00:01.13 | tuff | Hello guys |
00:01.39 | tuff | Please tell me where can I get the newest haret exe? |
00:02.48 | tuff | to play with HTC Alpine. All versions I find in internet don't support wpxa27xgpios command or don't support something else... |
00:03.17 | Marex | wgpio <time> |
00:03.29 | Marex | it's in handhelds.org wiki, use google |
00:04.03 | dcordes | somebody here with OE who can try building blueangel kernel? |
00:05.39 | tuff | the hh.org wiki documents some other version with another output format etc. |
00:08.08 | tuff | ok then could you tell me please how can I ignore some GPIOS in wgpio with this version I'm running? (HaRET pre-0.5.2-20080304_084227 ) |
00:09.35 | tuff | also, "dump gpio" gives me only 84 gpios, while my pxa270 has 120 |
00:13.14 | tuff | I mean, "wgpio 1" gives me a lot of changing gpios in the format "GPLR[77] changed to 1". I'm trying to ignore some of them using "ibit gpios 77". But it does not work, the 77 GPIO doesn't get ignored |
00:13.49 | Kevin2 | tuff: Things have change quite a bit. Read the docs at: http://www.handhelds.org/moin/moin.cgi/HaRET_20Documentation |
00:14.47 | Kevin2 | wgpio is the old command - the new command is "watch gpios". |
00:18.15 | tuff | Thanks for the google answer, but I'm googling the wikis and forums on this matter quite long, and I thought this place would give me some definitive answers... at http://www.handhelds.org/moin/moin.cgi/HaRET_20Documentation the output format od "watch gpios" is "000.998 GPIOS GPBDAT: GPB0(32)=1"; "000.998 GPIOS GPEDAT: GPE5(133)=0" and to ignore these bits i must do "ibit GPIOS 133 32". Here, the "watch gpios 1" output looks like: |
00:18.15 | tuff | "000999: mem GPIOS(2) aee00008=1001ebff (00001000)" and i just don't know how to write "ibit" for this |
00:25.58 | dcordes | tuff, well do you have the recent version now just to make sure? |
00:26.47 | tuff | dcordes: I have HaRET pre-0.5.2-20080304_084227 and official 0.5.1 as well, giving the same output |
00:28.19 | dcordes | 000999: mem GPIOS(2) aee00008=1001ebff (00001000) |
00:28.20 | Kevin2 | tuff: You need to use haretconsole - see the section on haretconsole at the link above. |
00:28.28 | dcordes | sorry |
00:28.50 | tuff | thanks I was using telnet. ok will try haretconsole |
00:29.00 | dcordes | haretconsole also uses telnet |
00:30.01 | Marex | htcsable_kbd.c ... are you serious ? |
00:31.24 | Kevin2 | dcordes: haretconsole will decode the raw memory dumps into something more readable. |
00:31.35 | tuff | Yes, much more readable! thanks! |
01:33.42 | *** join/#htc-linux ndavey (n=neil@pc075650.workstation.griffith.edu.au) |
01:44.42 | Marex | cr2, please test patch 07 |
02:46.19 | Marex | cr2, Im spamming kernel-bugs :p |
02:46.56 | *** join/#htc-linux joschkopf79 (n=joschi@85.179.50.38) |
03:14.55 | Marex | cr2, still awake ? |
03:33.25 | *** join/#htc-linux TimRiker (n=timr@rikers.org) |
04:52.21 | *** join/#htc-linux wireddd (n=wired@unaffiliated/wireddd) |
05:45.59 | *** join/#htc-linux hlbot (n=adm@iclem.net) |
05:46.21 | *** join/#htc-linux hlbot (n=adm@iclem.net) |
05:46.29 | *** join/#htc-linux joschkopf79 (n=joschi@85.179.50.38) |
05:55.49 | *** join/#htc-linux dereks (n=dereks@cpe-66-65-46-91.nyc.res.rr.com) [NETSPLIT VICTIM] |
05:56.06 | *** join/#htc-linux hlbot (n=adm@iclem.net) [NETSPLIT VICTIM] |
05:56.30 | *** join/#htc-linux cr2 (n=konversa@crpl6.physik.uni-wuppertal.de) [NETSPLIT VICTIM] |
06:51.52 | *** join/#htc-linux goxboxlive (n=goxboxli@208.84-48-176.nextgentel.com) |
06:54.25 | *** join/#htc-linux kiozen (n=kiozen@p5492A7E3.dip0.t-ipconnect.de) |
06:59.29 | *** join/#htc-linux BabelO (n=Fabrice@lun34-2-82-238-28-28.fbx.proxad.net) |
07:18.14 | *** join/#htc-linux kiozen (n=kiozen@p5492A7E3.dip0.t-ipconnect.de) |
07:33.48 | *** join/#htc-linux rmoravcik (n=rmoravci@gtsgw.ttc.cz) |
07:52.51 | *** join/#htc-linux Babel1 (n=fcr@lun34-2-82-238-28-28.fbx.proxad.net) |
07:55.56 | *** join/#htc-linux patoche (n=pat@77.197.28.145) |
08:21.29 | *** join/#htc-linux SmallR2002 (n=SmallR20@79-77-75-194.dynamic.dsl.as9105.com) |
09:00.32 | *** join/#htc-linux rob_w|laptop (n=rob@p549B9C6E.dip0.t-ipconnect.de) |
09:53.56 | cr2 | Marex: i'm not subscribed to kernel-bugs :) |
09:58.35 | kiozen | cr2: I have a gdal question |
10:00.30 | kiozen | I have 4 points easting/northing on a tif, and want to know the x/y pixel of other easting/northings |
10:00.52 | kiozen | the grid is not square pixel |
10:00.57 | kiozen | but rotated |
10:01.26 | kiozen | would you know what gdal func to use? GDALInvGeoTransform? |
10:08.18 | *** join/#htc-linux hlbot (n=adm@iclem.net) |
10:13.29 | *** join/#htc-linux Marex (n=marex@vasut.kolej.mff.cuni.cz) |
10:17.13 | kiozen | cr2: http://www.qlandkarte.org/shot2.png that's what I want to do |
10:19.15 | Marex | cr2, hi, why didnt the latest patches arrived to kernel-bugs yet? |
10:19.32 | *** join/#htc-linux goxboxlive (n=goxboxli@208.84-48-176.nextgentel.com) |
10:19.43 | Marex | I sent a few mails there and they havent arrived yet |
10:56.15 | kiozen | cr2: forget about it, have found it :) |
11:24.36 | *** join/#htc-linux Marex (n=marex@vasut.kolej.mff.cuni.cz) |
11:24.45 | *** join/#htc-linux Babel1 (n=fcr@lun34-2-82-238-28-28.fbx.proxad.net) |
11:24.53 | Babel1 | hi |
11:25.04 | Babel1 | kiozen: you shot look very nice |
11:44.52 | kiozen | BabelO_: thanks, hope the M version will be as nice somedays :) |
13:03.00 | *** join/#htc-linux dcordes (n=snoopdog@unaffiliated/dcordes) |
13:20.47 | dcordes | hi |
13:29.20 | BabelO_ | hi dcordes |
13:29.38 | BabelO_ | kiozen: so it use the grid of the map to set the good scale factor ? |
13:29.51 | dcordes | BabelO, hey BabelO.do you have OE local? |
13:30.28 | kiozen | BabelO_: well you need points with known geo position to georeference the bitmap |
13:30.31 | BabelO_ | yes i ve it on my home computer |
13:30.41 | kiozen | you can use road crossings |
13:30.51 | BabelO_ | kiozen: and the grid of the map are well know point ;) |
13:30.53 | kiozen | private waypoints |
13:31.00 | kiozen | or well, the grid |
13:31.19 | dcordes | BabelO, in OE something is very wrong with BA kernel. I tried like an hour to build yesterday after cr2 andyou commited but it didn't work |
13:31.20 | BabelO_ | kiozen: so maybe an autodetect softawre can use that ;) |
13:31.42 | BabelO_ | dcordes: what is your local.conf file ? |
13:31.49 | BabelO_ | dcordes: wait i port mine |
13:31.56 | BabelO_ | s/port/post/ |
13:32.02 | kiozen | hm, with correlation you might be able to detect that kind of stuff |
13:32.04 | dcordes | ok thnx |
13:32.33 | kiozen | but you will need a biiiiiiiig computer |
13:33.10 | kiozen | but if you define 3 grid crossings |
13:33.30 | kiozen | you can do it with a 568, too |
13:33.40 | kiozen | s/568/586/ |
13:35.11 | BabelO_ | kiozen: not so big computer, i ve a friend who works on software to align satellite picture to vector map |
13:35.17 | dcordes | kiozen, were able to scan your bigmap yet :) ? |
13:35.50 | kiozen | dcordes: yes, but with a A4 scanner it's still a PTIA |
13:36.27 | kiozen | BabelO_: satellite pictures have better crosses (more contras) |
13:36.59 | kiozen | the ones on the scanned paper map are hard to detect |
13:38.48 | BabelO_ | dcordes: my conf file -> http://pastebin.ca/946117 check PREFERRED_VERSION_linux-handhelds-2.6_* i think my settings are wrong ;) |
