00:09.42 | GSV | <PROTECTED> |
00:45.16 | *** join/#htc-linux SmallR2002 (n=SmallR20@88-107-101-180.dynamic.dsl.as9105.com) |
02:23.58 | *** join/#htc-linux the_sys0p` (n=the_sys0@cpe-66-75-196-173.bak.res.rr.com) |
02:25.42 | *** join/#htc-linux SmallR2002 (n=SmallR20@88-107-101-180.dynamic.dsl.as9105.com) |
02:25.53 | *** join/#htc-linux psokolovsky_ (n=psokolov@82.193.98.14) |
05:09.56 | *** join/#htc-linux |dion| (n=dion@inhex.net) |
05:11.45 | *** join/#htc-linux GSV (i=GSV@c-71-202-155-188.hsd1.ca.comcast.net) |
06:19.00 | *** join/#htc-linux goxboxlive (n=goxboxli@176.84-48-210.nextgentel.com) |
07:11.24 | *** join/#htc-linux dcsmith_ (n=dcsmith@ip68-1-166-115.ri.ri.cox.net) |
07:55.45 | goxboxlive | anyone awake |
08:01.23 | *** join/#htc-linux RoEn (n=roen@p54A655ED.dip.t-dialin.net) |
08:02.02 | *** join/#htc-linux _law_ (n=_law_@213.173.86.202) |
08:03.37 | goxboxlive | I have this on magician: cat /sys/class/power_supply/ds2760-battery.0/ but in the opie2 source the batterie is lsted as: cat /sys/class/battery |
08:04.04 | goxboxlive | Is there a way i can make something like if blabla, then blabla else ??? |
08:19.21 | *** join/#htc-linux SmallR2002 (n=SmallR20@88-107-101-180.dynamic.dsl.as9105.com) |
08:40.06 | *** join/#htc-linux \bash (n=sps@8.pool85-53-17.dynamic.orange.es) |
08:52.02 | *** join/#htc-linux pH5 (n=ph5@p5485D48E.dip.t-dialin.net) |
09:01.34 | *** join/#htc-linux pleemans (n=peter@d51A5E76A.access.telenet.be) |
09:03.11 | *** join/#htc-linux rob_w (n=bob@X17e6.x.pppool.de) |
09:04.03 | *** join/#htc-linux emiliano75 (n=emiliano@adsl-ull-116-216.42-151.net24.it) |
09:04.21 | emiliano75 | good morning |
09:04.44 | emiliano75 | i have problem.... |
09:04.59 | emiliano75 | <PROTECTED> |
09:05.18 | emiliano75 | during the loaded give me an error.... |
09:05.36 | emiliano75 | setty: ioctl() tiocspgrp call faile: inappropriate ioctlfor device... |
09:05.58 | emiliano75 | and then write familiar login |
09:06.10 | emiliano75 | What's happening? |
10:07.34 | *** join/#htc-linux tsdogs (n=twostupi@84.18.144.39) |
10:07.53 | tsdogs | emiliano75: what model are you using? |
10:08.37 | tsdogs | goxboxlive: hi, I think the best option right now it to modify source and have an evironment QWS_BATTERY pointing to the battery class |
10:11.12 | pH5 | arrgh, I don't get it to resume when booted from the htc bootloader, it resets instead. the only explanation I have is that something writes to those 52 words at 0xa0038000 so the checksum in PSPR isn't correct anymore. but what? I don't see any other branching going on in the resume path of htc bootloader / wince kernel. |
10:11.16 | pH5 | http://www.mail-archive.com/linux-arm@lists.arm.linux.org.uk/msg02648.html |
10:11.38 | pH5 | emiliano75: maybe there's a wrong entry in /etc/inittab and tinylogin tries to open a console on the wrong serial port? |
10:12.35 | emiliano75 | oh, thanks now it's ok..... |
10:12.46 | emiliano75 | i find the solution :-) |
10:13.41 | emiliano75 | i have a qtek 9090... now i'm finding how can use camera, wifi and other hardware |
10:15.41 | pH5 | qtek 9090 is htc blueangel, right? |
10:15.54 | tsdogs | emiliano75: I suggest you to use cvs kernel, they fixed many things in there. |
10:16.08 | tsdogs | pH5: yes |
10:16.16 | pH5 | tsdogs: yeah, I agree |
10:16.24 | emiliano75 | yes... blueangel |
10:16.37 | tsdogs | emiliano75: why you using familiar distro for BA? |
10:17.33 | emiliano75 | i discover linu two years ago and i am very sotisfed by the performance |
10:17.47 | emiliano75 | and i'd like use linux on my ba too |
10:19.18 | tsdogs | emiliano75: that's not what I meant. I think the Angstrom is better distro for BA. pH5 afayk there are images for angstrom? |
10:20.06 | emiliano75 | i see some angstrom imag |
10:20.34 | emiliano75 | i find in internet a tutorial how to install gpe |
10:20.39 | emiliano75 | and i try it |
10:21.18 | tsdogs | ok. |
10:22.50 | tsdogs | emiliano75: where did you get the image? If you are interested in testing/helping I think you should ask BabelO to give U updated images for BA. |
10:24.34 | emiliano75 | i take the GPE image handehelds.org.... |
10:25.07 | emiliano75 | ok, thanks tsdogs |
10:26.40 | tsdogs | it's from 2005. :) emiliano75 this one looks more recent http://gnulinux.biz/files/blueangel/people/rob_w/sd/ but I have no idea if it works |
10:27.10 | emiliano75 | ok i check immediatly at this link.... |
10:29.24 | emiliano75 | i'm downloading the image of gpe.... |
10:29.57 | tsdogs | :) I believe BabelO has an opie-image. |
10:30.33 | emiliano75 | as soon as available i change the image in my SD |
10:30.53 | tsdogs | ok |
10:31.14 | tsdogs | bbl |
10:31.15 | emiliano75 | thanks for your help :-) |
11:21.14 | *** join/#htc-linux Ranpha1 (n=Ranpha@hlm-n-73ed.adsl.wanadoo.nl) |
