IRC Logs for #crux-devel Tuesday, 2008-11-25

*** treach has quit IRC00:11
*** pedja has quit IRC00:23
*** tilman has quit IRC01:04
*** sepen has joined #crux-devel01:14
*** tilman has joined #crux-devel01:15
*** mike_k has joined #crux-devel03:34
*** DarkNekros has joined #crux-devel05:52
*** mike_ has joined #crux-devel06:38
*** mike_ is now known as Guest9421406:38
*** Guest94214 has quit IRC06:41
*** mike_ has joined #crux-devel06:41
*** mike_ is now known as Guest9968506:42
*** mike_k has quit IRC06:57
*** mike_k has joined #crux-devel07:49
*** sepen has quit IRC10:31
*** pitillo has joined #crux-devel11:22
*** sepen has joined #crux-devel11:37
*** jue has joined #crux-devel11:44
*** Rotwang has joined #crux-devel11:50
*** sip has joined #crux-devel12:57
sepenheyo sip12:58
tilmani'm running a hotpluggy x server13:02
tilmanbuilt with dbus and hal support13:02
tilmanworksforme, too13:02
tilmandeps are dbus, hal, hal-info13:03
sepensounds ok for me (and xfce)13:03
tilmanwhat i still don't know is: does it fix the input driver problems in xserver 1.5.3?13:04
juewell, if the hotplug stuff works correct we should add it to 2.513:07
jueat least this was our idea the last time we talked about that13:09
jaegertilman: when you say hotpluggy do you mean just without an xorg.conf or an xorg.conf that defines no input devices or what?13:09
tilmani started X without having xorg.conf13:11
tilmanwhen x was running, i plugged in my usb mouse13:11
tilmanand *ziomg* it worke13:11
jaegerhow does that affect binary blob video driver users?13:11
tilmani think the x people didn't want xserver to auto-detect the nvidia driver13:12
tilmans/detect/load/ i guess13:12
jaegerpresumably those of us with proprietary drivers can still use an xorg.conf13:14
jaegerI wonder if we can leave input sections out and get the best of both worlds13:14
jaegerthat would be cool13:15
tilmanyou'll probably want an xorg.conf too if you have a dual head setup13:15
tilmanby default my radeon comes up in clone mode, which is... silly13:15
tilmanjaeger: very probably we can just leave out the input sections, yes :)13:15
tilmani also stuck 'dbus' and 'hald' in SERVICES13:17
tilmanif you want to give it a try, too, then just build xorg-server while having dbus and hal installed13:18
tilmanit's enabled automatically then13:18
juewhat happens if haven't started the daemons?13:19
tilmani don't know :D13:19
tilmanlet me try that13:19
sipjue, I plan to drop qt3 for 2.513:20
sipI think you have qcad or something depending on it13:20
juesip, no problem, I'll drop qcad as well ;-)13:20
jueit's a nice CAD system, but I don't need it anymore13:21
sipunfortunately this means xxdiff is going away too13:22
sipstill struggling to find a valid alternative13:22
*** treach has joined #crux-devel13:22
tilmanthat wins you a free reboot via ssh :D13:23
tilmanie, i got neither keyboard nor mouse13:23
*** thrice` has joined #crux-devel13:23
jaegernice, I'll give it a try13:24
sipsomebody may point you to a plethora of virtualization solutions :-P13:24
* jue begins to hate all that blown desktop s***13:24
Rotwangnice avatars13:25
jueadding dbus and hald to SERVICES as default is a no-go IMO13:26
juebut, well, you win the same free reboot if you forgot to add evdev to your kernel options and run x after a X -config13:29
tilmanwell, i'm not saying we really should ship a hotplug-capable x server13:30
tilmanwe can well leave that decision to the user13:30
sipwell we should at least change the deps13:31
sipwhich is fine for me btw13:31
juetilman: seems that we need some kind of doc anyway, so ...13:31
tilmanif we're not encouring the user to start dbus and hal on boot, then it's kind of pointless to make xserver depend on those, isn't it? ;D13:31
sipyeah but activating it later would require re-compiling13:32
sipsepen, thanks. fatty deps tho ;-)13:33
juetilman: adding dbus and hald to SERVICES means htat they have to be core ports13:33
sepenthat would convert crux in a desktop based kiss distro13:34
sipit's bad enough when udev randomly renames my interfaces13:34
tilmanthat would suck :D13:35
tilmanit seems like we don't change anything wrt xserver? ;D13:37
tilmanof course i can document how to get a hotplug-capable xserver in the wiki ;)13:38
sipas I said, I vote for shipping a hotplug-capable server while having no changes in core ports / rc.conf13:39
sepenimho most people compile their own xserver after installation, so all you need is these deps before13:39
siptilman, how many ports wuld one have to recompile?13:39
tilmanjust xserver13:39
sipoh. nice. so xserver-hal is an option if you prefer to maintain the status quo13:40
tilmana separate port? :|13:40
sepenI prefer the nicetohave meta13:41
sipjust an idea. what happened to your crusade for keeping the iso small? :-)13:41
juewhat's nicetohave ?13:41
sepenjue, # Nice to have: hal13:42
tilman"optional" "dependencies" ;)13:42
tilmansip: :D13:42
jueI don't know that13:42
sepenwe discussed sometiem ago about Nicetohave or Optional meta too13:42
pitillowhy not the little referenfe/doc like tilman told? I think it's a good way to don't touch core and services13:43
pitilloand let the user the option of having old way Xserver or the new hotplug one13:43
tilmanyeah :P13:43
juetilman: what's your opinion?13:43
treachNote in the handbook could be nicetohave, too. ;)13:44
