IRC Logs for #crux-devel Saturday, 2012-07-14

Romsteri was expecting more work on xorg, is someone not git pushing what they have done yet?00:16
Romsteri can do some this weekend.00:16
*** mike_k has joined #crux-devel01:57
juegood morning02:55
*** Romster has quit IRC03:01
*** Romster has joined #crux-devel03:01
jueRomster: will do some updates now03:13
mike_kjue: hi! I assume there will be no easy/supported way to update existing x86 CRUX installations, right?03:26
juemike_k: you mean except by ISO?03:28
RomsterExcess Flood, as if silly ircd.03:33
Romsterrolling release updates than to have to use an iso.03:34
juea rolling release isn't that easy for a source-only distro that most people think03:37
Romsterrebuild tool chain in chroot install onto system? then sysup everything else?03:38
juewell, one example:03:38
Romsteri know not easy i've toyed with that idea.03:39
juegrep depends on libpcre, if your libpcre-update breaks grep your system is more or less unusable03:40
mike_kjue: even by ISO )03:40
Romsterhmm probably best to rebuild all of core in a chroot like the iso does.03:40
juemike_k: you mean from x86 to x86_64?03:40
mike_kyes03:41
Romstermike_k, i went from x86 to x86_multilib by updating core ports from the iso.03:41
Romsterthen editing kernel config for 32bit compat and then having to recompile everything else.03:42
Romsterbut i ididn't run the setup off the iso i used for p in *; do pkgadd.....03:42
Romsterin the core ports off the iso.03:42
Romsterwith the root path set to the system.03:43
mike_kRomster: mmm, that is what I mean by "no easy way" )03:43
Romsterifi can do it you should be able to :D03:43
Romstercrux is for experienced users03:43
mike_kyep, may be not so easy on VPS though03:44
mike_knevermind, the question really is "will it be officially supported or not?"03:44
Romsteri could knock up a chroot script to do such a upgrade. but why not pkgadd off the iso.03:45
Romsteron a VPS can't you rebuild it on your system then upload the new image to the VPS?03:45
mike_kif not, some of FS tickets could be fixed without thinking about backwards compatibility that much03:45
mike_kRomster: nah. I have to upload tarball and unpack it from special rescue mode to the XEN disk images03:47
mike_kjue: if it is not going to officially supported, we can use broader range for SYS_* variables in login.conf at least. FS#810 is my real intention here )03:49
mike_k*login.defs03:49
juemike_k: sorry, I'm a bit busy currently, but TBH I don't get what FS#810 has to do with your 'easy update'?03:57
juemike_k: why should the x86 x86_64 update not work if you do it via the ISO? Of course, you have to rebuild everything installed but not updated by the ISO03:59
juehave to run now, back in some hours04:03
Romstermike_k, should still be doable grab the portsdb file and throw it in a chroot rebuild ports throw into a directory and archive it with the newer portsdb file upload to server04:04
mike_kRomster: sure04:05
mike_kjue: on FS#810: that way we could set SYS_* to overlap current hardcoded defaults (as we do not set them explicilty now) and to rely on the fact user will do a clean install of 3.0 (where those changes to core/shadow might be introduced)04:08
mike_kit is critical issue if ID ranges would get mixed, though. point of the ticket is to allow "useradd" or "groupadd"  with "--system" switch just to work04:11
juere07:45
juemike_k: we talked already about FS#810 long ago, I simply  forgot to implement it with the latest shadow update. But I don't see a reason to wait with it until the next CRUX release?07:48
juemike_k: and again, I don't understand why this is releated to the 'easy update' issue07:50
mike_kjue: I thought without an update in mind it would be easier to extend SYS_* range to allow a lot more slots. that's all.08:01
mike_kexisting systems might already have IDs used within that (possibly) extended range08:02
mike_knot a bg deal to talk about so much, though!08:03
jueyeah, I see now. It sounds to me that you set the 'easy upgrade from x86 to x86_64' in some relation to FS#810, that was my problem08:05
juebesides the possibilty to use a broader range of ID's for a fresh install, which isn't doable in practice because we always have to deal with CRUX updates08:09
juemike_k: do you have the patch of FS#810 somewhere, the paste within the ticket doesn't work anymore08:11
mike_kjue: just a moment08:12
mike_khttps://gist.github.com/3a04349a14e0ef3e68b508:14
juethanks08:17
mike_knp08:17
jueRomster: I'm ready with xorg for today, I've done the update from top to xorg-font-xfree86-type108:35
jueRomster: I left out 'xorg' intentionally, so we can use that port as the final one08:36
Romsterk08:38
jueRomster: IIRC you had problems commiting to some other repo on crux.nu, let's try it with xorg as long as I'm here08:40
