IRC log for #htc-linux on 20070531

04:08.14*** join/#htc-linux KimuSan^ (n=kimusan@ip205.tankegang.dk) [NETSPLIT VICTIM]
04:35.03*** join/#htc-linux rmoravcik (n=rmoravci@pc-3s0zt5w2e4y0vzmhnrzq3a21zqajzfw.users.student.utc.sk)
05:01.26*** join/#htc-linux sadeness_ (n=LamersIn@nat2.rlan.ru)
05:27.28*** join/#htc-linux pleemans (n=peter@d51A5E76A.access.telenet.be)
07:49.44*** join/#htc-linux the_sys0p` (n=the_sys0@cpe-66-75-196-173.bak.res.rr.com)
07:53.20*** join/#htc-linux goxboxlive (n=goxboxli@176.84-48-210.nextgentel.com)
08:13.22*** join/#htc-linux DTMonterrey (n=evandro@cornelius.estig.ipb.pt)
08:15.28*** join/#htc-linux Foxdie (i=Foxdie@77-98-48-193.cable.ubr05.harb.blueyonder.co.uk)
08:39.27*** join/#htc-linux BabelO (n=Miranda@faim1.itesoft.com)
09:11.51BabelOhi
09:13.48BabelOcr2 : you and tsdogs work very hard on gps tools ;) to detect if a GPS is connected and then start it, there is a way to do that with htcphonehardaware, then setting a VS entry like GPS
09:19.35cr2hi BabelO
09:20.00cr2we are mostly trying to use the existing code, so it's easier than the phone.
09:20.16cr2BabelO: the qt4.3.0 is out ?
09:21.09cr2Trolltech Releases Qt and Qtopia Core 4.3  (May 30)
09:21.24BabelOcr2 : : don't know, but it is written that it is source compatible ;)
09:22.08cr2i need to compile qt4 for 32bit machine, so i've downloaded it.
09:22.55BabelOcr2: you can run a XEN domain on your portable then have a real 32 bits machine ;)
09:23.05BabelOportable is laptop , sorry :)
09:23.21cr2:)
09:23.36cr2qtopia is at 4.2.2
09:48.11BabelOcr2 : do you have some nice screenshot at the moment ?
09:50.48cr2no.
09:51.08cr2i need to resolve the segfault in buildmap.
09:52.03cr2hehe. can't mount the greenphone 4.2.* ext2 rootfs ?
09:52.58cr2recompiling qt4
09:55.50ljpthats not just a ext2 filesystem. its a partition system with two partitions :)
10:03.16cr2ljp: thanks. just wanted to have a look what#s new.
10:40.00*** join/#htc-linux bd2 (n=cbou@89.113.85.104)
10:48.24*** join/#htc-linux tsdogs (n=twostupi@84.18.144.39)
10:59.55*** join/#htc-linux gundam (n=gundam@slackware.it/staff/gundam)
11:00.35BabelOcr2 : what is  your sql interpreter you use this night for sqllite ?
11:01.42*** join/#htc-linux goxboxlive (n=goxboxli@ti100710a080-2108.bb.online.no)
11:01.55goxboxlivehowdie
11:03.00goxboxlivecr2 awake?
11:03.37goxboxliveBabelO: are you there?
11:03.49BabelOyes
11:03.53goxboxlivehi
11:03.59goxboxlivei borred
11:04.01BabelOhi
11:04.23goxboxlivechild is sleeping, there are plummers at my howm, there are now water and the sonk cant be used.
11:04.35goxboxlives/sonk/zink/&
11:05.00goxboxlivedo you know if qtopia-4.3.0 is beeing planned added to he opensource Opie?
11:05.37BabelOi think yes it is in list of port
11:06.06BabelOhttp://doc.trolltech.com/qtopia4.2/porting.html
11:06.26goxboxliveok, good. But why upgrade to 4.2.2 first. Cant it just be updated right to 4.3.0?
11:07.45*** join/#htc-linux TeringTuby (n=maarten@162-117-dyn.ipact.nl)
11:09.29BabelOi don't think because there is little change from 4.2.2 to 4.2.3
11:09.41BabelOgoxboxlive : you have no pin on you universal ? right ?
11:26.09goxboxliveno,  dont use it
11:26.23goxboxlivechild woken, have to go. bbl
11:26.26*** part/#htc-linux goxboxlive (n=goxboxli@ti100710a080-2108.bb.online.no)
11:34.26cr2BabelO: we use postgis. it's a package on top of postgres.
11:34.48cr2http://postgis.refractions.net
11:35.05BabelOcr2 : : ok, because i need to register call in and out and also missed to qtopia.sqlite
11:35.42cr2BabelO: it's only to prepare the map. in a custom binary format.
11:35.53cr2so no sql on the phone.
11:36.25cr2the 'buildmap' reads from the postgis database and creates the binary *.rdm map file.
11:37.10cr2it would be nice to have rdm2pgsql program for debugging, but we are not so far.