11:43.27 | *** join/#htc-linux pleemans (n=peter@d51A5E76A.access.telenet.be) |
11:46.52 | *** join/#htc-linux pwr (n=pwr@86.121.234.3) |
12:23.34 | *** join/#htc-linux BabelO (n=Fabrice@lun34-2-82-238-28-28.fbx.proxad.net) |
12:30.04 | pH5 | hi BabelO |
12:30.48 | BabelO | hi pH5 :) |
12:33.11 | BabelO | pH5: gpephone fail at compile time with OE, and same for GPE :( |
12:33.44 | BabelO | i haven't tried openmoko yet, because of screen |
12:34.16 | pH5 | BabelO: I know gpephone doesn't work until you use gtk 2.6 etc., but I'm building GPE (angstrom-x11-image) for magician right now - where did it fail for you? |
12:34.30 | pH5 | BabelO: are you going to build some images that we can put up at linuxtogo.org/~htcpxa? |
12:35.09 | BabelO | pH5: yes, i ve already opie image, but i think for gpe, i ve to take universal image and update it |
12:35.25 | BabelO | OE fail for me with bitbake gpe-image |
12:37.42 | pH5 | BabelO: I'm not sure if gpe-image is outdated maybe - try angstrom-x11-image or ask about gpe-image on #angstrom |
12:38.12 | BabelO | pH5: ok i have to try , thx |
12:39.00 | BabelO | maybe i ve to try special w100 X server |
12:39.51 | pH5 | I'm going to try and build an openmoko-devel-image later, but I expect that there's still a lot of work to do before it is presentable. |
12:41.06 | BabelO | ok |
12:41.53 | BabelO | i m working on opie2 on my side |
12:43.00 | pH5 | BabelO: very good! I'm hoping to get magician opie2 support for free now that goxboxlive has one, too ;) |
12:44.05 | pH5 | BabelO: btw, you can select the x server with XSERVER = "xserver-kdrive-w100" in conf/machine/htcblueangel.conf |
12:47.35 | BabelO | pH5: ok, last time i try this X server, it does not work, ( lookinjg for the correct word :) ) |
12:49.01 | BabelO | got lot of bar on the screen, like if modeline are bad |
12:52.44 | pH5 | I don't know much about w100/w3200/w3220, the ba chip is probably a bit different than w100 on zaurus or w3220 on hx4700 - there is also a newer xserver-kdrive-imageon by sirfred, but I think it doesn't even work on h4700 properly. |
12:53.22 | BabelO | ok, i ve to try all this one ;) |
12:56.37 | BabelO | nice librarie libsexy-0.1.11-r0 .... LOL |
12:57.25 | pH5 | :) |
13:10.35 | *** join/#htc-linux goxboxlive (n=goxboxli@176.84-48-210.nextgentel.com) |
13:15.16 | *** join/#htc-linux |dion| (n=dion@inhex.net) |
13:15.58 | *** join/#htc-linux SmallR2002 (n=SmallR20@88-107-72-104.dynamic.dsl.as9105.com) |
14:01.10 | *** join/#htc-linux scorpio16v (n=scorpio1@ACB3C069.ipt.aol.com) |
14:14.07 | *** join/#htc-linux pancake (n=xxx@55.pool85-52-227.static.orange.es) |
14:14.18 | pancake | hi cr2_ are you there? |
15:28.12 | cr2_ | pancake: hi |
15:36.18 | goxboxlive | hi cr2 |
15:44.58 | cr2_ | hi goxboxlive |
15:46.37 | goxboxlive | cr2_, magician is a fine phone, just missing wifi |
15:48.44 | cr2_ | you can use sdio (in theory). there is a pxa sdio host driver. |
15:48.55 | cr2_ | there was a sdio wifi acx100 card. |
15:49.45 | cr2_ | http://products.wi-fiplanet.com/wifi/sdio/1133890526.html |
15:50.40 | cr2_ | with the binary drivers for zaurus. |
15:54.03 | goxboxlive | cr2_, wow so it might work then |
15:54.04 | goxboxlive | cool |
16:02.25 | *** join/#htc-linux rob_w (n=bob@X17e6.x.pppool.de) |
16:02.43 | cr2_ | goxboxlive: you need to find one first. |
16:05.54 | goxboxlive | http://mmc.drzeus.cx/wiki/SDIO |
16:09.23 | *** join/#htc-linux rob__w (n=bob@X1248.x.pppool.de) |
16:29.36 | *** join/#htc-linux pleemans (n=peter@d51A5E76A.access.telenet.be) |
17:00.26 | *** join/#htc-linux |dion| (n=dion@inhex.net) |
17:24.09 | *** join/#htc-linux RoEn (n=roen@p54A655ED.dip.t-dialin.net) |
17:34.32 | *** join/#htc-linux tsdogs (n=twostupi@62.123.180.130) |
17:41.50 | cr2_ | hi tsdogs |
17:41.59 | tsdogs | hi cr2_ |
17:42.21 | tsdogs | cr2_: I think I fixed the label thing |
17:42.42 | tsdogs | can u try? |
17:42.55 | cr2_ | commit the code to CVS |
17:43.07 | tsdogs | already done |
17:43.10 | tsdogs | :) |
17:43.31 | cr2_ | i've learned a bunch of new things today. |
17:43.32 | cr2_ | ok. |
17:43.54 | tsdogs | :) like? |
17:45.45 | cr2_ | tell you a bit later. |
17:45.59 | *** join/#htc-linux pH5 (n=ph5@p5485D48E.dip.t-dialin.net) |
17:46.10 | tsdogs | ok |
17:46.56 | BabelO | pH5: angstrom x11 compiled... |
17:47.10 | BabelO | image is really lite |
17:49.58 | cr2_ | tsdogs: all the labels are not rotated on amd64. |
17:50.11 | tsdogs | :( on mine they are |
17:50.58 | tsdogs | cr2_: you using 2d right? |
17:51.11 | cr2_ | can't check qt4 on alpha |
17:51.16 | cr2_ | yes. |
17:51.35 | pH5 | BabelO: great, are you missing things? I'm just about to put my image on an SD. |
17:51.47 | cr2_ | tsdogs: does qt3 work ? |
17:52.04 | tsdogs | cr2_: nope. need to implement all the new functions. |
