15:10:52 <rdieter_laptop> #startmeeting kde-sig -- https://fedoraproject.org/wiki/SIGs/KDE/Meetings/2011-10-18
15:10:52 <zodbot> Meeting started Tue Oct 18 15:10:52 2011 UTC.  The chair is rdieter_laptop. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:10:52 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
15:10:59 <rdieter_laptop> #meetingname kde-sig
15:10:59 <zodbot> The meeting name has been set to 'kde-sig'
15:11:04 <rdieter_laptop> #topic roll call
15:11:08 <rdieter_laptop> hi, who's all here today?
15:11:09 <Kevin_Kofler> Present.
15:11:20 * rnovacek is here
15:11:48 * chuckf is here
15:11:48 <rdieter_laptop> than : ping
15:11:51 * than is here
15:12:12 <rdieter_laptop> #info rdieter Kevin_Kofler rnovacek chuckf than present
15:12:18 <rdieter_laptop> chuckf: hi
15:12:30 <rdieter_laptop> #topic agenda
15:12:48 <rdieter_laptop> no meeting page/agenda prepped ahead-of-time, so what to discuss today?
15:12:58 <Kevin_Kofler> F16 final polishing (change deadline coming Monday).
15:13:16 <rdieter_laptop> apper testing
15:13:34 <chuckf> how is F16 KDE coming along at this point?
15:14:15 <than> rdieter_laptop: security state
15:14:27 <than> rdieter_laptop: on agenda
15:14:32 <rdieter_laptop> than: in particular?  kdeutils?  other?
15:15:16 <than> rdieter_laptop: CVE-2011-3365 in kdelibs/kdelibs3 and CVE-2011-2725 kdeutils
15:15:28 <Kevin_Kofler> Seeing how Apper is faring so far, I'm not sure it's a good idea to rush it into the F16 release…
15:15:41 <Kevin_Kofler> And the thing is, I'm not sure whether we can pull it off as a post-release update either…
15:16:05 <rdieter_laptop> ok, got those topics tabulated, let's move on
15:16:11 <Kevin_Kofler> Though it's not really a bigger change than the kde-plasma-networkmanagement nm-compat → nm09 switch we did in F15.
15:16:17 <rdieter_laptop> #topic f16 status/polish
15:16:32 <Kevin_Kofler> So, what exactly is needed here?
15:16:47 <Kevin_Kofler> I see the kde-settings and kdepim/kdepim-runtime updates are now queued for stable.
15:16:51 <rdieter_laptop> in general I think we're in really good shape, no serious issues or blockers remain, that I'm aware of.
15:17:11 <Kevin_Kofler> A live-kickstarts change is needed to go back to not starting Nepomuk on the live image.
15:17:23 <rdieter_laptop> Kevin_Kofler: already done
15:17:28 <Kevin_Kofler> OK
15:17:38 <Kevin_Kofler> I'd also like the akonadi change to go back to MySQL as default to go into the release.
15:18:01 <Kevin_Kofler> (because Akonadi won't switch the backend by default if its default changes after installation)
15:18:05 <than> Kevin_Kofler: why not sqlite?
15:18:15 <rdieter_laptop> https://admin.fedoraproject.org/updates/FEDORA-2011-14281  currently in -testing
15:18:33 <Kevin_Kofler> than: Slower and does not support concurrent access, so everything blocks while using Akonadi with SQLite.
15:19:27 <Kevin_Kofler> I built an update for kde-plasma-networkmanagement to 0.9 beta1 (= 0.8.80), it now has karma +1 (from a proventester, even) and I queued it to stable.
15:19:32 <rdieter_laptop> now that we have good backend switching (per user in kdebase-runtime kcm and globally via site-specific /etc/xdg/akonadi/akonadiserverrc), my objections to switching back to sqlite are withdrawn
15:19:34 <than> Kevin_Kofler: i though it's fixed long time ago
15:20:30 <rdieter_laptop> now, when/if we do all ack going back to mysql, I think I would still like to use sqlite on the live image (requiring a minor kickstart modification)
15:20:37 <Kevin_Kofler> than: AFAIK, they said that SQLite is simply not thread- or concurrency-safe.