11:37.54cr2there is a "dumpmap" program but it's output is extremely cryptic.
11:41.22BabelOcr2 : ok
11:41.23BabelOme am looking into qtopia source to find Qtopia channel to call to display or query a contact
11:42.53cr2ok.
11:47.58cr2hit a bug with 4.3.0
11:48.01cr2<PROTECTED>
11:57.33cr2not on the 32bit.
11:57.40cr226 zoom levels.
11:59.40cr2or 25.
12:00.01cr2the 1:1 looks like setting level 20.
12:01.02BabelOcr2 : you are running it on your laptop ?
12:02.16cr2laptop is the x server.
13:22.33BabelOcr2 : so you startqroadmap on universal and display on X server of laptop ? ;)
13:42.20*** join/#htc-linux pH5 (n=ph5@e178234065.adsl.alicedsl.de)
13:54.16cr2BabelO: no. it's a qt4 program, so i compile it natively.
13:54.29cr2and hope that it will work on qtopia.
13:54.55cr2it's easier to iron out the bugs on a fast machine with a big display :)
15:04.21*** join/#htc-linux scorpio16v (n=scorpio1@ACAEBE7B.ipt.aol.com)
15:06.23*** join/#htc-linux WizMaui (n=WizMaui@62.112.90.250)
15:56.03*** join/#htc-linux RoEn_PC (n=roen@p54a647b1.dip.t-dialin.net)
16:25.28*** join/#htc-linux pleemans (n=peter@d51A5E76A.access.telenet.be)
16:28.08*** join/#htc-linux mr_vega (n=tom@i59F75FDF.versanet.de)
16:33.56*** join/#htc-linux bd2 (n=cbou@89.113.85.104)
16:34.22*** join/#htc-linux goxboxlive (n=goxboxli@176.84-48-210.nextgentel.com)
16:40.19goxboxlivei
16:40.20goxboxlivehi
16:40.43goxboxliveBabelO, have you seen this before? http://pastebin.no/1835
16:41.33goxboxliveThis happends on my laptop even though i am using same crosscompiler and a also tried to build a copy of the source on my build machine.
16:47.28BabelOgoxboxlive : it seem it do not found qlist.h file :( really strange
16:47.40goxboxliveyes, because it is there :-)
16:48.42goxboxliveAnd it is exactly the same source as on my build machine, so why suddenlyu doesnt it find it. I have also tried a fresh sourec, even qtopia-4.2.2
16:49.46BabelOgoxboxlive: , maybe a bad define detected from your machine... but which one ...
16:49.55BabelOback to home now
16:49.56BabelObbl
16:50.08goxboxlivemaybe, ok. cu
16:55.33cr2hi goxboxlive
16:55.40goxboxlivehi cr2
16:56.17cr2i've built a map for tsdogs on a 32bit machine and it works.
16:56.40cr2can't do anything with this map being a piece of shit, though ;)
16:57.27cr2i need to look for a really good map of Oslo :)
16:58.18cr2it can be done, but the route planning info will be lost. will need to generate some dummy one.
17:12.19*** join/#htc-linux the_sys0p (n=the_sys0@cpe-66-75-196-173.bak.res.rr.com)
17:29.15goxboxlivecool :-) It will work in OPie_II right?
17:33.02*** join/#htc-linux BabelO (n=Fabrice@lun34-2-82-238-28-28.fbx.proxad.net)
18:07.18tsdogscr2: I saw u builded the map :)
18:11.43*** join/#htc-linux TimRiker (n=timr@pdpc/supporter/bronze/TimRiker)
18:16.37*** part/#htc-linux scorpio16v (n=scorpio1@ACAEBE7B.ipt.aol.com)
18:24.31*** join/#htc-linux Foxdie` (i=Foxdie@77-98-48-193.cable.ubr05.harb.blueyonder.co.uk)
18:25.37*** part/#htc-linux TimRiker (n=timr@pdpc/supporter/bronze/TimRiker)
18:32.22cr2tsdogs: on a 32bit machine.
18:32.41tsdogs:) buggy 64bit code
18:32.59cr2goxboxlive: ask tsdogs is it works on qtopia :)
18:33.17cr2tsdogs: i have some minor bugs on 32bit too.
18:33.17*** part/#htc-linux mr_vega (n=tom@i59F75FDF.versanet.de)
18:33.19tsdogshaven't tryed it yet.
18:33.23tsdogsnext task
18:33.49tsdogscr2: is the screenshot from "my" map?
18:33.54cr2yes.
18:34.22cr2i did have a problem, but not a segfault.
18:34.26tsdogsIs it usable (u called it shit)
18:34.47goxboxlivecr2, let med see the screenshot
18:34.56cr2the problem was with the max seg number in a line.
18:35.13tsdogsu remember which one?
18:35.15cr2it's not usable for route planning.
18:35.19tsdogs:(
18:35.29cr2because the map itself is bad.
