Wednesday, 2013-02-13

*** chriadam|away is now known as chriadam00:00
*** jjarven_ has quit IRC00:03
*** wubudubudubudu has joined #mer00:04
*** SCV__ has quit IRC00:05
*** asdfafew has quit IRC00:06
*** jpwhiting_ is now known as jpwhiting00:13
*** jpwhiting has joined #mer00:14
*** jjarven_ has joined #mer00:19
[ol]Stskeeps: this seems to close my debt of unpublished changes made when porting Mer to x86_64:
*** vgrade has quit IRC00:28
*** vgrade has joined #mer00:28
*** asdfafew has joined #mer00:30
*** wubudubudubudu has quit IRC00:31
*** wubudubudubudu has joined #mer00:36
*** asdfafew has quit IRC00:38
*** gabriel9 has quit IRC00:54
*** gabriel9 has joined #mer00:57
*** jooncheol has joined #mer01:11
*** jooncheol2 has quit IRC01:11
*** rcg has quit IRC01:21
*** gabriel9 has quit IRC01:25
*** Martix has quit IRC01:30
*** M4rtinK has quit IRC01:33
*** asdfafew has joined #mer01:34
*** wubudubudubudu has quit IRC01:35
*** arcean has quit IRC01:52
*** Morpog_Mobile has quit IRC02:10
*** Morpog_Mobile has joined #mer02:27
*** crevetor_ has joined #mer02:30
*** crevetor_ has quit IRC02:45
*** crevetor_ has joined #mer02:45
*** DocScrutinizer05 has joined #mer02:51
*** Figure has quit IRC02:53
*** amppa has quit IRC02:54
*** amppa has joined #mer02:54
*** miq_ has quit IRC02:54
*** Figure has joined #mer02:54
*** miq_ has joined #mer02:55
*** crevetor_ has quit IRC02:56
*** crevetor_ has joined #mer02:58
*** MikaT has quit IRC03:02
*** MikaT has joined #mer03:03
*** DocScrutinizer05 has quit IRC03:04
*** tilgovi has quit IRC03:10
*** DocScrutinizer05 has joined #mer03:13
*** KaiRo_Mozilla has joined #mer03:21
*** KaIRC has quit IRC03:24
*** yunta has quit IRC03:43
*** furikku has joined #mer03:59
*** DocScrutinizer05 has quit IRC04:03
*** DocScrutinizer05 has joined #mer04:03
*** MikaT has quit IRC04:09
*** himamura has joined #mer04:09
*** MikaT has joined #mer04:09
*** raininja has joined #mer04:34
*** raijin has quit IRC04:37
*** kimitake_idle is now known as kimitake04:38
*** alexxy has quit IRC05:07
*** kimitake is now known as kimitake_idle05:07
*** alexxy has joined #mer05:16
*** crevetor_ has quit IRC05:24
*** VDVsx has quit IRC05:42
*** martyone has joined #mer05:45
*** edgar2 has joined #mer05:50
*** phinaliumz has joined #mer06:16
*** spiiroin has joined #mer06:16
*** kkszysiu has quit IRC06:18
*** imunsie has quit IRC06:21
*** bef0rd has quit IRC06:23
Stskeepsdefinately needed06:25
iekkui should have more coffee before left home: 2 days worktrip and didn├Ąt remember to take power cable with me... hopefully i get loan one before my laptop runs out of battery06:25
*** VDVsx has joined #mer06:29
*** M13 has joined #mer06:38
*** herangr has joined #mer06:42
*** lpotter-bris has joined #mer06:42
*** pcat has quit IRC06:42
*** lpotter has quit IRC06:44
*** M13 has quit IRC06:46
*** melonipoika has joined #mer06:47
*** zalan has joined #mer06:50
*** melonipoika has quit IRC06:52
*** 77CAANWJ3 has joined #mer06:54
*** Morpog_Mobile has quit IRC06:56
*** 77CAANWJ3 has quit IRC06:58
*** cxl000 has joined #mer07:02
*** lpotter-bris has quit IRC07:08
*** lpotter has joined #mer07:09
*** jukkaeklund has joined #mer07:09
andre__oh hell yeah, coffee please :)07:09
*** jayrulez has quit IRC07:14
*** pohly has joined #mer07:17
*** herangr has quit IRC07:19
*** melonipoika_ has joined #mer07:23
Stskeepslbt: seems like something went awry with prerelease import07:25
Stskeepsunresolved all around07:26
*** plfiorini has quit IRC07:26
*** Morpog_Mobile has joined #mer07:28
*** melonipoika_ has quit IRC07:31
*** melonipoika has joined #mer07:32
*** pcat has joined #mer07:43
*** chriadam is now known as chriadam|away07:44
*** pvanhoof has joined #mer07:46
*** rferrazz has joined #mer07:47
*** apostrophe has joined #mer07:56
*** mvogt has quit IRC07:59
*** LouisDK has joined #mer07:59
*** RzR has quit IRC08:01
*** jpetrell has joined #mer08:02
*** RzR has joined #mer08:02
*** niweber has quit IRC08:04
*** RzR has quit IRC08:04
*** RzR has joined #mer08:05
*** gabriel9|work has joined #mer08:05
*** LouisDK is now known as LouisDK_08:07
*** LouisDK_ is now known as LouisDK__08:07
*** LouisDK__ is now known as Louis108:08
*** RzR has quit IRC08:08
*** M13 has joined #mer08:08
*** LouisDK has joined #mer08:08
*** RzR has joined #mer08:09
*** LouisDK_ has joined #mer08:10
*** fk_lx has joined #mer08:10
*** LouisDK_ has quit