13:39.39 | dcordes | thanks |
13:44.11 | dcordes | BabelO, do you know what filesystem supports are written into ba kernel from alge? |
13:45.04 | BabelO_ | dcordes: i think exté/3 and vfat |
13:45.13 | BabelO_ | s/exté/ext2/ |
13:45.27 | BabelO_ | sorry too much wine :) |
13:45.34 | dcordes | ok I will make ext2 then. journaling not needed |
13:45.44 | dcordes | BabelO, on monday :D ? |
13:46.33 | BabelO_ | dcordes: yes, just come back from lunch with customer :) |
13:47.09 | dcordes | are you salesman? |
13:48.03 | BabelO_ | dcordes: no , just project/solution manager :) |
13:53.46 | dcordes | BabelO, in computer stuff? |
14:00.44 | dcordes | BA is a bit bulkier than I exppected it to be |
14:11.41 | dcordes | what does console=ttyS0,115200n8 make? |
14:12.07 | BabelO_ | dcordes: nothing ttyS0 is GSM modem |
14:12.30 | BabelO_ | dcordes: yes Computer OCR :) |
14:13.54 | dcordes | BabelO, I get kernel panic "can't mount rootfs on unknown block". I have sd with 2gb and 2 partitions: p1 ~150mb fat16 and kernel haret, haretconfig; p2 rest with ext2 and rootfs |
14:14.48 | dcordes | set cmdline "root=/dev/mmcblk0p2 rootdelay=10 console=ttyS0,115200n8 console=tty0" |
14:15.45 | BabelO_ | dcordes: you have two partition on your sd ? |
14:15.53 | BabelO_ | ok |
14:15.59 | dcordes | yes p1 fat16, p2 ext2 |
14:16.03 | dcordes | seems ok to me... |
14:16.13 | dcordes | I know the procedure from prophet try |
14:17.11 | BabelO_ | dcordes: is it your own compiled kernel ? |
14:17.19 | dcordes | no wait |
14:17.41 | dcordes | http://algepop.net/users/alge/angstrom-ba/images/htcblueangel/ |
14:18.26 | dcordes | statup.txt, zImage.bin, Angstrom-gpe-image-glibc-ipk-2007.12-r1-htcblueangel.rootfs.tar, |
14:18.34 | dcordes | haret is from kev's latest build |
14:20.28 | dcordes | BabelO, any idea? |
14:20.40 | BabelO_ | dcordes: i think partition is ext3 on my ba, wait ... |
14:20.42 | dcordes | both p1, and p2 mount perfect on my desktop |
14:20.45 | dcordes | ok |
14:22.40 | BabelO_ | dcordes: add mem=128M to your command line |
14:22.57 | BabelO_ | dcordes: you have wm5 or wm3 on device ? |
14:23.16 | dcordes | I don't know. it looks old |
14:26.16 | BabelO_ | dcordes: does it boot ? |
14:26.30 | dcordes | not tried I have problems with my sd |
14:26.47 | dcordes | shall I make ext3 or ext2 now? |
14:27.30 | BabelO_ | (15:22:40) BabelO_: dcordes: add mem=128M to your command line |
14:27.45 | BabelO_ | keep ext2 |
14:28.18 | BabelO_ | dcordes: when you reboot the ba do you havee to do calibration step or the copy / paste stuff ? |
14:28.35 | dcordes | http://algepop.net/users/alge/angstrom-ba/images/htcblueangel/ <- zImage and rootfs are both same size. are they symlink? |
14:28.45 | dcordes | BabelO, yes :) |
14:28.56 | BabelO_ | dcordes: ok so it is wm3 |
14:30.34 | dcordes | set cmdline "root=/dev/mmcblk0p2 rootdelay=10 console=ttyS0,115200n8 console=tty0 mem=128M" |
14:32.43 | dcordes | BabelO, is that ok? |
14:33.20 | PCrazee | where can i change the keyboard layout? or is there some way to define it myself? would be nice to have esc, alt, and ctrl on my universal keyboard... and there are some buttons i really dont need |
14:37.00 | dcordes | BabelO, now it dies on Jumping to kernel.. |
14:45.09 | dcordes | BabelO, booting. with mem=128M it does not work |
14:45.41 | dcordes | before the error was my sd card. it mounted ro because it is broken |
14:46.51 | *** join/#htc-linux TimRiker (n=timr@70-7-201-54.area7.spcsdns.net) |
14:53.55 | dcordes | BabelO, what's the name of keyboard module? |
14:54.18 | *** join/#htc-linux patoche (n=pat@77.197.28.145) |
14:55.07 | *** part/#htc-linux patoche (n=pat@77.197.28.145) |
15:03.43 | *** join/#htc-linux bernt__ (n=bernt@dslb-084-061-224-071.pools.arcor-ip.net) |
15:06.01 | BabelO_ | dcordes: remove your console=ttyS0 from command line and try the kernel from linuxtogo.org/~htcpxa jus to see if it boot |
15:06.16 | BabelO_ | dcordes: the kernel module is blueangel_kp, maybe built in |
15:07.57 | dcordes | BabelO, is kernel from there newer? |
15:12.48 | BabelO_ | dcordes: no, but i want to see a working kernel ;) and i think mine are compiled with ext2 |
15:28.09 | kiozen | cr2: now referencing is some kind of fun, svn up ! |
15:29.38 | dcordes | I killed ba keyboard module now second time |
15:33.29 | dcordes | BabelO, does alge images use gpsd? |
15:33.43 | dcordes | I wonder if I can take openmoko image and insert alge kernel and modules |
15:36.47 | *** join/#htc-linux SmallR2002 (n=SmallR20@79-77-75-194.dynamic.dsl.as9105.com) |
15:40.25 | *** join/#htc-linux noelcragg (n=noel@adsl-69-232-228-255.dsl.pltn13.pacbell.net) |
15:43.47 | *** join/#htc-linux Babel1 (n=fcr@lun34-2-82-238-28-28.fbx.proxad.net) |
15:45.46 | dcordes | Babel1, couldyou again pastebin your local.conf for ba? |
15:47.19 | Babel1 | dcordes: http://pastebin.ca/946117 from logs |
15:49.04 | dcordes | Babel1, I can reach pastebin.ca host :( |
15:49.11 | dcordes | s/can/can't/ |
15:49.26 | babelO_ | dcordes: ipv6 ? ;) |
15:49.50 | dcordes | me? don't know |
15:50.00 | dcordes | I have uboonto |
15:50.51 | babelO_ | dcordes: http://pastebin.com/m63c6c524 |
15:52.15 | *** join/#htc-linux the_sys0p (i=the_sys0@gateway/tor/x-a21ce95d9b34542b) |
15:53.56 | dcordes | BabelO, whats UTGAVE = "20080214"? |
15:57.52 | babelO_ | dcordes: don't know, maybe the release package date added on install |
15:58.22 | dcordes | I commented out all opie/qt/familiar stuff because I build openmoko |
15:58.35 | dcordes | BabelO, you think OM will work with gsmd? |
15:59.18 | babelO_ | dcordes: yes it works, i ve tried it some month ago |
15:59.23 | dcordes | cool |
15:59.32 | babelO_ | dcordes: and with the patched n_tihtc, should work better |
16:00.05 | dcordes | n_tihtc one of the patches cr2 commited? |
16:01.45 | dcordes | BabelO, do you have the local.conf setting that makes use of dualcore? |
16:03.35 | babelO_ | dcordes: no, but interesting by that :) |
16:04.09 | dcordes | it's twice fsat |
16:04.15 | dcordes | I will find option wait |
16:04.40 | dcordes | PARALLEL_MAKE = "-j 4" |
16:04.41 | dcordes | BB_NUMBER_THREADS = "4" |
16:04.46 | dcordes | that's it |
16:05.32 | *** join/#htc-linux noelcragg (n=noel@166.129.254.55) |
16:05.41 | babelO_ | dcordes: ok good so i set it to 8 :) |
16:06.37 | babelO_ | dcordes: for OM, maybe you need 2008.1 as distrib, because 2008.1 use qvga now |
16:07.05 | dcordes | ok |
16:07.10 | dcordes | openmoko-2008.1 ? |
16:07.38 | dcordes | INHERIT += "debian" what'S that? |
16:10.27 | babelO_ | dcordes: don't know, i ask goxboxlive for most of my settings ;) |
16:14.32 | dcordes | I think only important BA specific things in your local.conf are PREFERRED_VERSION_linux-handhelds-2.6_htcblueangel ?= "2.6.21-hh20" and MACHINE = "htcblueangel" |
16:22.36 | dcordes | BabelO, I just had problem making call with alge's gomunicator. it says element phoen not found |
16:22.53 | dcordes | and sounds comes very scratchy |
16:23.00 | cr2 | hi |
16:23.03 | dcordes | like krrrrrrrrrrrrrchhhlkrrrrrrrrrrrggrrhrkrkalsdö |
16:23.09 | dcordes | hi cr2 |
16:23.10 | cr2 | it was a long day today. |
16:23.24 | dcordes | you were sleeping til now? :) |
16:23.31 | cr2 | hehe. |
16:23.40 | cr2 | 8 hours of teaching. |
16:24.00 | cr2 | and the same tomorrow. |
16:24.06 | cr2 | no time for hacking :) |
