2010-11-25 00:03 xiangfu: I think I traced the python-sip problem down 2010-11-25 00:04 in packages/python-sip/Makefile, there is a HOST_FPIC variable 2010-11-25 00:04 it is dereferenced there with $(HOST_FPIC), but never set 2010-11-25 00:04 maybe it was introduced in a later backfire revision, or in trunk? 2010-11-25 00:05 wolfspraul: let me check. 2010-11-25 00:05 it brings us back to the old problem that the feeds are not in sync with the main OpenWrt tree 2010-11-25 00:05 also (unrelated), maybe we should sync the tracking_backfire branch every 2 weeks or so? I think we should probably have a fixed time interval. 2010-11-25 00:06 wolfspraul: the 'packages' is not in feeds. 2010-11-25 00:07 hmm, OK 2010-11-25 00:08 then it can neither come from a feed problem nor from infrequent tracking_backfire updates :-) 2010-11-25 00:08 HOST_FPIC should be set to -fPIC 2010-11-25 00:09 or in that Makefile, we should replace $(HOST_FPIC) with just -fPIC 2010-11-25 00:09 I did a grep over all feeds and packages, and nobody else is using HOST_FPIC 2010-11-25 00:09 wolfspraul: you compile today? got those error. 2010-11-25 00:09 wolfspraul: the build host compile fine in 1124. 2010-11-25 00:11 wolfspraul: sorry my bad. the python-sip is in feeds. 2010-11-25 00:14 wolfspraul: I grep all git source code. there is no HOST_FPIC either. 2010-11-25 00:17 wolfspraul: ok. the HOST_FPIC is define in 'trunk' rules.mk:50:HOST_FPIC:=-fPIC 2010-11-25 00:18 wolfspraul: 'backfire' upstream also no HOST_FPIC defined. 2010-11-25 00:20 hmm 2010-11-25 00:20 OK 2010-11-25 00:21 so for now I will just export HOST_FPIC=-fPIC manually 2010-11-25 00:22 even if the feature of tagging feeds is added to OpenWrt, we would still want to uptick them and run into a problem like this 2010-11-25 00:22 I guess it's unavoidable ;-) 2010-11-25 00:40 [commit] Xiangfu Liu: add fw_setenv_default for set all u-boot variable to default http://qi-hw.com/p/openwrt-xburst/49f1bbf 2010-11-25 00:43 wolfspraul: I got an error when I "git push" : hooks/post-update: line 24: echo: write error: Broken pipe 2010-11-25 00:44 wolfspraul: where is the hooks ? maybe I  check that a little. 2010-11-25 01:06 xiangfu: hmm 2010-11-25 01:07 I rebooted fidelio a few hours ago (after a dist-upgrade), and sometimes the NFS share on turandot gets stuck with a 'stale NFS file handle' 2010-11-25 01:07 maybe it's that 2010-11-25 01:07 I will look into it 2010-11-25 01:11 xiangfu (in absence) I remounted the nfs share, let's see on the next commit whether the broken pipe error is gone... 2010-11-25 02:59 xiangfu: (repost) I remounted the nfs share, let's see on the next commit whether the broken pipe error is gone... 2010-11-25 02:59 wolfspraul: thanks. 2010-11-25 03:05 xiangfu: I have a question/idea. see http://en.qi-hardware.com/wiki/Building_Software_Image#Feeds 2010-11-25 03:05 to update the feeds, we recommend to run "scripts/feeds update -a && scripts/feeds install -a" 2010-11-25 03:05 but in the root Makefile, there is a package/symlinks target that does exactly the same thing 2010-11-25 03:06 so we could also recommend to run "make package/symlinks" instead 2010-11-25 03:06 there are several pros and cons 2010-11-25 03:06 package/symlinks is cleaner/shorter, and if some of the parameters of scripts/feeds change it will be updated 2010-11-25 03:06 on the other hand package/symlinks is slower, because it will run through the whole openwrt prerequisites stuff 2010-11-25 03:07 also it will hide some of the stdout output you see when running scripts/feeds directly 2010-11-25 03:08 however, bottom line, on the wiki page I still think we should change it to "make package/symlinks" 2010-11-25 03:08 what do you think? 2010-11-25 03:11 wpwrak: I tried -p and -fp, no effect. But I'm using sysvinit 'halt'; I don't know what is there in openwrt 2010-11-25 03:12 agree. 