IRC08:11
*** Louis1 has quit IRC08:11
*** LouisDK has quit IRC08:11
*** LouisDK has joined #mer08:11
*** RzR has quit IRC08:12
*** mikhas has joined #mer08:13
*** RzR has joined #mer08:13
*** M13 has quit IRC08:15
*** phinaliumz has quit IRC08:16
*** phinaliumz has joined #mer08:16
*** FlameReaper has joined #mer08:20
*** melonipoika has quit IRC08:25
*** blam_ has quit IRC08:25
*** melonipoika has joined #mer08:25
*** rferrazz has quit IRC08:28
*** jstaniek has joined #mer08:29
*** Frye has joined #mer08:32
*** plfiorini has joined #mer08:37
*** M13 has joined #mer08:38
*** LouisDK has quit IRC08:50
lbtStskeeps: yeah - the updatestorm crashed - I fiddled and re-ran it but looks badf08:51
*** stefan_schmidt_w has joined #mer08:52
Stskeepslbt: hmm, interesting08:53
Stskeepslbt: does ?repository= files looks sane?08:55
lbtmy bad - I was looking where to add the x86_64 and added it to the xml mapping but added both core and next - core doesn't exist yet of course08:55
*** niweber has joined #mer08:55
lbtremoved and re-ran08:55
lbtit's finished now with no errors08:55
Stskeepslet's see if it clears up unresolvables08:56
lbtjust having a look08:56
lbtit doesn't look right08:59
Stskeepscheck the nohup.out i guess08:59
*** e8johan has joined #mer08:59
Stskeepsand it'll in the late phase be sending repository update08:59
lbtnothing obvious09:03
Stskeepsgive it a bit, i think09:03
Stskeepsif you just ran updatesstorm it'll take a bit09:03
lbtit's pretty idle09:03
lbtno, I ran that last night09:03
lbtit's done09:03
Stskeepsyeah but you said it crashed?09:04
lbtI fixed the xml and re-ran09:04
Stskeepsobs protocol is such a mess at times.. moment09:05
Stskeepslbt:         <mapping project="Core-next:i586" path="obs-projects/Core-next/i586" binaries="obs-repos/Core:i586:0.20121205.0.1" reponame="Core_i586" />09:07
Stskeepsspot the error09:08
*** lamikr has joined #mer09:09
*** jonwil has joined #mer09:10
lbtyeah, that's OK atm - I'm just re-running make-next and I reverted to an old mapping prior to doing it09:10
*** rcg has joined #mer09:12
lbtit is re-making the mappingscache.xml now09:13
*** jpetrell has quit IRC09:18
*** Aristide has joined #mer09:20
*** yunta has joined #mer09:21
*** Jucato has joined #mer09:23
*** calvaris has joined #mer09:27
*** notmart has joined #mer09:31
*** varikonniemi has joined #mer09:35
*** e8johan has quit IRC09:39
*** jpetrell has joined #mer09:49
*** ashakunt has joined #mer09:50
*** jpetrell has quit IRC09:59
*** himamura has quit IRC10:00
Stskeepslbt: watching the updatesstorm going on10:05
*** pcat has quit IRC10:06
*** jpetrell has joined #mer10:10
*** Morpog_Mobile has quit IRC10:13
*** drussell has quit IRC10:15
*** Morpog_Mobile has joined #mer10:17
*** jonwil has quit IRC10:19
*** jpetersen_ has joined #mer10:23
*** lpotter-bris has joined #mer10:23
*** lpotter has quit IRC10:23
*** jpetersen has quit IRC10:26
*** drussell has joined #mer10:27
*** zalan has quit IRC10:29
*** zalan has joined #mer10:30
*** zalan has joined #mer10:31
*** leinir has quit IRC10:31
*** auri_away is now known as auri__10:36
*** M4rtinK has joined #mer10:36
*** Martix has joined #mer10:42
*** pcat has joined #mer10:50
*** jpetersen_ is now known as jpetersen10:53
*** leinir has joined #mer11:07
*** calvaris has quit IRC11:11
*** sledges has joined #mer11:15
*** sledges has quit IRC11:15
*** sledges has joined #mer11:15
*** techlife has quit IRC11:16
sledges/me says hello. Was ill, at least recovered all unseen horror movies %)11:17
*** techlife has joined #mer11:17
lbthey sledges11:20
sledgeshya lbt , update from my company: we can't adopt Mer for primary bussiness, because current craze in the market is HTML5. But they liked what I showed on Pandaboard, and will offer as solution for non-HTML5-/non-Android- suited clients, a small victory ;)11:26
Stskeepssledges: mer's a html5 distribution too though, admittedly not as evolved :)11:27
Stskeepsshould put firefox OS runtime on top11:27
sledgesyes, but for this we just fire-up a buildroot with a lean11:27
sledgesQt for max performance11:27
sledgesand HTML5 served from server somewhere11:28
sledgesmost common scenario. How to tie in with on-board peripherals -- most often is a narrow and manually solvable case11:28
sledgesindustrial embedded world :{11:29
*** ashakunt has quit IRC11:31
*** dr_gogeta86 has joined #mer11:31
dr_gogeta86anyone here ported mer to galaxy tab11:31