15:21:06 <Kevin_Kofler> rdieter_laptop: But we aren't going to start up Akonadi by default anyway, are we?
15:21:16 <rnovacek> rdieter_laptop: +1, mysql on livecd is overkill
15:21:17 <rdieter_laptop> so yeah, while sqlite (still) has a lower runtime footprint, it's definitely slower
15:21:41 <rdieter_laptop> Kevin_Kofler: sure, but when/if any akonadi-using app is launched (like konversation)...
15:21:57 <Kevin_Kofler> Konversation triggers firing up Akonadi?!
15:22:20 * Kevin_Kofler starts wondering why we bothered doing all this work to not start it by default if everything requires it anyway.
15:22:23 <rdieter_laptop> it hooks into kaddressbook to lookup irc nicks with adressbook entries
15:22:42 <rdieter_laptop> I just didn't want akonadi starting on initial login
15:23:01 <Kevin_Kofler> (IMHO, starting Akonadi and Nepomuk by default is the right thing to do, even on the live image. It's also why I enabled it in spin-kickstarts without a second thought. But then you complained…)
15:23:15 <rdieter_laptop> have you actually tried using the livecd?
15:23:28 <Kevin_Kofler> Not starting Akonadi on login just makes any app using it take longer to start up. :-/
15:23:33 <rdieter_laptop> with both enabled (esp with akonadi/mysql), it was very painfully slow
15:23:41 <rdieter_laptop> now, it's quite fast
15:23:55 <rdieter_laptop> Kevin_Kofler: true, it's a perception thing
15:24:28 <rdieter_laptop> we've discussed it before, ideally, if we could come up with some way of pre-initializing these db's, that would help a bunch too
15:24:43 <rdieter_laptop> but that never materialized
15:25:20 <Kevin_Kofler> The akonadi update https://admin.fedoraproject.org/updates/FEDORA-2011-14281 needs karma in any case.
15:25:44 <Kevin_Kofler> We will not be able to push it before the change deadline if it doesn't get at least 1 positive karma point. :-(
15:26:05 <rdieter_laptop> so, now that we have fairly robust runtime/site switching of akonadi backend, I personally don't have any strong preference either way about default (my site/machines will use sqlite, regardless)
15:26:26 <rdieter_laptop> Kevin_Kofler obviously favors mysql, as does upstream.
15:26:32 <rdieter_laptop> other opinions?
15:27:07 <rdieter_laptop> I guess I'd have to say I still have a slight preference for sqlite, it's smaller and safer for most use-cases.
15:27:37 <rnovacek> I'm for sqlite for livecd and mysql otherwise
15:28:00 <rdieter_laptop> than: ?
15:28:46 <than> +1 for mysql as upstream does
15:29:04 <rdieter_laptop> ok, let's just karma++ what we have, and be done with it.
15:29:27 * rdieter_laptop will make the spin-kickstars mod to use sqlite on kde-live
15:30:03 <Kevin_Kofler> Karma'd by executive decision and queued for stable.
15:31:00 <rdieter_laptop> anything else f16 status/polish -wise?
15:31:11 <Kevin_Kofler> Another update I'd really like to see in F16 Final is the Qt one: https://admin.fedoraproject.org/updates/FEDORA-2011-14348
15:31:21 <Kevin_Kofler> * update from a snapshot to an official RC
15:31:33 <Kevin_Kofler> * fix for building stuff against qlist.h
15:31:36 <rdieter_laptop> oh, yeah, meant to mention that one too.
15:31:40 <Kevin_Kofler> * fix for running Qt under GNOME
15:31:41 <rdieter_laptop> lot's of good stuff there.
15:31:57 <rdieter_laptop> test/karma please
15:32:11 <Kevin_Kofler> * fix for leaked FDs when printing (memory/resource leak, plus can trigger SELinux denials if the thing then execs another process).
