SIGs/Desktop/Meeting-20071010

From FedoraProject

< SIGs | Desktop
Revision as of 16:32, 24 May 2008 by Admin (Talk | contribs)

(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to: navigation, search

Contents

Desktop SIG Meeting of 2007-10-10

Present

  • MatthiasClasen (mclasen)
  • John Nettleton (jnettlet)
  • JesseKeating (f13)
  • JeremyKatz (jeremy)
  • John Poelstra ('poelcat')
  • RayStrode (halfline)
  • DavidZeuthen (davidz)
  • WarrenTogami (warren)
  • Will Woods (wwoods)

Notes

  • Blocker bugs are being fixed

Log


Oct 10 14:00:59 *       mclasen_ has changed the topic to: Desktop SIG meeting
Oct 10 14:01:29 *       jeremy is lurking around
Oct 10 14:01:33 <mclasen_>      ok, so lets start, I guess
Oct 10 14:02:12 <mclasen_>      I posted two ideas for things to discuss
Oct 10 14:02:45 <mclasen_>      unless someone has better agenda items, I'd say we should start with those...
Oct 10 14:03:48 <jnettlet>      sounds good
Oct 10 14:03:58 <drago01>       mclasen_: so go ahead ;)
Oct 10 14:04:04 <mclasen_>      the first one is f8 blocker bugs
Oct 10 14:04:34 <mclasen_>      I wonder if we have all the blocker-worthy desktop-affecting bugs on F8Blocker ?
Oct 10 14:05:25 <jeremy>        I've been trying to make sure things I see get put there, but that's hardly a perfect approach
Oct 10 14:05:30 <halfline>      i noticed you did a brief run through earlier
Oct 10 14:05:31 <f13>   I'm here.
Oct 10 14:05:37 Oct 10 14:05:37 <halfline>      adding things to blocker list
Oct 10 14:05:52 <halfline>      mclasen_: ^^
Oct 10 14:05:54 <mclasen_>      halfline: i've been watching the bug lists and new incoming bugs for the last weeks
Oct 10 14:06:11 <mclasen_>      here is the current desktop blocker list: https://bugzilla.redhat.com/buglist.cgi?bug_id=298741,303801,326461,324301,325231,251996,318911,313811,312531,296671,302961,221974,296821,305441
Oct 10 14:06:11 <drago01>       mclasen_: some nm regressions
Oct 10 14:07:04 <mclasen_>      drago01: yeah, I noticed that nm is suspiciously absent from that list...
Oct 10 14:07:42 <mclasen_>      but I also know that dan has been pretty much working off his own list of must-fix issues, and ignoring bugzilla
Oct 10 14:07:51 <jnettlet>      did the nm-openvpn subversion build get pushed last night?
Oct 10 14:08:12 <drago01>       well hidden ssid support should be in current builds
Oct 10 14:08:33 <drago01>       but still no wpa/wpa2 enterprise/wep dynamic and leap
Oct 10 14:08:40 <mclasen_>      jnettlet: I can't see an openvpn build in koji right now
Oct 10 14:09:12 <warren>        ooh
Oct 10 14:09:22 <drago01>       we have a bug for it: https://bugzilla.redhat.com/show_bug.cgi?id=323371
Oct 10 14:09:32 <drago01>       its on the blocker list
Oct 10 14:09:39 <drago01>       but not on the desktop one
Oct 10 14:09:44 <mclasen_>      yeah
Oct 10 14:09:57 <mclasen_>      my tool is deficient
Oct 10 14:10:18 <mclasen_>      it doesn't know that dcbw is one of us...
Oct 10 14:10:29 <wwoods>        Technically, NM bugs aren't automatic release blockers
Oct 10 14:10:59 <drago01>       wwoods: not bugs but major regressions in one of the most used apps ;)
Oct 10 14:11:42 <wwoods>        drago01: call them whatever you want, it's still a non-default app that doesn't prevent installation or use of the system
Oct 10 14:12:01 <poelcat>       i'm seeing problems running system-config-{firewall,securitylevel} on x86_64 ... they never start after I provide root passwd and consume 100% cpu
Oct 10 14:12:08 <davidz>        wwoods: NM is installed by default on the live cd....
Oct 10 14:12:11 <poelcat>       anyone else seeing that?
Oct 10 14:12:14 <mclasen_>      wwoods: how is default defined in that sentence ?
Oct 10 14:12:33 <wwoods>        I'm not saying it's cool to have a busted NM, but just.. y'know. fair warning that other issues may take precedence.
Oct 10 14:12:40 <davidz>        wwoods: I disagree
Oct 10 14:12:44 <wwoods>        mclasen_: not enabled by default