2010-11-25 03:12 larsc: do you know if there is anything special to poweroff the nanonote? sysvinit halt does not work for me, and busybox halt either 2010-11-25 03:13 wolfspraul: "run through the whole openwrt prerequisites stuff"  is better.  if there is no ".config" it's will start the "menuconfig" etc.. 2010-11-25 03:15 wolfspraul: I found we can add "V=99" to "make package/symlinks" but it's give too much "WARNING ..." which is not good. 2010-11-25 03:15 wolfspraul: so just ""make package/symlinks"" is better. 2010-11-25 03:20 xiangfu: hmm. if make package/symlinks checks the config, then on the wiki page we have to move the Feeds section after the Configuration section 2010-11-25 03:21 wolfspraul: yes. 2010-11-25 03:21 I'll edit the page, you check whether you like it :-) 2010-11-25 03:23 it says on wiki that the release is not tagged 2010-11-25 03:23 but it is tagged now? 2010-11-25 03:25 done 2010-11-25 03:25 don't know the wiki needs lots of updates :-) 2010-11-25 03:26 one thing I am planning to add is a little table for each configuration file to list disk space requirements and build time (on a given reference system) 2010-11-25 03:26 because for minimal, xbboot, full_system or all_packages those will be dramatically different 2010-11-25 03:27 tuxbrain: welcome back! :-) 2010-11-25 03:27 good morning! 2010-11-25 03:27 hehehe I miss you all too :) 2010-11-25 03:27 how are the NanoNote sales? 2010-11-25 03:28 mine are super low, maybe 2-5 / week 2010-11-25 03:28 not espectacular but not bad as always, mmm same ratio aprox 2010-11-25 03:28 ok, not bad, at least a few people discover the gem once in a while :-) 2010-11-25 03:28 give me the chance to do stellar customer service for each one :-) 2010-11-25 03:29 hehehe that's right :), at least we have achieve one thing, that any costumer is complaining, that's a real good achievement :) 2010-11-25 03:32 kyak: Hi. 2010-11-25 03:33 kyak: when I start the "ash(default)" in gmenu2x , I always got "sh: can't access tty; jo..." 2010-11-25 03:33 kyak: and I found this: http://www.busybox.net/FAQ.html#job_control 2010-11-25 03:34 kyak: which "exec /dev/tty1 2>&1" make the warning gone. 2010-11-25 03:41 tuxbrain: did you see the Milkymist One wood case roh made? 2010-11-25 03:41 which case do you like? 2010-11-25 03:41 do you think you can sell the product with a wood case? 2010-11-25 03:41 roh is experimenting with an acrylic case next, and some other improvements (stability, labels, microphone, IR receiver, etc) 2010-11-25 03:42 uff, electronics and wood.... did it pass any CE certificate??? 2010-11-25 03:42 but in the end your opinion matters because you represent customers 2010-11-25 03:42 yeah that's the next question 2010-11-25 03:42 in order to avoid CE/FCC, can we sell it as a kit? 2010-11-25 03:42 so we put the bare board, plus some 'ikea style' stuff into the box 2010-11-25 03:42 I'm wondering what this would mean in terms of CE/FCC. 2010-11-25 03:43 not bad idea, I will ask victor to take a look at it, he is more used on that matters 2010-11-25 03:43 when Asus sells bare mainboards, what certs do they pass? 2010-11-25 03:43 can you look at the box of a bare mainboard sold in Spain, and let us know whether they have anything in the direction of CE/FCC? 2010-11-25 03:44 ok I will 2010-11-25 03:44 great, thanks 2010-11-25 03:44 but whatever what is the cost of the case? aprox? 2010-11-25 03:44 and how the linux port is going on? 2010-11-25 03:48 cost of the case not known yet 2010-11-25 03:48 the first wood one was ca. 20 EUR material cost 2010-11-25 03:48 12 EUR for the laser cutter, plus the wood 2010-11-25 03:48 roh's time was for free :-) 2010-11-25 03:48 now roh will experiment and improve some more, acrylic, etc. 2010-11-25 03:49 he is hoping to sell them, and I am 100% willing and ready to buy what he produces 2010-11-25 03:49 acrylic is more expensive than wood 2010-11-25 03:49 but there are many unknowns at this point 2010-11-25 03:49 because so many things in the case need