16:24.18 | dcordes | what a senseless week :D |
16:24.23 | dcordes | I was sleeping til 12 and fiddling with BA since then |
16:24.29 | cr2 | btw, i've forgot to commit the BA defconfig yesterday. |
16:24.50 | dcordes | ok my OE is just busy building depends for virtual/kernel |
16:24.57 | cr2 | well, it's also dull only to do hacking. |
16:25.23 | dcordes | do you teach physics? |
16:25.49 | cr2 | yes. experimental physics. |
16:26.49 | dcordes | I had the ability to visit uni wuppertal few weeks ago. maybe I would have seen one of your shows if I went there :) |
16:27.24 | cr2 | i was in stuttgart two weeks ago for the whole week. |
16:27.51 | dcordes | also lectures there? |
16:28.36 | cr2 | no, it was a linear algebra and supercomputing course there. |
16:29.13 | dcordes | I rule at linear algebra!! |
16:29.45 | cr2 | lol. |
16:30.28 | dcordes | but only at the stuff which is in langenscheidt lineare algebra |
16:31.35 | cr2 | then you'd be able to calculate the eigenvectors for a non-trivial Jordan normal form :) |
16:31.58 | cr2 | i remember it as a big PITA from my own uni time :) |
16:32.30 | cr2 | afaik GSL can't do that. |
16:32.39 | kiozen | cr2: read the log ;) there are easter eggs in GT |
16:32.48 | dcordes | normalenform yes I can do that it's easy. also durchstoÃpunkte and all. |
16:33.14 | cr2 | i need to check the eigenvalue calculation in GSL. |
16:33.28 | cr2 | easy to do on a computer ? |
16:33.38 | dcordes | I didn't say that. |
16:33.39 | cr2 | with free software. |
16:33.54 | cr2 | the math behind it is quite nice. |
16:34.10 | *** join/#htc-linux CVirus (n=GoD@82.201.222.52) |
16:34.17 | cr2 | kiozen: LOL, i'll svn up :) |
16:34.53 | kiozen | cr2: you will appreciate it with your 200 somewhat scanned maps ;) |
16:35.31 | cr2 | kiozen: do you know how to use a predefined palette with 'convert' / imagemagick |
16:36.10 | cr2 | gimp has 4 options for color reduction, but it's not practical to use gimp with so much images. |
16:36.13 | kiozen | nope, have done it with photoshop right after scanning, much easier ;) |
16:36.31 | cr2 | yes, it's also very easy in gimp too. |
16:36.42 | cr2 | just not scriptable. for me. |
16:37.18 | kiozen | you only need a common pallet if you want to stitch tiles together |
16:37.45 | cr2 | yes, i know. |
16:38.06 | cr2 | but it also reduces uglyness on the jpeg maps. |
16:38.39 | kiozen | brrr, jpg "weiche von mir" |
16:39.04 | dcordes | 200 maps? wow |
16:39.30 | dcordes | cr2, can you commit defconfig-htcblueangel? |
16:40.02 | cr2 | kiozen: not from you, and not my scans. but there are many jpeg maps in the net. |
16:40.30 | kiozen | hm, why cant people use sane format for maps |
16:40.38 | cr2 | kiozen: i'll also try to calibrate some historic Lambert-projected maps. it's was all GK so far. |
16:40.51 | cr2 | kiozen: size :( |
16:41.00 | kiozen | A4 600dpi scan ~100MB png, 18MB tif |
16:41.21 | *** join/#htc-linux edison (i=oystein@homelien.no) |
16:41.27 | kiozen | size ? |
16:41.28 | cr2 | tif should not be much less than png. |
16:42.01 | kiozen | it is much less as it only needs 8bit per pixel instead of 4*32bit |
16:42.06 | cr2 | yes, jpeg is smaller if you "just" want to pack 24bit |
16:42.39 | kiozen | oops 4*8bit :) |
16:42.49 | cr2 | it's too much knowledge for a simple scaning program sold with the scanner. |
16:43.08 | cr2 | i always save lzw tiff. |
16:43.22 | kiozen | me too |
16:43.23 | cr2 | because such programs usually don't support deflate. |
16:43.53 | *** join/#htc-linux Babel1 (n=fcr@lun34-2-82-238-28-28.fbx.proxad.net) |
16:44.59 | cr2 | kiozen: GT probably should not show the old map at startup. |
16:45.18 | kiozen | yes it should, that is what I need most |
16:45.47 | kiozen | but it's not save for privacy ;) |
16:45.49 | cr2 | calibrated or uncalibrated ? |
16:46.00 | dcordes | guys I have quite a problem with BA kbd. is there a layout for the german layout? |
16:46.01 | kiozen | the last, what ever |
16:46.25 | cr2 | i've exited from calibration several times, but the map pops up again and again. |
16:46.39 | dcordes | I can't make - / |
16:46.42 | dcordes | and | |
16:46.43 | cr2 | dcordes: it's a 'loadkeys' issue. |
16:46.59 | *** join/#htc-linux noelcragg (n=noel@adsl-66-127-210-245.dsl.snfc21.pacbell.net) |
16:47.02 | cr2 | dcordes: i was suffering the same. also on the uni. |
16:47.16 | cr2 | dcordes: i'll commit the defconfig RSN. |
16:47.23 | cr2 | kiozen: grid tool ? :) |
16:47.24 | kiozen | cr2: sure as it was the last, select another one and you will see that |
16:47.28 | kiozen | heh |
16:47.39 | cr2 | wow. |
16:47.50 | kiozen | referencing a whole tile by one point |
16:48.06 | dcordes | cr2, sorry, what's RSN? |
16:48.32 | cr2 | RealSoonNow |
16:48.52 | cr2 | kiozen: will it work with Lambert ? :) |
16:49.00 | cr2 | ok, it was a joke. |
16:49.03 | kiozen | don't know |
16:49.19 | kiozen | but works greate with utm grids |
16:49.20 | cr2 | very nice indeed. |
16:49.33 | cr2 | utm=TM=GK. |
16:49.39 | kiozen | and it leaves your brain in a sane state |
16:49.49 | dcordes | cr2, did you change anything about the keyboard there or was that just for the record? because I think I am running a kernel right now, configured with the defconf you are going to commiut |
16:50.06 | cr2 | dcordes: just fyi |
16:50.09 | dcordes | kk |
16:51.50 | cr2 | kiozen: the problem is that my 200 maps don't have the grid. i'd like to pick the crossroads from another georeferenced map, and adjust them on the page being calibrated. that's why i talked about DND. |
16:52.26 | kiozen | ok, will do that in the evening. |
16:52.58 | kiozen | scanning 200 maps without grid, was that for detention? |
16:53.32 | cr2 | kiozen: but i'll test it on the gridded map too. they are often distored in the middle, where skipped. |
16:54.53 | cr2 | kiozen: it's unusual that the city maps have the grid. look at the stadtplandienst.de |
16:55.52 | kiozen | havent had a city map in my hands for more than a year, have a 60SCx with CityNavigator :^) |
16:58.10 | cr2 | i'd like to make an overlay. the vector maps are often buggy. |
16:59.49 | cr2 | kiozen: got a segfault on exit. need to run it in gdb. |
16:59.49 | kiozen | cr2: yes vector overlay is still a big dream, on eday ... |
17:00.16 | cr2 | kiozen: when the maps are referenced, i can use qgis for overlay. |
17:01.07 | cr2 | GT with calibration, +M with moving map, and i'll be very happy. |
17:01.58 | kiozen | maybe M with vector overlay for routing ;) |
17:02.27 | kiozen | routing on the top10 will be fun |
17:02.37 | cr2 | the graph behind the map. |
17:03.01 | cr2 | vector overlay will only produce a lot of clutter. |
17:03.54 | cr2 | there was some basic qt program for route planning. |
17:04.01 | kiozen | less drawing it more using it for routing and displaying the route |
17:04.23 | cr2 | even with turn restrictions, if i'm not mistaken. |
17:04.25 | cr2 | ok. |
17:04.53 | cr2 | kiozen: http://www.mapserver.sk/~wonder/qnavigator |
17:05.08 | dcordes | can't get mixer element 'Phone' |
17:05.15 | dcordes | starting gomunicator |
17:06.03 | kiozen | nice map rendering |
17:06.14 | dcordes | note: hh.org build fine for BA w/o new defconfig |
17:07.09 | cr2 | dcordes: ok, but i'll run a diff between the old and the new. |
17:09.15 | *** join/#htc-linux pH5 (n=ph5@e178197198.adsl.alicedsl.de) |