18:35.49tsdogsI could add some route info with mapedit. only for few streets.
18:35.59cr2yes.
18:36.06cr2there is no street naming.
18:36.15tsdogsdidn't netgen generate them?
18:36.19cr2but these are the problems with the map itself.
18:36.24cr2names ?
18:36.39cr2how can it now ?
18:36.43tsdogsnot names, but route nodes
18:36.51tsdogs:) magic
18:37.02cr2i've generated the road net/graph with mapedit, but in an evil way.
18:37.19cr2by saving into its internal format.
18:37.56cr2so the lines were cut into pieces, therefore the problem with MAX_LINE
18:38.22tsdogsI run netgen on the map, and nodes where added.
18:38.24cr2the routing works, but it creates very strange ways.
18:38.33tsdogsho ok
18:38.43cr2netgen creates some "lines" with only 1 node, which is wrong.
18:39.02cr2of course you can enhance the map.
18:39.24tsdogsok, but I opened the map after netgen and it seemed to be ok
18:39.33cr2we can add the lake and rivers, it's not a very big problem
18:39.55tsdogsok. So  I need a 32bit machine
18:39.59cr2do you see the calculated route on the screenshot ?
18:40.08tsdogsyes.
18:40.15cr2is ot ok ?
18:40.37cr2i can send you the *.rdm files.
18:40.45tsdogsno idea, but could be, as there are lot of mountains here around
18:40.52cr2otherwise i need to setup postgis on amd64.
18:41.25tsdogsI see some pointer issues when building but don't know how to fix them
18:41.36cr2btw, if we had the altitude contours, we can add them as a "border" :)
18:41.55cr2i think it's a dglib problem.
18:42.14cr2because i was able to generate the map from DCW shapefiles.
18:42.27tsdogsI got the new snapshot from cvs, but it even segfaults before starting
18:42.31cr2but they were even more primitive ;)
18:42.35tsdogsno pointer issues though
18:42.56cr2you added qt4 dir to the latest CVS snapshot ?
18:43.13tsdogsno downloaded dgliib cvs snapshot
18:43.37cr2i've manually added 'from_travel_ref' and 'to_travel_ref' , but don't know how to fill them.
18:43.55tsdogshttp://grass.itc.it/dglib/source/snapshot/
18:43.56cr2where did you get it from ?
18:43.58cr2ok.
18:44.42cr2useless.
18:44.45cr2typedef long dglInt32_t;
18:45.04cr2i don't think it changed for a long time.
18:47.13tsdogssegmentation fault :(
18:47.47tsdogsNew version has long.
18:48.59cr2let's stay with the version included in roadmap_editor
18:49.12tsdogsok
18:49.17cr2it doubt it differs from the latest CVS snapshot ;)
18:49.23tsdogsdoes not change though
18:49.38tsdogsit does, the dglInt32_t now it's long :)
18:50.12tsdogsyou had the segfault with 32bit too ?
18:50.25cr2if you do typedef int dglInt32_t then the 64bit machine can at least work with the data .
18:50.28cr2no.
18:51.04tsdogsbuildplace_main.c: In function ‘buildplace_dsg_add’:
18:51.05tsdogsbuildplace_main.c:230: warning: cast to pointer from integer of different size
18:52.16tsdogsok. at home I have a 32bit machine I can work on.
18:54.23cr2it's not used.
18:54.36tsdogsyep. now I saw it.
18:55.57tsdogsbtw where are the xx_travel_ref used?
18:57.12cr2don#t know :)
18:57.31cr2but it's in the sql request, so it must be in the db.
18:58.13tsdogsI don't see any in there
18:58.50tsdogsyou mean fraddl ?
18:58.52cr2latest roadmap_editor CVS snapshot ?
18:59.23cr2http://roadmap.cvs.sourceforge.net/*checkout*/roadmap/roadmap_editor/src/buildmap_postgres.c?revision=1.17
18:59.35tsdogsho ok. you using that one?
18:59.50cr2from_travel_ref, to_travel_ref FROM segments
18:59.57tsdogsok
19:00.18cr2i#ve tried to adapt to this scheme. it's backwards compatible.
19:01.09cr2should i send you the *.rdm ?
19:01.49tsdogsyes.
19:02.04tsdogsbut it does not seem to be used in the source right now.
19:02.26cr2ok.
19:02.28cr2sent.
19:03.01cr2hm. true.
19:15.36*** join/#htc-linux kiozen (n=oeichler@rgnb-d9bde38b.pool.mediaWays.net)
19:16.10*** join/#htc-linux tsdogs_ (n=twostupi@62.123.180.130)
19:16.35tsdogs_sorry pc hung
19:19.16tsdogs_hmm what was the position?
19:19.31cr210.8 47.8
19:19.46cr2that's the problem ;)
19:20.41cr2there is no way to know what maps are there, their bounding boxes, zoom level, scale and current coordinates :)
19:20.57tsdogs_hmm :/ no map available :(
19:23.18cr2zoom.