to be improved, so it's not clear what the laser-cutter time will be in the end 2010-11-25 03:49 and whether it can be optimized 2010-11-25 03:50 or whether other production techniques will be added (for example the groove was manually carved out with a screwdriver on the first case) 2010-11-25 03:50 the material cost will change (up or down, when buying larger amounts it may also go down) 2010-11-25 03:50 and roh may need to charge some for his time 2010-11-25 03:50 bottom line: the case will be anywhere between 20 and 50 EUR, my feeling 2010-11-25 03:51 I'm not so much trying to bring this down right now, it's more important to get it done, improve details, optimize the process 2010-11-25 03:52 so , what was the price of production of a MM and the intended PVP? 2010-11-25 03:52 tuxbrain: linux port, well, Takeshi Matsuya keeps hacking, but lekernel thinks it's too early for the Linux kernel right now, and he wants to focus on the VJ application and RTEMS 2010-11-25 03:52 PVP? 2010-11-25 03:52 what is that? 2010-11-25 03:52 sorry is spanish price to public 2010-11-25 03:52 xiangfu: that's an awesome trick!!!! 2010-11-25 03:53 we are planning to sell the Milkymist One boards for 350 USD, if we can add the case I suggest we sell it for 499 USD 2010-11-25 03:53 xiangfu: this message was bugging me.. 2010-11-25 03:53 we want to add all sorts of goodies, like power adapter, jtag-serial board, etc. and all this costs money 2010-11-25 03:56 xiangfu: have you tested it? is it enough to start gmenu2x like this, and then all child processes will have the terminal? 2010-11-25 03:56 kyak: I just tested. seems not working 2010-11-25 03:57 @IF 2I =I3 2I H5, 2010-11-25 03:57 how do you do it? 2010-11-25 03:58 wait. maybe I just tested is wrong.  let me  try again. 2010-11-25 04:03 what is milkymist one for? 2010-11-25 04:08 viric: good question! 2010-11-25 04:08 I haven't figured out the one way to describe it yet 2010-11-25 04:08 first it's called an interactive VJ station 2010-11-25 04:09 it's like a computer, but totally geared towards processing audio, video, instruments, lighting systems 2010-11-25 04:10 basically if you would use it as your file server, that's exactly what it is not good at. it's the opposite. 2010-11-25 04:10 at the beginning it will run a real-time kernel, RTEMS, although a Linux kernel port exists already (but not in kernel.org yet) 2010-11-25 04:12 it's a nanonote with a dsp? 2010-11-25 04:12 viric: does that make sense so you? 2010-11-25 04:12 oh no 2010-11-25 04:12 :-) 2010-11-25 04:12 first it's not a mobile computer 2010-11-25 04:13 it needs a lot of power, so it has to be plugged into the wall 2010-11-25 04:13 close to xue? 2010-11-25 04:13 Ah 2010-11-25 04:13 it has no screen, no keyboard 2010-11-25 04:13 just a box 2010-11-25 04:13 who would usually use it? 2010-11-25 04:13 if it's interactive, it may have some way to get 'input' from a user... 2010-11-25 04:13 Is it termenvox-controlled? :) 2010-11-25 04:13 http://en.qi-hardware.com/wiki/File:Milkymist_One_wood_case.jpg 2010-11-25 04:14 I always wanted an open-hardware termenvox 2010-11-25 04:14 it's a computer, you can telnet into it, or ssh under Linux I would think 2010-11-25 04:14 ahh. 2010-11-25 04:14 it's a computer guys 2010-11-25 04:14 of course :-) 2010-11-25 04:14 meant to be used through a network 2010-11-25 04:14 sure 2010-11-25 04:14 Ethernet 2010-11-25 04:14 USB 2010-11-25 04:14 microSD 2010-11-25 04:14 all there 2010-11-25 04:14 ahh 2010-11-25 04:14 but if you see it as a replacement of your aging NSLU2, that's wrong 2010-11-25 04:14 totally wrong 2010-11-25 04:15 the connectors that make it special, as seen from the outside, are : 2010-11-25 04:15 video in 2010-11-25 04:15 vga out, audio in/out 2010-11-25 04:15 midi in/out 2010-11-25 04:15 DMX512 (3-wire) in/out 2010-11-25 04:15 and the hardware design inside is geared towards processing all that stuff 2010-11-25 04:16 I really need to think of a sentence or two to describe it well... 2010-11-25 04:16 http://en.qi-hardware.com/wiki/Milkymist_One 2010-11-25 04:16 is it used by musicians? or who? 2010-11-25 04:17 used by nobody right now, because it doesn't exist yet 2010-11-25 04:17 who is 'musician'? 