IRC Logs for #crux-devel Thursday, 2012-12-27

*** mike_k has joined #crux-devel01:06
juegood morning03:52
jueRomster: to view plain ps I mostly use gv03:53
juejaeger: I'll commit new versions of gawk, gettext and iptables today, probably worth to include into a new rc03:56
Romsteryeah found gv after awhile.03:58
jueRomster: if you don't like the cups web-interface you can globally set the default printer with lpadmin or for each user with lpoptions03:58
Romsteri couldn't get my direct cd print to work...03:59
Romsteri got it to print on A4 after setting default printer though...03:59
Romstereverytime i tried even after setting the default options ot cd tray, the mp640 cannon printer said processing then cancelling...04:00
Romsteri've resorted to windows on laptop to print onto cd for the time being.04:00
Romsteras i need to get this done.04:00
Romsteri would of preferred an all linux solution.04:00
Romstermy only guess is it might be gutenprint does not support cd tray printing perhaps.04:01
juenothing helpful in the cups logs?04:01
Romsterheck i never looked in those...04:01
RomsterBad driver information file "/usr/share/cups/model/Canon-PIXMA_MP640.ppd"!04:03
RomsterBad driver information file "/usr/share/cups/drv/cupsfilters.drv"!04:03
Romsterhmm04:03
Romster[Job 16] Failed to connect to system bus04:04
Romsternothing else interesting for today04:04
Romsterwhy would it say that i got that ppd file off the manufactures web site.04:04
Romsterand is printing fine on A404:05
Romsterweb interface isn't a issue, but i did have noscript blocking me changing printer to default for lpq to report.04:07
Romsteri'm too nooby with cups it'll take me some time to get over these hurdles.04:13
jueyeah, printing isn't allway funny with linux04:18
juelooks like your printer isn't full supported?04:19
Romsterno idea04:22
Romsterit's not even manufactured anymore so is considered an obsolete model now. but i bought it just before all the stocks ran out.04:22
Romsterit prints fine and i got this on it http://www.rihac.com.au/mp640-ciss-p-577.html04:24
jueRomster: what's defined as cupsFilter in your ppd?04:34
jueRomster: have you seen the cnijfilter, guess that's what you need for your printer?04:40
jueI've downloaded it from http://support-asia.canon-asia.com/contents/ASIA/EN/0100236704.html04:40
jueit includes a ppd file for your printer and a driver called pstocanonij (besides some other stuff)04:41
juethe source tarball is called cnijfilter-source-3.20-1.tar.gz04:43
Romsterjue, yeah i started to package that but i stopped when it required 32bits to work and then i got guttenprint to work04:55
Romsterhttp://dpaste.com/853434/04:55
Romsterseems i'll have todo aditional work to get it working04:56
jueRomster: why do you use the binary package?05:21
juefor a first test compile pstocanonij, cp that to /usr/lib/cups/filter and cp canonmp640.ppd to05:24
jue/usr/share/cups/model05:24
Romsteroh source... well i found some blog entry05:29
Romsterand i followed that....05:29
Romsterbut i gave up on that port05:30
Romsteri suck /usr/share/cups/model/Canon-PIXMA_MP640.ppd is zero bytes -_-05:31
Romsteri'll try your way later, when i'm less annoyed with it, and after i get all this stuff done.05:32
Romsterneed it for new years eve dj05:32
Romsterusing real cdj pioneer cd players05:33
jueok05:40
frinnsthm, udev fails to compile with stock glibc 2.1705:57
jueI read something related in lfs-dev06:00
* jue looks06:00
* frinnst too06:00
juehttp://linuxfromscratch.org/pipermail/lfs-dev/2012-December/067493.html06:01
juethough, no idea if this is valid for our udev06:02
frinnsthmm, no06:06
jue407 downloads for 3.0-rc1 :)06:07
frinnstnice06:08
frinnstthe only issues i've seen are the regular user-error kernel-compile stuff06:08
frinnstwe still build ext4, scsi-disk, and ahci as modules?06:09
frinnstif we switch those drivers to builtin in the kernel config from the iso we will probably reduce the support questions in #crux by 90% :)06:10
juedunno, but iso.git should show it06:12
juefrinnst: you remember our discussion about that -> http://crux.nu/Wiki/TODO2806:14
jue;)06:14
juewell, I still think that using the config from our boot-kernel is nonsense, because it compiles far too much stuff06:14
frinnstyeah but thats what you start with if you just run "make menuconfig" from the installed sources06:15
frinnstso, redo the config is probably a good idea06:16
