[00:01] <ScottK> infinity: I see you are a step ahead of me and already fixed libomxil-bellagio.
[00:11] <ScottK> Rejected the superseded xbmc binaries in any case.
[00:19] <ScottK> Checking qapt.
[00:25] <ScottK> Don't accept that one.
[01:50] <ScottK> We'll have a new upload in a bit.
[01:57] <infinity> ScottK: Yeah, a bit ahead.  I'm trying to sort out locally how to make xbmc armhf-friendly.
[01:58] <infinity> ScottK: Which is leading me to an ICE, so maybe it's a lost cause for now. :P
[02:13] <rsalveti> infinity: that's the weird thing, we got it to build at our PPA, probably a different config option
[02:13] <rsalveti> infinity: https://code.launchpad.net/~linaro-maintainers/+recipe/xbmc-eden-daily-precise
[02:14] <rsalveti> http://bazaar.launchpad.net/~linaro-maintainers/xbmc/xbmc-linaro-pkg-precise/view/head:/debian/rules
[02:14] <rsalveti> if you want to compare the build options with yours
[02:15] <infinity> Well, --enable-neon is wrong there, for Ubuntu.
[02:15] <rsalveti> sure, but don't think that this is what causing you the issue
[02:15] <rsalveti> I didn't yet check in details, just saw it
[02:15] <infinity> Upstream configure is also universally setting -mfloat-abi=softfp.
[02:15] <infinity> So, I patched that out.
[02:15] <infinity> And now I get an ICE later on.
[02:16] <infinity> Could just be entirely different versions between your PPA and the Debian packages?
[02:16] <rsalveti> infinity: well, we're using the released eden, probably a bit similar with the debian one
[02:16] <infinity> Although.  Only by about a day, if the date-as-version is to be believed.
[02:16] <rsalveti> and we got it to build for a few months already
[02:17] <infinity> Okay, you carry an armhf patch here.
[02:17] <infinity> And other bits.
[02:17] <infinity> I might just merge your packaging with the Debian one, and see how that works for me.
[02:18] <infinity> rsalveti: Thanks for the pointer.
[02:18] <rsalveti> yeah, true, avik is maintaining this package, and I remember he had to create an armhf specific patch for a ftbfs
[02:18] <rsalveti> quite a while ago
[02:18] <rsalveti> so that could be the fix you're looking for
[02:19] <infinity> Right, your armhf patch basically covers the configure fix I did locally, though.
[02:19] <infinity> So.
[02:19] <infinity> Hrm.
[02:19] <infinity> That probably still won't fix my ICE.
[02:19] <rsalveti> I think there's another one
[02:19] <rsalveti> let me look
[02:19] <infinity> You have other bits here that aren't in Debian's, though, I'll look.
[02:20] <rsalveti> infinity: http://bazaar.launchpad.net/~linaro-maintainers/xbmc/xbmc-linaro-pkg-precise/view/head:/debian/patches/0011-Fixing-MathUtils.h-for-armel.patch
[02:20] <infinity> Yeah, this source is a complete mess.
[02:21] <infinity> It's been a long time since I've seen a gcc-warnings-per-file ratio this dense. :P
[02:21] <rsalveti> you might also need http://bazaar.launchpad.net/~linaro-maintainers/xbmc/xbmc-linaro-pkg-precise/view/head:/debian/patches/0013-WinSystemX11GLES-Avoid-closing-X11-window-otherwise-.patch later on
[02:21] <rsalveti> but not critical, and we can fix once you get it to build
[02:21] <infinity> They desperately need someone anal-retentive upstream to go through this, warning-by-warning.
[02:21] <rsalveti> yup
[02:21] <infinity> I'm actually amazed it works on Windows.
[02:22] <rsalveti> hahah, yeah
[02:22] <infinity> Unless they're using mingw32 there.
[02:22] <infinity> Cause MSVC++ would probably just barf on this source and flip them the virtual bird.
[02:36] <infinity> rsalveti: If you've got a good handle on what needs to be done here, I won't mind terribly if you give me new xbmc packages for precise. :P
[02:37] <infinity> (And I can go back to planning my fpc transition)
[02:37] <rsalveti> infinity: sure, I can handle that
[02:37] <rsalveti> infinity: I'm planning to do a few things tonight already :-)
[02:37] <infinity> rsalveti: s/armel/armel armhf/ should be enough for debian/control and debian/rules.  And then all the Linaro patches.
[02:38] <rsalveti> infinity: ok
[02:38] <infinity> (At least, I hope that's enough to make it happy)
[02:38] <rsalveti> let me start my icecc enviroment :-)
[02:38] <infinity> Do you guys normally forward this stuff to upstreams and/or Debian?
[02:39] <rsalveti> infinity: yes, we didn't yet do to xbmc because we just enabled the build for precise, and for oneiric with have tons of other patches
[02:39] <rsalveti> but that's always the goal
[02:39]  * infinity nods.
[02:39] <infinity> Kay, just checking.
[02:51] <infinity> rsalveti: internal compiler error: in fp_immediate_constant, at config/arm/arm.c:13419
[02:51] <infinity> ^-- Same ICE I got locally.
[02:51] <rsalveti> infinity: hm, ok, will check
[02:51] <infinity> rsalveti: That MathUtils patch might fix it.  I haven't looked at what config/arm/arm.c does...
[02:53] <infinity> This is all obviously kinda fragile in a lot of places, given that the same source succeeded on Ubuntu/powerpc and failed on Debian/powerpc...
[02:53] <infinity> Can you spell b-r-i-t-t-l-e?
[02:54] <infinity> Okay, even more special.  It succeeded on Ubuntu's PPC buildds, but not my local Ubuntu/PPC build...
[02:54] <infinity> Best source EVER.