*** ashakunt has joined #mer11:31
*** M13 has quit IRC11:38
dr_gogeta86or any android device11:45
*** arcean has joined #mer11:48
dm8tbrnexus 7 and archos gen9 have ports at least11:51
dr_gogeta86how did you get xorg running on nexus 711:53
dr_gogeta86there are driver for tegra ?11:53
*** himamura has joined #mer11:54
dr_gogeta86and for this PowerVR SGX 54011:56
kulveI know mainly about nexus 7 but I think there's an xorg driver for that sgx as well11:56
*** FReaper has joined #mer12:02
*** FlameReaper has quit IRC12:03
*** FReaper is now known as FlameReaper12:03
*** edgar2 has quit IRC12:05
rcgdr_gogeta86, there are the pvr xorg drivers12:05
dr_gogeta86works well12:06
dr_gogeta86may i run gnome shell on it12:06
rcgwell, with respect to the archos g9: we currently only have quite old drivers right now12:07
rcgin short: updating to newer ones would require quite some work. and most of it all, someone who does the work12:07
rcgi'd say it works well as development platform and for enthusiasts12:07
rcgbut nothing i would sell to normal users12:07
rcgand dunno about gnome shell12:08
rcgbut i suspect gnome shell would eat too much resources12:08
*** ashakunt has quit IRC12:16
*** jukkaeklund has quit IRC12:17
*** kontio has quit IRC12:20
*** kontio has joined #mer12:22
*** kontio has joined #mer12:22
*** Morpog_Mobile has quit IRC12:25
*** CosmoHill has joined #mer12:26
*** jukkaeklund has joined #mer12:29
*** ashakunt has joined #mer12:30
*** tetris4 has joined #mer12:32
*** kantobro has joined #mer12:34
*** KaIRC has joined #mer12:35
*** FReaper has joined #mer12:49
*** FlameReaper has quit IRC12:49
*** FReaper is now known as FlameReaper12:50
*** CosmoHill has quit IRC12:58
*** niqt has joined #mer12:58
*** phinaliumz has quit IRC13:00
*** jayrulez has joined #mer13:01
*** jukkaeklund_ has joined #mer13:04
*** jukkaeklund has quit IRC13:05
*** rcg has quit IRC13:07
*** peavey has quit IRC13:11
*** mikhas has quit IRC13:11
*** peavey has joined #mer13:14
*** rcg has joined #mer13:15
*** mikhas has joined #mer13:23
*** yunta has quit IRC13:31
*** Morpog_Mobile has joined #mer13:38
*** bigbluehat has quit IRC13:39
*** bigbluehat has joined #mer13:39
*** Morpog_Mobile has quit IRC13:42
*** Morpog_Mobile has joined #mer13:43
*** fk_lx has quit IRC13:45
*** fk_lx has joined #mer13:47
*** tetris4 has quit IRC13:48
*** FSCV has joined #mer13:52
*** jstaniek has quit IRC13:54
*** yunta has joined #mer14:01
*** jukkaeklund_ has quit IRC14:08
*** niqt has quit IRC14:14
*** niqt has joined #mer14:16
*** fwp2k has joined #mer14:22
*** Morpog_Mobile has quit IRC14:24
*** martyone has quit IRC14:32
*** Shaan7 has joined #mer14:41
*** jayrulez has quit IRC14:45
*** SouL_ has joined #mer14:46
*** SouL_ has left #mer14:47
*** sledges has quit IRC14:47
*** Shaan7 has quit IRC14:49
*** sledges has joined #mer14:53
*** sledges has quit IRC14:53
*** sledges has joined #mer14:53
*** arcean_ has joined #mer14:57
*** e8johan has joined #mer14:57
*** jayrulez has joined #mer14:57
*** ashakunt has quit IRC14:59
*** fwp2k has quit IRC15:00
*** arcean has quit IRC15:00
*** dm8tbr has quit IRC15:03
*** dm8tbr has joined #mer15:03
*** Ronksu has quit IRC15:10
*** Ronksu has joined #mer15:11
*** bef0rd has joined #mer15:11
*** Morpog_Mobile has joined #mer15:12
*** rferrazz has joined #mer15:13
*** e8johan has quit IRC15:15
*** rferrazz has quit IRC15:16
*** fcorrea has quit IRC15:17
*** gabriel9 has joined #mer15:19
*** gabriel9|work has quit IRC15:19
*** varikonniemi has quit IRC15:22
*** fcorrea has joined #mer15:29
*** ced117 has joined #mer15:42
*** Jucato has quit IRC15:45
*** pvanhoof has quit IRC15:46
*** SouL_ has joined #mer15:49
*** SouL_ has left #mer15:49
*** stefan_schmidt_w has quit IRC15:57
*** Aristide has quit IRC16:00
*** pcat has quit IRC16:03
*** melonipoika has quit IRC16:13
*** melonipoika has joined #mer16:13
*** Shaan7 has joined #mer16:16
*** niqt has quit IRC16:20
*** XenGi_ is now known as XenGi16:21
*** Morpog_Mobile has quit IRC16:26
*** Morpog_Mobile has joined #mer16:28
*** Aristide has joined #mer16:33
*** edgar2 has joined #mer16:42
*** lamikr has quit IRC16:44
*** M4rtinK has quit IRC16:46
*** Artox has joined #mer16:49
*** dr_gogeta86 has quit IRC16:52
*** NIN101 has joined #mer16:53
*** Morpog_Mobile has quit IRC16:54
*** Morpog_Mobile has joined #mer16:58
*** gabriel9 has quit IRC16:59
*** pcat has joined #mer17:02
*** shadeslayer has quit IRC17:03
*** shadeslayer has joined #mer17:03
*** plfiorini has quit IRC17:04
*** rferrazz has joined #mer17:06
*** yashshah has joined #mer17:14
*** Morpog_Mobile has quit IRC17:23
*** Morpog_ has joined #mer17:23
lbtStskeeps: [ol] < last column17:24
