**** BEGIN LOGGING AT Wed Mar 21 02:59:56 2007 Mar 21 03:00:01 Much obliged. Mar 21 03:00:06 You work as an EE? Mar 21 03:00:14 yep, but hardware mostly Mar 21 03:00:18 Nice. Mar 21 03:00:41 Measurements are on the wiki; see the section on de-underclocking the slug and other related articles IIRC. Mar 21 03:00:43 I'm planning on building a cluster-in-a-suitcase from WRT54G(S/L)s and slugs. Mar 21 03:00:55 That's not when using full capacity, though. Mar 21 03:00:58 this software stuff is annoying (b/c software folks assume everyone knows all the commands by heart!) Mar 21 03:01:19 Agreed. Mar 21 03:01:30 It's full capacity of the slug -- you'll have to figure your own draw for the USB peripherals you add. Mar 21 03:01:32 I'm just in high school, but I'm looking at a career as an EE. Mar 21 03:01:39 but you know the USB port can't use but 500mA. The other port is unpowered (ferrite bead de-pop'ed) Mar 21 03:01:54 Right. Mar 21 03:01:55 Both are powered, neither is limited Mar 21 03:02:09 There's 3 ports that aren't brought out as well, apparently. Mar 21 03:02:15 Are those supported in Debian? Mar 21 03:02:37 yep. thse should be... but they're also unpowered (only D+/D-) Mar 21 03:02:42 Yep, unless you modify the hw, those don't take extra current, though. Mar 21 03:02:53 That's no problem. Mar 21 03:03:03 So all I need to do is add a connector? Mar 21 03:03:04 agreed Mar 21 03:03:37 Not bad. Mar 21 03:03:45 You'd think these things would cost more than $99, wouldn't you? Mar 21 03:03:50 dangit why is it I have BIOS of laptop set to PS2 mode (NOT ECP) but WinXP still shows ECP printer port (even after I deleted it and rebooted) Mar 21 03:04:12 RobNC, Because the BIOS settings are only for low-level, the OS can override any or all of it that it wants. Mar 21 03:04:25 theblue: yep but carefully. Dunno if you need power, but ground would be useful. Not sure if it's needed; can't recall if it's required per USB spec Mar 21 03:04:25 And with Windows, it can and often does. Mar 21 03:04:42 Once you're adding more ports, it's defenitely out of USB spec. Mar 21 03:04:55 true - how irritating... (windows overriding) Mar 21 03:05:10 Make the Switch, it's fun! Mar 21 03:05:23 theblue_: you should maybe take a look on msntv hacks... it could be easier for such a project (if you're new to embedded stuff and Linux) Mar 21 03:05:31 USB is differential signaling but wouldn't hurt to reference ground in connector. I wouldn't drive +5V since the brick only puts out 2A total. Mar 21 03:05:34 pierrelux, I'm not entirely new. Mar 21 03:05:43 pierrelux, But I don't know where I'd get an MSNTV or how capable they are. Mar 21 03:06:32 RobNC, Well, I could let it draw more and allow bus-powered on the other 3, right? Mar 21 03:06:44 too bad Linksys didn't pin out the serial port; could use extra serial port for other useful stuff Mar 21 03:06:46 I'm planning on throwing it in a suitcase with a big battery. Mar 21 03:07:22 theblue not a good idea. don't forget about the fuse and possibly current-limited components. I'll have to look into it (but that would be my concern). Mar 21 03:07:29 Makes sense. Mar 21 03:07:47 12V battery supplies the WRTs and USB HDDs, then into a DC-DC converter to the slugs. Mar 21 03:08:22 it's a nice project you have but there's a significant amount of work behind... I don't know if any openmosix stuff has been yet ported to the slug Mar 21 03:08:29 openmosix? Mar 21 03:08:45 gotcha... you could tap off the 5V (assuming you have another brick) to power the extra 3 devices (500mA is max current per USB spec) Mar 21 03:09:13 pierrelux, Even if it's not a true cluster, it'd be a nice little