15:32:24 <than> Kevin_Kofler: +1
15:32:45 <than> i will test it and  give karma+
15:33:02 <Kevin_Kofler> It has 1 karma, the threshold is set to 2, so it should get one more. (We could lower the threshold to 1, but I'd rather not do that.)
15:34:11 <rdieter_laptop> ok, let's move on
15:34:16 <Kevin_Kofler> What may or may not be fixed is the crash in KWrite settings.
15:34:21 <rdieter_laptop> #topic security stuff
15:34:26 <Kevin_Kofler> (It is reported no longer reproducible.)
15:34:28 <rdieter_laptop> than: ?
15:34:57 <than> CVE-2011-3365, input validation failure in KSSL
15:35:24 <than> it's fixed in kdelibs3 for f14/f15/f16/rawhide
15:35:53 <rdieter_laptop> woo
15:36:07 <than> rdieter_laptop: is the issue fixed in kdelibs for f16/f15/f14?
15:36:20 <than> rdieter_laptop:  i saw it in the changelog
15:36:30 <rdieter_laptop> than: f16 yes, f15/f14 still in updates-testing
15:36:37 <Kevin_Kofler> Fixed upstream for F16/Rawhide, fix backported for F14/F15.
15:36:42 <than> great
15:37:02 * rdieter_laptop saw the kdeutils/ark patches show up in bz today too
15:37:12 <than> i'm working on CVE-2011-2725
15:37:14 <Kevin_Kofler> And I backported it for kdelibs3, thanks to than for fixing a syntax error (copied and pasted multiple times) and building it.
15:37:19 <than> will finish it today
15:37:37 <rdieter_laptop> arg, glibc-2.14.90-12.999 got karma'd to stable updates. :(
15:37:43 <rdieter_laptop> means qt will FTBFS now.
15:37:47 <than> Kevin_Kofler: thanks for you patch!
15:37:54 <than> s/you/your
15:38:22 <rdieter_laptop> I guess I'll tag an older glibc into f16-override if I have to
15:38:36 <Kevin_Kofler> rdieter_laptop: Wait, it's still pending…
15:38:46 <Kevin_Kofler> If we get enough -1 votes, we might overturn it.
15:39:10 <rdieter_laptop> ok, didn't know that could happen
15:39:37 * rdieter_laptop is ok with f16-override tagging, the stuff that glibc fixes is nice too
15:39:42 <Kevin_Kofler> Well, it has +5, I think we need to get it to -3 to get anything to happen, if that works at all. :-(
15:39:59 <Kevin_Kofler> (It had +6, I voted it down.)
15:40:10 <rdieter_laptop> moving on...
15:40:12 <rdieter_laptop> #topic apper testing
15:40:31 <rdieter_laptop> https://admin.fedoraproject.org/updates/FEDORA-2011-14262
15:40:40 <rdieter_laptop> got apper in rawhide and f16-updates-testing.
15:40:56 <Kevin_Kofler> (It's annoying that testers vote up updates with already reported and -1'd regressions.)
15:41:02 <rdieter_laptop> it works fairly ok, but some quirks remain, I'll file some bugs today (for tracking purposes)
15:41:27 <rdieter_laptop> installing standalone rpms sometimes crashes apper-sentinel
15:41:34 <Kevin_Kofler> So what do we do with Apper in F16?
15:41:56 <Kevin_Kofler> I think rushing it into the release is going to be a bad idea if it has issues remaining, but how nice is pushing it as a post-release update?
15:41:57 <rnovacek> I tried to do some updates with apper and it works ok
15:42:12 <rdieter_laptop> and today (once) apper notified me of availble updates, but clicking on it's systray icon did nothing (I saw something flash in taskmanager, but it was very fast)
15:42:19 <Kevin_Kofler> It's not just a name change, they also switched to a new version of the PackageKit Qt bindings and made some UI changes.
15:42:49 <rdieter_laptop> so, the basic functionality generally works, I'm ok with working through the growing pains
15:42:53 <Kevin_Kofler> But then again, as I said in the introduction, we did something similar to kde-plasma-networkmanagement in F15.
