**** BEGIN LOGGING AT Sat Dec 30 02:59:58 2006 Dec 30 03:56:11 03bzhou * r4924 10optware/trunk/make/classpath.mk: classpath: explicit --disable-Werror Dec 30 04:54:16 03bzhou * r4925 10optware/trunk/Makefile: demoted gnokii for mss Dec 30 08:00:42 hello Dec 30 10:11:34 03oleo * r4926 10optware/trunk/make/uclibc-opt.mk: uclibc-opt: add Smash Stack Protection library Dec 30 11:04:20 03oleo * r4927 10optware/trunk/make/apr.mk: apr: change source Dec 30 11:10:57 03oleo * r4928 10optware/trunk/make/apr-util.mk: apr-util: change source Dec 30 11:15:05 03oleo * r4929 10optware/trunk/make/apache.mk: apache: change source Dec 30 15:09:36 03oleo * r4930 10optware/trunk/make/elinks.mk: elinks: 0.11.1->0.11.2 Dec 30 15:30:31 03oleo * r4931 10optware/trunk/make/ (sdl.mk snownews.mk xfixes.mk xrender.mk): bounce ipk to force uClibc rebuild Dec 30 16:52:04 03bzhou * r4932 10optware/trunk/make/snownews.mk: snownews: -lintl if uclibc Dec 30 18:36:17 03oleo * r4933 10optware/trunk/Makefile: optware: demote xrender and xfixes for uClibc - whole x11 libs collection needs update Dec 30 18:55:28 03oleo * r4934 10optware/trunk/ (3 files in 2 dirs): rrdtool: disable -std=gnu99 for uClibc Dec 30 19:12:38 03oleo * r4935 10optware/trunk/Makefile: optware: demote sdl for uClibc Dec 30 19:17:05 morning likewise Dec 30 19:17:31 NAiL: morning NAiL. Just replaced my caps, still the same boot problems here though; harddisk won't spin. Dec 30 19:19:14 hmm Dec 30 19:19:18 what PSU are you using? Dec 30 19:19:37 And does it spin up if you remove the IDE-cable? Dec 30 19:24:19 60W which came with a ITX casing for the EPIA. Dec 30 19:25:24 no same problem, sometime the hd starts to spin, then suddenly the 12 volts goes down to 11 or 10 and it stops Dec 30 19:25:41 ah Dec 30 19:25:52 I had the same problem when I was using a 60W Dec 30 19:26:02 the spike is too high for the PSU to handle Dec 30 19:26:10 using a 2.5" hdd worked fine for me Dec 30 19:31:02 03oleo * r4936 10optware/trunk/make/snownews.mk: snownews: strip target binary Dec 30 19:36:03 Hi oleo, ru there? Dec 30 19:40:06 likewise: 3.5" hdd's tend to suck a ginormous amount of power when spinning up, unfortunately. Dec 30 19:40:21 I believe there's a watt-calculator over at mini-box.org or something Dec 30 19:41:14 NAiL: hehe there are discs that have a jumper setting to limit the current, but these are often enterprise grade discs in big servers Dec 30 19:41:23 yeah Dec 30 19:41:41 and most often scsi Dec 30 19:41:50 and some support delaying spin-up, to avoid 20+ disks spinning up at the same time :) Dec 30 19:41:53 Nail: I have one of the power meters, it is called Kill A Watt. Is that what you were referring to? Dec 30 19:41:56 hehe jep Dec 30 19:42:21 chacko: no, I was talking about a web-based calculator thingy. Dec 30 19:42:24 and burn the psu :) Dec 30 19:42:28 exactly Dec 30 19:42:46 * NAiL is using a 450W PSU in his epia box Dec 30 19:42:51 wow Dec 30 19:43:02 I probably don't use more than ~100W when powering up Dec 30 19:43:20 NAiL: The drive reads 0.5A@12V and 0.6A@5V, so 9 Watts (not counting power-up peaks) Dec 30 19:43:23 After that, I guess the power is somewhere between 40-70W Dec 30 19:43:32 i have a dec vax machine with big (very big) discs and it sucks only 900W at normal, i'm curious how much it sucks when the discs are spinning up :) Dec 30 19:44:02 likewise: yeah, some of my drives claim that too. But I know for sure that some of my drives pull 35+W when spinning up. Dec 30 19:44:03 likewise: the spinup is then probably between 15-40Watts Dec 30 19:44:26 likewise: but these are often only ms peeks Dec 30 19:44:36 It is a function of how fast the disk spins up. The older big drives takes a bit of time to spin up, so the increase in power is not as drastic as the newer ones. Dec 30 19:44:39 but can trigger a psu shutdown :) Dec 30 19:44:49 yep Dec 30 19:45:17 strange thing is, if the drive spins up OK (for 20 secs or so), the 12 rail drops to 11 or 10. Dec 30 19:46:53 nail: who maintains the optware feeds for unslung? Are you the one? Dec 30 19:47:13 chacko: nope Dec 30 19:47:30 I do mess around with them from time to time Dec 30 19:49:03 I was thinking about how to automate the bootstrap of the optware feed for slugosbe -- the initial ddwon load abd install of ipkg-opt. Dec 30 19:49:53 I am going to use my big ATX PSU 12 Volts to drive the PSU in the EPIA... Dec 30 19:51:21 For slugosbe all we