17:52.08 | BabelO | pH5: there is nothing on this image... :( |
17:52.41 | BabelO | pH5: only settings folder , other are empty |
17:52.59 | cr2_ | tsdogs: ok. |
17:53.10 | pH5 | I see, it's a long time since I started from a clean image :) |
17:53.29 | cr2_ | tsdogs: not a big priority, actually. i#d better implement new cool ideas on qt4. |
17:54.00 | tsdogs | cr2_: yes, porting till now should simply be copying functions |
17:54.34 | cr2_ | tsdogs: i've hit some minor math problems. |
17:54.37 | *** join/#htc-linux GSV (i=GSV@c-71-202-155-188.hsd1.ca.comcast.net) |
17:54.56 | cr2_ | we should not use epsg:54004 probably. |
17:55.04 | cr2_ | i'll tell you why. |
17:55.18 | tsdogs | cr2_: well if you are able to fix it do it, you are the math master :) |
17:55.19 | tsdogs | ok |
17:55.24 | BabelO | pH5: i want to build minimo for it.. but it failed |
17:55.51 | cr2_ | this is "standard" epsg:54004 using WGS84 ellipsoid |
17:55.54 | cr2_ | +proj=merc +latts=0 +lon0=0 +k=1.0 +x0=0 +y0=0 +ellps=WGS84 +datum=WGS84 +units=m no_defs |
17:56.24 | cr2_ | google and others use a sphere and not the ellipsoid to project the data: |
17:56.33 | cr2_ | +proj=merc +latts=0 +lon0=0 +k=1.0 +x0=0 +y0=0 +a=6378137.0 +b=6378137.0 +units=m |
17:56.45 | cr2_ | the math is then easier. |
17:57.24 | cr2_ | to be tile-to-tile compatible, we should probably use this projection too. |
17:58.02 | tsdogs | ok. |
17:58.11 | cr2_ | we can also store the projected mercator coordinates in the *.rdm too. |
17:58.48 | cr2_ | it's not a big deal for the drawing routines, they use linear scaling anyway. |
17:58.51 | tsdogs | like geotiff does |
17:59.04 | GSV | hey guys |
17:59.04 | tsdogs | ok. gg bbl |
17:59.13 | GSV | any good news? |
17:59.21 | cr2_ | the only trick is to find and remove the multiplication by cos(lat) |
17:59.55 | cr2_ | and if you'd like to get the "real" lat/long you need to convert back from spherical mercator. |
18:00.03 | pH5 | BabelO: hm, I never tried to build minimo. doesn't gpe-minibrowser get built? |
18:02.28 | BabelO | pH5: i try to build svn version, because last time i try it, it always segfault after reading a page |
18:03.04 | *** part/#htc-linux scorpio16v (n=scorpio1@ACB3C069.ipt.aol.com) |
18:04.05 | tsdogs | cr2_: ok, seems easy. |
18:08.01 | cr2_ | tsdogs: i need to recompile postgis with proj4 support. |
18:08.58 | cr2_ | then it will be possible to request the DB data reprojected to mercator directly and get rid also of the 10^6 coordinate scaling. |
18:09.46 | cr2_ | the 54004 is easy to do, but i need to read the docs how to use a custom proj4 definitions. |
18:10.22 | cr2_ | tsdogs: this should be changed then: |
18:10.25 | cr2_ | <PROTECTED> |
18:10.25 | cr2_ | <PROTECTED> |
18:10.25 | cr2_ | <PROTECTED> |
18:10.25 | cr2_ | <PROTECTED> |
18:10.25 | cr2_ | <PROTECTED> |
18:11.03 | cr2_ | <PROTECTED> |
18:11.03 | cr2_ | <PROTECTED> |
18:11.17 | cr2_ | <PROTECTED> |
18:11.17 | cr2_ | <PROTECTED> |
18:11.17 | cr2_ | <PROTECTED> |
18:11.31 | cr2_ | is not needed anymore. |
18:12.05 | cr2_ | because the vertical and horizontal scale are equal. |
18:12.12 | tsdogs | ok |
18:13.06 | cr2_ | static int buildmap_line_calc_length |
18:13.24 | cr2_ | this will be replaced by simple sqrt(dx*dx+dy*dy) |
18:13.59 | cr2_ | and not the spherical geometry monster as it is now :) |
18:14.11 | tsdogs | :) |
18:14.53 | cr2_ | grep 1000000 * |
18:15.04 | cr2_ | shows many places that need to be fixed. |
18:15.45 | tsdogs | ok |
18:16.00 | *** join/#htc-linux goxboxlive (n=goxboxli@176.84-48-210.nextgentel.com) |
18:16.15 | cr2_ | don't know what will happen near -180/+180 line. but i doubt the current code is better in this respect. |
18:16.50 | cr2_ | hmm. spherical trig should be capable to deal with it. |
18:18.18 | tsdogs | ok bbl. |
18:20.05 | *** join/#htc-linux kalski (n=Miranda@217.75.132.159) |
18:20.21 | cr2_ | INSERT INTO "spatial_ref_sys" ("srid","auth_name","auth_srid","srtext","proj4text") |
18:20.31 | cr2_ | does not seem to be very difficult. |
18:26.25 | *** part/#htc-linux kalski (n=Miranda@217.75.132.159) |
18:30.37 | regulate | hermes? |
18:31.50 | cr2_ | AddToPROJ4SRSCache: Cannot find SRID (54004) in spatial_ref_sys |
18:31.54 | cr2_ | that's good. |
18:44.36 | cr2_ | http://docs.codehaus.org/display/GEOSDOC/Custom+projection+definition+in+PostGIS |
18:48.59 | *** join/#htc-linux rototo (n=lilliput@82-47-146-102.cable.ubr11.brad.blueyonder.co.uk) |
19:33.32 | *** join/#htc-linux SmallR2002 (n=SmallR20@88-107-72-104.dynamic.dsl.as9105.com) |
19:44.01 | *** join/#htc-linux Ranpha1 (n=Ranpha@hlm-n-73ed.adsl.wanadoo.nl) |
20:02.41 | cr2_ | tsdogs: i'm trying to understand the coordinate transform inside roadmap. |
20:02.55 | tsdogs | cr2_: good luck |
20:02.59 | tsdogs | :) |