19:24.03cr2i've seen a screenshot, but can#t find it. for the zoom visualization and scale bar.
19:25.31kiozenany expert around to answer some question what it would take to linuxfy a new device?
19:26.30cr2kiozen: what cpu do you have ?
19:26.39kiozenI suspect some ARM inside an ASIC
19:27.00cr2do you know the software that controls it ?
19:27.22kiozenproprietary Garmin
19:27.40cr2linux on a garmin device ?
19:27.43cr2which one ?
19:27.46kiozenyep
19:28.06cr2wince/palm/something else.
19:28.42kiozenjust a minute , phone, sorry
19:28.47tsdogs_bbl
19:30.17cr260Cx , 76Cx may be interesting. they use sirf3, so it's higly likely the main ASIC is smaller and simplier. and sirf is just on a serial port.
19:31.29cr2and internally 60Cx==76Cx afaik. at least they have the same firmware.
19:32.05kiozenok back from phone, sorry, GPSMap60SCx
19:32.10*** join/#htc-linux Jonathan_Watkins (n=chatzill@adsl-227-5-38.jan.bellsouth.net)
19:32.44kiozenthis one displays maps. there must be some calculation power
19:34.12cr2ok, so i doubt it's wince.
19:34.32cr2do you know something about the rom update structure ?
19:34.50cr2or anything about the device internals ?
19:35.11cr2afaik it has 16MB internal flash for the basemap.
19:35.33cr2which can be reflashed with a faked "rom update".
19:36.33kiozenno I know nothing, just start to investigate.
19:36.33kiozenI guess Garmin uses it's own OS
19:36.33cr2ok, will you bother to open the device ?
19:36.34kiozenno
19:36.34cr2i can see the parts of the sirf firmware in the rom update.
19:36.54cr2is it possible to reflash the rom and log the usb traffic ?
19:37.27cr2i _may_ think about opening 76Cx.
19:37.35kiozenyou mean load a new frimware and log it?
19:37.43cr2yes.
19:37.50kiozencan do
19:38.18cr2i'll try to find out the links/information about reflashing the basemap.
19:39.14cr2the LCD is most likely 8bit,
19:39.24BabelOhi
19:39.38cr2if we will find where its framebuffer is mapped, then we can "draw" on the screen :)
19:39.44kiozenyes it shows 8 bit windows bitmap
19:40.11cr2hi BabelO, i've generated the italian map for roadmap :) at it works.
19:40.24cr2s/at/and/
19:40.53kiozeni try to find a firmware to load
19:40.55cr2kiozen: we need to find out the framebuffer ram location.
19:41.10cr2ok, that will be a big step forward.
19:41.17BabelOcr2: better italian map, there is some mountain for test ;)
19:41.36BabelOany idea of garmin processor ?
19:42.00cr2roadmap is too primitive for that. garmin is better right now ;)
19:42.15cr2BabelO: it's ARM, and the sirf3 is also arm.
19:42.20cr2and brf6150
19:42.26cr2and acx100
19:42.35cr2they are all arm-based controllers.
19:42.54cr2like omap850.
19:43.04cr2all undocumented.
19:43.30cr2but porting to garmin is much more cool than to omap850.
19:43.48cr2at least i have a device :)
19:44.17BabelOcr2: which garmin do you have ?
19:44.30cr276Cx
19:44.40BabelOmy last test with Qlandkarte and 1.8Mo img file was very slow on Blueangel
19:45.02cr2QLandkarte does some unnecessary things.
19:47.18cr2it loads the whole img into RAM buffer.
19:47.18kiozenwhat things?
19:47.18cr2probably some strange heritage from gpsmapedit :)
19:47.20kiozensure it loads the complete map. for speed.
19:47.20kiozenok that is no good idea for embedded devices :)
19:47.21cr2kiozen: on embedded device, you don#t want it.
19:47.21cr2:)
19:47.21kiozenI guess so
19:47.21cr2why not just mmap it ?
19:47.22*** part/#htc-linux DTMonterrey (n=evandro@cornelius.estig.ipb.pt)
19:47.26cr2hm. not good for the XORed maps.
19:47.41kiozenI tried to be speedy on a pc
19:47.47cr2ok.
19:47.48kiozenmore speedy than mapsource
19:48.24cr2sure.
19:48.34cr2i don't use mapsource.
19:48.44kiozennow, me either
19:48.48cr2just put the *.img on the microSD.
19:49.12kiozenI like to have it more confy
19:49.56cr2the gui certainly needs some changes for a smaller screen.
19:50.09kiozen:)
19:50.09cr2BabelO has 240x320
19:50.19cr2and the biggest is VGA.
19:50.29cr2and these are quite nice devices.
19:50.33cr2with 128MB RAM.
19:50.47cr2don#t know how much 60/76Cx has.
19:50.54cr2certainly much less.
19:51.06cr232 or 48, is my guess.
19:51.13cr2and 16bit color.