17:09.42 | kiozen | cr2: think I killed the segfault |
17:09.54 | dcordes | cr2, ok. I don't know if OE fetched a new BA or if it took the one which lies in OE metadata... |
17:10.04 | dcordes | all I know is it just build with recent mtn |
17:10.08 | dcordes | hi pH5 |
17:10.31 | dcordes | s/BA/BA defconfig/ |
17:10.41 | pH5 | hi |
17:13.47 | BabelO_ | tf rsn |
17:13.56 | BabelO_ | ~wtf rsn |
17:14.05 | BabelO_ | hi pH5, cr2 |
17:14.36 | kiozen | bbl |
17:16.38 | cr2 | dcordes: the changes are big, but i'll commit it anyway. |
17:17.03 | *** join/#htc-linux mxer (n=sativa@ip68-230-62-137.ph.ph.cox.net) |
17:17.51 | cr2 | dcordes: done. |
17:17.55 | dcordes | cr2, great |
17:27.09 | cr2 | where is kiozen when you need him ? :) |
17:28.15 | cr2 | kiozen: the program should do error attenuation instead of the error amplification. i think some changes are needed in the grid tool. |
17:30.18 | dcordes | cr2, OE takes its own defconfig linux-handhelds-2.6_2.6.21-hh20.bb, l15: file://defconfig" |
17:31.10 | dcordes | maybe it would be better to remove this from the .bb |
17:31.25 | dcordes | since optimal defconfig is in cvs |
17:32.48 | dcordes | pH5, in case I remove that line from the bb, will bitbake then build with the right defconfig from cvs? |
17:34.42 | pH5 | dcordes: no, it won't. |
17:34.55 | cr2 | dcordes: i've said it many times before, but the opinion is that OE is not a tool for some bleeding edge kernel development, but a stable config tool. |
17:35.24 | cr2 | dcordes: that's why it's a good idea to build the kernel without oe. |
17:35.54 | cr2 | dcordes: and to build the 'bootbundle' instead of the zimage+other+haret+nastyness :) |
17:36.14 | pH5 | dcordes: I do agree with cr2 here. I'm using the oe-built toolchains, but for kernel development there's just too much overhead with bitbake. |
17:37.37 | pH5 | the correct OE way is to fix the blueangel defconfig recipes in defconfigman from which the defconfigs in oe are built. |
17:37.57 | cr2 | once we will have a stable working kernel and can reflash the wince, this will change. |
17:38.03 | dcordes | I think everything else beside defconfig is ok now.. |
17:38.49 | pH5 | hrghn. we can reflash wince, it just doesn't boot :/ |
17:38.53 | cr2 | dcordes: actually, i don't see any reason why BA needs a separate setup from, for example, magician |
17:38.59 | cr2 | lol |
17:39.19 | cr2 | pH5: vpp on ba ? |
17:39.54 | cr2 | dcordes: btw, you may check the docg3.c driver on ba |
17:40.32 | cr2 | it may be a bit buggy on ba, because of the 8bit vs. 16bit access. then we should fix it. |
17:40.37 | dcordes | what's that? |
17:41.02 | cr2 | ~/htcuniversal/docg3.c |
17:41.12 | dcordes | I'm on ba |
17:41.43 | cr2 | does not matter. |
17:41.56 | dcordes | can I try this out on the alge image? |
17:42.00 | cr2 | it's the same mdocg3 chip. |
17:42.09 | cr2 | compile it as a module. |
17:42.18 | dcordes | ok |
17:42.20 | cr2 | there is nothing uni-specific there. |
17:42.30 | dcordes | I think suspend doesn't work for me |
17:42.35 | dcordes | I got a white screen after resume |
17:42.44 | cr2 | boardid 6 ? |
17:42.53 | dcordes | how do I check again? |
17:42.57 | cr2 | dmesg |
17:43.07 | cr2 | dmesg | grep -i boardid |
17:43.10 | cr2 | after boot |
17:43.56 | dcordes | ok |
17:44.43 | *** join/#htc-linux LunohoD_ (n=alex@e180075224.adsl.alicedsl.de) |
17:46.07 | dcordes | tippen sie auf den touchscreen um ihren pocketpc einzurichten. |
17:51.19 | dcordes | bb iab |
18:04.42 | *** join/#htc-linux Marex (n=marex@vasut.kolej.mff.cuni.cz) |
18:18.54 | dcordes | cr2, wm3 keeps telling me something about emergency battery being empty |
18:21.50 | *** join/#htc-linux tsdogs (n=tsdogs@net-84-18-144-39.c.bkom.it) |
18:22.56 | dcordes | hi Marex, tsdogs |
18:23.02 | Marex | hi dcordes |
18:23.10 | Marex | cr2, hi, I managed to start beetles bootloader in qemu :p |
18:23.55 | tsdogs | dcordes: hi |
18:25.45 | dcordes | cr2, dmesg | grep board empty |
18:27.29 | tsdogs | cr2: to create a common palette with imagemagic: |
18:27.47 | tsdogs | convert *.tif +map -colors 256 common/palette.tif |
18:27.58 | tsdogs | mogrify -map common/palette.tif *.tif |
18:28.01 | *** join/#htc-linux diogene31_ (n=rj@mur31-2-82-243-122-54.fbx.proxad.net) |
18:28.15 | tsdogs | it's pretty slow and it takes forever with a lot of files. |
18:30.12 | pH5 | dcordes: cat /proc/cpuinfo |
18:30.17 | pH5 | revision should contain the board id |
18:30.50 | BabelO_ | dcordes: can you pastebin your dmesg output ? and how did you manage to start it ? |
18:31.40 | BabelO_ | ola tsdogs |
18:31.42 | *** join/#htc-linux edison (i=oystein@homelien.no) [NETSPLIT VICTIM] |
18:31.50 | tsdogs | hi BabelO_ |
18:32.05 | dcordes | BabelO, it was all about my sd card: it was ro |
18:32.13 | dcordes | hi again |
18:33.13 | dcordes | 0006 |
18:33.15 | dcordes | Revision |
18:33.25 | dcordes | is that the screwy board id? |
18:33.56 | pH5 | yes |
18:34.13 | BabelO_ | cr2: i already try docg3 on ba without success |
18:34.20 | *** part/#htc-linux noelcragg (n=noel@adsl-66-127-210-245.dsl.snfc21.pacbell.net) |
18:34.53 | BabelO_ | dcordes: if you want resume work then you need to remove some gpio for suspend mode and thenit works, but do not save power :( |
18:35.21 | dcordes | BabelO, that's death sentence for my BA long term use |
18:35.27 | BabelO_ | dcordes: your sd card was ro , but you put image on it ? |
18:36.03 | dcordes | BabelO, a bit complicated: the small swith on the card is broken so I taped the card. In my card reader it seems to mount rw fine, but not in BA |
18:36.42 | dcordes | is there a script for boardid 6 gpio changing? |
18:41.37 | dcordes | pH5, is there a way to dial with bash? |
18:43.01 | *** join/#htc-linux rmoravcik (n=rmoravci@ip-89-102-255-171.karneval.cz) |
18:43.48 | pH5 | dcordes: do you have gsmd running? libgsmd-tool --mode atcmd, ATD012345; |
18:44.25 | pH5 | argh, the memory corruption I'm seeing since some weeks is only when booting from the bootloader, haret booting works :/ |
18:44.55 | pH5 | ..sometimes |
18:46.04 | dcordes | pH5, what are you booting from which bootlaoder? |
18:46.52 | pH5 | dcordes: whenever I'm babbling it's most probably magician |
18:47.03 | pH5 | (booting from htc bootloader) |
18:47.06 | dcordes | oh ok. so you can |
18:47.18 | dcordes | ah that's the project in ltg git? |
18:47.44 | pH5 | htc-tools? yes |
18:48.01 | pH5 | something is subtly wrong though |
18:48.28 | pH5 | I can boot the hanging kernel if I remove a certain debug printk :) |
18:50.44 | dcordes | the _hanging_ kernel? |
18:51.41 | pH5 | I have a kernel that locks during boot if I boot with htc-tools. But it works when booting from haret. It also works with htc-tools if I remove that printk. |
18:53.20 | pH5 | I guess I have to look for another memory region to upload the kernel to. |
18:57.26 | dcordes | good luck |
19:08.53 | dcordes | pH5, is there a mirror for http://www.linuxtogo.org/~htcpxa/? |
19:09.13 | dcordes | whole OE stuff seems pretty overloaded today |
19:10.43 | pH5 | load ~15%, looks like it is chewing on the mailing lists again |
19:10.57 | pH5 | I'm not aware of a mirror, unfortunately. |
19:13.36 | dcordes | also the angstrom feeds are extreme slow |
19:13.38 | pH5 | heh right, moving kernel+atags from 93000000 down to 92020000 makes it boot again |
19:13.50 | pH5 | dcordes: it's the same server |