20:03.28 | tsdogs | I'm trying to draw a scale for zoom. but it does not seem to take the commands :( |
20:04.34 | cr2_ | to draw a scale you need to understand the math :) |
20:04.52 | tsdogs | not really, I only use pixels for the zoom scale :) |
20:05.00 | Kevin2 | cr2_ / pH5: FYI, I sent a generic cpld driver to kernel-discuss. |
20:05.44 | cr2_ | it seems to me that the .zoom_x/y are the pixel sizes in "microdegrees" |
20:06.03 | cr2_ | using mercator will actually simplify the code. |
20:06.28 | cr2_ | Kevin2: that's great! didn't get the e-mail though. |
20:06.58 | cr2_ | tsdogs: somehow i prefer meters to microdegrees :) |
20:07.08 | Kevin2 | cr2_: I just sent it. It will be along shortly I'm sure. |
20:07.10 | tsdogs | :) |
20:07.22 | tsdogs | Kevin2: I just got it :) |
20:08.01 | cr2_ | tsdogs: and the zoom_x will be equal to zoom_y with mercator. |
20:08.05 | pH5 | Kevin2: great indeed! |
20:11.53 | cr2_ | Kevin2: i vote for applying the patch right now, we are the only users anyway. the only thing i'd like to change are the c++ comments. |
20:13.13 | cr2_ | tsdogs: a spherical mercator to latlong function is needed for your purposes. |
20:13.56 | Kevin2 | What's wrong with c++ comments? They are standard in c99 and gcc has always supported them. |
20:16.47 | Kevin2 | I'll be back later. |
20:16.55 | pH5 | Kevin2: magician uses 32bit reads/writes, not 16bit |
20:19.42 | cr2_ | Kevin2: ok, i had a wrong impression that // is not widely used in the kernel. doing grep proved that i'm wrong. |
20:20.33 | cr2_ | tsdogs: the zoom_x can be easily rescaled back from microdegrees to meters. zoom_y is more complex. |
20:20.49 | bd2 | Kevin2, hey. patch looks good. I don't have any htc device, thus my comments are only about cosmetic: it would be great to correspond to linux coding style (including c++ comments). |
20:21.22 | tsdogs | cr2_: first I need to be able to draw something on the canvas. dunno why but it is not drawing my polygons :/ |
20:21.35 | bd2 | cr2_, c++ comments forbidden by CodingStyle, though yes, they're anyway widely used. |
20:21.36 | cr2_ | tsdogs: polygons ? |
20:22.01 | tsdogs | didn't you want a scale on the canvas? |
20:22.06 | tsdogs | for zoom |
20:22.39 | cr2_ | bd2: you are right. then we should not make things worse. |
20:22.42 | cr2_ | Linux style for comments is the C89 "/* ... */" style. |
20:22.43 | cr2_ | Don't use C99-style "// ..." comments. |
20:23.09 | cr2_ | linux/kernel26/Documentation/CodingStyle |
20:23.27 | bd2 | yup. so if HTC guys also dreaming about mainline inclusion, it would be great to be close to it from the start ;-) |
20:24.08 | cr2_ | bd2: is there any sed/whatever tool to autoconvert the files ? |
20:24.21 | cr2_ | i think i've never used // in my own c code. |
20:24.26 | tsdogs | cr2_: btw I have no idea how garmin does it, I've only seen the plus and minus buttons on screenshots. |
20:24.30 | cr2_ | i like to use gcc -ansi -pedantic |
20:24.46 | tsdogs | :) |
20:24.53 | cr2_ | tsdogs: i'll look for a screenshot for you. |
20:25.01 | tsdogs | cr2_: ok. |
20:25.34 | bd2 | cr2_, for converting comments? probably no. for indent there is `indent' utility, but it's not perfect |
20:26.09 | *** join/#htc-linux goxboxlive (n=goxboxli@176.84-48-210.nextgentel.com) |
20:26.32 | cr2_ | bd2, it should be some bison-style parser |
20:26.59 | cr2_ | i think something must be available. this problem is quite formal. |
20:27.47 | pH5 | I'm not sure I like the huge htc_egpio_pinInfo structure. do we really need 16 bytes per pin? |
20:28.04 | bd2 | google found the reverse tool http://people.scs.fsu.edu/~burkardt/cpp_src/recomment/recomment.html ;-) |
20:28.12 | cr2_ | doing // text -> /* text */ for each line is ugly. |
20:28.15 | cr2_ | :) |
20:29.39 | Kevin2 | I'm back. |
20:30.03 | cr2_ | Kevin2: // is against the linux/kernel26/Documentation/CodingStyle |
20:30.08 | Kevin2 | cr2_ / bd2 : Yes, I see CodingStyle says that. I that is a stupid restriction, but so be it - I'll convert them. |
20:30.18 | Kevin2 | s/I that/I think that/ |
20:30.23 | cr2_ | ok. |
20:30.40 | cr2_ | but then i#d like to convert all "my" files too. |
20:31.16 | Kevin2 | BTW, there are 26185 c++ style comments in hh.org linux tree. :-) |
20:31.32 | bd2 | Kevin2, I also think that _any_ restrictions is stupid. But on the other hand, code unification is wise, and much better than the mix of many "own" coding styles. ;-) |
20:33.56 | Kevin2 | pH5: Is magician's use of 32 reads cosmetic, or does it actually need them? |
20:37.45 | cr2_ | bd2: http://vim.sourceforge.net/tips/tip.php?tip_id=1497 |
20:38.56 | cr2_ | hmm. it's very simple too. |
20:39.56 | pH5 | Kevin2: not really, the registers are 8bit on magician |
20:40.20 | *** join/#htc-linux goxboxlive (n=goxboxli@176.84-48-210.nextgentel.com) |