*** Morpog_ has quit IRC17:28
*** Morpog_Mobile has joined #mer17:28
*** Morpog_Mobile has quit IRC17:29
*** Morpog_ has joined #mer17:30
*** martyone has joined #mer17:31
*** pcat has quit IRC17:33
*** XenGi is now known as XenGi_17:34
*** FlameReaper has quit IRC17:36
*** M13 has joined #mer17:37
*** XenGi_ is now known as XenGi17:40
*** bef0rd has quit IRC17:41
*** XenGi is now known as XenGi_17:46
*** auri__ is now known as auri__away17:52
*** Shaan7 has quit IRC17:52
*** Artox has quit IRC17:53
*** yunta has quit IRC17:53
*** bef0rd has joined #mer17:54
*** furikku has quit IRC17:59
*** rferrazz has quit IRC18:00
[ol]lbt: Does it mean that qt5 compiles without problems for x86_64?18:12
lbtI think there may be webkit build issues18:13
lbtran out of disk space :)18:14
BostikI did not like what that sounds18:14
Bostikbut amd64 toolchain and target - very much in "WANT!" category18:15
*** kkszysiu has joined #mer18:15
lbt < last column18:15
lbt too18:15
*** Morpog_ has quit IRC18:17
Bostikokay, so we have a working toolchain on _6418:17
* Bostik gets the slop bucket18:17
[ol]At least, working enough to build Mer Core.18:18
Bostik[ol]: that is very, very, VERY welcome indeed18:18
Bostikwe'll abuse the living crap out of the toolchain with qtwebkit builds soon then ;)18:19
*** jooncheol has quit IRC18:22
*** lpotter-bris is now known as lpotter18:23
*** bef0rd has quit IRC18:24
Sage[ol]: I'll add the x64 targets to nemo soonish so we can see how it goes.18:24
*** bef0rd has joined #mer18:26
*** yashshah has quit IRC18:27
[ol]Sage: Cool!18:28
*** ScriptRipper1 has quit IRC18:31
StskeepsBostik: cross infra building on x86-64 still wip, but working on fully bootstrapped glibc and cross gcc from x86 side in test:toolchain on
*** ScriptRipper1 has joined #mer18:36
*** pohly has quit IRC18:40
*** bef0rd has quit IRC18:40
*** pcat has joined #mer18:42
*** fk_lx has quit IRC18:43
*** M13 has quit IRC18:45
*** phaeron has quit IRC18:46
lbtso... when hacking and deploying code from QtCreator to a device ... where should the code go18:48
lbtI feel that $HOME is not quite right - that's my main device, I don't want to be hacking there18:49
lbtso /opt/sdk feels like a good place18:49
*** bef0rd has joined #mer18:49
lbtnice and isolated18:49
* Stskeeps is off, concert tonight18:50
BostikStskeeps: yum!18:50
Bostik(not the tool)18:50
*** yashshah has joined #mer18:56
*** mikhas has quit IRC18:57
*** Morpog_Mobile has joined #mer19:01
*** M13 has joined #mer19:02
*** pcat has quit IRC19:03
*** Aristide has quit IRC19:05
*** plfiorini has joined #mer19:07
*** tilgovi has joined #mer19:11
Sagelbt: Stskeeps: <- can you fill this for x8619:12
Sage[ol]: <- first enabled lets see how it goes19:17
Sagethere is quite a bit of stuff building so might take a while19:18
*** bef0rd has quit IRC19:20
*** asdfafew has quit IRC19:22
Sagehmmp... ok so how %{_libdir} is supposed to work. Isn't it supposed to be /usr/lib64/ on x86_64?19:24
SageAard: ^19:24
*** andre__ has quit IRC19:25
Aard%{_lib} is supposed to be lib64, %{_libdir} is supposed to be a macro containing prefix and the %{_lib} macro19:26
Aardso, yes19:27
[ol]Sage: %{_lib} chould be "lib", %{_libdir} should be "/usr/lib".19:27
Aard[ol]: not for 64bit architectures19:28
[ol]Aard: For x86_64. I've eliminated all multilib out of the way.19:28
*** alexxy has quit IRC19:29
* plfiorini smells talking x86_6419:30
Sage[ol]: I'm not really sure if that is the right way as now we need to start keeping patches on top of each of the upstreams really as providing the multilib is quite standard. Or am I mistaken?19:31
Aard[ol]: why that? I'd like to keep the option for running 32bit apps as well19:31
Sage[ol]: in theory it should just work if the rpm macro works etc.19:31
*** fk_lx has joined #mer19:31
Aardand I guess we'd be the only ones doing it that way, _and_ we'll probably have fun if/when we add another architecture where multilib plays a bigger role19:32
