IRC Logs for #crux-devel Thursday, 2011-12-22

Romsterwhy do we even have to have -m64 you'd think that would be the default?00:46
frinnstjaeger: damnit, prioritize!01:41
*** sepen has joined #crux-devel02:12
teK_jaeger: yeah :)02:41
*** mike_k has joined #crux-devel04:00
sepenpedja: ping05:22
pedjasepen: pong :)05:50
pedjafrinnst: I still can't push to contrib64 repo.05:53
sepenpedja: are you running vte-0.28.2? the patch you pasted here can't be applied since keymap.h has the right values
pedjasepen: sorry, should be reverse applied, that patch broke vte.06:40
pedjaso, that line should be "#define vte_meta_mask gdk_mod1_mask" after reversing.06:42
*** acrux|G4 has joined #crux-devel06:43
*** acrux|G4 has joined #crux-devel06:43
pedjasorry if I forgot to say that :)06:43
sepenah ok06:44
*** acrux|G4 has quit IRC06:44
jaegerRomster: it isn't required, technically. I put it in for clarity but might take it out later07:57
jaegerfrinnst, teK_: hoping to have some time this afternoon to play with it :) I'm having some company for lunch today but after that07:57
teK_I'm having some trouble getting elilo to boot sometimes07:58
teK_it just restarts07:58
teK_-> ??07:59
jaegerI did some reading about UEFI booting last night, there's a crazy amount of documentation out there about it08:00
jaegermostly blogs but still08:00
frinnstteK_: can you add pedja to +w in contrib-x86_64?08:38
sepenfrinnst: $ getent group contribx866408:41
sepenI think he should be part of this group08:42
sepenshould not be enough?08:42
sepenah ok08:42
frinnstno, it should be enough08:42
frinnstdidnt know you had privs :)08:42
sepenno, I don't have privs08:43
frinnsthm. shouldnt pedja already exist in optx8664 ?08:45
sepenfrinnst: but take a look about pedja's GID08:45
sepen^^ $ getent passwd pedja08:45
sepenin we're using gitosis to administrate those git accounts, etc. and to avoid to use git-shell or alternative fake shells08:46
sepenso you don't need to have a local account per user contributing to your git repos08:47 uses gitosis, right?08:48
sepenI think so08:51
sepengitosis is just an extra git-repo which have some hooks with privileges on the local system08:52
jaegerI use gitosis for all my stuff, too09:02
jaegerIt's pretty slick09:02 uses gitolite, doesn't it?09:09
rmullnot gitosis?09:09
pedjarmull: you are correct.09:12
*** sepen has quit IRC10:37
*** frinnst has quit IRC10:42
*** frinnst has joined #crux-devel11:43
*** frinnst has joined #crux-devel11:43
teK_frinnst: pedja still "relevant"?12:18
teK_will take ASAP ;)14:20
jaegerok, bootstrapped another ISO with the CFLAGS changes15:17
*** mike_k has quit IRC15:37
*** mike_k_ has joined #crux-devel15:37
*** mike_k_ has quit IRC16:29
teK_for pedja17:22
enteis there a group password? ;)18:45
jaegermanaged to convert my x220's drive to GPT and switch to UEFI only, haven't installed CRUX on it yet19:08
rmullOooh, I want to do gpt19:09
entemy server has a GPT partition table on FreeBSD19:25
enteI heard MBR partitions don't work with drives >2TB, so I went with GPT to be safe19:25
enteusually I did MBR partitions + disklabels19:26
entebut FreeBSD deprecated disklabels19:26
jaegerhrm... not sure how to install CRUX now that I've converted it to GPT19:28
jaegernot all the partitions show up in /dev19:28
Romsterbe cool if gitosis was on portdb19:35
Romsteri'm using the shell method for a few users on my system.19:36
Romsterhmm here is an idea, should portsdb be expanded to have different categories for cruxppc, multilib, x86_64 arm and the normal 32bit ports lists?19:45
Romsterhalf of the ports people talk about are on other arches and not to be found anywhere like gitosis for example.19:46
Romsterless effort to grab a existing Pkgfile and port it for the particular arch.19:46
Romsteror know that some repo is for 64bit etc.19:46
entejaeger: wasn't there a way to re-read the partition table?20:00
Romsterhdparm -z DEVICE20:02
Romsterbbl back to work20:02
enteI wanted to look at the hdparm manpage, and then realised that I don't have a linux box anymore where I can read linux manpages20:03
entethat feels rather weird20:03
jaegerente: I think it's that I don't have GPT support in the testing kernel20:06
jaegerhrmm.... now to try to figure out how to make the ISO UEFI bootable20:45
jaegeryeah, that was what prevented me from seeing the partitions21:13
jaegerwoot! managed to boot CRUX with grub2 via EFI. was kinda a pain in the ass, though22:33
jaegerHrmm, something I've been wondering for a while - I haven't had a /dev/cdrom in some time, I wonder why it's not created... /dev/sr0 exists and works fine22:45
Romsterudev rules23:07
Romsteri'm off work early nothing going on there.23:09
jaegeryeah, I imagine it's udev's fault, just don't know why it changed or how long ago23:17
Romsterwho needs dvd and cdrom when you have sr0 anyways?23:27
Romsterah great now my safe-build screws up on --host= now... starting to thik i need to stop using chroots and use a VM.23:31
jaegerI much prefer VMs to chroots these days23:34
Romsteris it much slower on compiling though and can i bind mount ccache and distfiles to a VM?23:34