network, eh? Mar 21 03:09:44 I wouldn't use the board power for that, for sure! devices such as USB hard drives draw a significant current during spinup and could cause voltage drooping and thus IX420 reset due to power-on reset circuitry Mar 21 03:09:59 But the USB HDD is powered by its own 12V supply. Mar 21 03:10:12 ah no problem then! Mar 21 03:10:16 It could probably get away w/o a +5V line. Mar 21 03:10:34 yeah true but I wouldn't leave the ground unconnected - just to eliminate ground loops Mar 21 03:10:39 Defenitely. Mar 21 03:10:49 But if you have ground, don't you have to have the +5V line? Mar 21 03:10:55 no Mar 21 03:10:59 theblue_: yes it would, but a little expensive though http://en.wikipedia.org/wiki/Openmosix Mar 21 03:11:00 Sounds good to me! Mar 21 03:11:17 pierrelux, Could be a lot more expensive. Mar 21 03:11:42 ground is just insurance, in case the jokers that made the 12V supplies didn't let the outputs float (which would float to 120V/2!!) Mar 21 03:12:10 No, what I'm saying is that everything's just hooked up to a UPS battery inside a briefcase. Mar 21 03:12:19 And the slugs go through a voltage converter to 5V. Mar 21 03:12:45 No AC inside at all,. Mar 21 03:13:14 oh okay so long as 12V and 5V both have same ground, then, yes, no need to run the USB ground either Mar 21 03:13:31 Not bad! Mar 21 03:14:22 Should I expect much RF irritation, with 2 or 3 slugs and 4 or 5 routers and a switch in really close to each other? Mar 21 03:15:22 no, but I hear that the NSLU2 included some ethernet cable with a ferrite bead, probably to pass part 15 FCC. I don't think it'd be a problem. I'd be more worried about heat! Mar 21 03:15:44 Yeah, it did. Mar 21 03:16:08 What do you recommend for heat, then? Mar 21 03:16:17 I know that most air-cooling is out of the question? Mar 21 03:16:55 I hear that metal enclosures do best. I would avoid anything with a fan (as they wear out and produce noise). Best to have a thick case that can thermally dissipate heat via conduction and surface area of case Mar 21 03:16:58 IMHO of course Mar 21 03:17:11 That's gonna be tricky. Mar 21 03:17:37 HDDs don't like heat that can't escape. Mar 21 03:17:42 I've noticed. Mar 21 03:17:52 This particular HDD/enclosure runs nervewrackingly hot. Mar 21 03:18:30 I once placed the KB3NNJ Universal Cooling Device (a zip-loc bag of ice cubes) on it, it was melted and the water was warm in 20 minutes. Mar 21 03:18:39 ah I would be cautious then. I hear that 2.5" HDDs run cooler than their 3.5" brothers Mar 21 03:18:59 Yeah, but I'm never going to find a 300GB 2.5" HDD. Mar 21 03:19:35 that's quite a bit of heat. If that's the case, you may need to get a HDD enclosure with fins (more surface area to dissipate heat) Mar 21 03:20:00 Or otherwise play with a Dremel? Mar 21 03:20:30 pierrelux, Can I just run any application on openmosix as if it were on a single machine? Mar 21 03:20:38 that's too much work, unless you're bored :-) Mar 21 03:20:48 True. Mar 21 03:21:01 Or what if I did away with the enclosure entirely? Mar 21 03:21:16 And just used the USB/IDE board attached to the back of this thing? Mar 21 03:22:20 no that is actually worse... better to tightly screw the case to a metal enclosure so the heat can get out of the case - otherwise it might not get out if you know what I mean. Mar 21 03:22:41 I mean heat getting out through conduction from HDD to enclosure. Mar 21 03:22:57 So you're saying that the more surface area that comes in contact with the HDD itself, the cooler it gets? Mar 21 03:23:04 OR! Mar 21 03:23:11 Could I use a small watercooling system? Mar 21 03:23:56 Or perhaps a Peltier element? Mar 21 