2010-11-25 04:17 if you like to play the piano, you probably have no use for it 2010-11-25 04:18 if you are Keith Richards, you probably also have no use for it 2010-11-25 04:18 whos "Keith Richards"? :) 2010-11-25 04:18 but sure, like I said, you can plug MIDI instruments into it, and 'do stuff' 2010-11-25 04:18 i mean, it will be used by whom? 2010-11-25 04:18 i don't really udnerstand 2010-11-25 04:18 Keith Richards is an old rocker (guitarist) from the Rolling Stones 2010-11-25 04:18 VJ station 2010-11-25 04:19 you can process audio and video 2010-11-25 04:19 wolfspraul: is it outputting 320x240 video? 2010-11-25 04:19 uh, Rolling Stones i've heard about :) 2010-11-25 04:19 I'm sure it can go much higher 2010-11-25 04:19 what is there doing the heavy video processing? 2010-11-25 04:19 (electronics) 2010-11-25 04:19 right! good question 2010-11-25 04:19 it's all based around a Xilinx Spartan-6 FPGA 2010-11-25 04:19 process audio and video.. right now i'm thinking about connecting it to a TV.. am i wrong? 2010-11-25 04:19 quite powerful FPGA 2010-11-25 04:19 can the fpga be reprogrammed at will? 2010-11-25 04:19 running a GPL licensed CPU design, called... Milkymist! :-) 2010-11-25 04:20 ahh 2010-11-25 04:20 'at will' maybe not yet, but that's one of the exercise to make that more open, yes 2010-11-25 04:20 a spartan6 is better than a dsp there :) 2010-11-25 04:20 hmm 2010-11-25 04:20 that's a very good question 2010-11-25 04:20 the two cannot easily be compared like that I think 2010-11-25 04:20 wpwrak could probably give you a better analogy than me 2010-11-25 04:20 ok 2010-11-25 04:21 I design for spartan3... I know the spartan3 family 2010-11-25 04:21 I'm going to get into spartan6 soon 2010-11-25 04:21 a high-end DSP can be geared towards certain things, and manufactured in a high-end process, so then it can probably beat any FPGA at whatever it was optimized for 2010-11-25 04:21 on the other hand an FPGA is programmable logic, so it's much more flexible 2010-11-25 04:21 viric: oh wow. well then you should definitely consider taking a closer look at Milkymist One. 2010-11-25 04:21 check that link I pasted above 2010-11-25 04:22 http://en.qi-hardware.com/wiki/Milkymist_One 2010-11-25 04:22 here it is again 2010-11-25 04:23 I'll spam about it among people with free time :) 2010-11-25 04:24 thank you 2010-11-25 04:33 that page needs to be MUCH BETTER at explaining what the product is 2010-11-25 04:33 in the first line, first paragraph, etc. 2010-11-25 04:33 but explaining is not so easy, need to zoom out mentally, use the right words to describe it, etc. 2010-11-25 04:33 your feedback is very appreciated 2010-11-25 04:34 we will probably go about marketing it strongly next year (it's not available yet) 2010-11-25 04:37 ok 2010-11-25 04:46 kyak: I think we should remove the first line in /etc/profile "[ -f /etc/banner ] && cat /etc/banner" 2010-11-25 04:54 wait  if we remove the /etc/banner never show up. :(. 2010-11-25 04:54 yes! 2010-11-25 04:54 :-) 2010-11-25 04:55 it's too much clutter 2010-11-25 04:55 I think we should remove the banner entirely. 2010-11-25 04:55 we have way too much text showing up all the time 2010-11-25 04:55 banner, warnings, error messages, what not 2010-11-25 04:55 the less the better, imho 2010-11-25 04:59 when we start the 'gmenu2x', it' always try to run "/etc/profile" it make the banner shows too much. which is not good. 2010-11-25 05:00 sure, first remove the banner from /etc/profile - +1 from me 2010-11-25 05:00 after that, delete /etc/banner entirely - +1 2010-11-25 05:00 :-) 2010-11-25 05:06 yeah, perhaps this banner is not necessary in /etc/profile 2010-11-25 05:12 [commit] Xiangfu Liu: files: remove display banner in /etc/profile http://qi-hw.com/p/openwrt-xburst/011732f 2010-11-25 05:13 viric: (halt -fp) hmm, then busybox is exonerated for once. maybe your kernel just doesn't have the power-off code enabled. i think it lives somewhere in power management. 