20:41.11 | Kevin2 | pH5: I'm confused - is it 32 bits or 8 bits? |
20:41.57 | Kevin2 | pH5: The pin info structure is only for those pins that need more info - input pins that you need to line up with irqs, all irq pins, and output pins that have a non-zero start value. |
20:41.58 | pH5 | the drivers use 32bit reads/writes, but only the 8 lsb are used, so it's 32bit-addressed 8bit registers. |
20:42.15 | pH5 | Kevin2: ah, cool. |
20:42.51 | cr2_ | pH5: most of htc CPLDs i've seen use ldrh/strh to access the registers. |
20:43.30 | cr2_ | tsdogs: http://www.gpsw-support.co.uk/reviews/garmin-gpsmap-76/screen1.jpg |
20:43.39 | cr2_ | tsdogs: "Map page" picture |
20:44.07 | Kevin2 | pH5: I can easily shrink the structure. I can also remove the requirement to list all irq pins. Not sure it is a big deal though. |
20:44.27 | pH5 | cr2_: magician does not, it's the same for the ds1wm's 8bit registers - they are just not packed here. |
20:45.16 | cr2_ | pH5: strange. i need to check the aic2/3 code for ba/hima. |
20:45.24 | *** join/#htc-linux skodde (n=skodde@unaffiliated/skodde) |
20:45.36 | tsdogs | cr2_: bottom left corner ? if you click on it it will zoom change, or it's sufficent to display actual zoom? |
20:46.29 | Kevin2 | pH5: magician_cpld uses __raw_readb / writeb - that causes an an 8-bit access. |
20:48.56 | cr2_ | tsdogs: only display. think about devices without ts/mouse. like garmin or the phones. |
20:49.39 | tsdogs | cr2_: did you change the setPen(Qt:NoPen) in the polygons draw in qt_canvas? |
20:50.31 | cr2_ | tsdogs: i don't touch any qt gui code. |
20:51.00 | tsdogs | ok, then it was wither me or did miss that one :( |
20:51.48 | pH5 | Kevin2: I have some output pins to enable, but no irq/input pins, so I don't think the size is too important - the bigger deal is to have the shared infrastructure at all! :) |
20:52.08 | pH5 | Kevin2: oh, either that was an 'optimization' from the wince driver's ldr + and 0xff, or I'm just remembering incorrectly. Any way, those 8bit registers are spaced 4 bytes apart. |
20:54.06 | cr2_ | pH5: looking at the athena CPLD1 i have an impression it also uses only the lower 8bits. but wince accesses the registers with ldrh/strh |
20:56.45 | Kevin2 | pH5: You have to list all your irq sources with the pin info struct. |
20:57.48 | Kevin2 | pH5: So, does the generic driver need to know about 8/16/32 read/writes, or do you think 16 bit accesses will work on magician? |
20:58.21 | cr2_ | tsdogs: microdegrees to meters -> zoom_x*(2*3.1415926*6378137)/360/1000000 |
20:58.47 | tsdogs | ;) ok |
20:59.44 | cr2_ | hmm. that means divide by 8.98315 |
21:00.24 | cr2_ | need to add some printf() to check if that makes sense. |
21:02.33 | cr2_ | tsdogs: since we are working with integers only, this number should be approximated by some rational fraction. |
21:03.22 | tsdogs | cr2_: which integers? |
21:03.38 | tsdogs | zoom or coords? |
21:03.43 | cr2_ | both. |
21:03.59 | tsdogs | Ok. |
21:04.24 | cr2_ | even the cos is an integer between 0 and 32767 |
21:08.35 | cr2_ | cos(60°)=1/2 ? |
21:09.36 | cr2_ | is oslo i have z_x=32 z_y=16 |
21:11.12 | cr2_ | tsdogs: if i'm right with my microdegrees formula, then z=2 means 2/9 m/pix = 22 cm/pix |
21:11.58 | tsdogs | hmm could be. |
21:13.39 | pH5 | Kevin2: I'll try tomorrow. I believe ldrh/strh should work, but I need a way to use 32bit-aligned registers instead of tightly packed ones. |
21:14.16 | cr2_ | tsdogs: google is ok at 1.2m/pix, it's between z=8 and z=16 |
21:14.30 | tsdogs | ok. |
21:14.32 | Kevin2 | pH5: Okay. I think you should be able to just space your pin descriptions out. |
21:15.19 | cr2_ | tsdogs: the roads are too wide at z=8. |
21:15.48 | cr2_ | but maybe it's my screen, need to think about dpi/size issues :) |
21:15.56 | tsdogs | :) |
21:16.22 | pH5 | Hm. With 4 irqs and estimated 8 or more initially set output gpios I'd have to use >= 192 bytes just for pin configuration.. |
21:17.18 | pH5 | Kevin2: I'd prefer a bus_shift variable over manually spacing the pin descriptions - I have 7 registers, so I'd need 13 'registers' altogether with every second unused when trying to view them as packed 16bit :( |
21:18.25 | cr2_ | pH5: is .bus_shift a big performance hit ? |
21:19.38 | pH5 | cr2_: we are talking about gpios here, so whatever the performance hit is (an additional ldr and an add?), it won't be noticeable. |
21:19.59 | cr2_ | the code look much nicer, but then dividing by two happens on each access. |
21:20.15 | cr2_ | pH5: it's <<1 for each ldr. |
21:20.23 | pH5 | thats a lsl by one bit, you don't even need an extra insn for that. |
21:20.31 | cr2_ | ok. |
21:20.46 | pH5 | yay for arm's barrel shifter. |
21:20.54 | cr2_ | yes. |
21:21.08 | cr2_ | forgot about nice arm arch. |