19:51.25cr2not good for the battery, but looks nice :)
19:51.40cr2garmin's 16bit palette sucks.
19:51.40kiozen160x240
19:51.53cr2it's the LCD site.
19:52.02cr2i mean the main RAM on garmin.
19:52.43kiozenshrug ?
19:53.14cr2i've seen some software on sf.net that deals with the garmin roms. what was it ...
19:53.53kiozeni plunge into windows to do this trace, i'll beeee baaack
19:54.06cr2garmin has flash (==ROM) and RAM. i'd like to know the sizes and base addresses.
19:54.08cr2ok.
19:55.04BabelOcr2: so you did not need to go to qlandkarte mailing list ;)
19:55.15cr2BabelO: let me know which "square" do you need for roadmap.
19:55.24cr2BabelO: ok :)
19:55.45cr2it may be quite fun to port linux there.
19:56.00cr2a bit more difficult though :)
19:56.12cr2because there is no haret.
19:56.48cr2it's like booting linux from the HTC diagnostic SD card.
19:57.00cr2which we can do. in principle.
19:57.32BabelOok, but if flash is small, first 2Mo will be  assembler boot and init stuff ;)
19:57.48cr2BabelO: so, how do we advance the phone on opie2 ? we have a lot of gps software already :)
19:57.58cr2i don't remember.
19:58.06cr2it's 16MB, or maybe 8.
19:58.51BabelOcr2: i am spliting code into 3 class, qgnokii - qdialer and base phone
19:59.05cr2ok.
19:59.16BabelOsome conflict, i have phone part running in a single cpp file, but really dirty
19:59.36cr2the bootstrap process should be so simple that even such dumb guy as me will be able to do it :)
19:59.40BabelOmaybe one or two days more
19:59.52cr2ok.
20:00.17cr2i'd like to help, but need to build opie2 first.
20:06.38BabelOhumm ;) close to compile new source, maybe have to comit them to xanadux
20:07.21BabelOtsdogs: do you know how can i add an external include path from pro file ?
20:08.09*** join/#htc-linux kiozen (n=oeichler@rgnb-d9bde38b.pool.mediaWays.net)
20:08.25kiozengot it
20:09.18kiozenhow to send big files? (shht, I am a total noob to IRC)
20:10.00BabelOkiozen: how big is file ?
20:10.27kiozenthe trace 4.5 MB the FW file 3.2MB
20:11.43BabelOcr2, kiozen : i can share my ftp server
20:11.53cr2ok.
20:12.29cr2kiozen: is it 029201000310.rgn ?
20:12.34kiozenyep
20:12.37cr2ok.
20:12.40kiozenno
20:12.50kiozen<PROTECTED>
20:13.01cr2330 ?
20:13.10kiozen3.30 version of the fw
20:13.11cr2i thought 3.10 was the latest.
20:13.15kiozenno
20:13.20cr2ok.
20:13.29kiozenyou missed nothing :)
20:13.34cr2;)
20:13.50cr2i didn#t update anyway.
20:13.56cr2no windows.
20:14.19kiozenno Garmin soft would be better
20:14.25cr2btw, qlandkarte does not honor the map codepage.
20:14.27cr2lol.
20:14.41kiozenwhat code page?
20:14.47BabelOlinuxbox.hd.free.fr:5060 login/password is invite
20:14.55cr2for the map.
20:15.05cr2which is in the .img
20:16.26kiozenis there? I guess we have to put knowledge together
20:17.21kiozenBableO: no ping no promt?
20:18.33cr2kiozen: there should be. there is one in the cgpsmapper.
20:19.22cr2there is even the multibyte encoding. and the taiwan firmware for 69Cx.
20:19.23cr2s/69/60/
20:19.37kiozenI always wonder how they do asian signs :)
20:19.53cr2yeah, that was my idea too.
20:20.16kiozenstill no connection to the ftp ?
20:20.19cr2afaik there is a software written by russians to replace the fonts. with the font editor.
20:20.29cr2i'll try.
20:21.01cr2BabelO: nothing from here too.
20:21.30cr2oh. it works without :
20:21.34BabelOhm, really strange i can access it from my work computer :(
20:21.54cr2kiozen: ftp host 5060
20:22.37kiozen:) messed it up i am in
20:23.15cr2ok.
20:23.18kiozencan't open file
20:23.41BabelOok, it is pure ftp, so i ve maybe miss some write access 2s , i ll look
20:24.45cr2kiozen: one more btw :) the latest gpsmaedit source has some additional data about .net and .nod
20:25.59kiozenyeah and seems to decode the encrypted map level. I wonder if it decrypts or gathers the data from subdivision
20:26.17kiozenI wrote a mail to the author lets see
20:26.28BabelOit is ok now ;)
20:26.58kiozenworking
20:28.26cr2kiozen: i don't mean the xor. i mean the routing. what you are talking about is available since 2 years. but you don't want to implement it.