2010-11-25 05:15 so, the kernel has a syscall for poweroff? 2010-11-25 05:16 and all userland is expected to call that, right? 2010-11-25 05:16 I mean there is a linux-abi defined way of powering off 2010-11-25 05:19 viric: yup, linux-*/kernel/sys.c:reboot (and kernel_power_off) 2010-11-25 05:20 hm I'll check what I have there. 2010-11-25 05:20 thank you! 2010-11-25 05:23 it's not so easy :) then you have to trace what machine_power_off does 2010-11-25 05:23 but that's at least the abi 2010-11-25 05:24 I mean I will read that. 2010-11-25 05:28 xiangfu: for the reflash_ben.sh problem that $? == "8" looks suspicious. Maybe you just try the .ubi whenever the ubi.bz2 wget fails? That would be more common I think... 2010-11-25 05:30 is qi-kernel.git still up to date at all? I'm a bit worried if we leave stale trees around that people will get lost. 2010-11-25 05:33 wolfspraul: I'm using the upstream kernel 2.6.35 + openwrt instead of the qi-hardware git; I remember something determined that choose, but I can't remember. Some talks here, I imagine. 2010-11-25 05:33 kyak: I just found , the new release in our openwrt-xburst.git openwrt-package.git is not tag. it's a new branch. :( 2010-11-25 05:34 wolfspraul: the 'master' branch is totally outdate. 2010-11-25 05:34 wolfspraul: others I am not sure. 2010-11-25 05:34 xiangfu: yeah.. isn't it the same? 2010-11-25 05:35 kyak: I think not same. at least the type is not same. one is tag another is branch :) 2010-11-25 05:36 xiangfu: i thought that it is the same in terms of git :) 2010-11-25 05:36 a branch is kind of "bookrmark" 2010-11-25 05:38 kyak: yes. it's kind of same. but when you use "gitk" "gitg", the display is different. 2010-11-25 05:38 okay.. 2010-11-25 05:39 kyak: IMO. branch which is develop on it. new feature or less feature.  tag just a "bookmark" 2010-11-25 05:39 I just want make less confuse :) we should use tag in release. 2010-11-25 05:41 xiangfu: so who is still using qi-kernel.git ? 2010-11-25 05:41 hmm... but for the last release . make it as tag is not easy. since it cherry-pick some commit. :( 2010-11-25 05:41 or should anyone still use it? if not, why don't we delete it? 2010-11-25 05:42 wolfspraul: hmm... larsc use that develop the 2.6.36. 2010-11-25 05:44 oh really? 2010-11-25 05:44 well then larsc needs to tell us whether it was just a temporary working repository, or whether there is any future use for it, or in which cases we should direct people to use it, or not use it 2010-11-25 05:48 wolfspraul:  then we lost all lars commit history. 2010-11-25 05:48 sure I am not advocating to delete it, I am just trying to understand its function right now. 2010-11-25 05:49 when should someone use it? and how do we maintain it? 2010-11-25 05:49 some people may just find that repository accidentally and basically get 'trapped' 2010-11-25 05:49 and we don't seem to know clearly what to tell people :-) 2010-11-25 05:49 specially those expecting the 'master' branch being "the stable kernel" 2010-11-25 05:50 sure 2010-11-25 05:50 when we develop patches for openwrt. we may need such a repo. 2010-11-25 05:51 I think we can mark it as 'private' 2010-11-25 05:52 when we develop kernel with other peoples . we may need such a repo. 2010-11-25 05:53 private is not good 2010-11-25 05:53 we just need good documentation 2010-11-25 05:53 if the 'master' branch is auto sync with openwrt. that will be great. 2010-11-25 05:53 that is my plan . setup a auto sync in qi-kernel.git. 2010-11-25 05:54 I am a little lazy on qi-kernel.git :) 2010-11-25 05:54 since I am not using it much 2010-11-25 05:55 we just need to document what it is used for, and what to expect from it 2010-11-25 05:55 no need for 'auto sync', that sounds like asking for a lot of trouble 2010-11-25 05:57 xiangfu: well, it's not that easy. 