Oct 10 14:12:49 <davidz>        wwoods: it is!
Oct 10 14:12:53 <davidz>        wwoods: install the live cd
Oct 10 14:12:56 <mclasen_>      wwoods: on which spins ?
Oct 10 14:13:18 <wwoods>        in the typical installation cases (upgrade, install from media, install from network)
Oct 10 14:13:45 <davidz>        so - there was some talk that we should point Fedora users who wants a desktop to use the Fedora Desktop Live CD
Oct 10 14:13:47 <jeremy>        wwoods: based on numbers from the torrents, that's only about half of our users these days
Oct 10 14:13:52 <mclasen_>      poelcat: there have been a number of bugs where people have been seeing issues with X apps that seemed to be x86_64 specific; might be worthwhile to try sudo gdb /usr/sbin/system-config-foo --sync and see if the run into an X error
Oct 10 14:13:56 <davidz>        and I'd argue this would be the "typical" use case
Oct 10 14:14:03 <jeremy>        wwoods: and in fact, with test3, we have had _more_ live downloads than non-live (via torrent)
Oct 10 14:14:21 <wwoods>        davidz: I'd agree that it should be the typical case, and I want to see NM rock-solid (especially as it's a Feature for F8)
Oct 10 14:14:23 <poelcat>       mclasen_: you got it. thanks for the tip... strace wasn't telling me anything :)
Oct 10 14:15:07 <davidz>        wwoods: sure; so you agree that NM is one of those things that are typically installed by default? And if it is, the logical conclusion is that we need to fix it if it's broken..
Oct 10 14:15:09 *       davidz thinks anyway
Oct 10 14:15:14 <wwoods>        this is just a gentle reminder that, if push comes to shove and the only remaining blocker bugs are things like "WPA2 doesn't quite work right".. I'd be very hesitant to slip the release over that
Oct 10 14:15:33 <wwoods>        not that it's my decision, but that'd be my opinion
Oct 10 14:15:42 <mclasen_>      poelcat: s-c-firewall starts up fine for me on i386, fwiw
Oct 10 14:15:48 <f13>   that's the problem with trying to hard define this stuff
Oct 10 14:16:04 <f13>   what is a "blocker" and what isn't changes as the release date gets closer (or farther behind)
Oct 10 14:16:35 <f13>   and there is a difference between "a fix we'll take in" and "A fix we won't ship without"
Oct 10 14:16:54 <f13>   currently we try to call both of those "blockers" as in "only blocker fixes accepted" but that's not fair.
Oct 10 14:16:59 <mclasen_>      but at this point there is no individual build-examination by rel-eng, or is there ?
Oct 10 14:17:10 <f13>   mclasen_: right now?  No.
Oct 10 14:17:12 <mclasen_>      ie stuff we fix right now ends up in rawhide automatically
Oct 10 14:17:27 <wwoods>        yeah, in theory we should be keeping better separation between Blocker and Target
Oct 10 14:17:31 <f13>   mclasen_: once we enter the final freeze on the 16th there will be.
Oct 10 14:17:57 <mclasen_>      so we have one week left
Oct 10 14:18:18 <wwoods>        and, seriously, refuse to ship until Blocker is empty. but that's.. far too rigid
Oct 10 14:18:22 <f13>   of unrestricted build, then some more time of restricted builds.
Oct 10 14:18:37 <mclasen_>      wwoods: well, there is 2 ways to empty blocker...
Oct 10 14:18:48 <drago01>       wwoods: so you consider stuff that worked in fc5,fc6 and f7 that stopped working completly not a blocker?
Oct 10 14:19:03 <f13>   wwoods: you can be far more strict from the get go about what is a "blocker" vs a "fix we'd take during a freeze".
Oct 10 14:19:26 <halfline>      mclasen_: 1) fix it 2) klugden it until it's lower severity that it no longer is a blocker?
Oct 10 14:19:27 <f13>   drago01: that's too general of a statement.
Oct 10 14:19:29 <davidz>        wwoods: drago01 got a good point; NM was default in F7 live spins too
Oct 10 14:19:34 <wwoods>        drago01: not everything that someone calls a regression is a stop-ship issue
Oct 10 14:19:42 <mclasen_>      halfline: I had 2) declare it a non-blocking issue