20:29.21kiozenxfer done
20:29.42cr2and you don't need to ask the author about it. it's all already there.
20:29.55cr2ok.
20:30.33kiozeni did not find the *inc file
20:30.55cr2it's not needed.
20:31.29cr2just remove the == 150.
20:31.54kiozenjust a minute, let me see
20:32.53cr2ok, downloaded.
20:38.33*** join/#htc-linux goxboxlive (n=goxboxli@176.84-48-210.nextgentel.com)
20:39.37cr2BabelO: i forgot to tell you that i've found one more G3 mdoc register. or maybe two.
20:39.53goxboxlivecr2, I got my new battery today, and since i use ssh so much i dont want to let the battery charge when connected. What do i have do comment in the power driver to disable charging?
20:40.09cr2and got some idea what are the undocumented registers are used for.
20:40.24cr2goxboxlive: the charge enable gpio.
20:40.51goxboxliveok
20:41.40BabelOcr2: so we have luck soon top have an open source ttfs driver ;)
20:42.16goxboxlivecr2, should i commnt all this, or just remove it? Is there anything else?
20:42.19goxboxlivestatic void charge_on(int flags)
20:42.19goxboxlive{
20:42.19goxboxlive<PROTECTED>
20:42.19goxboxlive}
20:43.28cr2set it to 1<<GPIO_CHARGE_EN instead of 0.
20:43.44cr2so it will stay disabled.
20:43.47goxboxlivethank you
20:44.10cr2kiozen: found the project for editing the fonts in garmin firmware http://sourceforge.net/projects/flawia
20:45.02goxboxlivecr2, but it is '1' . Look at the line, should it be '0' then?
20:46.01BabelOgoxboxlive: at end of line
20:46.46goxboxliveBabelO, ok
20:47.18cr2asic3_set_gpio_out_b(&htcuniversal_asic3.dev, 1<<GPIOB_CHARGE_EN, 1<<GPIOB_CHARGE_EN);
20:47.52goxboxlivedanke
20:48.45cr2BabelO: you are right. there is "header.bin" , "loader.bin" , "bootseg.bin" and "fw_all.bin"
20:49.07kiozenok, I am too slow to get this today, I'll have a look at this tomorrow.
20:49.36kiozenif it is no real encryption I can take it into QLandkarte
20:50.07cr2ok.
20:50.31cr2it's neither rsa nor des ;)
20:51.14kiozenIt  must be just trivial to comply to German law
20:51.43cr2wether it's "effective" of not, i don't know.
20:52.06cr2rot13 may be effective too ;)
20:52.34kiozenmost likely I will keep it out of QLandkarte, there is already an interface for such things.
20:54.41BabelOcr2: any any about processor from this firmware ? loader.bin seem interesting
20:54.42BabelOgarmin do thing clear :)
20:55.31cr2no. maybe it's better to split the firmware first.
20:55.57cr2kiozen: do it like libcss.
20:56.19kiozenthat's what I did already you just have to open the gates of Moria ;)
20:56.29cr2:)
20:57.39kiozenand you need a Sweedish server
20:59.27cr2<PROTECTED>
20:59.34cr2didn't know about it.
21:00.48cr2Copyright SiRF Technology, Inc. 2005
21:00.57cr2that should be the firmware.
21:01.52cr2heh. html inside.
21:02.54cr2arm. who has doubted that.
21:02.57cr2Stack Overflow: r0=task(%s), r%d=limit, r13=sp, r15=pc
21:06.06kiozenhtml? found some xml stuff for GPX file
21:06.25cr2WARNING: Do not remove batteries while this feature is active.
21:06.45cr2When you think you know where the beast is, from an adjacent spot, press ENTER to initiate aiming.
21:07.01cr2a lot  of junk there.
21:07.39cr2i don't know where is the code page, but maybe it's good to let the user select it in the gui.
21:07.52cr2but eventually we'll find it.
21:08.04kiozenat 308fe0.. is that a key?
21:10.21cr2don't know, looking at the fw now.
21:11.05cr2BabelO: looking for "Abandoned wreck" in french :)
21:12.13goxboxlivehttp://www.boygeniusreport.com/2007/05/24/htc-kaiser-unboxing/
21:14.13cr2goxboxlive: the Kaiser's RAM, 128MB of it. 75MB are usable to the user for programs. Incredible!
21:14.21BabelOcr2: there is a special feature for that  :)
21:14.32cr2m$ is very space-efficient when it comes to RAM.-
21:14.52goxboxlive:-), wonder what kond of cpu it has
21:14.59cr2a mini blog publishing tool built right in!
21:15.16cr2i want to know the LCD size.
21:16.01goxboxlive32bit Qualcomm MSM7200 Chipset
21:16.13goxboxlivehttp://pdadb.net/index.php?m=specs&id=733
21:16.25goxboxliveqvga
21:16.41cr2athena rulez :)
21:17.03goxboxliveno wrong, universal still rules.
21:17.13goxboxliveThe kaiser has microsd.