need to do is to put the ipkg-opt.pak in the regular slugos/be native feed, and then install that... Dec 30 19:52:22 likewise: haha Dec 30 19:53:21 is anyone fluent in init.d scripts on unslung here? :) Dec 30 19:53:53 jollino: I have used them, modified them some time ago... Dec 30 19:55:11 chacko: i made two, one for my self-written irc bot (which works perfectly) and one for tor; the one for gets executed (i confirmed it by having it write on a file and checking it after boot), but it just does not run tor. i tried both giving the full path and by issuing a cd to its directory and then having it launch ./tor... Dec 30 19:55:21 it seems that tor isn't run at all, because i have no traces of the boot time in its logs :( Dec 30 19:57:32 am i missing something? Dec 30 20:00:11 jollino: both of you running the same vers of unslung? I think we need to do a full path in all executables, that is how I have done. Dec 30 20:00:42 both of me? i'm one. :) Dec 30 20:01:22 What is tor? thought it was another person :( Dec 30 20:01:36 tohttp://en.wikipedia.org/wiki/Tor_%28anonymity_network%29 Dec 30 20:01:38 this :) Dec 30 20:01:40 http://en.wikipedia.org/wiki/Tor_%28anonymity_network%29 Dec 30 20:03:09 NAiL: My big PSU can provide 15A@12V, so I tried using it as my external PSU. Same problem. So it seems my internal (12V -> ATX) PSU is broke. Dec 30 20:03:22 broken, rather Dec 30 20:04:09 jollino: perhaps you can open a private channel to me and then post both the scripts, I can take a look -- no promises :) Dec 30 20:04:53 a query? Dec 30 20:06:48 jollino: on my icechat it is DCC->chat and then select my nickname Dec 30 20:07:28 oh, alright, a dcc Dec 30 20:07:34 jollino: for some reason I cannot open private channels to you, but I have had others connect to me... Dec 30 20:08:32 likewise: Does the board boot up fine if you connect your "large" psu to it? Dec 30 20:08:41 NAiL: yes Dec 30 20:08:44 ok Dec 30 20:09:00 the 60W one is probably not feeling very well then :-P Dec 30 20:09:08 chacko: i have some problems offering DCCs too. how about a 'normal' query, or even a temporary channel? :) Dec 30 20:09:26 NAiL: seems the internal PSU cannot provide enough power (on the 12 line presumably), I think the 60W external PSU is fine though. Dec 30 20:10:01 jollino: I am willing to try, i have very limited experience with irc... Dec 30 20:10:21 chacko: please join #jollinochacko for the time being Dec 30 20:12:03 likewise: the power brick is probably ok. The 12V->ATX sucks if it's the same as mine :-\ Dec 30 20:12:33 NAiL: Morex Procase D2D v2.0.1 Dec 30 20:12:46 hmm Dec 30 20:12:59 Can't remember what the PSU was named. It sucked though :-P Dec 30 20:13:02 NAiL: probably I will go for a picoPSU then Dec 30 20:13:29 NAiL: comes in handy for the EFIKA (in car) project as well :-) Dec 30 20:14:41 03oleo * r4937 10optware/trunk/make/snownews.mk: snownews: reconfigure if snownews.mk changes Dec 30 20:16:12 I'm looking at building a carpc later as well Dec 30 20:16:21 The voompc enclosure looks very nice Dec 30 20:17:11 together with a 7" touchscreen, accelerometers, gps, 802.11, gsm/gprs, bluetooth ;) Dec 30 20:19:19 tomtom on steroids? ;) Dec 30 20:20:09 I'm looking for good accelerometers, but I can't find any with high resolution Dec 30 20:21:18 It'd be fun with proper datalogging ;) Dec 30 20:22:47 I won't be buying it until after I buy a new car though :-P Dec 30 20:23:08 Saving up for a car that is a bit newer than my '86 Ford ;) Dec 30 20:30:10 NAiL: what OS do you plan to run on that carputer? Dec 30 20:32:34 Linu Dec 30 20:32:35 x Dec 30 20:33:30 is there any good software for it on linux? on windows they have road runner, etc, etc.. Dec 30 20:33:39 gps software? Hardly. Dec 30 20:34:40 I heard something about Garmin hiring linux devs not long ago though Dec 30 20:34:51 might be interesting Dec 30 20:37:54 https://my.garmin.com/iRecruitment/jobdetails.htm?vacancyId=265 infact Dec 30 20:57:41 * Jollino is away: taking care of personal hygiene (aka "unbearding and showering") Dec 30 21:13:13 What is a "cumulative GPA" in this vacancy? Dec 30 21:13:56 In The Netherlands, in the MSc. program we get ranks from 0 to 10, 10 highest, for each course. Dec 30 21:14:31 I think GPA of 4.0 is top in the US. Dec 30 21:31:02 rwhitby: thanks. not that I was soliciting. :-) Dec 30 21:32:35 Some U's in US have a 5.0 max based gpa also. Dec 30 21:33:49 rwhitby: Who is doing