15:43:02 <rdieter_laptop> dantti was going to test it today and get back to me
15:43:19 <Kevin_Kofler> (and here too, we might get pressured, I don't know for how long rhughes plans to supports the old PackageKit-Qt 1 bindings)
15:43:43 <Kevin_Kofler> (They're already gone in upstream PackageKit 0.7.)
15:43:51 <Kevin_Kofler> (but F16 is still on 0.6, I think)
15:44:07 <rdieter_laptop> dantti isn't going to support the older stuff much anymore, we may as well push forward
15:44:25 <Kevin_Kofler> KPK 0.6.x has been out of support for a while already, really.
15:44:35 <Kevin_Kofler> What was missing was just a stable release of Apper, and that's there now.
15:44:50 <rdieter_laptop> so, test test test.  have I mentioned test?
15:45:00 <Kevin_Kofler> The problem is that we didn't manage to get it tested early, so now F16 is coming and there are issues. :-(
15:45:02 <rdieter_laptop> the mustard indicates progress!
15:45:48 <rdieter_laptop> yeah, dantti was a little late releasing, and our reviewing it took a little longer than expected...
15:46:25 <rdieter_laptop> I'd still propose sticking with it, at least until we find anything blocker-worthy
15:46:51 <rnovacek> sorry about the delay with the review, I'm quite busy nowadays
15:47:15 <rdieter_laptop> no worries, everyone's busy.
15:47:25 <Kevin_Kofler> The reviewing wasn't the real delay, getting it packaged and submitted for review in the first place was. (dannti had asked us at least twice to package up snapshots.)
15:47:37 <than> why can't we push it later in f16 update (when it's well tested)?
15:47:52 <rdieter_laptop> Kevin_Kofler: really?  I asked him about a snapshot, and I thought he said to wait...
15:47:57 <rdieter_laptop> oh well, doesn't matter now.
15:47:59 <Kevin_Kofler> But dannti was too busy to do a release earlier and we were too busy to get this though review earlier.
15:48:27 <rdieter_laptop> than: we could I suppose...
15:48:51 <rdieter_laptop> I'd rather not though, if we can avoid it.  it's much easier to make the switch now than post-release.
15:49:32 * rdieter_laptop wonders, is the kpk -> apper switch release notes worthy?
15:49:45 <Kevin_Kofler> I'm worried about shipping the release with a broken updater…
15:49:47 <than> rdieter_laptop: yes, but it's little riskant
15:50:07 <rdieter_laptop> Kevin_Kofler: updating works
15:51:16 <rdieter_laptop> I'd rather base the decision on the results of testing.  can we at least try it?
15:51:26 <Kevin_Kofler> Are we sure it works for everyone?
15:51:35 <Kevin_Kofler> Things like:
15:51:36 <Kevin_Kofler> <rdieter_laptop> and today (once) apper notified me of availble updates, but clicking on it's systray icon did nothing (I saw something flash in taskmanager, but it was very fast)
15:51:45 <Kevin_Kofler> don't sound encouraging to me at all.
15:51:55 <rdieter_laptop> Kevin_Kofler: easy enough to work around that (e.g. run apper by hand did work)
15:52:01 <Kevin_Kofler> We don't want our users to not be able to update to a working version.
15:52:18 <than> Kevin_Kofler: +1
15:52:26 <rdieter_laptop> when is the change deadline ?
15:52:26 <Kevin_Kofler> TBH, I think Apper is F17 material.
15:52:32 <Kevin_Kofler> Monday.
15:53:25 <rdieter_laptop> ok, if we don't have confidence by friday, let's revert back to kpk for f16 ga
15:54:31 <rdieter_laptop> (what I propose)
15:54:54 <rdieter_laptop> anyone else able to help test besides me?
15:56:04 <than> rdieter_laptop: i will test it
15:56:50 <rdieter_laptop> (running short on meeting time), is that plan agreeable?
15:57:14 <rnovacek> yes
15:57:18 <than> rnovacek: +1
15:57:37 <Kevin_Kofler> +1
15:57:53 <rdieter_laptop> ok, let's give it shot.  thanks everyone.
15:57:56 <rdieter_laptop> #endmeeting