15:02:54 <rdieter> #startmeeting kde-sig
15:02:54 <zodbot> Meeting started Tue May 26 15:02:54 2015 UTC.  The chair is rdieter. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:02:54 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
15:03:00 <rdieter> #meetingname kde-sig
15:03:00 <zodbot> The meeting name has been set to 'kde-sig'
15:03:04 <rdieter> #topic roll call
15:03:06 * heliocastro here
15:03:12 <rdieter> hi all, friendly kde-sig meeting, who's present today?
15:03:18 <jgrulich> hi
15:05:02 <rdieter> jgrulich: may be a short meeting :)
15:05:14 <rdieter> #info heliocastro jgrulich rdieter present
15:05:18 <rdieter> #chair heliocastro jgrulich
15:05:18 <zodbot> Current chairs: heliocastro jgrulich rdieter
15:05:21 <Kevin_Kofler> Present.
15:05:26 <rdieter> #info Kevin_Kofler present
15:05:28 <rdieter> #chair Kevin_Kofler
15:05:28 <zodbot> Current chairs: Kevin_Kofler heliocastro jgrulich rdieter
15:06:39 <heliocastro> No topics opened ?
15:06:42 <rdieter> #topic agenda
15:06:55 <rdieter> good question, any topics for discussion today?
15:07:06 * pino|work here
15:07:07 <rdieter> I can give a brief update on kde-apps-15.04.1 progress
15:07:12 <rdieter> #info pino|work present
15:07:15 <rdieter> #chair pino|work
15:07:15 <zodbot> Current chairs: Kevin_Kofler heliocastro jgrulich pino|work rdieter
15:08:00 * tosky late
15:08:44 <rdieter> #info tosky present
15:08:45 <rdieter> #chair tosky
15:08:46 <zodbot> Current chairs: Kevin_Kofler heliocastro jgrulich pino|work rdieter tosky
15:09:02 <rdieter> #topic kde-15.04.1 status
15:09:34 <rdieter> brief status report, I was holding off on doing the reset of kde-apps-15.04.1 waiting for f22 release and freeze to drop
15:09:54 <rdieter> already did the core kde4 bits, and submitted updates for those (kdelibs, kdepim, kde-workspace)
15:10:01 <rdieter> the rest I'll try to do today
15:10:31 <rdieter> that's all, any questions or comments?
15:11:22 <Kevin_Kofler> Good.
15:11:56 <pino|work> rdieter: +1
15:12:08 <rdieter> ok, moving on
15:12:12 <rdieter> #topic open discussion
15:12:16 <rdieter> anything else for today?
15:12:30 <rdieter> we can celebrate a little for f22 release today
15:12:35 <rdieter> \o/  yay
15:12:41 <Kevin_Kofler> Great!
15:12:57 <Kevin_Kofler> Hopefully the flood of complaints will be smaller than for F9…
15:13:20 <rdieter> I suspect we'll get some amount of "OMG where did my legacy systray icons go?"
15:13:27 <Kevin_Kofler> We did receive a bunch of complaints about Plasma 5 regressions, or Plasma 5 in general.
15:13:36 <Kevin_Kofler> The systray one is a big one.
15:13:37 <heliocastro> Can we get drunk ?
15:13:54 <rdieter> <nod>, anyone with expectations that plasma5 have 100% parity with kde4 will be disappointed too
15:13:59 <Kevin_Kofler> Sadly, no short-term solution in sight, they want all the systray applets to be ported.
15:14:03 <rdieter> heliocastro: +1 , maybe a little :)
15:15:37 <rdieter> Kevin_Kofler mentioned earlier in #fedora-kde about considering using calamares for f23, looking forward to that ml discussion
15:17:24 <rdieter> if there's nothing else to discuss today, let's wrap up and get back to work
15:17:41 <jgrulich> is it stable enough to use it instead of Anaconda?
15:17:46 <Kevin_Kofler> So why should that be on the ML rather than IRC?
15:18:08 <rdieter> ml has wider audience
15:18:39 <Kevin_Kofler> jgrulich: This is Fedora, we always replace working software with incomplete unstable alphas. ^^
15:19:19 <Kevin_Kofler> More seriously, Calamares works fine in the scenarios I tested, but it doesn't support all the weird corner cases Anaconda does.