03:24:53 yeah was just thinking that. But then, what to do with the heat. It has to get out somewhere... if you take the heat out to a radiator, then where does the heat go then? Anything you add with a heat exchanger will actually add to the total heat, because it takes work to transfer heat. Mar 21 03:25:15 I would stay clear of peltier. When they fail, they tend to excessively overheat and cook things Mar 21 03:25:27 Well, the radiator could go into the outside case of the briefcase. Mar 21 03:25:34 Ok, noted. Mar 21 03:26:02 plus, peltier, just like in thermodynamics, nothing is 100% efficient, so you'll be effectively removing more heat with the peltier than without it. Mar 21 03:26:02 Or would being silly and just getting some liquid nitrogen squirted into this thing work? Mar 21 03:26:24 Isn't that the purpose of Peltier? Mar 21 03:26:37 um, that would be even worse, because condensation inside HDD. Mar 21 03:26:52 True. Mar 21 03:26:58 And HDDs hate condensation Mar 21 03:26:59 . Mar 21 03:27:02 Peltier cools one side but it heats up the other side, much more than the heat coming without the peltier Mar 21 03:27:15 Oh. Mar 21 03:27:18 you'll learn that in thermo classes :-) Mar 21 03:27:22 So I'd wind up generating more heat than I really need. Mar 21 03:27:35 yep. Mar 21 03:27:51 liquid cooling would have the least added heat, IMHO Mar 21 03:28:15 I could probably hack something together... Mar 21 03:28:33 And then use my Palm Zire 31 as a terminal, on a USB cable to a slug? Mar 21 03:28:42 if you could get them, I'd say get 5400 or 4200 RPM drives. They run cooler than 7200s Mar 21 03:29:09 Ok. Mar 21 03:29:44 dunno if that is possible. I saw there is a hack to use the NSLU2 as USB slave versus master, but I can't recall... not for the faint hearted! Mar 21 03:30:05 Not interested. Mar 21 03:30:15 http://www.nslu2-linux.org/wiki/Peripherals/Palm Mar 21 03:30:21 http://www.nslu2-linux.org/wiki/HowTo/AddDeviceSideUSBPort Mar 21 03:30:33 u got it pierrelux Mar 21 03:30:37 it shows how to use that with a serial palm. Mar 21 03:30:50 but mine's got a standard mini USB port on it. Mar 21 03:31:07 So I could just use a normal cable, and then use that as a terminal, with no added adapters, right? Mar 21 03:31:09 that's just a cabling, not interface diff Mar 21 03:31:15 Right. Mar 21 03:31:23 But the software side is pretty much the same? Mar 21 03:31:43 dunno, I am not sure how the palm is used (some are used as slave, some as master, for terminal emulation) Mar 21 03:31:58 I'll look into it and let y'all know. Mar 21 03:32:04 Well, I'm off to pass out, night all! Mar 21 03:32:15 ok good luck! Mar 21 03:32:24 Thanks. Mar 21 03:32:27 03bzhou * r5802 10optware/trunk/make/py-axiom.mk: py-axiom: 0.5.0 -> 0.5.20 Mar 21 03:32:33 03bzhou * r5803 10optware/trunk/make/py-epsilon.mk: py-epsilon: 0.5.0 -> 0.5.6 Mar 21 03:32:42 03bzhou * r5804 10optware/trunk/make/py-mantissa.mk: py-mantissa: 0.5.0 -> 0.6.1 Mar 21 10:46:45 03blaster8 * r793 10kernel/trunk/patches/2.6.21/ (3 files): Refresh cmdline fixup patches for timer freq fixup Mar 21 10:50:11 03blaster8 * r794 10kernel/trunk/patches/2.6.21/88-nslu2-dflt-cmdline.patch: Update whitespace-damaged cmdline fixup patch Mar 21 13:17:49 NAiL: good morning. When you get a chance, please check in rev1 of the openpbx package, and the new build dependency sox-native http://bugs.openembedded.org/show_bug.cgi?id=2023 Mar 21 13:18:17 ah, you're getting it to do the sound-thingies? :) Mar 21 13:18:24 yeah Mar 21 13:19:12 and moved some extraneous stuff to -dev and fixed the sample files dependency on the MusicOnHold feature by reintegrating it into the main package