the bootstrap installation of the ipkg-opt for the new slugosbe optware stream? Dec 30 21:34:46 chacko: so then a 3.0 average is not that good :-) Dec 30 21:35:03 chacko: dunno Dec 30 21:35:47 likewise: but most often it is 4.0 max, and 3.0 for that is considered a B, where A is the top. Dec 30 21:37:33 I have always wondered what my MSc degree and marks are worth in the US. I'm above 75% for my facultative (i.e. specialization) courses, and quite random between 60% and 100% for my other courses. Dec 30 21:37:38 rwhitby: I know eno is doing the cross build stream. Who should I ask? I have some ideas for automating the intial install of the new pkg. I have experimented with a manual bootstrapping, it appears to be working. Dec 30 21:38:24 With HopsNBarley on the Angstrom build, will this mean we will target PowerPC's as well? Dec 30 21:38:53 Likewise ain't the only one who will work on ppc ;) Dec 30 21:38:56 urhm Dec 30 21:39:00 HopsNBarley Dec 30 21:39:22 Well, the IXP42x seems more popular so far. Dec 30 21:39:40 NAiL: but I know there are some who have PPC devices :-) Dec 30 21:39:47 likewise: we'll be doing Angstrom for NAS devices, not specifically ixp4xx Dec 30 21:40:01 (we already have Optware targets that are PPC, for instance) Dec 30 21:41:40 likewise: There is no easy direct translation. For university admissions, each university have their own translation, depending on where the degree si from. The same 75% from a top University might be considered a gpa of 3.5 or even 4.0. For example when I graduated from engineering in India, that year there were only 2 out of 1200 who got about 70%... So the grading curve has to be normalized. Dec 30 21:41:51 likewise: yeah. PPC is nice ;) Dec 30 21:44:43 For slugosbe optware, there is already a ipkg-opt.pak. With some minor modifications, if we put that package in the native stream of slugos/be, then just doing an ipkg install of ipkg-opt can be made to setup optware ipkg installation. Dec 30 21:46:51 likewise: I was a Prof at Berkeley long time ago :) Dec 30 21:47:02 chacko: we won't be doing that - too easy for someone who doesn't understand what they are doing to install that package by mistake Dec 30 21:47:27 wget /.ipk, then ipkg install foo.ipk is the way that we do bootstrapping. Dec 30 21:48:18 chacko: I should see which US institute accredited the MSc program I followed. Might come in handy when I move abroad. Dec 30 21:50:18 rwhitby: once people have tried the new stream out and it is working ok, what is the the risk of making the newstream just like the native stream. The only minor problem is that the same pkg may be available from two streams --- but I think that can be address almost like people trying back versions of a pkg. Rod, what is the risk you envision? Dec 30 21:51:26 * Jollino is back (gone 00:53:44) Dec 30 21:53:37 likewise: Usually the normalization or translation is not formalized or published. Each dept establishes own private guidelines based on the students they have seen previously from a foreign university. If you have a good feeling for the relative standing of your university with respect to the top us schools, then perhaps I can make a guess :) Dec 30 21:55:19 chacko: what is wrong with the two command version I just wrote? It's a suitable barrier to weed out the users who cannot even set their PATH correctly to get to the existing ipkg command. Dec 30 21:55:53 The risk is user confusion. And that is risk enough not to do what you are suggesting. Dec 30 21:58:29 rwhitby: there is nothing wrong or too complicated about the version given. I think once you get familiar with this, and nobody comes up with some potential risk problems, my view is that this is no dfferent than two streams like a native build, and cross build. The easier we make things, provided there is no risk, the better it is for users, IMHO. Dec 30 21:59:13 chacko: 2.5 years of experience with this community says otherwise. Dec 30 22:00:10 the difference between the two streams case is that those two streams do not contain the same packages. Dec 30 22:00:15 whitby: That is good enough reason to proceed with caution :) I respect your experience :) Dec 30 22:00:55 the confusion is exactly caused by there being two versions of the same package available for the one firmware distribution, and the interactions between packages from different feeds. Dec 30 22:01:22 e.g. what happens if someone