[02:55] <infinity> Oh, hah: Illegal instruction
[02:55] <infinity> I'm guessing it's building with altivec or something.
[02:55] <infinity> (Which the G5 buildds could handle, my local machine can't)
[02:56] <infinity> I'll look at that later, after you make armhf work.
[02:56] <rsalveti> hm, weird, are you using a clean local build?
[02:56] <rsalveti> ok
[02:57] <infinity> rsalveti: Oh, a hint on upstream's configure, you might want to call it with the "we're using tegra" flag (I forget what the flag was, but grepping for tegra should do).
[02:57] <rsalveti> infinity: yeah, I remember that
[02:57] <rsalveti> probably because of neon
[02:57] <infinity> rsalveti: Since that's pretty much exactly Ubuntu's target.  arch=armv7-a, tune=cortex-a9, neon off.
[02:58] <infinity> (we might actually tune for a8 by default, but we should be tuning for a9, IMO, so it wouldn't hurt for something CPU intensive like xbmc)
[02:59] <rsalveti> yeah
[03:02] <infinity> rsalveti: On a vaguely related note, if you and your minions are currently carrying any other arm-specific patches in PPAs that aren't in Precise, could you spend the next week or so finding those and getting them sponsored before we go into final freeze?
[03:02] <infinity> rsalveti: Things that aren't too scary to review, anyway, and obviously can't break other arches. ;)
[03:03] <rsalveti> infinity: sure, that's something I requested already last week, but I'll keep my eye on it
[03:03] <infinity> (I realise xbmc is kinda special here, since it wasn't *in* Ubuntu until yesterday)
[03:03] <rsalveti> yeah
[03:04] <infinity> rsalveti: Also stop slacking and become a core-dev.
[03:04] <infinity> kthx. ;)
[03:05] <rsalveti> infinity: I'll be applying for next cycle, I didn't want to become a core-dev because then I'd be forced to review and apply all patches from others
[03:05] <rsalveti> and I could become a blocker easily
[03:05] <rsalveti> what kind of worked, wookey pushed a bunch of fixes, marcin is now motu
[03:05] <infinity> rsalveti: There's a whole sponsorship team (and queue, and process) for that sort of thing.  Get your people to talk to our people and smooth that out if it's not working for them/us.
[03:06] <infinity> rsalveti: (wookey seems to have learned this cycle).
[03:06] <rsalveti> yeah
[03:06] <rsalveti> that was my goal
[03:06] <rsalveti> we're improving
[03:06] <rsalveti> :-)
[03:06] <infinity> Still, I think *you* should be core-dev.  You have the skills, and we'd like you. :P
[03:07] <infinity> I'm happy with wookey continuing to be sponsored, while he learns our quirks.
[03:07] <rsalveti> yeah, maybe will apply for the meeting we usually have at uds
[03:07] <rsalveti> will be at uds this time as well
[03:07] <infinity> \o/
[03:07] <infinity> I should be at Connect too, I think.
[03:07] <infinity> So, I can bug you twice in a row about it.
[03:07] <infinity> With violence.
[03:07] <rsalveti> cool :-)
[03:07] <rsalveti> lol
[03:12]  * infinity decides to do something non-work-related for a bit on his long weekend.
[03:12] <infinity> rsalveti: Ping me if you get xbmc happy (or if you give up and decide to have a weekend).
[03:13] <rsalveti> infinity: sure, will fix it ;-)
[03:13] <rsalveti> enjoy
[03:18]  * ScottK looks at qapt again.
[08:25] <Laney> infinity: omxil-bellagio> You don't mind the circular depends?
[09:41] <infinity> Laney: It's mildly unfortunate, but it does the trick.
[09:41] <infinity> Laney: It pretty much guarantees they go in the same dpkg run so, in practice, it does what I want it to do.
[09:45] <infinity> Laney: And undoing the lib/bin split, while a viable solution right now, is a no-go if anyone plans to multi-arch the library.
[09:45] <infinity> Laney: (Then again, it may be that the plugin registry it maintains would need rethinking in the face of multiarching too, I dunno)
[09:46] <infinity> (accepting obviously-correct base-installer fix)
[09:49]  * infinity flip-flops between sleep, or a run to Subway for a 4am snack...
[17:29] <rsalveti> infinity: got xbmc to build, will create the proper debdiff and make it available at a ppa to be sure it's building for all archs
[19:35] <tumbleweed> could an archive-admin please unblacklist xnee. bug 973872
[19:35] <ubot2> Launchpad bug 973872 in ubuntu "Sync xnee 3.11-1 (universe) from Debian testing (main)" [Wishlist,Triaged] https://launchpad.net/bugs/973872
[19:40]  * ScottK looks at ubuntu-defaults-zh-cn and purges unseeded.
[20:53] <CareBear\> porn?
[20:53] <CareBear\> hm, wrong channel, but ok
[21:50] <superm1> rsalveti: once you verify, would you mind sharing it with amejia (the debian maintainer)?  could easily put it in debian then and resync
[21:50] <superm1> oh and thanks for helping out on alternate architectures btw :)
[21:59] <rsalveti> superm1: sure
[22:02] <cjwatson> tumbleweed: done
[22:03] <tumbleweed> thanks
[22:41] <ScottK> superm1: does that mean you're good with accepting what we have now (xbmx)?
[23:04] <superm1> ScottK: yeah -4 can be accepted
[23:04] <ScottK> superm1: Thanks.
[23:04] <superm1> ScottK: if rsalveti's patch lands upstream we'll resync with that, but it's just for fixing FTBFS for arm* arch'
[23:05] <ScottK> Right.  There was a note not to accept it and check with you,so I'm just closing the loop.  Thanks.
[23:06] <ScottK> DOne.
[23:11] <superm1> thanks