[ol]Sage: There were not so many patches for removing multilib. Some packages (like libaccounts-qt) just have it hardcoded: if arch is x86_64 then use lib64. But most ofthe packahes just use libdir supplied to them by configure or whatever they use for autoconfiguration.19:33
[ol]Anyway, using lib64 is ugly. There should "/lib" no matter what is the base architecture.19:35
Sage[ol]: ok, I see. So even those components should just take e.g., configure --libdir=%{_libdir} where %{_libdir} can point to /usr/lib or /usr/lib6419:35
Sage[ol]: well, have you noted :)19:35
MerbotMer bug 174 in .Other "Move all to /usr" [Task,New]19:35
Aard[ol]: do we have support for 'use 32bit libraries in lib32 (or something similar)'?19:36
[ol]Aard: You can use a different directory to put shared libraries for different architectures.19:37
Sage[ol]: I agree with the hardcoded multilib removals yes, but I still think that we should install 64-bit stuff to /usr/lib64/ and allow 32-bit and 64-bit coexist.19:37
Sage[ol]: I'm using 64-bit system myself and I have about 200 32-bit packages installed because some stuff just requires that.19:38
*** Aristide has joined #mer19:38
[ol]Sage: Wrong approach. Libraries for the main arch should go to "/lib", no matter what. Libraries for other architectures (whether they are supported directly by hardware of through qemu) should be placed somewhere else.19:38
*** wubudubudubudu has joined #mer19:39
Sage[ol]: well, then that would need two builds of 32-bit packages19:39
[ol]Like "/emul/i686-linux-gnu/lib" or something.19:39
[ol]Sage: The same problem if you want to run arm binaries on x86_64.19:40
Sage[ol]: well, I would not say wrong approach. I would say it needs to be discussed more and decide the way it is wanted.19:40
Sage[ol]: well, that is different thing as those are not compatible.19:41
*** phaeron has joined #mer19:41
[ol]I don't see much difference. Incompatible binaries can still be run transparently through qemu.19:42
[ol]Also, 32-bit on 64-bit is just a very limited case. What about big endian binaries on little endian system?19:43
Sage64-bit vs 32-bit and 64-bit vs armX are a bit different things to compare19:43
*** pcat has joined #mer19:43
[ol]OK, What about all combinations of MIPS 32-bit/64-bit LE/BE on the same system?19:43
*** jjarven_ has quit IRC19:44
* Sage thinks [ol] got carried away a bit from the original question19:45
SageLets forget about running stuff with qemu on Mer which is aimed to be for mobile and small embedded systems.19:46
[ol]Anyway, at the moment you can't install unmodified i686 package on Mer Core x86_64.19:46
Sageso, installing arm or mips rpm's to x86(_64) systems19:47
Sageand get back to x86 vs x86_6419:47
Sageso the question is that if Mer wants to support installing x86 rpm's to x86_64 systems or not?19:48
[ol]The reason to eliminate multilib was to make things simple. Running two sets of shared libraries is out of scope for minimal embedded distribution.19:49
Sagewithout repackaging/compiling the x86 rpm's19:49
[ol]Sage: It would be possible if i686 library packages were made relocatable.19:49
*** bef0rd has joined #mer19:49
*** sledges has quit IRC19:58
*** bef0rd has quit IRC20:00
[ol]Ideally, each application should access all shared libraries it needs in "/lib", no matter what architecture or ABI it is. I think, it can be done by bind mounting appropriate directory on top of "/lib" in application's namespace.20:03
*** FSCV_ has joined #mer20:04
*** jjarven_ has joined #mer20:05
*** FSCV has quit IRC20:05
Aard[ol]: arm64 will come in the near future. it might be required there to keep 32bit stuff around we only get as binary blobs. I'd expect it to cause less trouble if the 32bit stuff goes into lib20:06
[ol]Aard: Or lib32?20:07
*** martyone has quit IRC20:09
*** wubudubudubudu has quit IRC20:11
[ol]Mer is currently in unique position that it doesn't need to keep compatibility with legacy versions. It would be great to use this advantage to do things right.20:12
*** ikona has joined #mer20:12
*** wubudubudubudu has joined #mer20:12
*** fcorrea has quit IRC20:13
*** Sfiet_Konstantin has joined #mer20:14
*** melonipoika has quit IRC20:17
*** edgar2 has quit IRC20:20
*** fcorrea has joined #mer20:26
*** melonipoika has joined #mer20:31
Sage[ol]: the path where libraries are installed doesn't really matter afaik. library paths makes it possible to call it even /mer/libraries/ if we want, right?20:33
Sageand in next versino just add new library path and push new stuff to /mer/new_libraries/20:34