19:14.12 | BabelO_ | pH5: rocket science like cr2 says often ? ;) |
19:14.17 | pH5 | 92040000 doesn't work. ARGH |
19:14.24 | pH5 | this is a minefield |
19:15.44 | dcordes | don't you have a complete memory dump to sort out what's used what's not? |
19:17.34 | pH5 | dcordes: good point, I don't. I thought I know which regions the bootloader uses from the asm, but obviously I missed something. |
19:19.33 | *** join/#htc-linux dcordes (n=snoopdog@unaffiliated/dcordes) |
19:19.59 | tsdogs | BabelO_: is the library building working in the git tree? |
19:20.10 | tsdogs | tapioca I mean :) |
19:21.11 | BabelO_ | tsdogs: yes perfect ;), now need to move it to third party :) |
19:21.19 | tsdogs | good :) |
19:21.40 | dcordes | do I want to build with arch/arm/configs/htcblueangel_defconfig Babel0? |
19:22.11 | BabelO_ | dcordes: yes, or the own i provide yesturday ;) |
19:22.23 | tsdogs | dcordes: I think _static_ is better ... |
19:22.24 | dcordes | BabelO, cr2 commited it |
19:22.33 | dcordes | I don't know which cr2 commited :( |
19:22.47 | BabelO_ | pH5: can't you write 0x1 to all memeory region then reread them after, and all it htc bootloader ? |
19:22.50 | tsdogs | afair static is for devel and the other is for release (or the other way around :?) |
19:23.36 | BabelO_ | tsdogs: static is produced by psokolovsky tools |
19:23.49 | tsdogs | ok so it's the other way around :) |
19:24.42 | tsdogs | BabelO_ you sure? checking ... |
19:25.55 | tsdogs | BabelO_ ok it's like u say |
19:29.09 | dcordes | I will just kernel26$ cp arch/arm/configs/htcblueangel_defconfig . |
19:30.25 | tsdogs | dcordes: "make htcblueangel_defconfig" should do the trick |
19:31.47 | *** join/#htc-linux kiozen (n=kiozen@rgnb-4db1df82.pool.einsundeins.de) |
19:32.03 | kiozen | cr2: ok, what did you break? |
19:32.09 | pH5 | BabelO_: yes, I should try something like that |
19:32.29 | pH5 | for now I'll try to clean up and submit my magician patches quickly before it hangs again :) |
19:33.01 | dcordes | how can I set output from make to english? |
19:41.23 | tsdogs | dcordes: LANG=C make |
19:41.27 | dcordes | thanks |
19:43.34 | dcordes | http://pastebin.ca/946529 basically it's blueangel.c:547: error: 'blueangel_phone' undeclared here (not in a function) |
19:43.39 | dcordes | BabelO, can you take a look? |
19:43.58 | dcordes | looks like something went wrong while cr2 commited |
19:46.35 | cr2 | kiozen: i have some problem with the points. |
19:46.45 | kiozen | ok |
19:46.51 | dcordes | cr2, can you look at error? |
19:47.00 | cr2 | tsdogs: i have the palette as a gimp .gpl file |
19:47.27 | cr2 | kiozen: i should put the 4 points to what places exactly ? i have 1 km grid. |
19:47.32 | tsdogs | cr2: hmm, no idea on how to use that with imagemagik |
19:47.50 | kiozen | like in the little picture at the grid crossings |
19:48.04 | cr2 | kiozen: i have 522 grid crossings :) |
19:48.23 | kiozen | like in the example picture |
19:48.37 | cr2 | any 1 km square ? |
19:48.53 | kiozen | yes, any |
19:48.59 | cr2 | kiozen: i don't quite understand the logic behind it. |
19:49.09 | cr2 | tsdogs: ok, i'll read the docs. |
19:49.16 | tsdogs | :) |
19:49.21 | kiozen | lol, docs ? |
19:49.53 | tsdogs | cr2: well having a file which ends with .gpl is not an easy search on google :) |
19:50.10 | kiozen | the idea is to define 4 points on the grid to calculate the transformation param |
19:50.18 | cr2 | kiozen: it's better to add the points as far as possible from each other. |
19:50.44 | kiozen | next GT will calculate equally spaced points to all other grid crossings |
19:50.54 | kiozen | finally you can fine tune the placement |
19:51.10 | tsdogs | gg bbl |
19:51.20 | cr2 | kiozen: i do mark 1km square in the NW corner, but then i have terrible distortion in SE corner. |
19:51.51 | cr2 | it's like 200m shift. |
19:52.00 | kiozen | ok, did you give the top left esating/northing and the spacing? |
19:52.09 | kiozen | in [m] |
19:52.18 | cr2 | 1pixel error in SE will be amplified when you have 5000+ pixel image. |
19:52.40 | kiozen | that is why you have to move all points to their final location |
19:52.56 | cr2 | yes, and the NW is ok, but the error increases with distance. |
19:53.09 | kiozen | you can take a 3000 or 5000 m spacing if there are too many points |
19:53.33 | cr2 | yes, i have 1 km grid spacing =522 GCPs |
19:53.55 | kiozen | you should reduce that by increasing the spacing |
19:54.00 | cr2 | isn't it more logical to add 4 points close to the image corners ? |
19:54.24 | cr2 | yes, but i like having so much gcps. |
19:54.39 | kiozen | where ever, it's just 4 gcps to calculate the transformation |
19:54.41 | cr2 | because then i can correct the distortion in the middle. |
19:55.01 | dcordes | cr2, BabelO said he already the thing you wanted me to test. |
19:55.07 | cr2 | ok, but i can enter the coordinates for only 1 of them |
19:55.15 | kiozen | you won't be happy with a gcp every 3000 m? |
19:55.33 | kiozen | yes because the spacing defines the other 3 |
19:55.33 | cr2 | no. think of the map skipped in the middle. |
19:55.51 | cr2 | or a book. |
19:56.28 | kiozen | such a map sucks anyway ;) |
19:56.45 | cr2 | yes, but sometimes you have no choice. |
19:57.04 | cr2 | with old books you don't want to stress them mechanically. |
19:57.21 | cr2 | and the libraries will be unhappy anyway if you'll do. |
19:57.40 | kiozen | then you have to scan the first page and next the 2nd and reference them |
19:58.09 | kiozen | or you have to move the calculated gcps a bit more than a few pixel |
19:58.12 | cr2 | kiozen: i'm looking for a wikipedia sample for you :) |
19:59.34 | cr2 | and you've obviously seen the npemaps site. |
19:59.50 | kiozen | no |
20:00.04 | cr2 | with the "photoshop" rectification, that looks like shit. |
20:00.08 | cr2 | hmm. |
20:00.26 | cr2 | http://www.npemap.org.uk |
20:01.53 | cr2 | dcordes: yes, i've just remembered. |
20:01.55 | kiozen | it isn't that bad after all, photoshop works in many cases, gimp let's you look dump |
20:02.28 | cr2 | kiozen: it's not about photoshop vs. gimp |
20:02.38 | *** join/#htc-linux CVirus (n=GoD@82.201.222.52) |
20:02.58 | cr2 | it's about the old scanned maps with severe nonlinear distortions |
20:03.54 | kiozen | you should always correct you input as much as possible before you start referencing it |
20:04.09 | cr2 | referencing by 2 image corners is better than by 1 corner. |
20:04.13 | kiozen | gdal isn't that good to fix all these errors at once |
20:04.40 | cr2 | correcting is exactly the task for gdalwarp. by entering as much gcps as possible |
20:04.53 | cr2 | add the tps transform |
20:05.07 | cr2 | it leaves all gcps at their places. |
20:05.34 | cr2 | -tps option instead of -order N |
20:05.52 | BabelO_ | dcordes: just command the line ;) |
20:05.58 | kiozen | but it's a PITA to define more than 1 point |
20:06.16 | dcordes | BabelO, #? |
20:06.23 | kiozen | can you upload one of your tiles? |
20:06.25 | cr2 | use my 2 point calibration code. already included :) |
20:06.59 | cr2 | hmm. then you need my gcps too. |
20:07.28 | kiozen | isn't there a grid? |
20:08.02 | cr2 | kiozen: yes. but you reference only 1km square in the NW corner. |
20:08.15 | cr2 | and extrapolate to the whole image. |
20:08.34 | kiozen | what happens if you place it more in the middle? |