21:17.56cr2like hermes.
21:18.08cr2the CPU is not qualcomm ;)
21:18.14cr2CPU Clock:   384 MHz
21:18.19cr2strange clock.
21:18.43goxboxliveyes i saw
21:19.03cr2i suspect it's not sirf3 there.
21:19.37cr2they put 128MB in hermes. that#s all.
21:20.05cr2and enabled gps in msm chipset.
21:21.22kiozenI have to get some sleep to be nice to customers tomorrow :)
21:21.24kiozencu
21:21.29cr2good night.
21:21.34BabelOgood night
21:30.20cr2Firmware image at 00015F57h
21:30.21cr2Can not find string offset
21:32.12cr2the firmware is 3MB.
21:33.33BabelOreally small ;)
21:33.36cr2loader.bin looks like arm asm.
21:34.17cr2ea 00 00 02
21:34.17cr2it#s some b 0x1000 or something like that,
21:34.23cr2the loader is 73K
21:42.13cr2BabelO: look like normal arm asm. not THUMB.
21:42.45cr2the 0xfffexxxx addresses ? was it omap ?
21:43.29BabelOcr2: maybe omap, it is less power
21:43.52cr2xxxx is ccxx
21:44.06cr2need to look at the omap headers :)
21:44.36cr2the code looks like as written in raw asm. no compiler fluff.
21:44.56cr2just writing in the hardcoded registers.
21:45.02BabelOsoit is really optimized for size ;)
21:46.57cr2then .code 16
21:47.00cr2thumb .
21:47.29cr2stack pointer at 0x080A00xx
21:47.54cr2RAM at 0x08000000 ?
21:49.00cr20x0801xxxx used.
21:49.56cr20x0804xxxx
21:50.58cr2some fiddling with the msr/mrs. maybe switching virtual mapping on. not good ;)
21:59.39cr2the offsets are unrealistic.
22:00.08cr2probably necessary to check what are these mrs/msr doing.
22:02.48*** join/#htc-linux FossiFoo (n=Fossi@e176102138.adsl.alicedsl.de)
22:17.54*** part/#htc-linux rmoravcik (n=rmoravci@pc-3s0zt5w2e4y0vzmhnrzq3a21zqajzfw.users.student.utc.sk)
22:26.21tsdogscr2: how did you generate the shp prj files?
22:26.59cr2i did a hack.
22:27.13tsdogsI've taken the Israel map
22:27.19cr2i've generated mapinfo files from .mp
22:27.23RoEn_PC;)
22:27.27RoEn_PCoops ;)
22:27.33cr2and then converted them to .shp with ogr2ogr.
22:27.39tsdogsok
22:27.41cr2actually it's not very smart.
22:27.52cr2it's easier to generate sql directly.
22:28.03cr2but it's historical now.
22:28.12cr2map of israel ?
22:28.14cr2which one ?
22:28.19tsdogsfrom mapsource
22:28.29cr2is it good ?
22:28.31tsdogsit seems to be the more complete there
22:28.40tsdogsit's the same that was used for testing
22:28.50cr2?
22:29.07tsdogshttp://mapcenter.cgpsmapper.com/maplist.php?cnt=100&rgn=&cat=2
22:29.22tsdogsthe one we downloaded from roadmap_editor site
22:29.29tsdogsonly it is the source :)
22:29.42tsdogsso we can build new one to be used with new version.
22:30.13cr2are you sure ?
22:30.56tsdogsyep
22:30.57cr2but ok, it's good for testing.
22:31.21cr2260k is not much for a good map.
22:31.30tsdogs:)
22:31.33cr2a good map is about 10MB for one city :)
22:32.05tsdogsdid not find any on those site
22:32.46cr2look for something at mapcenter2.cgpsmapper.com
22:37.55cr20x081e0000
22:38.33cr2at least 2MB
22:39.06tsdogshttp://mapcenter2.cgpsmapper.com/maplist.php?country=118&category=2
22:39.20tsdogsthis are 8 maps here in the sorroundings
22:40.36tsdogshmm no routing
22:40.37cr2it's a bike map.
22:40.41tsdogs:)
22:40.54cr2http://www.garda-gps.de
22:41.00cr2this is the source
22:41.33cr2yes, for the bike routes
22:42.35cr2ok, i know where to get better ones. but without routing too.
22:43.29cr2but at least with the normal roads :)
22:43.56tsdogs:)
22:44.36cr2but maybe it's better to take a small map, but with good routing.
22:44.48tsdogsyes
22:44.58cr2i'll look at this 260K map
22:45.00tsdogswhat's the command line with ogr2ogr?
22:45.15cr2what do you want to do ?
22:45.27tsdogscreate shp files to be imported in db
22:45.37cr2from what ? :)
22:45.44tsdogscreating the map
22:46.00tsdogsfrom MP
22:46.09cr2i'll send you the awk script .
22:46.12tsdogsok
22:46.38cr2but it should output sql instead of what it's doing right now.