Mar 21 13:19:15 the sounds are the same as asterisk? Mar 21 13:19:23 or it is a different set? Mar 21 13:19:24 oh and added LDAP support option Mar 21 13:19:35 osas: it's a different set Mar 21 13:19:39 I see Mar 21 13:19:45 multiple codecs? Mar 21 13:19:51 of course Mar 21 13:19:56 cool Mar 21 13:20:04 then break them down per codec Mar 21 13:20:16 * hillct doesn't screw around when it comes to this stuff Mar 21 13:20:18 this is what I did in optware for asterisk 1.4 Mar 21 13:20:24 :-) Mar 21 13:20:27 hmm Mar 21 13:20:55 I'll consider that for the next rev Mar 21 13:21:04 ok Mar 21 13:21:21 it should be pretty straight forward Mar 21 13:21:37 it seems to me that all codecs should be in a single package as this permits the best codec to be selected based on the codec used by the incoming call Mar 21 13:21:39 just sort them based on extension (codec) and package them Mar 21 13:22:08 providing all codecs in a single package reduces the potential need for transcoding Mar 21 13:22:32 no transcoding is important since the slug doesn't have a hefty CPU Mar 21 13:22:34 if one want's to force a single codec, it doesn't need all the other codecs Mar 21 13:23:02 forcing a single codec would be a bad practice Mar 21 13:23:04 and 729 for instance is pretty difficult to transcode on a small platform ... Mar 21 13:23:16 but I suppose we could make it an option Mar 21 13:23:26 it is up to you .... Mar 21 13:23:49 I'd rather consider it further and run it by the developers of OpenPBX before doing that Mar 21 13:24:19 sure Mar 21 13:24:25 lets keep it all as one sound package for this rev and if we go that route, do so in the next rev Mar 21 13:24:32 oh, and Mar 21 13:24:52 anoter reason to keep them all together is so the sample config works right out of te box Mar 21 13:25:36 one thing that I did for aterisk was to package all the sample files in a separate sample directory (for reference) Mar 21 13:25:50 eg /opt/etc/asterisk/samples Mar 21 13:26:00 so there you have all the samples Mar 21 13:26:05 It could get pretty messy, dynamically inserting Allow= lines into [general] to deal with sound file availability changes Mar 21 13:26:12 after an upgrade, all the samples are up to date Mar 21 13:26:13 hillct: where's sox-native? Mar 21 13:26:24 in the sox package Mar 21 13:26:33 I added sox_13.0.0.bb Mar 21 13:26:44 and sox-native_13.0.0.bb Mar 21 13:27:05 no maintainer was listed for sox Mar 21 13:27:17 so I don't know who's responsible for it Mar 21 13:27:42 it's in the dependancy ticket 2019 Mar 21 13:28:14 http://bugs.openembedded.org/show_bug.cgi?id=2019 Mar 21 13:28:34 thanks Mar 21 13:29:34 np Mar 21 13:30:36 hillct: btw, in sox, you have a couple of VARIABLENAME="value". I'm correcting those to VARIABLENAME = "value". someone in #oe will get nasty on you if you don't ;-) Mar 21 13:31:21 03nail 07slugos-3.10-beta * r413 10slugos/openembedded/packages/sox/ (sox-native_13.0.0.bb sox_13.0.0.bb): Add updated sox-package and new package sox-native Mar 21 13:31:54 NAiL: what's wrong with them? still out of order? Mar 21 13:31:56 03nail 07slugos-3.10-beta * r414 10slugos/openembedded/packages/openpbx/ (5 files in 3 dirs): openpbx: Update to rev1, add openpbx.org-perl, undo all the renaming-stuff I did for the last revisions Mar 21 13:32:04 I thought I squared them away Mar 21 13:32:06 no, there's an OE styleguide ;) Mar 21 13:32:16 yeah Mar 21 13:32:24 variable assignments must have spaces before and after the "=" Mar 21 13:32:25 I read it and corrected the order Mar 21 13:32:32 LOL Mar 21 13:32:37 I missed the spaces? Mar 21 13:32:39 wow Mar 21 13:32:42 yeah ;) Mar 21 13:32:45 thought I cought them all Mar 