20:08.37 | cr2 | it's exactly as bad as referencing the whole image by these 4 gcps |
20:09.06 | kiozen | as told you have to finetune the resulting gcps |
20:09.11 | cr2 | ok, it will probably half the extrapolation error. |
20:09.37 | cr2 | if the middle is distorted symmetrically. |
20:10.14 | kiozen | the grid tool is just about placing equal spaced gcps at approx their final location on the map |
20:10.38 | kiozen | the exact position has to be done by the user |
20:10.56 | cr2 | usually you want to push the gcps as far as possible for the least-square fit. |
20:11.46 | cr2 | ok. i'll think about it. |
20:11.53 | dcordes | BabelO, /*&blueangel_phone, */ |
20:14.24 | cr2 | kiozen: imo, it's better to drop the points 2 and 4, and add the box to enter the coordinates of the point 3. but it's up to you to decide :) |
20:15.03 | cr2 | you know, i'm more interested in the entering the gcps by clicking at a referenced geotiff. |
20:15.57 | kiozen | afaik you need 3 gcps to calculate a complete transform, don't you? |
20:16.53 | cr2 | kiozen: no. your transform is identical to what i suggest (mathematically). but my approach reduces the calibaration error. |
20:17.11 | cr2 | we have square pixels on a gridded map. |
20:17.21 | cr2 | i'm talking about GK of course. |
20:17.34 | kiozen | but how do you calculate the rotation without 3 points? |
20:17.34 | cr2 | not Lambert or such. |
20:18.07 | cr2 | the "squareness" of a pixel is the additional constraint. |
20:18.25 | cr2 | funny that you ask it :) |
20:18.34 | cr2 | because it's already working. |
20:18.50 | cr2 | you can calibrate the map by 2 GCPs already. |
20:18.50 | kiozen | but what if your scan is rotated a bit |
20:18.59 | kiozen | only if it is not rotated |
20:19.11 | cr2 | should i write you the formula ? |
20:19.26 | cr2 | 2 GCPs give you 4 values. |
20:19.36 | cr2 | x1,x2, y1, y2 |
20:19.45 | cr2 | so you can have 4 unknowns. |
20:19.56 | cr2 | X0, Y0 for the offset |
20:20.06 | cr2 | alpha for the rotation angle |
20:20.23 | cr2 | and S for the square pixel size (m/pix) |
20:21.01 | *** part/#htc-linux BabelO_ (n=fcr@lun34-2-82-238-28-28.fbx.proxad.net) |
20:21.18 | kiozen | ok i always assume scale x != scale y |
20:21.35 | cr2 | try to reference your rotated map with 2 GCPs yourself. |
20:21.47 | cr2 | it's a pretty wrong assumption. |
20:22.04 | kiozen | ot neccessarily |
20:22.08 | cr2 | even the google/osm tiles have square pixels, so scale x = scale y. |
20:23.31 | cr2 | the nonsquare distortion on GK maps is only the printing error. |
20:23.48 | cr2 | nobody in their right mind print the maps with nonsquare pixels. |
20:24.16 | kiozen | adfGeoTransform1[0] = 278627 |
20:24.16 | kiozen | adfGeoTransform1[1] = 4.23941 |
20:24.16 | kiozen | adfGeoTransform1[2] = -0.153687 |
20:24.16 | kiozen | adfGeoTransform1[3] = 5.1624e+06 |
20:24.16 | kiozen | adfGeoTransform1[4] = -0.189805 |
20:24.17 | kiozen | adfGeoTransform1[5] = -4.25749 |
20:24.19 | cr2 | unless it's some shitty supermarket drive directions. |
20:24.31 | kiozen | that is what I get from my scans |
20:24.40 | *** join/#htc-linux LunohoD (n=alex@e180075224.adsl.alicedsl.de) |
20:25.03 | cr2 | 0.25% error ? |
20:25.18 | cr2 | 4.23941 vs 4.25749 ? |
20:25.31 | cr2 | ok, 0.5 :) |
20:25.41 | kiozen | well if you take it by 5000 px |
20:26.35 | cr2 | can you pastebin the gdalinfo output (without the palette) of the _ref.tif ? |
20:28.00 | cr2 | afaik gdalwarp always produces square pixel, no rotation output. |
20:28.16 | kiozen | that is right http://pastebin.ca/946610 |
20:28.40 | cr2 | Pixel Size = (6.158934954000642,-6.158934954000642) |
20:29.12 | cr2 | so your non-squareness is just compensated by gdalwarp. |
20:29.40 | kiozen | but still leaves me to the problem to have a simple way to place gcps as close to the grid crossings as possible |
20:30.20 | cr2 | hmm. ok. |
20:30.47 | cr2 | then go for entering 4 GCP with coordinates |
20:30.57 | cr2 | and use affine transform. |
20:31.15 | cr2 | to add the rest. |
20:31.39 | kiozen | better to do one for good scans, and two for lousy ones |
20:31.49 | cr2 | yes, agreed. |
20:31.55 | *** join/#htc-linux tsdogs (n=tsdogs@62.123.180.130) |
20:33.48 | cr2 | i'm just curious how long it will take to do a TPS warp with 552 gcps :) |
20:34.13 | cr2 | it's not something you'd like to do on an old 586. |
20:34.48 | kiozen | I can code that in 5 min and you can try |
20:35.01 | kiozen | you said -tps instead of -n? |
20:35.26 | cr2 | from gdal_tps.cpp -> |
20:35.29 | cr2 | It can be quite an expensive operation |
20:35.29 | cr2 | <PROTECTED> |
20:35.29 | cr2 | <PROTECTED> |
20:35.58 | cr2 | yes, '-tps' instead of '-order N' |
20:36.02 | kiozen | piece of cake :) |
20:36.25 | cr2 | 10s is bad for gui response. |
20:36.40 | kiozen | cr2: what is the min num of gcps |
20:36.52 | cr2 | i think 3 |
20:37.06 | cr2 | but i need to think about it. |
20:37.14 | cr2 | how much does a cubic spline need ? |
20:37.21 | cr2 | 4. |
20:37.46 | cr2 | ok, it's not a very simple question. |
20:37.53 | cr2 | i need to think about it :) |
20:39.33 | kiozen | wasn't -n was -order |
20:39.46 | cr2 | [21:35] <cr2> yes, '-tps' instead of '-order N' |
20:40.20 | cr2 | internally '-1' is used in gdal. afair. |
20:42.17 | kiozen | cr2: svn up |
20:42.30 | kiozen | hope that works |
20:42.31 | cr2 | pH5: it seems that i need to look once again at the BL ramboot problem. |
20:42.36 | cr2 | kiozen: ok. thanks. |
20:47.08 | cr2 | kiozen: it worked. comparing to -order 1 and -order 2. 7 GCPs were used on a Lambert map warped into mercator. |
20:48.00 | kiozen | never have seen a lambert map |
20:48.17 | kiozen | or a polynomial one |
20:48.35 | kiozen | the EADS mpr/mph spec talk about them |
20:49.01 | cr2 | interesting. tps is closer to the quadratic. |
20:49.19 | cr2 | which is not surprising, actually. |
20:49.41 | cr2 | wasn't this indian map with the mapinfo calibration a lambert projection ? |
20:50.05 | kiozen | hm, don't know, just tried it :) |
20:50.44 | kiozen | drag-n-drop has to wait. seems to more compicated than to do on an evening |
20:52.06 | cr2 | kiozen: ok. |
20:52.09 | cr2 | http://www.lib.utexas.edu/maps/historical/shepherd_1911 |
20:52.13 | cr2 | http://www.lib.utexas.edu/maps/historical/shepherd_1911/shepherd-c-057.jpg |
20:53.03 | kiozen | ah, ok it's those maps with circular latitude lines |
20:53.48 | cr2 | yes, but depending on the parameters the circles can be very flat |
20:54.29 | kiozen | and this map covers a quite big area |
20:54.50 | cr2 | that's why they are circlic :) |
20:54.59 | kiozen | you ever tried to warp it to meractor? |
20:56.07 | cr2 | yes, but not this one. |
20:56.27 | cr2 | it's interesting to check what will happen. |
20:56.28 | dcordes | <PROTECTED> |
20:56.52 | kiozen | maybe tomorrow if work gets too boring ;) |
20:56.57 | dcordes | cr2, BabelO with line 547 in blueangel.c commented out it works |
20:56.57 | cr2 | :) |
20:57.47 | dcordes | maybe change that in cvs |
20:57.57 | cr2 | kiozen: makes a good low-zoom map :) |
20:58.11 | cr2 | dcordes: what's that ? |
20:58.28 | kiozen | yes for my 1:25000 map collection |
20:58.33 | dcordes | cr2, I explained/pasted on top |
20:58.43 | dcordes | <dcordes> http://pastebin.ca/946529 basically it's blueangel.c:547: error: 'blueangel_phone' undeclared here (not in a function) |