Romsterso were following the versions in http://www.x.org/releases/X11R7.7/changelog.html for R7.7 than the all updated versions that i did to my xorg fork.08:40
Romsterneed to clone it first.08:40
teK_romster's access rights should have been fixed if you mean that, jue08:41
jueteK_: thanks, no idea what I meant, just remember that there was something ... :)08:43
Romsterok got git cloned and on 2.7 branch08:43
Romsteryeah some object permissions last time.08:43
Romsterin opt-x86_6408:43
Romsterjue, do you want me to bump xorg port to 7.7? or some other ports still need checking for versions.08:45
Romsterhmm next port after that from top down is xorg-gccmakedep08:46
jueRomster: no08:46
jue15:36 < jue> Romster: I left out 'xorg' intentionally, so we can use that port as the final one08:46
Romsterright get ya now08:47
Romsteri can change the maintainer on xorg-gccmakedep no version bump on that port.08:47
Romsterok seems fine.08:49
jueyeah, works :)08:55
Romsteri'll do a bunch of ports until i get sick of it.08:57
frinnstwireshark broke with the new glib? or is it just me?09:00
teK_moo09:01
teK_testing.09:01
Romsterbusy atm to test09:01
Romsterjue, are you phased on the commit string update to x.x.x than the one i'm used to using a.a.a -> x.x.x09:03
jueRomster: no, do what you want09:05
juefrinnst: works for me09:05
frinnstok, my toolchain then09:05
frinnstbtw, is there an "official" repo for the next-toolchain?09:06
juehmm, good question, I'd say no or not yet09:07
teK_works here, too09:09
Romsterwireshark is fine confirmed it here.09:10
Romsteron 32 bit at least09:10
juesometimes a test repo would be nice for all kind of preliminary stuff09:11
Romstersandbox.igt perhaps.09:12
Romstergit*09:12
teK_I could arrange that as soon as I finished this friggling paper..09:13
Romstergetting too tired up to xorg-inputproto i'll do more after sleeping.10:17
*** mike_k has quit IRC15:26
*** acrux has quit IRC18:21
*** acrux has joined #crux-devel18:23
*** joe9 has quit IRC20:58
*** mavrick61 has quit IRC21:47
*** mavrick61 has joined #crux-devel21:48
Romsterresuming xorg updating i can't push this until i've bumped all the major versions as this and that depends on this and i can't build that without bumping this. if that makes any sense :D22:20
prologichmm22:29
prologicwhat's the conclusion of crux-next so far?22:30
prologicI've sort of been following22:30
prologicwe going pure 64bit?22:30
prologic(we should)22:30
prologicdrop all 32bit support ideally22:30
Romsteri think it's undecided still22:30
prologichmm22:30
prologicfair enough22:30
prologicI think we should adopt frinnst's 64bit repos and merge it all into the main repos22:30
Romsterme and a few others need or would like a minimal multilib gcc but i think the majority would like pure6422:31
prologicand what's current rename/move to x3222:31
prologicI think it's too much work to be honest22:31
Romsteri'd assume majority wins and have some mehod to rebuild with an overlay for multilib22:31
prologicthe way forward is to drop backwards compatibility22:31
prologicif anyone needs it, we'll leave old unmaintained 32bit repos around22:31
Romsternot always an option though22:31
prologiclike what's happened with i586 support22:31
Romsterx86 apps are going to be around alot longer than the i586 chipset ever did.22:32
prologicabsoultely22:32
Romsteranyways i'm working on xorg atm.22:32
prologic32bit is going away quite rapidily though22:32
Romsterwe'll have R7.7 before next crux is out.22:33
prologicso we should not focus too much energy in maintaining 32bit ports22:33
Romstertrue22:33
prologicihmo22:33
jaegerIt'll be years yet before 32-bit is gone, if ever22:33
prologicnah22:33
prologicnot years22:33
prologicmaybe 2-3 at most22:33
jaegerI doubt it will ever fully go away while CPU manufacturers continue to support it22:33
prologiceven arm is quickly going 64bit22:33
jaegerI think you're being extremely optimistic about that but I guess we shall see22:33
Romstersmethings actually are better done in 32bit too. heck small micros will probably stay 32bit for meany years22:33
prologicin any case22:33
prologicI think I will try to contribute more to crux's efforts in a few months22:33
prologicwhen I buy a new desktop :)22:34
prologicthis one is 6+ years old22:34
prologicand I run 32bit crux 2.6 on it22:34
prologicso much stuff out of date22:34
Romsterand perhaps clean update your ports prologic ?22:34
Romsteryour on 2.6 still O_O22:34
prologicprobably rm -rf a lot of it22:34
Romsterthat expalns the lack up port updates22:34
prologicyup22:34
prologicneed to start over :)22:35
prologicfresh install on fresh hardware22:35
Romsteri've pinched a bunch off your ports in contrib too anyways22:35
prologiceh nps :)22:35
Romsterprobably a few i'll drop that i'm not using that much22:35
prologicI think going forward I'm going to separate out my personal ports a bit22:35