Oct 10 14:19:58 <mclasen_>      anyway, lets not get hung up on blocker criteria
Oct 10 14:20:01 <davidz>        wwoods: that's of course right - but quite a few folks are using e.g. WPA2
Oct 10 14:20:06 <davidz>        right
Oct 10 14:20:09 <f13>   drago01: generally a horrible regression in a lesser used app isn't a stop-ship item.  NM however is a marquee feature and should be treated differently.
Oct 10 14:20:16 <f13>   drago01: not ever package is equal.
Oct 10 14:20:20 <mclasen_>      looking at the current list of blocker bugs, I think we haven't fixed the gnome-keyring problems yet
Oct 10 14:20:25 <mclasen_>      despite alex' efforts
Oct 10 14:20:28 <jnettlet>      does anyone else run EXA?  I have seen some regressions there, but haven't had a chance to debug them
Oct 10 14:20:41 <drago01>       f13: but here its the case... I would prefer "fix it or revert to 0.6.5 and ship a update later"
Oct 10 14:20:47 <bpepple|lt>    tibbs: ping.
Oct 10 14:20:52 <drago01>       than shiping with suchs regressions
Oct 10 14:21:00 <tibbs> bpepple|lt: Hmm?
Oct 10 14:21:05 <bpepple|lt>    d'oh! wrong channel. sorry.
Oct 10 14:21:21 <davidz>        drago01: (Yeah. I would love just to have vpnc with NM working)
Oct 10 14:21:28 <f13>   drago01: that's still somewhat the plan, but in the mean time none of the existing 0.6.5 bugs have been fixed in 0.6 code.
Oct 10 14:21:42 <f13>   davidz: uh, that's working fine for me.
Oct 10 14:21:48 <davidz>        as mclasen said we can deal with the blocker bugs individually
Oct 10 14:21:56 <davidz>        f13: ok, I'll try again (tried 4 days ago)
Oct 10 14:22:12 <davidz>        and we can fight over what's a blocker and what isn't on a per-bug basis
Oct 10 14:22:27 <mclasen_>      drago01: I just asked dcbw, and he is confident that we can have most of the regressions fixed early next week
Oct 10 14:22:43 <f13>   davidz: lets talk about your vpnc stuff in another chan, nothing has changed that should make that better/worse in the last 4 days.
Oct 10 14:22:54 <davidz>        f13: ok; we'll take it offline
Oct 10 14:22:58 <drago01>       mclasen_: ok this are good news... should make this a non issue ;)
Oct 10 14:23:19 <davidz>        also dcbw is working very hard on this; so we should cut him some slack
Oct 10 14:24:35 <f13>   indeed
Oct 10 14:24:40 <f13>   this has been the known plan for a while.
Oct 10 14:24:51 <drago01>       ok what about "the gdm restarts with different theme after update on shutdown bug" (not a blocker but dunno if its filled at all has been discussed at the ml)
Oct 10 14:24:54 <f13>   core functionality by test2/3, the rest before final.
Oct 10 14:25:04 <halfline>      drago01: i'm pretty sure it's been filed
Oct 10 14:25:09 <halfline>      i remember seeing it flying by at one point
Oct 10 14:25:13 <mclasen_>      drago01: I believe that may be a side effect of some recent rhgb/initscripts issues ?
Oct 10 14:25:17 <halfline>      i haven't investigated it though
Oct 10 14:25:35 <halfline>      mclasen_: I don't *think* it would be
Oct 10 14:26:02 <mclasen_>      or maybe it is another mysterious greeter crash ?
Oct 10 14:26:10 <f13>   was the gdm not starting correctly bug on the blocker list?
Oct 10 14:26:13 <halfline>      could be
Oct 10 14:26:15 <f13>   where you have to vt switch for it?
Oct 10 14:26:33 <mclasen_>      f13: there is https://bugzilla.redhat.com/show_bug.cgi?id=318911
Oct 10 14:26:35 <drago01>       I only saw it once ... just thought "wtf is goind on" ... only got my attention after I saw that other people hd seen it too
Oct 10 14:26:37 <drago01>       had
Oct 10 14:26:52 <mclasen_>      f13: but we have totally failed to reproduce that so far
Oct 10 14:27:16 <f13>   mclasen_: I can do it every time on my laptop
Oct 10 14:27:30 <halfline>      f13: there is a "you have to vt switch" bug that we thought we fixed a while ago, but turned out to not be completely fixed until yesterday
Oct 10 14:27:43 <mclasen_>      f13: what is 'it' here ?