20:59.01 | dcordes | babel told me to comment out the line and kernel builds fine with it |
20:59.18 | cr2 | dcordes: ok. |
20:59.31 | dcordes | cr2, I made "make htcblueangel_defconfig" |
20:59.39 | dcordes | did that give me the defconfig you commited (from babel0)? |
21:00.03 | cr2 | dcordes: done. |
21:00.10 | dcordes | thanks |
21:00.24 | cr2 | yes, the defconfig was copied from the BabelO's site |
21:01.12 | dcordes | ok so I can assume I pretty much built the kernel which is available at alge's. I will try it out right now with openmoko image |
21:02.43 | BabelO | back from work, so dcordes build is complete now, also for openmoko ? |
21:02.59 | BabelO | dcordes: what bitbake did you use ? |
21:03.54 | dcordes | BabelO, I built image normally without OE. Used angstrom toolchain. It worked fine |
21:04.00 | dcordes | with the line commented out. |
21:04.09 | dcordes | But also my OE built finished perfectly |
21:04.12 | BabelO | ok |
21:04.16 | dcordes | BUT there is dorky defeconfig. |
21:04.22 | dcordes | not your defconfig. |
21:04.34 | dcordes | cr2, and pH5 said I want to use my manually built kernel |
21:05.55 | dcordes | NOTE: package linux-handhelds-2.6-2.6.21-hh20: completed; my bitbake version is 1.8.10 |
21:05.57 | *** join/#htc-linux skodde (n=skodde@unaffiliated/skodde) |
21:06.18 | dcordes | where do modules and kernel go on manual build? |
21:06.32 | BabelO | dcordes: sorry, i mean the bitbake command line to build openmoko ;) |
21:06.42 | dcordes | bitbake openmoko-image |
21:06.44 | BabelO | i have to wait your image hten :) |
21:06.49 | dcordes | distro is "openmoko" |
21:07.24 | dcordes | it took me like 4 hours for complete build |
21:07.37 | dcordes | 3,7k tasks |
21:07.58 | cr2 | BabelO: http://www.lib.utexas.edu/maps/historical/shepherd_1911/shepherd-c-061.jpg |
21:08.38 | dcordes | can somebody tell me where I find the modules after make modules? |
21:09.14 | BabelO | cr2: lol, lol, there is only real town here ;) |
21:09.42 | BabelO | cr2: Rodez is lost in mountain ;) |
21:09.43 | cr2 | BabelO: we are looking for the nice low-zoom maps for M :) |
21:09.44 | dcordes | lol I got a haret startup.txt how cool |
21:09.53 | dcordes | you guys are genious |
21:10.21 | BabelO | dcordes: how do you build modules ? with hh kernel or from OE bb files ? |
21:10.29 | dcordes | BabelO, no, manually |
21:10.37 | dcordes | I made "make modules" |
21:10.47 | cr2 | dcordes: argc, **argv ? |
21:10.49 | dcordes | in hh.org cvs 2.6 tree |
21:10.59 | BabelO | dcordes: so do export INSTALL_MOD_PATH=/where is you root install |
21:11.04 | BabelO | then make modules_install |
21:11.16 | dcordes | cool :D |
21:12.09 | dcordes | e.g. / of my openmoko filesystem lives in ~/openmoko/ I do export INSTALL_MOD_PATH=~/openmoko/ ? |
21:12.16 | Marex | cr2, hi, check this http://disk.jabbim.cz/marex@jabber.cz/ipaq.png :P |
21:13.08 | BabelO | dcordes: example is, i mount my installed root fs under /media/sdcard/ so i export INSTALL_MOD_PATH=/media/sdcard |
21:13.44 | dcordes | great |
21:14.00 | BabelO | Marex nice shot, what did you change in qemu ? |
21:14.03 | cr2 | Marex: lol. but you don't have asic3 driver for qemu ? |
21:14.14 | Marex | cr2, nah |
21:14.35 | Marex | BabelO, I used openmoko qemu, ripped spitz.c and made some modifications |
21:14.35 | cr2 | Marex: but you can trace hw accesses |
21:14.36 | Marex | hang on |
21:14.52 | cr2 | Marex: is it possible to run wince in qemu ? |
21:15.03 | Marex | cr2, I just managed to get the bootloader running |
21:15.10 | Marex | hang on, Ill release the source |
21:16.45 | Marex | or maybe patch will be better |
21:18.55 | Marex | cr2, BabelO http://disk.jabbim.cz/marex@jabber.cz/ipaq.patch here you go |
21:19.11 | Marex | (just download openmoko qemu from svn and apply this patch) |
21:19.30 | Marex | then "obtain" some bootloader (eg for magician) |
21:19.45 | Marex | I used beetles_bootloader.bin ... that's the bootloader image |
21:20.12 | Marex | and run qemu-system-arm -M ipaq -kernel somefakekernelfile -m 192m |
21:21.41 | cr2 | Marex: what about wince ? |
21:22.16 | Marex | havent got so far yet |
21:22.53 | cr2 | the host machine is arm ? |
21:23.00 | Marex | moreover I was unable to add GPIO tracking there as well |
21:23.12 | Marex | I had that in hackndev-qemu which was based on some older checkout |
21:23.37 | Marex | cr2, my hostmachine is x86-64 ;D |
21:23.43 | Marex | and Im qemulating arm/pxa27x |
21:24.00 | Marex | configure that with --target-list=arm-softmmu |
21:24.03 | cr2 | ok. |
21:24.04 | dcordes | BabelO, what does reflash.ctl do? |
21:24.18 | Marex | cr2, please check PM |
21:24.30 | cr2 | Marex: what's missing for wince ? |
21:24.52 | cr2 | Marex: i've used the update from hp.com |
21:24.52 | Marex | no idea ... half of hardware probably ? :D |
21:25.01 | Marex | cr2, to get whole rom ? |
21:25.08 | Marex | with bootloader etc ? |
21:25.10 | cr2 | yes. |
21:25.33 | cr2 | it's different from beetles |
21:25.46 | Marex | I see |
21:25.48 | cr2 | because the spl is copied over to ram. |
21:26.11 | cr2 | and stays there. |
21:26.17 | cr2 | so the layout of different. |
21:26.36 | cr2 | s/of/or/ |
21:26.53 | BabelO | dcordes: not good at the moment it is like on hx4700 to reflash, but not for ba |
21:26.56 | cr2 | Marex: but there is a wince image for the m$ emu ? |
21:27.16 | cr2 | dcordes: we need to get some more data about the vpp. |
21:27.16 | dcordes | BabelO, replace wince in flash with linux? |
21:27.25 | Marex | cr2, hmm .. no idea, I have no m$ emu |
21:27.38 | BabelO | dcordes: yes, but not for ba it is for ipaq |
21:27.42 | Marex | I think m$ emu will be the same as palm's emu - that is wm compiled for x86 |
21:27.48 | cr2 | Marex: i#ve checked it a year ago or so. |
21:27.53 | Marex | (in palm's case palmos compiled for x86) |
21:28.07 | Marex | cr2, can you point me to something ? |
21:28.11 | cr2 | Marex: no. you are loading some wnce rom into the emu |
21:28.20 | Marex | I see |
21:28.39 | cr2 | i knew the .msi name and used google. |
21:28.59 | cr2 | to avoid m$ pa$$port or how it is called |
21:29.07 | cr2 | but know i've forgot the .msi name |
21:29.27 | Marex | :D |
21:31.06 | dcordes | loool openmoko psplash |
21:31.10 | dcordes | looks like vga |
21:31.47 | dcordes | hope rest is for qvga |
21:31.57 | cr2 | Marex: pm does not work for me. |
21:32.04 | cr2 | dcordes: moko is vga |
21:32.16 | Marex | cr2, you have to register with freenode |
21:32.29 | Marex | cr2, do you use jabber ? |
21:32.30 | dcordes | cr2, neo1973 is but not whole openmoko. there are themes and stuff for qvga ;) |
21:32.35 | *** join/#htc-linux noelcragg (n=noel@cn-sfo1-natout.cnet.com) |
21:32.57 | cr2 | Marex: no jabber here |
21:33.15 | dcordes | asic3 demux interrupt processing ovverrun |
21:33.23 | noelcragg | thanks cr2 and Marex for the help yesterday; got kernel booting on my HTC Sable |
21:33.48 | cr2 | noelcragg: any bugs to report ? |
21:33.55 | cr2 | noelcragg: does it resume ? |
21:33.58 | Marex | noelcragg, then hang on |
21:34.15 | noelcragg | cr2: not yet; still assembling ramdisk of useful stuff so I can do some testing |
21:34.29 | dcordes | fsck looks like lockup |
21:34.30 | cr2 | sd image ? |
21:34.42 | noelcragg | cr2: should have some info later tonight (yes using SD) |
21:34.51 | cr2 | ok. |
21:35.11 | cr2 | noelcragg: will be a pita with the 240x240 screen. |