21:24.39 | pH5 | oh, except that it only works for fixed shifts. with bus_shift configurable, that'd be an extra lsl instruction. |
21:27.05 | cr2_ | ok. |
21:28.00 | goxboxlive | tsdogs, do you know if there is complicated to make a reboot/shutdown/restart aplet in Opie2? On magician we dont have keuboard, and since it isent any console in Opie2 we have to ssh to the device to shut it down. |
21:29.00 | BabelO | goxboxlive: add a definition to apm -s to a button :) |
21:29.06 | cr2_ | goxboxlive: lol. talk about linux stability. |
21:29.28 | cr2_ | can't we trap gpio1 aka soft reset ? |
21:29.43 | BabelO | goxboxlive: there is a /opt/etc/defaultbutton.conf i think |
21:30.07 | goxboxlive | BabelO, ok i'll take a look |
21:30.09 | goxboxlive | thx |
21:30.58 | pH5 | goxboxlive: I've put today's angstrom-x11-image(gpe) and zImage up at linuxtogo.org/~ph5/magician - should we link/copy that to ~htcpxa? |
21:31.06 | cr2_ | tsdogs: at z=32768 it's 3.6km/pix |
21:31.22 | tsdogs | ok |
21:31.30 | pH5 | goxboxlive: for the power off key I'm using http://en.pastebin.ca/570778 as the key thing in gpe (keylaunch) seems to be buggy, too. |
21:31.34 | goxboxlive | pH5, i think we should add it to htcpxa. I can do it. |
21:32.07 | cr2_ | tsdogs: we will really know when we overlay with rater maps :) |
21:32.17 | cr2_ | s7rater/raster/ |
21:32.20 | cr2_ | s/rater/raster/ |
21:32.28 | pH5 | goxboxlive: ok |
21:32.41 | tsdogs | :) |
21:33.04 | Kevin2 | bd2 / cr2_ / pH5 - I've sent a new email to kernel-discuss with latest patch. (Remove c++ comments and a couple of minor changes.) |
21:34.09 | cr2_ | tsdogs: on a 160pix screen 160*3.6km=576km |
21:34.38 | tsdogs | Half italy. |
21:34.54 | cr2_ | on 240 - 864km, on 480 - 1728 km |
21:35.14 | tsdogs | I think it's enough then |
21:35.22 | cr2_ | and in vga landscape - 2304km |
21:35.26 | tsdogs | unless we want to display the whole world |
21:35.40 | cr2_ | but it's the highest zoom. |
21:35.55 | cr2_ | world is 40000km |
21:36.16 | pH5 | Kevin2: ok |
21:36.20 | cr2_ | we may bump up the upper zoom value. |
21:36.41 | tsdogs | cr2 have U tryed ? |
21:36.55 | Kevin2 | I'll be back later. |
21:37.07 | cr2_ | you'll lose performance with the current code. |
21:37.16 | tsdogs | ok |
21:37.22 | cr2_ | on an opteron it's probably not a problem. |
21:37.49 | cr2_ | rdm does not support zoomlevel dependent submaps. |
21:37.54 | cr2_ | now. |
21:38.04 | cr2_ | garmin img does that. |
21:40.11 | cr2_ | tsdogs: i've lost the yellow streetname popup color now. |
21:40.45 | tsdogs | hmm.I have it |
21:41.40 | cr2_ | Info.Background: yellow |
21:41.52 | tsdogs | but there was a bug in polygon drawing. |
21:42.02 | tsdogs | I'll commit changes. |
21:42.05 | cr2_ | ok. |
21:43.34 | tsdogs | commited. |
21:54.10 | cr2_ | ok. |
21:54.53 | tsdogs | cr2_: how do I get the zoom level from the zoom (32768 = 15, etc?) |
21:55.22 | cr2_ | zoom_x=zoom level |
21:56.02 | cr2_ | 15 is more like garmin/google are using, but there are other differences. |
21:56.27 | tsdogs | i'd need this to put the scale in the right position . |
21:58.53 | cr2_ | i'll look how others have done that. |
21:59.35 | cr2_ | wow. the street names are rotated now. |
21:59.50 | tsdogs | I know I could simply divide by n 2 but ain't there a function for this? |
21:59.52 | tsdogs | :) |
22:00.00 | cr2_ | and the background is yellow. |
22:00.23 | cr2_ | >> ? |
22:00.24 | goxboxlive | BabelO, how do i know wich FXX my powerbutton are connected to? |
22:01.13 | tsdogs | goxboxlive: does the kernel say anything complaining about an unmapped key? |
22:01.14 | cr2_ | tsdogs: if you'll change the 'endline' status will it be rendering slower ? |
22:01.23 | BabelO | look in htckbdhandler |
22:01.25 | goxboxlive | tsdogs, no |
22:01.31 | goxboxlive | ok |
22:01.40 | tsdogs | there is a qt guide somewhere |
22:01.48 | tsdogs | cr2_: what you mean |
22:02.01 | goxboxlive | so i have to rebuild, it's not just a editing conf file |
22:02.41 | tsdogs | goxboxlive: if kernel is not complaining then it should be mapped to something. |
22:02.56 | goxboxlive | tsdogs, hmm ok |
22:03.21 | tsdogs | goxboxlive: I have a utility to read the eventX and tell you the keycode if you need it |
22:03.35 | BabelO | goxboxlive: i think it is missing it is GPIO 116 |
22:03.38 | cr2_ | tsdogs: Qt::RoundJoin |
22:03.44 | goxboxlive | kbdtest? |
22:03.46 | BabelO | sorry not gpio keycode |
22:04.02 | tsdogs | cr2_: no idea. |
22:04.13 | tsdogs | goxboxlive: yes |
22:04.57 | cr2_ | tsdogs: right now it's not very nice on low zoom. but it can be the very wide linewdth. |
22:05.35 | tsdogs | cr2_: I saw it. |
22:06.07 | goxboxlive | tsdogs, : Testing keyboard </dev/input/event0> events 16 |
22:06.08 | goxboxlive | read: type=1, code=116, value=1 |
22:06.09 | goxboxlive | read: type=0, code=0, value=0 |