'master' reflects the final source code, while openwrt only has bunch of patches over mainline. 2010-11-25 05:57 I don't think it is easy to sync. It's manual work. 2010-11-25 05:58 additionally there are "the openwrt patches", and "the patches to make the linux work on nanonote" (generic vs xburst-qi_lb60) 2010-11-25 05:58 I imagine larsc tries to make changes so they can go upstream, that is, unrelated to openwrt 2010-11-25 06:00 - well, this is the picture I have in mind, that may be wrong, because I'm not doing any kernel development :) 2010-11-25 06:04 xiangfu: set up a sync with qi-kernel.git - good idea! 2010-11-25 06:05 we need a repo for co-work in kernel develop. that's why there is qi-kernel.git 2010-11-25 06:05 so that when we build from openwrt-xburst.git, we will get the latest kernel, right? 2010-11-25 06:05 (high-end DSP) i need some for GPS correlation ;-) 2010-11-25 06:06 kyak: no. 2010-11-25 06:06 it's like : work on qi-kernel.git --> generate patches for openwrt. 2010-11-25 06:07 this is already how it works, i think? 2010-11-25 06:07 but at sometime. lars direct generate the patches for openwrt. that's why there are different. then I try to sync the openwrt to qi-kerne.git 'mater' because I am also in kernel develop :( 2010-11-25 06:08 because I am slow in kernel develop (sorry for typo) 2010-11-25 06:09 it's like  10 commit in qi-kernel.git --> one patch in openwrt. that is the origin plan. 2010-11-25 06:09 s/origin/original 2010-11-25 06:10 but since lars is working on 2.6.36, but openwrt is 2.6.32, there is some additional work backporting those patches? 2010-11-25 06:10 i suggest going to 2.6.36 in openwrt-xburst :) 2010-11-25 06:11 kyak: maybe. I am not look into the detail. 2010-11-25 06:12 kyak: then we far away from 'backfire' :) 2010-11-25 06:12 xiangfu: that's what i was going to say :) let's follow openwrt trunk 2010-11-25 06:24 kyak: back later (sorry) 2010-11-25 06:24 xiangfu: np! 2010-11-25 06:44 [commit] Xiangfu Liu: new package, tunec, A command line guitar tuner for linux http://qi-hw.com/p/openwrt-packages/1404776 2010-11-25 07:56 kyak: Hi I am back. 2010-11-25 07:56 hey xiangfu 2010-11-25 07:57 kyak: hmm... about track 'trunk', I am not sure. I think we better track 'backfire' because I guess 2010-11-25 07:57 If we track upstream. we may fall into DEBUG OpenWrt. 2010-11-25 07:57 heh 2010-11-25 07:57 guess you are right 2010-11-25 07:58 the 'backfire' is their most stable and last release.  then we can focus on NanoNote :) 2010-11-25 07:59 yeah, makes sense 2010-11-25 07:59 and in addition, mirko is holding out the trunk for us 2010-11-25 07:59 he's our trunk pioneer :-) 2010-11-25 08:00 I will give up 'private' qi-kernel.git, which is not good. yes we need document the qi-kernel.git well. 2010-11-25 08:00 and he says openwrt trunk can build bootable images for the Ben, and he will continue to watch that that remains so 2010-11-25 08:00 xiangfu: but didn't you say we loose some of larsc history then? 2010-11-25 08:01 thanks to Mirko. :) 2010-11-25 08:01 did larsc develop in that qi-kernel.git repository or is the history also elsewhere? 2010-11-25 08:01 wolfspraul: in fact, I am not sure. I just check the commit log. I don't know is there some else. 2010-11-25 08:02 xiangfu: when you say 'give up', what do you mean? do you plan to delete it or make it private? 2010-11-25 08:02 if so, you should announce that on the mailing list first, and wait for feedback 2010-11-25 08:03 since I guess some people have started to use that repository as the starting point for their work, and it would be quite unfriendly if it now suddenly disappeared, even if we realized that we need to clean it up, and maybe remove it 2010-11-25 08:05 wolfspraul: no. no. I mean we should document qi-kernel.git more. make it clear. 