Romsterbut feel free to join contrib and maintain them again.22:36
prologicmake my prologic collection just that, a collection of crap - not necessarily up-to-date22:36
prologicI want to maintain a separate mythtv repo for example22:36
Romsterwhy not in contrib for mythtv22:36
prologicyeah I plan to22:36
Romsterthen again you could do that and keep using your hg than git22:36
prologicbut it makes contrib a big massive unorganized collection you see22:36
Romsternot really unorganized is it?22:37
prologicno structure :)22:37
prologicit would be good if contrib has sub-repos22:37
Romsterif everyone pitches in to maintain there ports22:37
prologicof22:37
prologicofc22:37
Romsterhrmm22:37
prologicaynway22:37
prologicthat's my plan22:37
Romsterexpalin this sub-repos idea22:37
prologica tree of repos22:38
prologicmercurial supports repos within repos for example22:38
prologicsurely git does22:38
Romsterlike meta ports that add other foo.rsync files22:38
prologicso you'd have for example multiple repos22:38
prologiceg: http://hg.crux.org/contrib/mythtv22:38
Romstersure git has sub-modules22:38
Romsterbut the problem there is you'll find dependencies between sub-repos22:39
Romsterthe majority being what i got in contrib22:39
prologicthat shouldn't matter22:39
Romstersince i've been trying to support the major stuff.22:39
prologicbecause clonging contrib will clone all subrepos too22:39
prologichmm22:39
prologicI should register cruxlinux.org22:40
Romsterwhy?22:40
prologicwhy not ? :)22:40
Romstergiven your history of domains coming and going i'm not really enthusiastic22:41
prologicheh22:41
Romsterwhy nto jsut a crux.shortcircuit.net 3rd level domain22:41
prologicI'll have you know I've kept on to quite a few22:41
prologicshortcircuit.net.au softcircuit.com.au softcircuit.net.au sahriswiki.org sahriswiki.com circuitsweb.com circuitsframework.com22:42
prologic:)22:42
Romsteryeah but i do also see dead links everywhere to your stuff too22:42
prologicyeah that's true22:42
prologicthere's a good reason for that22:42
prologicwho has the time :)22:42
Romsterlike i could of suggested your bot irc bot to someone looking for a cool bot but the links are dead.22:42
prologicyou spend 12hrs a day at work including transportation22:42
prologicno one has time for anything22:42
Romsteryeah makes it hard22:43
prologicsure does22:43
prologicyou mean kdb?22:43
Romsteryeah22:43
prologickdb's code is still available22:43
Romsteri'd hate to see that die off.22:43
prologicnot sure if it works though :)22:43
Romsterthe pypy site needs updating.22:43
prologicit probably needs a bit of a compatibility update22:43
prologicas circuits' API has changed a bit over the years22:43
Romstersee what i mean, i could of suggested that to a person that wants to get rid of another bot that's really a pain to maintain.22:44
prologicyeah22:45
prologicwell kdb is pretty easy to maintain22:45
prologicbut I think it's core is broken atm22:45
prologicI should fix it22:45
Romsterprobably is22:45
prologicmight try to do a few things on the bus in the coming weeks22:45
prologicjust setting up my new macbook air with Xcode, Python and all my development tools22:45
Romsteri haven't got alot of time either but i try my best to help out at crux.22:46
prologicheh22:46
prologictry working, having a family life and baby on the way :)22:46
Romsteroh you never said anything about you being married again22:47
Romsterand a child... onthe way22:47
prologiclol22:47
Romsteryou really keep to yourself too much thesedays22:47
prologicmarried again?22:47
prologicdude I was never married22:47
prologicbut soon will be I guess22:47
Romsteroh that last one you were not married too?22:48
prologicyeah I don't have much of a public/online life heh22:48
prologicthings like Facebook, Google+, etc are a waste of my time22:48
prologicnope22:48
prologicand that was about 5 years ago :)22:48
Romsteryeah i don't use them socal sites22:48
Romsteronly irc22:48
prologicI've afraid these days - since about 5 or 6 years, I keep very little online that's considered private/family/etc22:49
prologicI find the whole concept of a social network stupid and arcane22:49
Romsterbest way to go really22:49
prologicthing is we've had social networks for the past 15 years22:49
prologicevery since the creation of irc in 199422:49
Romsterthought it was 9322:50
prologic'94 was when it really took off22:50
Romsterah22:50
Romsternot sure why some of these xorg modules are not mentioned anymore like libxkbui22:54
Romsteron http://www.x.org/releases/X11R7.7/changelog.html22:54
Romsteri'd assume it's obsolete, will have to come back to those and see what really needs them anymore.22:54
*** c0x` has joined #crux-devel23:23
*** c0x has quit IRC23:24

Generated by irclog2html.py 2.11.0 by Marius Gedminas - find it at mg.pov.lt!