Oct 10 14:27:50 <f13>   halfline: oh yesterday.  Ok, so that should hit rawhide tomorrow.
Oct 10 14:28:05 <halfline>      f13: i don't know if it's the bug you're seeing
Oct 10 14:28:07 <f13>   mclasen_: X for gdm wouldn't start up until I manually vtswitched.
Oct 10 14:28:13 <halfline>      the bug that was fixed yesterday was a very rare race
Oct 10 14:28:32 <mclasen_>      f13: only after boot, or after every logout ?
Oct 10 14:28:56 <f13>   mclasen_: I only noticed it on boot, I haven't done (m)any logouts.
Oct 10 14:29:12 <halfline>      well, try upgrading rhgb and see if it fixes things for you
Oct 10 14:29:53 <f13>   I will.
Oct 10 14:30:36 <mclasen_>      halfline: your split artwork will also hit rawhide tomorrow, right ?
Oct 10 14:31:34 <halfline>      tomorrow yea
Oct 10 14:31:42 <halfline>      there's one package left to build
Oct 10 14:31:51 <halfline>      and then i have to do a fedora-logos fix too
Oct 10 14:32:55 <halfline>      not entirely which package is the right one to put Obsoletes: redhat-artwork in though
Oct 10 14:33:21 <halfline>      the alternative is to leave redhat-artwork a stub package that just Requires all the others
Oct 10 14:34:16 <mclasen_>      or just leave it an empty corpse; but then we still have the problem to get rid of it for F9
Oct 10 14:38:23 <mclasen_>      I guess we are done talking about F8 blocker bugs...
Oct 10 14:38:58 <halfline>      i'm adding the Obsoletes in fedora-gnome-theme for now
Oct 10 14:39:12 <halfline>      it might make sense to add it to bluecurve-icon-theme since a lot of redhat-artwork is for the icons
Oct 10 14:39:19 <mclasen_>      is that going to hurt the kde spin ?
Oct 10 14:39:35 <halfline>      i'm not sure
Oct 10 14:39:41 <mclasen_>      no, probably not
Oct 10 14:40:04 <mclasen_>      it means that people who upgrade a kde-only installation will get fedora-gnome-theme
Oct 10 14:40:15 <mclasen_>      but considering they had even more gnome stuff before in redhat-artwork...
Oct 10 14:40:45 <drago01>       talking about icons what happened to bluecurve based icons that are used in the fedora icon theme?
Oct 10 14:40:53 <drago01>       they look "blury" in f8
Oct 10 14:42:39 <halfline>      drago01: in which setting?
Oct 10 14:42:42 <halfline>      drago01: panel?
Oct 10 14:42:47 <mclasen_>      the inheritance of the default icon theme is Fedora -> Mist -> gnome -> hicolor
Oct 10 14:42:48 <drago01>       yes
Oct 10 14:42:57 <halfline>      mclasen_ has a hack to fix that
Oct 10 14:43:02 <halfline>      but is hesitant to apply it
Oct 10 14:43:13 <mclasen_>      so the only bluecurve based icons should be ones that get installed into hicolor
Oct 10 14:43:15 <drago01>       the hack is?
Oct 10 14:43:26 <halfline>      ask for 23 pixel icons in the panel
Oct 10 14:43:32 <halfline>      since some of our icons are 22 and some are 24
Oct 10 14:43:41 <halfline>      it turns out they all look sharp then
Oct 10 14:43:43 <mclasen_>      there is some braindeadness going on
Oct 10 14:43:51 <mclasen_>      where half the icon themes use the meaningful 24
Oct 10 14:44:01 <mclasen_>      and the other have use the less meaningful 22
Oct 10 14:45:56 <drago01>       so better be apply the hack and remove it after fixing the icons?
Oct 10 14:48:51 <mclasen_>      there is no real fix, I fear
Oct 10 14:53:40 <mclasen_>      so, should we talk for a bit about what the plans are for post-f8 rawhide ?
Oct 10 14:53:45 <mclasen_>      if anybody is interested in that...
Oct 10 14:57:08 <mclasen_>      ok, that seems not the case
Oct 10 14:57:22 <mclasen_>      lets just close for today then
Oct 10 14:57:28 <mclasen_>      and get back to blocker bugs...
Oct 10 14:57:53 *       mclasen_ has changed the topic to: Channel is used by various Fedora groups and committees for their regular meetings | Note that meetings often get logged | For questions about using Fedora please ask in #fedora | See http://fedoraproject.org/wiki/Communicate/FedoraMeetingChannel for meeting schedule