22:06.33 | tsdogs | :) ok it's keycode 116. is it in /etc/keymap ? |
22:06.42 | BabelO | goxboxlive: so you need to add it |
22:07.12 | goxboxlive | hold on i'll check. dont know if i have a keymap |
22:07.52 | goxboxlive | i dont have a keymap. So i should add it to the rootfs image like on universal then |
22:08.12 | tsdogs | goxboxlive: no wait, probably it's standard |
22:08.21 | goxboxlive | ok |
22:08.22 | BabelO | goxboxlive: if it is opie2 you have to add it to plugin |
22:09.48 | tsdogs | goxboxlive: if it's on tty0 then is not needed anymore. |
22:09.50 | goxboxlive | BabelO, so i should take a look at htckbdhandler even though i dont have a keyboard, not htcbuthandler ight |
22:10.03 | tsdogs | new kernel muxes it correctly |
22:10.17 | tsdogs | the htcbut can be taken away |
22:10.34 | goxboxlive | ok so i can be removed. |
22:10.50 | BabelO | goxboxlive: sorry it is a mistake from me |
22:10.57 | BabelO | look into htcbuthandler.cpp |
22:11.02 | goxboxlive | BabelO, could you soon commit htcphone to opie2 tree |
22:11.06 | BabelO | power button is F34 |
22:11.07 | goxboxlive | ok, now i am confused |
22:11.18 | goxboxlive | where did you find that BabelO |
22:11.24 | BabelO | :q |
22:11.26 | goxboxlive | htcbuthandler.cpp |
22:11.49 | BabelO | sorry :) devices/htcphone/src/plugins/qtopiacore/kbddrivers/htcbuttons/htcbuthandler.cpp |
22:11.59 | tsdogs | goxboxlive: try removing the QWS_KEYBOARD, and see if it works, otherwise use keymap to map it. |
22:12.36 | goxboxlive | tsdogs, i use this: export QWS_KEYBOARD="htcbuthandler:/dev/input/event0,GPIO" |
22:12.52 | BabelO | goxboxlive: there is PWR_BUTTON define that is used to grep keycode 116 and simulate a F34 |
22:12.56 | goxboxlive | and only one button do something in opie2 |
22:13.00 | tsdogs | goxboxlive: I think it's not needed. |
22:13.06 | goxboxlive | i'll try |
22:14.52 | tsdogs | found it: http://doc.trolltech.com/qtopia4.2/syscust-keypadarch.html |
22:14.59 | goxboxlive | thx |
22:16.57 | tsdogs | BabelO: htcbuthandler I don't think it's needed anymore, last time I checked the default /dev/tty0 handling worked for me. |
22:17.31 | tsdogs | also this way keyboard on console was not fsk** |
22:17.40 | BabelO | tsdogs: but i ve other thing not mapped to tty0 on BA |
22:18.01 | tsdogs | BabelO: does it have an event? |
22:18.38 | BabelO | i ve 3 event from button and another from keyboard |
22:19.32 | tsdogs | but the kernel is intercepting them right? |
22:21.04 | BabelO | don't know didn t check |
22:21.12 | tsdogs | On universal we should only need some keyboard filter. that's it. I have no idea on BA, but I think it could be handled with default and keymap at least for buttons |
22:21.27 | tsdogs | BabelO: ok. I need to check it again myself. :) |
23:11.02 | cr2_ | tsdogs: i've modified a lot of code, but more changes are needed. |
23:11.50 | tsdogs | :) |
23:11.52 | cr2_ | the most unpleasant change is the conversion of NMEA. |
23:12.01 | tsdogs | hmm |
23:12.07 | goxboxlive | this is strange. When i held the contacts key in for some sec. it raise beam conatct, but when i just touch the contact button nothiung happends. |
23:12.15 | cr2_ | it will cost some CPU cycles. |
23:12.39 | cr2_ | but it's unavoidable if we want the raster overlay. |
23:12.40 | *** join/#htc-linux wireddd (n=wired@97-81-78-105.dhcp.athn.ga.charter.com) |
23:12.56 | tsdogs | ok |
23:13.17 | cr2_ | tsdogs: each NMEA lat/long should be converted |
23:13.29 | tsdogs | goxboxlive: in opie there was a short/long button press, maybe they ported it also |
23:14.11 | goxboxlive | tsdogs, yes it is correct, but contact wil not launch, only beam as it should with long button press. |
23:14.14 | tsdogs | cr2_: that should be only from map, or there is also into the code? |
23:14.25 | goxboxlive | [Button1] |
23:14.26 | goxboxlive | Name[]=Contacts Button |
23:14.26 | goxboxlive | Key=F10 |
23:14.26 | goxboxlive | PressedActionService=Contacts |
23:14.26 | goxboxlive | PressedActionMessage=raise() |
23:14.26 | goxboxlive | HeldActionService=InfraredBeaming |
23:14.28 | goxboxlive | HeldActionMessage=beamPersonalBusinessCard() |
23:14.45 | cr2_ | tsdogs: but the latlong itself is an artifical thing. |
23:14.52 | cr2_ | tsdogs: what do you mean ? |
23:15.02 | tsdogs | goxboxlive: than I agree it's strange. |
23:15.27 | tsdogs | the lat/long should only be in the map. |
23:15.45 | cr2_ | no, there will be no lat/long anymore. |
23:15.49 | tsdogs | then we should convert all to mercator for map drawing, and related |
23:15.53 | BabelO | goxboxlive: does your calendar button start ? |
23:15.59 | cr2_ | but since NMEA is latlong it should be converted. |
23:16.26 | goxboxlive | BabelO, how do i know? It is active, i tested with tsdogs kbdtest |