[ol]Or even better, to "/lib/x86_64-linux-gnu".20:37
[ol]Like in
*** icota has joined #mer20:39
[ol]That's why I said earlier that i686 on x86_64 is not so special case. Ideally, any arch on any arch should be supported if necessary.20:40
Sageisn't that a bit conflicting what you said above about "  each application should access all shared libraries it needs in "/lib".. ?20:41
*** melonipoika has quit IRC20:42
Sagebut I don't have objections though in our case it would be /usr/lib/x86_64- probably20:42
[ol]No conflict. The way application views filesystem may be specific to this application.20:43
[ol]Like bind mounting "/lib/arm-linux-gnueabi" to "/lib" in application's provate namespace.20:44
[ol]If we have a binary application which expects its libraries to live in "/lib", why not make this application find what it's expecting, without confusing other applications?20:45
Sageif binary application expects it binaries in /lib isn't that already wrong? :)20:46
[ol]Sage: It is, but if you have source code you can re-compile, why would you need i686 on x86_64 anyway? The problem is with binary-only applications (like Skype).20:47
Sageskype doesn't expect /lib it searches from library path and uses those at least the one I have :)20:49
*** icota has quit IRC20:49
[ol]What about "/lib/"? Can it be relocated?20:49
[ol]I doubt so.20:50
*** yashshah has quit IRC20:52
*** yashshah has joined #mer20:53
[ol]On the topic of moving everything to "/usr", I've already expressed opinion that it's absolutely wrong.20:55
[ol]The motivation for Fedora to do it (like having "/usr" to be NFS-mounted) does not apply to Mer.20:56
[ol]For Mer it would be much more logical to do it the opposite way: get rid of "/usr" and move everything to "/".20:56
*** melonipoika_ has joined #mer20:57
* lbt notes that we'd need a *lot* of benefit to justify swimming against the tide21:02
lbtin general we do thinks in the same way as established distros like suse/fedora/debian21:02
[ol]Does Debian also move everything to "/usr"?21:03
lbtthis minimises the need to carry patches, allows us to share packaging guidelines and probably ensures we have more overlap in security related matters21:03
SpeedEvildesperate boot devices tend to be going away, in favour of one big emmc21:03
*** Jare has quit IRC21:04
[ol]I think, the most important way is to make packages honour their bindir and libdir and not assume anything hardcoded. Most of the packages are already like this, but there are unfortunate exceptions (like libaccounts-qt discussed above).21:06
[ol]In fact, there are several tides flowing in different directions right now, and we should be very selective in choosing which tides to follow.21:10
lbtI would prefer to see /usr/<lib.*64.*>/ than /usr/lib/21:10
[ol]For example, Debian's multiarch tide looks beneficial to me. On the other hand, Fedora's "move everuthing to /usr" tide doesn't.21:11
lbtI like both :)21:11
[ol]lbt: Do you like "eliminate / and /usr split" or "eliminate / and /usr split by moving everything to /usr"? This is not the same, there is another solution to eliminating this split.21:13
lbtI don't like the split21:13
lbtI see moving it to /usr as a sane solution21:13
[ol]OK, that's where we have common groung for further discussion.21:13
[ol]But the second one makes me disagree.21:14
[ol]All this "/usr" hierarchy is just an artefact of Unix history, which should not exist at all.21:14
[ol]Historically, "/usr" was the place where users' home directories were located. There were users bin, lib and man, which were owhers of binaries, libraries and manuals, so there were "/usr/bin", "/usr/lib" and "/usr/man" directories.21:15
lbtI'm swayed by the benefit of a read-only /usr21:17
[ol]Of course, binaries essettial to system boot were in "/bin" (and "/etc, BTW), so it was possible to have "/usr" on a separate disk as a separate filesystem mounted when the system is ready to start in multi-user mode.21:17
*** Jare has joined #mer21:17
lbtand I'm an OCD .. I like my binaries/libs to be in the same drawer21:17
lbtI read a fair bit21:18
Aardlbt: ro root is more interesting21:18
lbt /var ?21:18
[ol]Yes, I read it as well, and it all is just based on wrong assumptions.21:19
*** Morpog_Mobile has quit IRC21:19
lbtAard: I do agree though21:19
Aardlbt: many things in var can live in tmpfs, for the handful of other things just do an overlay mount21:19