tilmanif we make the port depend on dbus/hal, then we should also run those daemons on startup13:44
tilmanto be consistent13:44
tilmansince we don't want that, i think i wouldn't want to make the port depend on those two13:44
tilmanso yeah, documenting howto-get-a-hotplug-capable-xserver in the wiki sounds good to me13:45
sepenfor me Wiki page and Nicetohave meta are ok13:45
jueok, fine for me, but without that nicetohave, please13:46
pitillotilman, for me too, there are WM which has hal/dbus deps or to add some features, but I can't see the point of adding those services to get a clean little server running. Documenting that, can be the best way (IMO)13:46
tilmanjue: yes13:47
sepenjue, ok but atm we have 11 ports in opt with this meta13:48
tilmanwe do?13:48
tilmani grepped before and didn't find anything13:48
sepen0 for xorg atm13:48
sepen$ grep '# Nice to have' /usr/ports/opt/*/Pkgfile | wc -l13:48
tilmanoh, i fail at running grep, hahaha13:48
sepen0 # Optional too13:48
juebut none of them are on the iso, right?13:49
sepenjust I tried to be consistent too ;-)13:49
sepenjue, yep, forme its also ok13:49
tilmananyhow, i agree that # Nice to have should be used sparingly ;)13:49
treachwhy not decide on one form, and make prt-verify fail/warn on abberations?13:50
sepenyep, also is not used by prt-get for now13:50
sepentreach, what kind of abberations?13:51
juewell, I think we should stop here, we have more important stuff left13:51
tilmanokay, what's next?13:51
jueI've been running 2.9 here today13:51
jueyep, I thought it's a must to do some test at least13:52
juewell, what should I say, all core ports build fine13:53
juehaven't had the time to do more tests, mc is partly broken13:53
sipmaybe a RC with glibc 2.9 would help13:53
juedidn't know why13:54
juesip: if I had not found the mc problem, I'd say yes, but so13:56
jueTBH, I'm unsure what's best13:57
sepensqlite appears in the iso, so is it fixed in 3.6.3?13:58
jueno, upstream don't accept it13:58
jueif we want a rock-solid 2.5 _soon_, we have to stick with 2.8 IMO14:00
sepenwe can ship 2.9 from ports14:01
sipthat would be a mess14:02
sepenafter iso was released14:02
tilman2.8 in 2.5 is okay for me14:03
sepens/was/will be/ :P14:05
jueyep, my thoughts as well, but howling quietly too14:07
jueno objections?14:10
tilmannope :]14:10
juecool, so we build a new iso, probably a 2.5-rc?14:11
sepenI there are no mayor changes for me its ok14:12
tilmanyes please :)14:12
jue.oO jaeger ...14:12
tilmanjaeger: the mtab issue that you saw was with an old iso, right?14:13
sipoh, I've also seen some problem with loading the ide-generic kernel module on test314:13
sipI think it was 'Could not allocate memory'14:14
sipduring squashfs magic14:14
juehmm, jaeger has reported something similar14:15
sipI was ready to archive it under the 'don't buy cheap cds' file :-)14:15
jueI've no fear to remove ide-generic module at all14:17
jaegertilman: yeah14:17
jaegeralso, I'm building a 2.5 iso right now if you want to use it as the rc14:17
jaegerI'll test it and see if it has the squashfs issue14:18
jueyes, we are waiting for it :-)14:18
sipdoes it include gnome?14:18
jaegerwhy would it include gnome? :P14:18
* sip hides14:18
sepenjue, the cpio issue appears newly?14:18
jaegerI might add a target to the Makefile to build static unsquashfs as well if someone hasn't already done that14:18
juehopefully fixed it14:19
juejaeger: thanks, that was on my todo list14:19
sepenhmm was that14:22
sepennice :-)14:22
juetilman: any news on #380?14:22
* tilman hides in a corner14:23
tilmanjue: no :|14:23
tilmani'll poke cptn about it again :D14:23
sipmmh the handbook may need a revamp if we're building a rc14:26
juewhich parts of it?14:28
sipreference to crux-2.4.iso14:29
juecptn added a chapter about local generation already14:29
sipor we can replace it with crux-x.y and be ok for the next 20 years ;-)14:30
sipsame thing for kernel references14:31
*** tri has joined #crux-devel14:33
sepenwell also its possible to replace 2.4 for 2.5 when using the 'get_wiki_handbook' script14:34
sipgood call14:34
sepenbut webpage need also to be maintained14:35
jaegerdoes that look ok?14:35
jaegerah, that doesn't fix the location from which unsquashfs gets copied but I'll add that14:37
sepenalso the .PHONY14:38
jaeger <-- this one's better14:38
jaegershould be trivial to move to .7, I'd think14:41
jaegershould I go ahead and commit the unsquashfs diff?14:46
tilmancptn: maybe we can have a look at #380 together tomorrow or so?14:47
tilmani'm off for now, poke me hard when i need to fix something14:47
*** thrice` has left #crux-devel14:49
juesip: if you are there, our release notes need some love as well14:51
juejaeger: have to leave now as well, do you change the kernel stuff or should I do it tomorrow?15:00
jaegerI will do it if you like15:01
juethanks, please do it15:01
juenight all15:02
jaegertake care15:02
*** jue has quit IRC15:02
*** mike_k has quit IRC15:41
*** sip has quit IRC15:43
*** tri has quit IRC16:01
*** DarkNekros has quit IRC16:03
*** Rotwang has quit IRC16:20
*** pitillo has quit IRC16:26
*** Rotwang has joined #crux-devel16:34
*** Rotwang has quit IRC17:12
*** tilman has quit IRC17:15
*** tilman has joined #crux-devel17:18
*** sepen has quit IRC18:22
*** treach has quit IRC19:49

Generated by 2.11.0 by Marius Gedminas - find it at!