juedo we cp our boot-kernel config to /usr/src/.. ?06:16
jueyeah, we do, setup line 24406:20
juehmm06:20
juehaving a nice/minimal default .config would be better, I guess06:22
jueor even nothing like that ;)06:23
pitillohey frinnst, did you try e17 finally?06:25
frinnstnot yet :)06:25
pitillorelated to kernel, may be a wiki page inside one the first points (installation or other one), writting about the basics for all those people having problems to boot the new kernel and don't explain it always in the channel can avoid srepeat the same answer06:27
juefrinnst: if we stick with our boot-kernel config, we should ask jaeger if he sees drawbacks if we build the suggested stuff not as modules06:28
frinnstyep06:28
frinnstbut just including a quick-n-dirty standalone config is probably a better idea06:28
frinnstwith sane defautls06:28
jueyeah, would be nice, but we have to find a volunteer to maintain such a config ;)06:31
juepitillo: indeed, the handbook is very short wrt kernel building, having a wiki-page with requirements like devtmpfs etc. would be helpful06:37
pitilloyeah, nothing extended but at least those requeriments if people isn't used to build a kernel, as this seems the biggest problem people find (here on irc, ML, ...)06:43
Romsterbut, but, that's our natural filter to keep the would be from the persistent that get it to work :D06:44
frinnstjue, I can do it. should be pretty low maintenance unless we start releasing every other month :)06:45
Romsteri'd say no config and a small README on the required options for udev and a mention of ia32 if compat-32 ports is to be used06:45
Romsterguide to compiling kernel, lspci -k ... how to search in make menuconfig06:46
jueRomster: I agree basically, but using our bootkernel-config, which is a very special thing, as the default in /usr/src/.. seems like a bug to me06:49
jueRomster: but as you said, better no .config than that :)06:51
frinnstwas ages since i started out with defconfig.. maybe time i see what it looks like06:52
frinnstwell, atleast ahci is selected :)06:54
frinnstyeah seems pretty sane06:56
frinnsta few redundant stuff selected but as a whole pretty good06:56
frinnstlets just remove the iso-config07:03
juelet's wait for a comment from jaeger, but a +1 from me07:08
Romsteryeah07:09
Romstermaintaining a config is effort, unless someone is going todo that.07:10
jueRomster: as a simplification measure here's a Pkgfile for pstocanonij -> http://dc70bae29ab6f510.paste.se/ :)08:05
juenot tested with cups and a printer of course08:07
juehmm, cnijfilter is probably a better name for the port08:20
jaegerLast time we discussed including a default kernel config I think the consensus was that CRUX users should be advanced enough to build their own without help. That's obviously turning out not to be the case.08:23
jueRomster: http://0020fd2e34a48a7b.paste.se/08:24
juejaeger: I think that starting with kernel's defconfig is better that using our .config for the iso-kernel08:25
jues/that/than/08:25
jaegerthe kernel's defconfig is messy as hell if I recall correctly08:25
jaegerperhaps it's not as bad these days, don't know08:25
juefrinnst: ?08:26
jaegerI see frinnst's comments above, maybe they've done some work to clean it?08:26
Romsterjue, thanks i'll try that after i sleep/08:27
jueno idea, if I ever used it that was a long time ago08:27
Romster--program-suffix=*08:28
Romsterno idea what that's supposed to do.08:28
jueRomster: that selects all ppd files08:28
Romsteroh it's a selection afaik --program-suffix is for appeding something onto the end of files in bin on autoconfig08:29
Romsterbad name for selecting ppd files on there behalf08:30
jueindeed08:30
jueRomster: --program-suffix=<Printer Model Name>08:31
Romsteryou know how wrong that is don't you <_<08:31
Romsterprogram-suffix, what were they thinking.08:31
Romsteranyways i'll leave that in and make a comment on that in the Pkgfile08:32
juethat's from their INSTALL :)08:32
Romsterheck i'll build it now...08:32
juebetter go to bed ...08:33
Romsterme?08:33
jueyep, you said that above :)08:34
Romsteryeah i did, your right building when tired is not a wise idea08:34
juejaeger: maybe I'm wrong, but with our all-modules kernel config it's easy to overlook a [M] setting, I got the impression that most of the problems reported here are that kind of error08:38
jaegerPerhaps08:40
jaegerWe might cause some confusion if the default kernel doesn't support all the same stuff as is listed in the wiki for the ISO image08:41