installs a package that wants to use port 80 from *both* feeds? Dec 30 22:01:43 (i.e. apache from slugos and apache from optware/slugosbe) Dec 30 22:01:46 Yes, I do not yet know fully what interaction problems will be there... There is a possibility. Dec 30 22:02:42 so someone has to do something out of the ordinary (the wget, ipkg sequence) to make them realise they are doing something non-standard. Dec 30 22:03:32 it's called a "small barrier to entry" Dec 30 22:03:50 Well, the way I see it is, the paths will be setups such as ...:/usr/sbin:/opt/sbin so if it is available in the native stream, that will get used, otherwise the optware version. Dec 30 22:04:06 chacko: Believe me. rwhitby is right when he talks about the community. Dec 30 22:04:41 nail: I am not questioning that at all :) Dec 30 22:04:56 chacko: we can't even get optware users to set their path correctly with one feed. who knows what will happen with two feeds. Dec 30 22:05:28 hence the small barrier to entry. you're just gonna have to live with typing an extra command, cause that's how it's gonna be. Dec 30 22:06:08 rwhitby: I am not thinking about me, I have alreay done that :) Dec 30 22:07:19 we simply doesn't want it to be "dead simple". "Dead simple" incurs a lot of support headaches. Dec 30 22:07:39 nail: I understand Dec 30 22:09:07 NAiL: ah, but we have chacko in the channel now to handle all the support headaches :-) Dec 30 22:09:44 While I'd like it to be just "ipkg install optware-support" or something like that since I know what I'm doing, somewhere between 90-95% of the users probably don't. Dec 30 22:13:52 Hey, in almost all places I was, I was the most cautious person; I have a lot of grey hair :) Dec 30 22:15:57 The reason I stuck with unslung for a long time, instead of trying slugos was that I thought linksys/cisco will erect enough cautions barriers instead of the possible uncontrolled changes in the other area :) Dec 30 22:22:14 Another topic: where is the setting to change the slugos/be boot to go from the current turned up disk partition, say /dev/sda1 to use another partion /dev/sda2? I am saving multiple copies of my boot drive image for experiments. Dec 30 22:23:15 I assume this boot startup drive setting is one in the flash somewhere? Dec 30 22:23:24 rwhitby: Is it possible to change "openembedded/packages/linux/unslung-kernel/defconfig" so that the "pwc" module is build? Dec 30 22:23:24 I have build it and get a "kernel-module-pwc_2.4.22.l2.3r63-r16_ixp4xxbe.ipk" so that could be useful for the next release. Dec 30 22:25:09 marceln: it is, but I thought there was a pwc package that was more recent than the in-kernel module ? Dec 30 22:25:29 Hi marceln: ! Dec 30 22:25:32 chacko: /initrd/linuxrc Dec 30 22:25:38 use turnup to change it Dec 30 22:26:09 $ ipkg list | grep pwc Dec 30 22:26:09 setpwc - 1.1-1 - Set and list settings of WebCams with the 'PWC chipset' Dec 30 22:26:19 I can't find a kernel module for it. Dec 30 22:27:10 marceln: have a chat to mwester about changing the defconfig Dec 30 22:27:31 rwhitby:I thought of using turnup to do it. But here is my paranoia... I do not know what else turnup might change. But if I can do this manually, I can confine it to things I know :) See I am extra cautious :) Dec 30 22:27:37 Oke. Dec 30 22:30:27 rwhitby: ok I see the /initrd/linuxrc line. For the boot to change, I will have to do this change, when booted without any disk, right? Dec 30 22:32:54 chacko: have a read of the turnup script Dec 30 22:33:58 Is there any reference where I can readup on this? Is the UUID used by the booter? Eventually I want to make a script which will on next boot use /dev/sdaN. Where is the turnup script located? Dec 30 22:34:39 sbin/turnup on the disk? Dec 30 22:38:17 rwhitby: Thanks! It looks like turnup already has exactly what I want. I just need to get comfortable with it :) Thanks again. Dec 30 23:12:22 night all Dec 30 23:46:18 03oleo * r4938 10optware/trunk/ (5 files in 2 dirs): reconfigure when .mk changes Dec 31 00:10:35 03oleo * r4939 10optware/trunk/ (7 files in 3 dirs): optware: rrdcollect initial import Dec 31 01:21:43 Looking for a reliable way for a script to tell the Unslung release. Current issue is the sd? vs /share/?? mapping. But should be a stable method. As a hackque, I suppose /etc/motd existing implies 6.* (and can parse), but Yuck! Does somethng exist? **** ENDING LOGGING AT Sun Dec 31 02:59:57 2006