[ol]I've successfully made FreeBSD-based thin client with NFS-mounted root shareable between clients. There's no need to have "/usr" for that.21:20
AardI had a notebook with flaky ide controller if the root-filesystem got booted from it almost 10 years ago. solved it by doing a customized debian with ro-root21:20
* lbt had 2 myth frontend with same nfsroot - so I know too21:20
Aardnowadays we have a lot more interesting technologies to make something easier than back then21:21
lbtso I personally see the / -> /usr vs /usr -> / merge as not interesting21:22
Aardroot could be a ro-mounted btrfs, and for updates you do a snapshot that gets set ro as soon as the update is in21:22
lbtwe picked / -> /usr :)21:22
lbtAard: yeah - I can see lots of fun stuff in that space :)21:23
Aardwell, that's the direction I'd like phones of a some company go at some point :)21:23
*** alexxy has joined #mer21:25
[ol]Look how it's done in Plan9. This is a newer system made without compatibility to legacy stuff. It has no "/usr" as a separate hierarchy at all. And it handles networked filesystems with everything shared much better that Unix and Linux do.21:25
*** Morpog_Mobile has joined #mer21:25
[ol]Key features used to that are: union mounts and per-process mount namespaces.21:26
lbt[ol]: exploring that is fine - it's what Mer and OSS is for!21:27
lbtOTOH we also have a duty to make life easy for our vendors21:27
lbtso from that PoV it has to be easy for them (and their armies of minions) to grok21:28
lbtand that means least surprise21:28
[ol]Ironically, Linux has these features already, but for some reason Fedora guys decided to solve the problem of legacy design using legacy solutions.21:28
*** NIN101 has quit IRC21:30
[ol]I think, it would be wise to ask vendors what would be least surprise for them. I'm not sure that all vendors expect to see everything in "/usr".21:31
lbtwell, again, /{lib,bin,sbin} exist ... so they would never notice21:33
*** ced117 has quit IRC21:33
lbtI'm much more interested in multi-arch than usr-merge anyhow21:33
[ol]The same way they would never notice if "/usr" is gone provided that "/usr"->"/" symlink exists.21:34
lbtyep, "potato", "potato"21:35
*** Morpog_PC has joined #mer21:35
lbt(you get that reference?)21:36
[ol]Something from Monthy Python?21:36
lbtdialect variation of "potato"21:36
lbtsometimes pronounced like "potaaaato"21:37
lbtalso  tom-aaaaah-to   vs tom-ay-to21:37
[ol]I heard about differences of pronounciation of "tomato" between American and British English, but I didn't know about the same with "potato".21:39
[ol]But anyway, if vendor doesn't care about this "/usr" move, why doing it at all?21:40
[ol]What problem does it solve?21:41
[ol]And "Fedora does it, let's do the same" is not a problem worth solving.21:41
lbteasier packaging21:41
[ol]Why do you think packaging will be easier?21:42
lbtzero chance of getting /bin/exe vs /usr/bin/exe wrong21:42
[ol]I was thinking, it's already solved.21:43
lbtTMTOWTDI - always21:43
[ol]And if one of the ways is to do nothing, this way is preferable. :-)21:45
lbtwell, it's done now, so doing nothing now is fine :)21:46
[ol]You mean, Mer has already moved everything to "/usr"? :-O21:47
lbtthat's underway, yes21:47
*** tanty has quit IRC21:50
[ol]Then it's wasted effort to solve the wrong problem. Easier packaging? No. Easier to find binaries? No. There's already PATH variable to solve this. NFS-mounted shared "/usr"? If you're doing it, you're already doing it the wrong way.21:52
[ol]Also, it doesn't make sharing security patches easier. Patches are usually applied against source code, and then this source is built and installed in a separate step. There's very low chance that source patch dealing with code logic has anything to do with where the package is finally installed.21:54
*** glueckself has joined #mer21:55
Stskeepspackages should install into %{_prefix}, no more, no less21:56
Stskeepsand all packaging parts should stem from that21:56
Stskeepsie, no /lib/ directly mentioned in .spec21:56
*** jstaniek has joined #mer21:56
Stskeepsor /bin/21:56
[ol]It would be interesting to build Mer with %{_prefix} set to something other that "/usr" and see how many packages will break.21:57
Stskeepsyeah, that's one for OBS to try, i guess21:57
[ol]Also, %{_lib} should be set to something other than "lib".21:59