23:16.28 | tsdogs | cr2_: If I want to find a position, I need lat/long, also GPS is giving me lat/long |
23:16.33 | cr2_ | tsdogs: i've done many of such conversions already. but it's #ifdef'd. |
23:16.51 | cr2_ | tsdogs: yes. |
23:16.54 | BabelO | goxboxlive: in opie2 just push it ;) it works for me |
23:17.06 | tsdogs | ok |
23:17.42 | cr2_ | tsdogs: right now the situation is similar. the long is just scaled. the lat is multiplied by cos(lat). |
23:18.40 | cr2_ | with mercator the lat uses ln(tan) |
23:18.43 | cr2_ | http://en.wikipedia.org/wiki/Mercator_projection |
23:19.05 | cr2_ | which is a bit more involved than just cos from a table. |
23:19.39 | goxboxlive | BabelO, i am pushing it short, and nothing happends. But when i hold it (long push) the beam contacts opens as it is defines in defaultbuttons.conf |
23:21.14 | tsdogs | cr2_: http://84.18.144.44/tsdogs/screenshots/roadmap_zoom.png what do you think? |
23:21.30 | BabelO | goxboxlive: for me short and long press are working with calendar button, but not with contact |
23:21.49 | goxboxlive | BabelO, same as mine |
23:22.26 | tsdogs | BabelO: can you check if there is really a dcop event for contacts called raise ? |
23:22.54 | cr2_ | tsdogs: is it transparent ? |
23:22.56 | goxboxlive | BabelO, i think the contact is wrong, do you know where the overview over the diffrents options i can map in defaultbuttons.conf |
23:23.04 | tsdogs | cr2_: yep |
23:23.23 | cr2_ | ok, looks fun. |
23:23.24 | tsdogs | also the info boxes should be transparent (for you too) |
23:24.04 | cr2_ | btw, do you see these ~[0x1b] things in labels ? |
23:24.33 | BabelO | goxboxlive: i found this http://doc.trolltech.com/qtopia4.2/syscust-keypadbutt.html |
23:24.39 | BabelO | looking for description |
23:24.54 | tsdogs | cr2_: yes |
23:25.19 | cr2_ | tsdogs: actually they have a special meaning, and should not be rendered. |
23:25.43 | cr2_ | i need to lookup the details, but they are rendering hints. |
23:25.45 | tsdogs | cr2_: I know, but either we handle it nor we remove them. |
23:25.52 | tsdogs | :) |
23:27.17 | cr2_ | tsdogs: it's nice for the zoom level, but does not say anyhting about metric properties. |
23:27.32 | cr2_ | what is the window size ? |
23:27.33 | tsdogs | cr2_: one step at the time :) |
23:27.38 | cr2_ | :) |
23:27.45 | tsdogs | the same you see on the screen |
23:28.07 | cr2_ | but not vga/qvga |
23:28.18 | cr2_ | i'm thinking how we can test these. |
23:28.22 | tsdogs | nope. |
23:28.29 | cr2_ | 240x240 will be interesting too. |
23:28.34 | cr2_ | and 176x220 |
23:28.46 | cr2_ | these are 4 standard resolutions. |
23:28.54 | tsdogs | I putted a limit, if it's width is less than 48 then will be fixed |
23:29.24 | cr2_ | nokia770 with its 9:16 screen and other iphones do not counts. |
23:29.37 | tsdogs | but we can do this tests later, it's only pixel drawing calculations. |
23:29.44 | cr2_ | ok. |
23:31.01 | cr2_ | goxboxlive: what do you think about this screenshot ? |
23:31.04 | tsdogs | BabelO: how is it called the calendar app in qtopia? |
23:32.10 | BabelO | tsdogs: datebook |
23:35.42 | goxboxlive | cr2_, what, where when |
23:36.38 | tsdogs | goxboxlive: http://84.18.144.44/tsdogs/screenshots/roadmap_zoom.png |
23:37.28 | BabelO | tsdogs: what is bar graph on bottom of your screenshot ? gps reception level ? |
23:37.50 | goxboxlive | tsdogs, cr2_ cool |
23:38.12 | tsdogs | BabelO: nope, zoom :) |
23:38.18 | goxboxlive | BabelO, i have the contact button working some how now |
23:38.19 | BabelO | tsdogs: ok :) |
23:38.27 | BabelO | goxboxlive: :( |
23:38.30 | tsdogs | It's just a test, but could be used for volume :) |
23:38.55 | goxboxlive | BabelO, it just to fill in some of this: http://doc.trolltech.com/qtopia4.2/contactsservice.html |
23:40.56 | goxboxlive | haha i got the cam button working :-) |
23:41.41 | goxboxlive | cr2_, do the magician and universal share the same cam? |
23:41.47 | cr2_ | tsdogs: this is the opinion of the endusers :) |
23:42.14 | cr2_ | goxboxlive: maybe not exactly the same but close enough |
23:42.17 | tsdogs | :) |
23:42.28 | cr2_ | it may depend on the boardid. |
23:42.29 | goxboxlive | ok, good |
23:42.32 | goxboxlive | i c |
23:42.43 | cr2_ | the differences are minimal. |
23:42.59 | cr2_ | i think most of the devices uses the same cam. |
23:45.19 | tsdogs | cr2_: commited. for today I'm done. good night all |
23:45.40 | goxboxlive | night |
23:45.48 | BabelO | goxboxlive: good night |
23:46.05 | cr2_ | tsdogs: found |
23:46.09 | cr2_ | <PROTECTED> |
23:46.09 | cr2_ | <PROTECTED> |
23:46.09 | cr2_ | <PROTECTED> |
23:46.10 | cr2_ | <PROTECTED> |
23:46.10 | cr2_ | <PROTECTED> |
23:46.10 | goxboxlive | oh no, that was for tsdogs . |
23:46.29 | tsdogs | cr2_: cool |
23:46.34 | cr2_ | ok. good night. |