2010-11-25 08:06 yes, for sure 2010-11-25 08:06 thinking about if we need sync with openwrt. 2010-11-25 08:06 I don't think so, too much work. 2010-11-25 08:08 hmm... 2010-11-25 08:09 then we can try remove the 'master' branch. and document qi-kernel.git like: 2010-11-25 08:11 "this git is only for xburst cpu upstreaming works. nothing relate to openwrt-xburst.git, please use openwrt-xburst/target/linux/xburst for the last NanoNote kernel" 2010-11-25 08:12 (my bad English, maybe someone write a good one :) 2010-11-25 08:12 sure, if that documentation is accurate (I don't know) 2010-11-25 08:12 English is good, start with that 2010-11-25 08:12 we only find out what it's used for if the people that use it start to speak up (like you just did) 2010-11-25 08:13 i silently use it to build latest 2.6.36 for Ben :) 2010-11-25 08:20 will modify the desc for qi-kernel.git for now. 2010-11-25 08:22 kyak: have you tried to use the latest kernel.org mainline and see how much is missing? 2010-11-25 08:23 wolfspraul: nope, haven't tryed it 2010-11-25 08:25 http://projects.qi-hardware.com/index.php/p/qi-kernel/ 2010-11-25 08:25 some more 'Project Summary ' , hope it getting better. 2010-11-25 08:26 looks pretty underground - nice 2010-11-25 08:30 there is still some message error message come out in tty1(before we start gmenu2x) , 2010-11-25 08:30 all start with "ALSA lib .... : error ..." 2010-11-25 08:30 ah yeah, remember that... 2010-11-25 08:32 kyak: do you know why? how can we track down this error message? 2010-11-25 08:42 xiangfu: not sure.. haven't looked at it closely yet 2010-11-25 08:48 viric: wich spartan3 board do you have? are u coding in verilog? 2010-11-25 08:48 (i wonder if i asked this in the past) 2010-11-25 08:48 no, vhdl. 2010-11-25 08:48 you did not ask, I think 2010-11-25 08:49 I never did verilog. I'm in the old continent :) 2010-11-25 08:49 hehe 2010-11-25 08:52 i did some *very* basic vhdl then discover verilog and start learning about ti 2010-11-25 08:53 ok 2010-11-25 09:03 [commit] Xiangfu Liu: change the way of check wget error, which is more common http://qi-hw.com/p/openwrt-xburst/d3e685d 2010-11-25 09:03 wolfspraul: the " hooks/post-update: ...." message gone. 2010-11-25 09:04 good, thanks 2010-11-25 09:15 kyak: will add them message to my TODO list :) 2010-11-25 09:16 s/them/those ALSA LIB/ 2010-11-25 09:23 xiangfu: great ;) 2010-11-25 10:10 roh: hi 2010-11-25 10:12 roh: i'm doing a quick presentation tonight at atelier überall (near Schlesisches Tor)- do you have some pics of the acrylic case (or even the case itself :)? 2010-11-25 10:16 lekernel i havent done milling yet. and it didnt come out as nice. wrong speed etc... 2010-11-25 10:16 have the correct params now (and a process to find them (testpatterns)) 2010-11-25 10:17 but i need more acryllic and to complete milling. need to build some tool to help me align the lasered parts 2010-11-25 10:38 ok, i'll take the wooden case then :) 2010-11-25 13:14 [commit] kyak: hide warning messages on gmenu2x start http://qi-hw.com/p/openwrt-xburst/ade739d 2010-11-25 13:15 [commit] kyak: Revert "hide warning messages on gmenu2x start" http://qi-hw.com/p/openwrt-xburst/f0ca65e 2010-11-25 13:18 kyak: regret? why 2010-11-25 13:19 kristianpaul: when a shell is started afterwards, it's redirected to /dev/null, too 2010-11-25 13:19 must be some other way.. 2010-11-25 13:19 ahh 2010-11-25 13:19 :) 2010-11-25 13:20 i wonder if is just a debug parameter in gmenu 2010-11-25 16:44 good openqrt xbusrt finished compilation with on error (btw i was commits from yday not today) 2010-11-25 16:44 is GMT-5 here. 2010-11-25 16:44 wonder if commits are in UTC ;) 2010-11-25 23:20 grmbl. this seems to be disk failure season. the one in the pc controlling my lab instruments is making scary noises and now the one in the pc controlling the mill doesn't mount :-(