21:35.21 | cr2 | you need to adjust all themes. |
21:35.42 | noelcragg | cr2: that's ok, used to assembling this kind of stuff from my Zaurus days |
21:36.24 | cr2 | noelcragg: the gps is the same as on moko1, but goes over spi. the protocol is not completely cracked afair. |
21:36.46 | dcordes | BabelO, my moko image with self-made kernel boots, but hangs before X |
21:37.05 | cr2 | noelcragg: there is a test program for the phone, but i don't know its status. |
21:37.58 | BabelO | dcordes: you have to edit your /etc/network/interface and add auto usb0 then fill ip adress / mask and add /etc/modutils/g_ether file ... then you can connect to ba over usb and look more |
21:39.15 | dcordes | BabelO, I see the kernel loading, then I see how rootfs is mounted, nifty openmoko psplash then it hangs close to the end |
21:39.20 | dcordes | I suspect it is just X |
21:39.27 | dcordes | BabelO, have no usb :( |
21:39.41 | dcordes | auto wlan0 would also do I guess |
21:39.44 | BabelO | dcordes: a BA without usb ? |
21:40.01 | dcordes | BabelO, I only see strange flat port for charger. |
21:40.19 | BabelO | dcordes: no cradle ? |
21:40.23 | dcordes | no |
21:40.29 | dcordes | got shitty charger with it |
21:41.23 | BabelO | dcordes: should be hard to do something without usb or serial cable |
21:41.39 | dcordes | udev.. ip bla network blabla starting to configure packages, nothing to be done, finished configuring packages. now I see cursor. |
21:42.00 | dcordes | keyboard backlight reacts but I can't type. no console. |
21:42.47 | dcordes | BabelO, is everything right with the kernel if I get this far? |
21:42.50 | dcordes | w/o panic? |
21:43.02 | dcordes | maybe I did modules install wrong |
21:43.16 | BabelO | dcordes: yes everything is ok, |
21:43.37 | dcordes | I will try networking. |
21:43.38 | BabelO | dcordes: no x can mean a segfault somewhere |
21:43.49 | Marex | cr2, n560 bootloader http://pastebin.ca/946713 |
21:43.57 | dcordes | BabelO, maybe because of qvga vs vga? |
21:44.14 | BabelO | dcordes: no, don't think |
21:45.00 | Marex | cr2, beetles loader http://pastebin.ca/946714 |
21:45.37 | Marex | cr2, btw beetles loader is really piece of nice code, check what's pxa lcd controller's 0x204 0x208 0x20c (r/o registers) ;D |
21:48.28 | dcordes | BabelO, I try with auto wlan0 and set post-up udhcpc -i wlan0 do make sure I get ip |
21:50.08 | dcordes | is there any way to turn off wm3 configuration wizard after reset? It is annoying me a lot. |
21:52.13 | BabelO | dcordes: lol, i install wm5 just for that, and pH5 do its htc-tools just for that too ;) |
21:52.40 | BabelO | maybe not just, but a little ;) |
21:52.44 | dcordes | BabelO, it is definetly one annoying thing ba is online |
21:53.08 | BabelO | dcordes: you have open access point ? |
21:53.18 | dcordes | no |
21:53.24 | dcordes | but wep only |
21:53.28 | dcordes | so yes :D |
21:53.52 | BabelO | dcordes: iwconfig wlan0 essid <ssid name> key abcdef0123 |
21:53.56 | BabelO | this is my key ;) |
21:54.01 | dcordes | BabelO, I am already online ;) |
21:54.11 | dcordes | I have clear text |
21:54.11 | Marex | cr2, sable hates me :C |
21:54.31 | dcordes | roflmao /etc/init.d/xserver-nodm start and it's done :D |
21:54.52 | BabelO | dcordes: ok so you sssh on it ? |
21:55.03 | dcordes | yep |
21:55.28 | BabelO | if you start x by hand , is it ok ? |
21:55.35 | dcordes | hahaha |
21:55.47 | dcordes | yes :D it looks shity though |
21:56.26 | dcordes | dialer error: no network |
21:57.27 | dcordes | BabelO, what module is needed for telephone? |
21:57.28 | BabelO | dcordes: hmm, you have to try the wiki about gsmd |
21:57.34 | dcordes | nope. |
21:57.37 | BabelO | dcordes: ps axf | grep gsmd |
21:57.40 | dcordes | hh wiki? |
21:57.49 | dcordes | gsm is new ground for me |
21:58.24 | dcordes | BabelO, http://paste.ubuntu-nl.org/59972/ |
21:58.25 | BabelO | cr2: it is strange that psokolovsky did not connect here since your ba commits ;) |
21:58.47 | dcordes | BabelO, he alredy fixed the wrong line |
21:59.08 | BabelO | dcordes: no for that, , i speak for ds1wm commit |
21:59.13 | *** join/#htc-linux skodde (n=skodde@unaffiliated/skodde) |
21:59.40 | BabelO | pH5: wht is the magic gsmd command line to speak with ligsmd -cmd atcmd ? |
21:59.56 | *** join/#htc-linux Raven_coda (n=Raven_co@64.50.35.182) |
22:01.07 | BabelO | dcordes: there is a command with libgsmtools - <something> atcmd |
22:01.14 | BabelO | i don't remember well |
22:01.20 | BabelO | it is in openmoko wiki |
22:01.25 | dcordes | I'll check it |
22:04.54 | cr2 | Marex: 114 and 116. i need to understand who controls them. |
22:05.46 | pH5 | BabelO: --help might help ;) I think it is --mode atcmd |
22:06.12 | cr2 | BabelO: lol. he is probably very busy :) |
22:10.22 | BabelO | cr2: it seem yes, past year, he already been here :) |
22:11.23 | BabelO | pH5: i have a new Howto :) |
22:11.33 | BabelO | pH5: how to keep women busy, lol |
22:11.50 | BabelO | pH5: scummvm and n800 with indiana jones atlantis :) |
22:13.00 | BabelO | think : " now just need something for the cat, maybe the Sony electronics dog" :) |
22:20.24 | dcordes | I'm getting loads of asic3_irq_demux system messages |
22:21.20 | BabelO | dcordes: yes ;) me tto depend of the boot |
22:21.31 | BabelO | it is random electronics :) |
22:21.41 | dcordes | not important? |
22:21.56 | BabelO | dcordes: no, i remove them |
22:22.11 | BabelO | next step is to understand why |
22:22.21 | dcordes | all 4 seconds there are like 10 messages |
22:25.37 | dcordes | pH5, somehow OE gave me an openmoko-image with VGA themes and icons. is there a way to correct this before building? |
22:26.08 | dcordes | I installed qvga theme and icons but somehow fscked whole gui now. |
22:27.30 | dcordes | BabelO, is it needed to set any variables for make modules_install? |
22:29.00 | BabelO | dcordes: yes if you build it by hand, else you install all modules on your host desktop tree |
22:30.08 | dcordes | BabelO, I'm aware I need INSTALL_MOD_PATH but do I need crosscompile and arch for this? |
22:30.15 | pH5 | BabelO: :) |
22:30.19 | pH5 | dcordes: http://en.pastebin.ca/946747 |
22:30.47 | dcordes | pH5, you just commited this :) ? |
22:31.17 | pH5 | not yet :) |
22:31.37 | BabelO | dcordes: i just export arm-* into my PATH and INSTALL_MOD_PATH too |
22:32.17 | dcordes | pH5, +THEMES_htc..... I guess that's the important line right? what about the settings? I mean today needs to know which theme to put |
22:32.28 | dcordes | BabelO, ok I just make both |
22:33.49 | BabelO | pH5: picture resize is already done in 2008.1 ? |
22:35.47 | pH5 | BabelO: for the theme. I didn't work on it anymore though. last time I looked at it, some apps had their own vga-sized icons... |
22:36.24 | dcordes | pH5, as for the icons it looked ok. at least in openmoko-today2 |
22:36.42 | dcordes | the vga icons were replaced with qvga ones. |
22:36.47 | dcordes | but not the themes. |
22:36.54 | dcordes | s/themes/theme |
22:47.28 | BabelO | good night |
22:59.27 | *** join/#htc-linux tuff (n=phil@tuffnatty.dialup.corbina.ru) |
23:10.12 | dcordes | pH5, does it work? |
23:27.34 | *** join/#htc-linux tsdogs (n=tsdogs@62.123.180.130) |
23:38.09 | *** join/#htc-linux _mxer_ (n=sativa@ip68-230-62-137.ph.ph.cox.net) |
23:59.29 | *** join/#htc-linux noelcragg_ (n=noel@adsl-66-127-210-245.dsl.snfc21.pacbell.net) |