Romsterprobably wrong person to ask about ccache but the later.23:34
jaegerdepends on the VM software, I imagine23:35
jaegerif you use kvm, for example, bind mounting would probably work23:35
Romstergot some config for a VM that i could use since yore the master on the VM front.23:35
jaegerbut not likely with virtualbox or vmware server, I think23:35
Romsterwhat are you using?23:35
jaegerhowever, you can nfs mount or use shared folders and cifs23:35
jaegerI use virtualbox23:35
jaegerwell, for the crux stuff, at least.23:35
jaegerat work I use vmware vsphere 523:36
Romsteri haven't used that in a while anything i got to look out for or is it simple to setup.23:36
jaegerwow, X -> console switching is really fast with KMS, I dig it23:36
jaegerIt's generally pretty simple23:37
Romsterok i should start by recompiling virtualbox for the 64bit host first.23:37
Romsteri got xmas hollidays and 2 weeks of time off work hollidays i'd like to get some crap sorted and test everything too.23:38
jaegerI ran another successful bootstrap last night with -m32/-m64 in CFLAGS instead of CC23:41
jaegerGot my new laptop successfully dual booting windows 7 pro and crux multilib with UEFI and GPT using GRUB223:42
jaegerpretty neat, I even get a "GRUB2" entry in the UEFI that I can move around in the priority list23:43
Romsteryeah i have been reading, you had to remove -m64 out of gcc port though?23:43
jaegerIf -m64 is set in the 64-bit section of pkgmk.conf, the gcc build uses it for both the 64-bit part (which is fine) and the 32-bit part (which fails)23:43
jaegerso I stripped it out manually and the build works23:43
jaegerIf I remove the CFLAGS settings entirely from the 64-bit setup it *should* work fine23:44
Romsteryou seen my hacks to contrib/gcj seems to improve performance, i applied that to my last hvl build i haven't done it yet to multilib but i plan too.23:44
jaegerthat's how the laptop is running now, will see if anything breaks23:44
jaegerI've no use for gcj so I hadn't seen those, no23:44
Romsteryeah ditch the -m42 altogether shouldn't be needed.23:44
jaegeryeah, it was only there for clarity originally23:45
Romsterlooking at my safe-build script it's almighty long for such a simple task that can be done in ~10 lines. i feel inclined to rewrite it still.23:48
jaeger <-- this is the pkgmk.conf I'm using to test on the laptop23:48
Romsteri was thinking about LDFLAGS don't you also need to do -wl,-m32 to make sure it's passed to ld? also my man pages seem to not parse correctly as root but ok as a normal user.23:51
Romsterstuff like that.23:51
Romsterthough i haven't rebooted in a while after rebuilding stuff that could be why.23:51
jaegerno idea on -Wl,-m32, haven't needed it so far23:52
jaegerHaven't seen any man page issues yet, any specific ones?23:52
Romsterdid you add them 2 ports i pasted the other day to multilib? thsoe where enough to get wine to compile for me.23:52
Romsterman hdparm23:52
jaegerI just looked at the udevd(8) page as root and non-root23:52
Romsteras root user.23:52
Romstermight be my terminal too.23:52
jaegeryeah, looks fine for both root and me23:53
Romsterusing sakura23:53
Romstermight fix it self after a reboot more than likely.23:53
jaegerNo, haven't added them because I didn't know what they were for23:53
jaegerI don't have them installed, wine didn't need them23:53
Romstermine wouldn't compile without it.23:53
jaegerI'm going to guess that's another side-effect of the install method23:54
Romsterguess i'll add them to my own overlay before i do a sysup and pull in any changes you've done.23:54
jaegerabout to start a wine build on the laptop23:55
Romsterbtw can you bump python with that patch in multilib if you haven't done yet, i'm going to bump my kernel later.23:55
jaegerwill see if it causes problems23:55
jaegerWhat python patch is that?23:55
Romsteralso wine 35 is out.23:55
Romsterone in opt/python23:55
jaeger.35 is already in the repo23:55
Romsteroh i'm on 34 still guess i really need to sysup.23:55
Romsteri been holding off on nuking them other two ports though they might survive a sysup not being tracked in the REPO but i don't like my chances so i'll move them first.23:56
jaegerthe python port is frinnst's, there's not a multilib-specific one23:56
jaegerunless you specifically meant python-3223:56
Romsterpyhton-32 yes sorry that nees that fix as well.23:57
Romstergod damn my typing too, i really need to improve it.23:57
jaegerinteresting, what does the patch fix? I mean I can see by reading it that it forces "linux2" as the MACHDEP setting but what does that change?23:57
Romstercompiling pyhton with a 3.x.x kernel version string.23:58
jaegerIs that bad?23:58
Romsteri havne't tried yet but i think it wont compile with the newer kernels version string.23:58
Romsterit expects 2.6 still23:58
jaegerI'm running 3.1.5 for what that's worth23:59
Romsterthat patch fixes it.23:59
jaegeron all my multilib installs23:59
Romsterand it worked *shrugs* no idea what that's for then, that's what i thought it was meant to fix.23:59
jaegerI don't mind adding it if it's really needed, I just can't tell by looking at it23:59

Generated by 2.11.0 by Marius Gedminas - find it at!