15:19:33 <jgrulich> yeah, but installer is not just a simple app and that it works for few people doesn't mean it would work for everyone, this is something what needs to be heavily tested
15:19:46 <Kevin_Kofler> One thing it did choke on last I checked was installing over an Anaconda LVM-based installation.
15:20:25 <Kevin_Kofler> There's some logic that unmounts everything that does not look for LVM, I guess that was the issue there.
15:20:42 <Kevin_Kofler> Anyway, I think such issues can be fixed, if we commit to Calamares.
15:20:55 <rdieter> one evil baby step to consider... add calamares to f23 kde live (in addition to anaconda)
15:21:06 <heliocastro> +1
15:21:09 <rdieter> either as official, or unofficial/secondary installer
15:21:24 <Kevin_Kofler> I think having 2 installers on the spin would be a mess, to be honest.
15:21:41 <rdieter> Kevin_Kofler: if you want more testing, that's a good way to do it
15:21:47 <jgrulich> but it's a safer option
15:21:50 <rdieter> other than completely replacing anaconda
15:22:20 <rdieter> just a thought anyway
15:22:43 <rdieter> ideally and long-term, I agree including 2 installers is not ideal
15:24:53 <Kevin_Kofler> The Fedora way of getting testing has always been to force the users to use the new stuff.
15:25:47 <rdieter> sometimes there are no other viable options than "forcing" like that, in this case, we do have options
15:26:26 <Kevin_Kofler> Things like that "empty locale" fiasco really make me fed up of Anaconda.
15:26:38 <Kevin_Kofler> And all the other GTK+ and Anaconda bugs we had.
15:26:49 <Kevin_Kofler> How many blockers only because the installer uses GTK+ on a KDE spin?
15:27:24 <rdieter> true, but I'd rather those stay blockers and get fixed rather than ignoring them
15:27:28 <jgrulich> Kevin_Kofler: how many blockers we can have due to switching to Calameres? That's also a thing that should be considered
15:28:04 <Kevin_Kofler> It is also going to depend on what criteria it will be expected to comply.
15:28:30 <Kevin_Kofler> If it needs to pass all the current installer/Anaconda test cases, it will likely fail several of them.
15:28:52 <Kevin_Kofler> E.g. it doesn't support iSCSI, or at least it doesn't do anything to support it. (If it works, it's just by accident.)
15:29:07 <Kevin_Kofler> It also currently doesn't do LVM, LUKS, etc.
15:29:40 <Kevin_Kofler> Some of that stuff can and will be added eventually, other stuff might never be added.
15:29:53 <Kevin_Kofler> (I don't think iSCSI setups are really the target userbase for Calamares at this time.)
15:30:25 <Kevin_Kofler> That said, if it gets added by our developers, they'll accept it, they're very open to contributions.
15:30:33 <Kevin_Kofler> But what's sure is that I can't implement all these things alone.
15:30:49 <Kevin_Kofler> I'm currently the only Fedora maintainer for Calamares.
15:31:10 <Kevin_Kofler> I'm fine with it staying that way, but then don't expect it to support all the obscure partitioning corner cases.
15:32:06 <Kevin_Kofler> For what it's worth, one of the reasons Anaconda is such a mess is because of all the weird corner cases it tries to support.
15:33:36 <Kevin_Kofler> Oh, and one thing that would likely be considered a blocker is that Calamares currently does NOT support "Secure Boot".
15:33:57 <Kevin_Kofler> Installing GRUB-EFI should work now, but only directly, without shim, and thus without "Secure Boot" support.
15:34:29 <Kevin_Kofler> That's one thing that can and should be fixed in the F23 time frame though.
15:35:48 <Kevin_Kofler> One thing I'd need to know is what needs to be done before Calamares can even be considered, and what can be ignored.
15:36:51 <rdieter> Kevin_Kofler: good point, maybe start making an initial/work-in-progress calamares feature/change page with such details
15:37:05 <Kevin_Kofler> If you say it needs to support everything, that's obviously more work than if we agree on a sizable set of setups that we can actually test with.
15:37:44 <Kevin_Kofler> By the way, for "Secure Boot" setup, even if I do the work, I'll need somebody else to test it because I don't have any EFI-enabled machine.