jaegerThough it would be ridiculous to enable all that stuff08:41
jaegerAt the very least I feel like a default kernel would need: ata_piix, pata_jmicron, ahci, sata_nv, ext2, ext3, ext4 (or ext4 with the legacy support enabled), reiserfs, xfs, jfs08:42
jaegerThat leaves out a lot of older hardware but theoretically if we're mainly supporting 64-bit capable stuff that should also cut out a lot of said older hardware08:42
*** Romster has quit IRC08:47
juehmm, never looked from that side, but if we want a default that has more or less the same features like our iso kernel we should change nothing at all08:47
jaegerThe down side to changing nothing is that the kernel is huge if all is built in08:48
jaegerAny more commits today before I bootstrap again?08:48
juejaeger: no, a static kernel is a no-go than, IMO08:49
juejaeger: no pending commit here08:50
jue+s08:50
jaegerIt wouldn't be much work to maintain one with a few drivers like I listed above, it's just that we have to keep in mind that it's different from the list in the installation docs08:50
juejust looked at the handbook, here we have "The kernel used during installation ..."08:51
juebut well, I've no strong opinion on that matter at all, maybe it's not worth to spend a lot of time for it08:54
jaegerI'm not sure what the best solution is08:55
jaegerPerhaps the best is a compromise - include a short list as <Y> to cut down on the "kernel panic: unable to mount root vfs" questions at install time but leave the ISO kernel intact for installation support08:59
jaegerSome people may still ask about the kernel panics but they would either have hardware outside the list or have changed it manually or the like09:00
juewhere is the short list located? In our handbook?09:02
jaegerI would probably not publicize the short list to avoid confusion09:02
juesorry, I don't get it ;)09:03
jaegerok, I'll rephrase09:04
jaeger1) Leave the ISO kernel/initramfs as it is09:04
jaeger2) include a kernel config with only the list above (or whatever we decide is correct) that gets installed by setup09:04
jaegerI really like having support for almost everything available for the install ISO but a separate config with just the short list built in as <Y> would cut down on a lot of the questions as mentioned09:05
jueok, that sounds like a plan. So we create a .config with defconfig and enable everything you mentioned above as <Y> (if not enabled already)09:10
jaegerworth a try, we'll test it with the rc09:10
jaegersome of these are already enabled in the defconfig09:16
jaegerAHCI and ATA_PIIX are enabled. SATA_NV and PATA_JMICRON are not09:22
jaegerext4 is enabled with 2 and 3 support by default09:22
*** Romster has joined #crux-devel09:22
*** Romster has joined #crux-devel09:22
jaegerreiserfs, jfs, and xfs are not09:22
jaegerwill test it today but need to go AFK for a while. started a bootstrap09:29
*** joe9 has quit IRC10:31
jaegeroooh.... the USB EFI magic works if I just make /dev/sdb a fat32 filesystem instead of using partitions10:46
jaegerI wonder if xorriso knows some magic to make UEFI and isohybrid play nicely14:01
jaegernot very elegant but did something like this for now:14:24
jaegercp kernel/linux-3.7.1/arch/x86/configs/x86_64_defconfig kernel/linux-3.7.1.defconfig14:24
jaegerfor C in SATA_NV PATA_JMICRON JFS_FS REISERFS_FS XFS_FS; do \ echo "CONFIG_$C=y" >> kernel/linux-3.7.1.defconfig; \14:24
jaegerdone14:24
frinnstnice15:35
frinnstdevtmpfs is not included in the defconfig, maybe enable that too?15:36
jaegergood call, should do that15:37
jaegerhttp://pastebin.com/yxLKXDVV <-- thoughts, objections, etc.?15:55
jaegeroops, sec15:55
jaegerhttp://pastebin.com/JgkxsHXz <-- better15:56
jaegertakes me 2m11s to compile the defconfig kernel on my build VM for what that's worth16:21
*** mike_k has quit IRC16:44
Romsterhmm is it wise editing the config directly?19:39
horrorStruckadding the possibility to use localyesconfig in the handbook might also be an option19:42
jaegerRomster: defconfig is a trimmed version of the config anyway, hence the yes "" | make oldconfig21:23
jaegerit constructs the real config from the skeleton defconfig21:23
jaegerhorrorStruck: might be, though I've not looked into that at all yet21:23
Romsterah k jaeger21:28
*** __mavrick61 has quit IRC21:46
*** __mavrick61 has joined #crux-devel21:47

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