21 13:32:57 no worries Mar 21 13:32:57 already had one OE guy get real pissy about thay Mar 21 13:32:59 that Mar 21 13:33:04 hehe, I know Mar 21 13:33:18 ok, updated in 3.10-beta Mar 21 13:33:20 his comment was useless though which is why I didn't catch all that he wanted corrected Mar 21 13:33:37 thanks Mar 21 13:35:11 NAiL: why the hell isn't OE bugzilla configured to autodetect .bb files as text/plain? That's ridiculous Mar 21 13:35:24 I know, and it's very annoying Mar 21 13:35:34 ask in OE ;) Mar 21 13:35:59 I was goind to add a comment to my ticket that if bugzilla is broken that's not my problem Mar 21 13:36:10 but I didn't want to get in a pissing match with them Mar 21 13:38:25 03nail 07slugos-3.10-beta * r415 10slugos/openembedded/packages/sox/src-Makefile.am.patch: sox: Add missing makefile patch Mar 21 13:43:07 anyway, committed and closed Mar 21 14:00:18 NAiL: thanks. Is trhe feed updated imediately only based on a cron job every few hours? Mar 21 14:01:06 cron job every now and then Mar 21 14:01:25 ah Mar 21 14:01:43 rebuilding it manually now Mar 21 14:02:16 one thing I'm concerned about is it's unclear whether sox-native expects to find libgsm or not Mar 21 14:03:37 I may need to create a libgsm-native package to assure it's there. It works in by build enviroment without issue but that may be because it's installed as part of my distro Mar 21 14:04:04 we'll find out soon enough Mar 21 14:04:13 yup Mar 21 14:17:29 sox-native bult Mar 21 14:17:30 built Mar 21 14:43:24 hillct: both built, so should be in feed shortly Mar 21 14:43:42 cool Mar 21 14:43:44 thanks Mar 21 15:32:03 03oleo * r5805 10optware/trunk/ (make/thttpd.mk sources/thttpd/mime_types.patch): thttp: disable charset force for text/html pages Mar 21 15:33:18 03oleo * r5806 10optware/trunk/make/analog.mk: analog: add default paths Mar 21 15:35:00 NAiL: two items. openpbx-org-musiconhold should be removed from the feed as it's no longer relevent since it's been reintegrated into the main package. openpbx-org-perl has it's own .bb file so it presumably needs to be added separately to the autobuild system Mar 21 15:37:56 03nail 07slugos-3.10-beta * r416 10slugos/openembedded/packages/meta/slugos-packages.bb: slugos-packages: Add openpbx.org-perl to feed Mar 21 15:38:43 remind me later to remove musiconhold from the feeds ;) Mar 21 15:38:53 K Mar 21 15:38:55 is there an alternative/better way I should have handled openpbx.org-perl? Mar 21 15:39:10 perhaps make it a build dependency? Mar 21 15:39:32 even though strictly speaking it isn't Mar 21 15:41:16 What exactly is openpbx.org-perl? Mar 21 15:42:24 musiconhold is removed from the feeds. Packages isn't updated yet though. Mar 21 15:42:29 a package included in contrib in the openpbx tarball, which adds a perl wrapper to the Openpbx gateway interface Mar 21 15:42:51 then openpbx.org-perl is fine. It's not necessary. Mar 21 15:42:56 it uses cpan.bbclass to build Mar 21 15:44:00 I would have liked to be able to include it in the main recipe but you can't really do that aparently, where you initiate subsidiary builds in deeper root source directories Mar 21 15:44:05 K Mar 21 15:50:38 cat /home/repvik/builds/openslug-3.10-beta/openslug-nslu2.tmp/work/openpbx.org-perl-1.0-r0/temp/log.do_configure.3437 Mar 21 15:50:41 Can't open perl script "Makefile.PL": No such file or directory Mar 21 15:52:48 hmm Mar 21 15:53:27 I wonder if I switched the order of S = when I was told it wasn't style-compliant Mar 21 15:53:32 lemme see... Mar 21 15:54:52 could have sworn I tested this successfully after that... Mar 21 15:54:54 checking Mar 21 15:57:10 works fine for me: http://www.pastebin.ca/405180 