15:38:50 <heliocastro> Kevin_Kofler: Can be done on qemu
15:39:19 <heliocastro> The current fedora package for virt-manager and edk is not good enough, but i will talk to cole to try help and fix this
15:39:26 <Kevin_Kofler> QEMU EFI support is not in Fedora.
15:39:30 <Kevin_Kofler> The firmware is not Free Software.
15:39:32 <heliocastro> Yes, i knoe
15:39:46 <heliocastro> And i will help to put in proper way
15:40:18 <Kevin_Kofler> The only way to fix the problem properly is to replace the FAT driver by one not derived from M$ specs (but from the existing Free drivers, e.g. the OpenBSD one).
15:40:19 <heliocastro> Cole robinson is the virt + qemu packager, i will try to talk to him
15:40:27 <Kevin_Kofler> So this is coding work.
15:40:34 <Kevin_Kofler> It's not just a packaging thing.
15:40:42 <heliocastro> Kevin_Kofler: No no, just prepare package to use ovmf
15:40:51 <Kevin_Kofler> That doesn't solve any problem.
15:41:01 <heliocastro> And then add ovmf in rpmfusion
15:41:06 <heliocastro> I foresee the issue already
15:41:08 <Kevin_Kofler> This needs to Just Work, and for that, it needs Free firmware.
15:41:18 <heliocastro> and there's a free firmware
15:41:22 <Kevin_Kofler> It's not Free.
15:41:26 <Kevin_Kofler> The FAT driver is encumbered.
15:41:34 <Kevin_Kofler> I'm not going to install that.
15:41:40 <heliocastro> Ok ok, i will see through that
15:41:43 <Kevin_Kofler> It's also never going to be in Fedora.
15:42:21 <Kevin_Kofler> That said, I don't think you are the right person to work on the driver.
15:42:43 <Kevin_Kofler> This needs somebody familiar with file system drivers.
15:43:02 <Kevin_Kofler> It's sad that Red Hat hasn't managed to do anything about that issue for years now.
15:43:16 <Kevin_Kofler> The problem has been known for a long time.
15:43:54 <heliocastro> ok, forget about, i'm not the right person.
15:44:06 <Kevin_Kofler> Of course the real problem is that some people at Intel thought the right way to write a FAT driver was to write it from scratch based on M$ specs, instead of just porting a working Free driver.
15:44:13 * rdieter forgot to mention  earlier, looks like qt 4.8.7 release happened today too, I'll update rawhide later
15:44:16 <pingou> so much encouragements
15:46:27 <Kevin_Kofler> heliocastro: Trying to package things so that the firmware can be pulled from a third-party repository is something you may be able to do. Actually fixing the problem, I think there ought to be people at Red Hat who can do it much faster.
15:46:45 <Kevin_Kofler> (or even outside of Red Hat, for that matter)
15:48:06 <pingou> and you know heliocastro so well that you know he/she cannot be that person
15:48:37 <pingou> great show of faith and support
15:49:20 <Kevin_Kofler> pingou: Look at the history of PackageKit-hif comps support…
15:49:53 <rdieter> enough, can we stay ontopic please?
15:50:02 <Kevin_Kofler> I don't want to offend anybody.
15:50:29 <Kevin_Kofler> I just don't want somebody to sign up for a task without knowing what it entails.
15:50:53 <Kevin_Kofler> I did look into the issue and I know what needs to be done, I also know that I'm definitely not the right person to do it.
15:51:02 <Kevin_Kofler> (Does that make me incompetent?)
15:52:19 <rdieter> Kevin_Kofler: it's wise to know what one cannot do.  though it may not be polite to point out what others can or cannot do
15:52:59 <rdieter> esp when making an effort to help couldn't hurt anything
15:53:38 <Kevin_Kofler> Meh, whatever I say, it's always wrong.
15:54:51 <rdieter> I'd encourage you to document the requirements/nice-to-have features you mentioned, and I'm fairly certain adding calamares as a secondary installer option to kde live shouldn't be (too) controversial
15:55:38 <rdieter> when requirements and sufficient testing is satisfied, then, we can consider more
16:00:14 <rdieter> our time's up, let's close thing, thanks everyone
16:00:17 <rdieter> #endmeeting