22:46.48cr2do you know awk ?
22:47.25cr2it was thought of as a dirty hack,
22:47.26tsdogsyes
22:47.40cr2so i need to think how to do it properly.
22:47.55cr2maybe something between netgen and osm2pgsql.
22:48.13cr2netgen is c++ and osm2pgsql is c.
22:48.37tsdogsc+ ?
22:48.41tsdogs:)
22:48.53cr2:)
22:49.41tsdogsisn't there any tool to simply import mp files into postgis ?
22:52.28BabelOcr2: some progress, but now i have splitted all in many file, i lose some GUI stuff ;)
22:52.31BabelOhttp://linuxbox2.free.fr/blueangel/ScreenShot/capture15.png
22:52.54tsdogsnice :)
22:53.36cr2tsdogs: not yet. netgen+osm2pgsql=mp2pgsql :)
22:53.50tsdogsok.
22:53.55cr2don't forget that mp is a frontend to cgpsmapper
22:54.00BabelOtsdogs: yes, take many times to split all in many file but now i can go faster ;)
22:54.04cr2something that should generate money.
22:54.26cr2overybody thinks that by doing route planning they will make money fast.
22:54.31cr2including nokia.
22:54.38tsdogs:)
22:54.42cr2http://www.smart2go.com
22:54.42tsdogsain't we?
22:55.05tsdogsBabelO: nice work.
22:55.05cr2we do 'destroy the ability to make money fast in this way' :)
22:55.22tsdogs;)
22:55.48cr2BabelO: looks very good.
22:55.54cr2tsdogs: sent the script.
22:56.04tsdogsI see.
22:56.09cr2you can dramatically strip it down :)
22:56.31cr2and make output sql.
22:57.03tsdogsok
22:57.18cr2but maybe let me do it.
22:57.27tsdogssure
22:57.46cr2and make some gui changes instead.
22:58.07BabelOcr2: i plug gnokii in a qgnokii class, because i know only how to detect gsm event using qtopia ;)
22:58.13tsdogsThat's why I want the map
22:58.18cr2i think this 'no map available' and the inability to do anythng is very annoying for the enduser.
22:58.21cr2ok.
22:58.47tsdogsand gps sim
22:58.48cr2BabelO: it's not a problem, we need a new AT parser anyway.
22:59.10cr2tsdogs: ok, i'll think about the simulator.
22:59.19tsdogsok thx.
23:02.06BabelOok, keyboard is to small...need to fill screen and same for number view
23:09.36cr2BabelO: looks like omap730.h
23:09.41cr2at least some addresses in the same range.
23:11.16BabelOwith a nice picture of an opened garmin it should be better ;)
23:11.25cr2lol.
23:11.40cr2it may have some custom asic number.
23:11.46cr2and omap730 inside.
23:12.15cr2ATI on blueangel has an OTG USB host, btw.
23:12.29cr2which is not done by ATI.
23:12.43cr2i've found a datasheet from the company that designed it.
23:12.59cr2and the SD part of ASIC3 is done by toshiba.
23:13.10cr2with an unreadable datasheet ;)
23:14.17*** join/#htc-linux tsdogs (n=twostupi@62.123.180.130)
23:14.39cr2hehe. there is a bit more registers than documented in omap730.h
23:14.51cr2but i'm looking at the hh.org CVS.
23:15.03cr2maybe the native omap svn will be better.
23:15.31BabelOATI capture port sand register are interesting
23:17.07cr2i know :)
23:17.20cr2but it's not easy.
23:17.58BabelOmaybe someone with a 64 channel logic analyzer ;)
23:18.30BabelOthere is some device really powerfull in this domain
23:20.05cr2Kevin2 could trace ATI chip on hermes
23:20.11BabelOis there a security for garmin flashing ?
23:20.28cr2i doubt it. only obscurity :)
23:21.39cr2omap730 also has the phone.
23:21.57cr2it's probably something close to omap730 in design.
23:22.01cr2some custom asic
23:22.30BabelOcr2: so you need the VHDL source ;)
23:22.32cr2but we really need only the LCD control, serial port and suspend/resume.
23:22.36cr2LOL.
23:22.47cr2and the 5 buttons.
23:22.55cr2there is nothing more on the device.
23:23.04BabelOi see that ds1wm is only VHDL not a regulate l chipset
23:23.14cr2yes.
23:26.18BabelOso BA version have maybe more bug than universal or hx4700 version ?
23:26.44cr2why ?
23:26.56cr2it's just different
23:27.27BabelObecause on BA, it works only if we reread interrupt register before leaving an interrupt
23:27.36BabelOelse, all freeze
23:28.16BabelOsince i discover Kevin2 tools, i can make more test, without recompililng everything
23:32.49cr2ok.
23:33.53BabelOlater now
23:33.56BabelOgood night
23:38.15cr2good night.

Generated by irclog2html.pl Modified by Tim Riker to work with infobot.