications and not libs and so on.
[15:32] <dobey> but saying "you have N% battery left" is entirely useless
[15:33] <dobey> batteries are like snowflakes :)
[15:34] <mvo> asac: right
[15:36] <dobey> tedg: so really, all the battery level icons need to be based on time left, not percent left
[15:37] <tedg> dobey: That's not much better...  and as it gets lower the time gets less accurate, because the models have less information at those lower levels.
[15:37] <mclasen> do the icons say 30% ?
[15:37]  * mclasen doesn't follow
[15:38] <tedg> dobey: What we've discussed is combining them and doing, 100, 80, 60, 40, 20 percent and then 10 min, 5 min, 1 min.
[15:38] <tedg> mclasen: No, but they icons are chosen by percentage.
[15:38] <dobey> tedg: what do you do when 100% ~= 10 minutes for a user then?
[15:39] <Tm_T> interesting debate that was
[15:39] <tedg> dobey: We playing the "laughing hyena" sound.
[15:39] <dobey> lovely
[15:40] <tedg> dobey: With our deal with dell we offer them a deal on a new laptop with Ubuntu pre-installed :)
[15:41] <tedg> I think it's a rare enough case, and in that case the person probably wants the 10 min icon anyway.
[15:44] <dobey> right, but i also want icon names to have meaningful titles for names, not foo-[somenumbers]
[15:47] <pochu> my brother has a lenovo with winxp that has a green battery icon that has the % battery charge, I find it quite useful and easy to read
[15:48] <dobey> yeah, well i don't care if g-p-m has a text label that says NA%/0:05:00 or whatever, next to the icon
[15:56] <Keybuk> seb128: should do
[15:57] <dobey> i just want the icons to be meaningful
[15:57] <seb128> Keybuk: it's working now after reinstalling thanks
[15:57] <seb128> Keybuk: btw did you do a new bootchart yesterday after moving the .desktop to see if you still had the 5 seconds timeout?
[15:59] <Keybuk> I didn't
[15:59] <Keybuk> after putting jaunty back on, the machine won't boot
[15:59] <Keybuk> X hangs/crashes
[15:59] <Keybuk> or all I get is a black screen
[16:06] <rickspencer31> calc: ping
[16:43] <calc> rickspencer3: pong
[16:43] <calc> rickspencer3: whats up?
[16:46] <rickspencer3> calc: Did you have plans in Jaunty to improve the experience for OOo users who request a feature that depends on java when they don't have java installed?
[16:47] <calc> rickspencer3: lots of places that need java don't even give an error box at all, so even if i did hook into some spots it wouldn't catch all of them
[16:48] <calc> pretty much OOo needs Java in general
[16:48] <calc> we just rip it out because it won't fit on the cd
[16:48] <rickspencer3> I saw a proposal to change the errors to a box that provides a button to install java
[16:48] <rickspencer3> calc: ACK why java is not there
[16:49] <calc> yea, that is just one spot that happens to throw an error box
[16:49] <rickspencer3> did you have any thoughts on the button idea?
[16:49] <rickspencer3> are there lots of spots that  throw those?
[16:49] <calc> there are lots of spots that need java, but not all of them (not sure what percentage) even let you know why its not working
[16:50] <calc> i found one spot last night that doesn't work without java and it doesn't indicate any problem at all without java it just doesn't work
[16:50] <rickspencer3> so lots of failing silently
[16:50] <calc> yes
[16:50] <seb128> vuntz: do you also get a speaker beep when using shutdown or reboot in gnome-session?
[16:51] <calc> i'm going to add to the error message that it shows on the terminal to install the ooo-java-common as well
[16:51] <rickspencer3> how much effort would it be to adjust the existing error dialogs to add a button to install java?
[16:51] <calc> i think that part always shows, but only if they run it from the terminal
[16:51] <calc> i'm not sure, since OOo doesn't really use gtk/qt i would have to look into it
[16:52] <vuntz> seb128: not on my laptop, but I still have the 2.24 one
[16:52] <calc> they have their own gui framework just for OOo
[16:52] <seb128> vuntz: ok
[16:52] <rickspencer3> sweet
[16:52] <rickspencer3> I suppose it's a complicate cross platform affair that requires tons of tweaking on each platform anyway? :P
[16:53] <calc> i think the issue is more that OOo is ~ 20 years old
[16:53] <calc> so gtk/qt didn't exist and they don't want to put in the effort at this point to port it
[16:54] <calc> novell is doing some work on getting dialog layout to work despite using vcl (OOo's framework)
[16:54] <calc> but i think that is going to take a long time
[16:55]  * calc hopes he doesn't break suitesparse when he splits it up in a few min :)