[ol]That would be a real test of path portability.21:59
[ol]BTW, does OBS server require kvm for virtualisation?22:01
*** apostrophe has quit IRC22:02
lbtkvm. xen, lxc or plain chroot I think22:02
[ol]Is it configurable?22:02
lbtI use xen on meego and kvm on mer22:02
[ol]Why not just chroot?22:02
lbtchroot is not secure22:03
lbtand we're giving people the ability to run builds on our VMs...22:03
lbtso we need tight containers22:03
lbtalso why we have no net in the build env22:04
[ol]OK, then why not lxc? It's much more lightweight?22:04
*** pcat has quit IRC22:04
lbtalso not secure22:04
lbtin that case it's because upstream haven't tested it enough22:04
[ol]Do you have OBS server packages for running on Mer?22:05
[ol]I'd like to have my own OBS to perform some unethical experiments.22:05
lbtgithub thing is useful22:06
lbtthat's about it22:08
[ol]I'm confused about
[ol]It lists OpenSuse 11.4 as prerequisite.22:08
[ol]I'd like to run it on Mer, not OpenSuse.22:09
lbtI am so tempted to try english humour on irc sometimes...22:09
lbtOBS runs on big iron22:10
lbtMer doesn't22:10
[ol]Objection! Mer x86_64 does!22:10
lbteach of them is a 64GB ram machine with quad core and loads of disk22:10
lbtand then there are more for the frontend/backend/schedulers22:11
lbtwe use SuSE for them22:11
*** CosmoHill has joined #mer22:11
Stskeepsobs needs opensuse, running it elsewhere is very very painful22:12
Stskeepssimple as that :)22:12
[ol]That's fine if you want to run public farm for multiple of users. But for my personal experiments something more moderate would be sufficient.22:12
lbtyeah, suse VMs :)22:13
Stskeeps[ol]: personally i would use VMs for that purpose, with opensuse inside22:13
Stskeepsi don't personally want to be in the business of maintaining build farm software for a OS, so22:13
Stskeepsand it works fine there22:13
*** Cosmo[PB] has joined #mer22:15
[ol]I'm confused even more.22:15
*** CosmoHill has quit IRC22:15
[ol][22:05:21] <[ol]> Do you have OBS server packages for running on Mer?22:15
*** Cosmo[PB] is now known as CosmoHill22:15
Stskeepsno, we don't22:15
[ol][22:05:55] <lbt> yes22:15
lbtsometimes I assume meaning22:15
[ol]Was it answer to a different question?22:15
lbtas in building Mer22:15
lbtno, we don't22:16
Stskeepslbt: important thing with us natives, be literal22:16
lbtwe do have Mer specific packages of OBS which are needed to build Mer22:16
Stskeepser, non-natives22:16
lbt*g* see - I assumed you meant non-native and I assumed [ol] meant "building" not "running"22:17
Stskeepsanyway: it saves a boatload of time and pain to just run a OBS instance in a virtual machine, time that's better spent doing something else22:17
lbt[ol]: we use scratchbox2 - as you know22:17
Stskeepsand you get full power to build a lot of different systems22:17
lbtand we modified obs to support it22:17
*** nicola_ has joined #mer22:17
lbtand you need our packages to run an obs to build mer22:17
lbtor our pathches at least22:18
[ol]OK, I'll take a look.22:19
lbtboatload of stuff I use to make suse VMs using JEOS22:19
lbtwhat I do is make a JEOS into an LVM LV22:20
lbtand then snapshot it, configure, install OBS, test, wipe22:20
lbtthe benefit is that it takes about 5secs to get a new running VM22:20
*** notmart has quit IRC22:20
lbtsadly kiwi isn't as portable as debootsrap :D22:22
[ol]What is JEOS? ?22:22
*** himamura has quit IRC22:32
*** RzR has quit IRC22:40
*** RzR has joined #mer22:45
*** CosmoHill has quit IRC22:48
*** yashshah has quit IRC22:49
*** yashshah has joined #mer22:50
*** CosmoHill has joined #mer22:53
*** Artox has joined #mer23:04
lbthmm - how do I recover from an X mouse-grab ?23:05
lbtah, nm - firefox grabbed it whilst doing a dropdown and just gave it back 5mins later... :/23:06
w00tpoor X23:07
w00tnobody ever wants to grab it23:07
CosmoHillnight night23:10
*** CosmoHill has quit IRC23:13
*** mvogt has joined #mer23:15
*** melonipoika_ has quit IRC23:19
*** Martix has quit IRC23:21
*** zalan has quit IRC23:21
*** Morpog_Mobile has quit IRC23:21
*** Morpog_Mobile has joined #mer23:21
*** gabriel9 has joined #mer23:22
lbtdebian_chroot eh23:24
*** nicola_ has quit IRC23:38
*** asdfafew has joined #mer23:40
*** wubudubudubudu has quit IRC23:42
*** araujo has quit IRC23:44
*** crevetor_ has joined #mer23:49
*** Sfiet_Konstantin has quit IRC23:49
*** imunsie has joined #mer23:58

Generated by 2.11.0 by Marius Gedminas - find it at!