Mar 21 15:57:18 is that what got checked in? Mar 21 15:59:11 I checked in the attachment in bugzilla Mar 21 15:59:14 looks the same Mar 21 16:00:27 well, it's missing from here http://www.openembedded.org/filebrowser/org.openembedded.dev/packages/openpbx Mar 21 16:00:37 did you give it it's own package directory? Mar 21 16:00:57 wouldn't expect that to make a difference though Mar 21 16:02:13 am I looking at a cached page or is the tree browser just not up to date? Mar 21 16:02:50 the error suggests it's simply not changing the source directory right? Mar 21 16:04:04 I'm using the 3.10-beta release of everything. Might there be a change to bitbake that requires S to be set prior to inheriting the cpan class? Mar 21 16:06:44 you changed the version when you checked it in Mar 21 16:06:48 it's no 1.0 Mar 21 16:07:15 it's version 0.0.1 Mar 21 16:07:58 so you'd need to add a PV line or manually specify te source directory Mar 21 16:08:07 NAiL: ^^ Mar 21 16:09:31 so, openpbx.org-perl_0.0.1.bb then? Mar 21 16:10:14 yup Mar 21 16:10:17 I didn't change the version, since no version was specified Mar 21 16:10:27 bugzilla doesn't show the filenames Mar 21 16:10:27 not in the file Mar 21 16:10:31 ah Mar 21 16:10:35 sorry Mar 21 16:10:44 I should have specified it in a comment Mar 21 16:10:56 been using trac recently which doers specify filenames Mar 21 16:11:08 03nail 07slugos-3.10-beta * r417 10slugos/openembedded/packages/openpbx/ (openpbx.org-perl.bb openpbx.org-perl_0.0.1.bb): openpbx.org-perl: Rename bb to correct version Mar 21 16:14:48 ok, I've cleaned out the old version and started the build again Mar 21 16:28:10 yup, built no Mar 21 16:28:12 w Mar 21 16:39:07 hi Mar 21 17:33:08 Could some one else try to compile this file for the NSLU2 Mar 21 17:33:09 http://www.kanoistika.sk/bobovsky/archiv/umts/huaweiAktBbo.c Mar 21 17:33:11 I only get a lot of strange errors. Mar 21 17:54:50 03bzhou * r5807 10optware/trunk/make/findutils.mk: findutils: 4.2.29 -> 4.2.30 Mar 21 17:55:21 LarsC: It would help if you could paste the output somewhere Mar 21 17:58:48 well the output is a long list of strange errors would be great if somebody could compile it and send the binary to me :-) lars159@gmail.com Mar 21 17:58:50 The program will make the Huawie e220 modem be detected as an usb serial instead of usb storage Mar 21 18:00:47 someone might be able to help if you pastebin that "long list of strange" errors somewhere. Mar 21 18:00:59 It would also be helpful to know what distribution you run Mar 21 18:01:17 infact, read this first: http://www.google.com/url?sa=t&ct=res&cd=3&url=http%3A%2F%2Fwww.catb.org%2F~esr%2Ffaqs%2Fsmart-questions.html&ei=4HIBRuKZOaHi0ASs6J3mCA&usg=__yjXjC9NiaBiwv6eFAjHhzwApI_c=&sig2=-e8O-cEeRC5rjAbjwM0y_w Mar 21 18:01:22 dang Mar 21 18:01:27 wrong url Mar 21 18:01:31 http://www.catb.org/~esr/faqs/smart-questions.html Mar 21 18:02:18 ok thanks, im running NSLU debian Mar 21 18:03:24 that's still kind of ambigous. Are you running debonaras (BE debian) or "standard" debian? Or even "debianslug"? All of those three are different. Mar 21 18:05:50 i think is debian Slug I followed the http://www.cyrius.com/debian/nslu2/ Mar 21 18:06:19 ok, that's not debianslug. It's Debian/NSLU2. But the confusion is understandable. Mar 21 18:06:23 Are you trying to compile natively? Mar 21 18:06:48 (ie, on the slug itself) Mar 21 18:08:00 thanks for the explenation was googeling on natively:-) yes on the slug Mar 21 18:08:42 ok, in that case, you will want to send a mail to the debian mailinglists... Mar 21 18:08:52 or there might be a debian-arm channel or something here Mar 21 18:09:11 apparently not... Mar 21 18:09:43 Point is. You're running debian, which is different from slugos/unslung. So you should ask debian for support. Mar 21 18:10:14 but aret they using the same compiler? Mar 21 18:11:00 no Mar 21 18:11:50 ok Mar 21 18:13:53 Of course, without any output we don't know if the problem is the compiler itself -- or more likely missing libraries or headers? Mar 21 18:16:15 * NAiL guesses missing kernel sources Mar 21 18:17:05 LKG3519CB:~# Mar 21 18:17:06 LKG3519CB:~# gcc -I /usr/src/linux-headers-2.6.18-4/include/linux huaweiAktBbo.c Mar 21 18:17:08 In file included from /usr/include/_G_config.h:44, Mar 21 18:17:09 from /usr/include/libio.h:32, Mar 21 18:17:11 from /usr/include/stdio.h:72, Mar 21 18:17:12 from huaweiAktBbo.c:9: Mar 21 18:17:14 /usr/include/gconv.h:72: error: expected declaration specifiers or '...' before 'size_t' Mar 21 18:17:15 /usr/include/gconv.h:88: error: expected declaration specifiers or '...' before 'size_t' Mar 21 18:17:17 /usr/include/gconv.h:97: error: expected declaration specifiers or '...' before 'size_t' Mar 21 18:17:18 /usr/include/gconv.h:174: error: expected specifier-qualifier-list before 'size_ t' Mar 21 18:17:20 In file included from /usr/include/stdio.h:72, Mar 21 18:17:21 from huaweiAktBbo.c:9: Mar 21 18:17:23 /usr/include/libio.h:354: error: expected declaration specifiers or '...' before 'size_t' Mar 21 18:17:26 /usr/include/libio.h:363: error: expected declaration specifiers or '...' before 'size_t' Mar 21 18:17:29 /usr/include/libio.h:479: error: expected '=', ',', ';', 'asm' or '__attribute__ ' before '_IO_sgetn' Mar 21 18:17:32 In file included from huaweiAktBbo.c:9: Mar 21 18:17:34 /usr/include/stdio.h:302: error: expected declaration specifiers or '...' before 'size_t' Mar 21 18:17:37 /usr/include/stdio.h:309: error: expected declaration specifiers or '...' before 'size_t' Mar 21 18:17:40 /usr/include/stdio.h:351: error: expected declaration specifiers or '...' before 'size_t' Mar 21 18:17:43 /usr/include/stdio.h:353: error: format string argument not a string type Mar 21 18:17:45 /usr/include/stdio.h:355: error: expected declaration specifiers or '...' before 'size_t' Mar 21 18:17:48 /usr/include/stdio.h:357: error: format string argument not a string type Mar 21 18:17:50 /usr/include/stdio.h:603: error: expected '=', ',', ';', 'asm' or '__attribute__ ' before 'fread' Mar 21 18:17:51 ... Mar 21 18:20:06 sorry for floding Mar 21 18:36:21 03bzhou * r5808 10optware/trunk/make/zsh.mk: zsh: 4.2.6 -> 4.3.2 Mar 21 20:17:51 03bzhou * r5809 10optware/trunk/make/rc.mk: rc: use a better SITE Mar 21 20:31:54 hi all, i wonder if anyone can help me. i currently am using slugosBe, and use scsi-idle to send one of my disks to sleep when its not in use. but in my log files ( or serial console if its connected ) i get the following error. Mar 21 20:31:56 program scsi-stop is using a deprecated SCSI ioctl, please convert it to SG_IO Mar 21 20:32:24 any ideas how i can get rid of this? it isnt actually stopping scsi-idle from working though. Mar 21 20:33:12 no reason to get rid of it Mar 21 20:33:19 "it speaks the truth" Mar 21 20:34:49 you would have to patch the kernel Mar 21 20:34:55 http://lxr.linux.no/source/drivers/scsi/scsi_ioctl.c Mar 21 20:35:11 knock out lines 208-209 there Mar 21 20:35:23 if that's too much trouble, then just get used to the error :) Mar 21 20:39:58 lol. will give it a try. how do i use patch files? ( bit of a newbie to linux, but am learning bits ) Mar 21 20:40:06 oh heck Mar 21 20:40:27 do you know how to set up the build system for SlugOS Mar 21 20:40:44 do you have a spare linux x86 computer to cross-compile Mar 21 20:41:10 ah, well i have tried. and seem to fail on bits and pieces. i have been playing arround with virtual machines. Mar 21 20:41:21 may give it a try this weekend/next week. Mar 21 20:41:32 is it best to ask when i have that all set up? Mar 21 20:42:00 seriously it's a mammoth undertaking to 'fix' an 'error' that is in fact a perfectly valid warning with no functional effect Mar 21 20:42:20 other than teaching you how to use the build system, I have to point out that this is a spectacular waste of time :) Mar 21 20:42:37 nevertheless, here's a pointer Mar 21 20:42:40 lol. i like learning by doing. Mar 21 20:42:42 ;) Mar 21 20:42:48 guessed as much Mar 21 20:42:49 anyway Mar 21 20:42:58 install debian etch on a virtual machine Mar 21 20:43:14 update everything to the latest versions Mar 21 20:43:27 download http://www.nslu2-linux.org/Makefile to a new directory Mar 21 20:44:11 run 'make slugosbe-3.10-beta' Mar 21 20:44:16 that last bit is vital Mar 21 20:44:30 otherwise you make the HEAD version which is a 'bad move'(tm) Mar 21 20:44:46 ( i am using slugosBe 4.3 alpha on my slug, is it the same? ) Mar 21 20:44:49 aah Mar 21 20:44:54 how did you get that? Mar 21 20:45:04 of the snapshot site thing. Mar 21 20:45:08 like it so far. Mar 21 20:45:11 there's a snapshot site? Mar 21 20:45:13 news to me Mar 21 20:45:22 coredump set up Mar 21 20:45:23 let me just dig it out Mar 21 20:45:32 hentges.net somewhere iirc Mar 21 20:45:44 I wouldn't recommend running that version atm Mar 21 20:46:12 thats the one. Mar 21 20:46:17 yeah... Mar 21 20:46:19 it works, and i am happy with it. Mar 21 20:46:30 well anyway, then you're on your own Mar 21 20:46:34 hmmm.... I thought that site was going to be time limited..... Mar 21 20:47:00 I can't tell you how to patch kernels for HEAD (i.e. 4.3) as its too complicated to easily explain Mar 21 20:47:12 can i still set up a build-environment by using "make slugosbe-3.3-beta"? Mar 21 20:47:25 well, then that will be SlugOS 3.3 not 4.3 Mar 21 20:47:38 why 3.3? Mar 21 20:47:52 "spikes-nslu2: can i still set up a build-environment by using "make slugosbe-3.3-beta"?" Mar 21 20:48:16 doh, typo. meant 4.3 Mar 21 20:48:26 there is no 4.3-beta release, that comes from OE HEAD Mar 21 20:48:35 which is usually dev-only land Mar 21 20:48:52 so a friendly warning that what you are trying to do is *really* hard Mar 21 20:49:02 ah, i see. is there a rough time scale as to when 4.3 is in "proper" beta land? Mar 21 20:49:22 how long is a piece of string? ;) Mar 21 20:49:27 it's not quite that bad Mar 21 20:50:03 lol. well i seem to remember reading that its in feature hold now. so i take it its a case of a few months? and not 6months+? Mar 21 20:50:03 but there's no strict timetable yet Mar 21 20:50:10 i know its asking alot. lol. Mar 21 20:50:22 no it's fine, we definitely need a new release soon Mar 21 20:50:32 but there are some issues that need agreeing and fixing first Mar 21 20:51:12 yep, i tend to read the irc logs when i am bored at work, so have a bit of an idea. ( albeit a very slim knowledge ). Mar 21 20:51:27 but as i said, i am learning and enjoying it. Mar 21 20:51:48 thanks for the help anyway, at least i now know that its nothing to be concerned about. Mar 21 20:52:01 If you have a working system, stick to it for now Mar 21 20:52:06 *looking forward to the next release* :) Mar 21 20:52:11 so do I Mar 21 20:52:14 and I Mar 21 20:52:19 yep. thanks again. Mar 21 20:52:27 no worries Mar 21 20:52:31 any bugs you hit, tell us Mar 21 20:53:00 will do. so far thats the only one. am doing daily 'ipkg update / upgrades' anyway. Mar 21 20:53:22 another bad plan... Mar 21 20:53:24