| 11:01.57 | *** join/#htc-blueangel cp15 (i=rbedlf@p549A17E0.dip0.t-ipconnect.de) |
| 11:01.57 | *** mode/#htc-blueangel [+o cp15] by ChanServ |
| 11:02.13 | cp15 | Good morning |
| 11:02.20 | bhima | good morning. |
| 11:04.07 | cp15 | Hi bhima, I have a question regarding the fpu-less sphinx. Is it somewhere available for download? |
| 11:04.57 | bhima | cp15: I got an e-mail pointing me to CVS. I can forward it to you; it was sent to a fairly open mailing list. |
| 11:05.10 | cp15 | Yes, please |
| 11:05.21 | bhima | address...? |
| 11:05.25 | cp15 | Have you already played with it? |
| 11:05.41 | cp15 | Send it to martin.schaller@gmx.de |
| 11:10.12 | bhima | sent. I found anewer e-mail also mentioning a URL. I may try to get it installed on my nokia770, though I don't think I can yet directly access audio input. |
| 11:11.18 | bhima | actually, I can get bluetooth headset audio, cause it's outside of the system's normal audio framework,a nd that's what I want anyway. :) |
| 11:25.38 | *** join/#htc-blueangel magnet (n=magnet@vsg94-1-82-234-118-20.fbx.proxad.net) |
| 11:25.49 | magnet | hi |
| 11:26.09 | cp15 | Hi magnet |
| 11:27.33 | magnet | I almost get oe installed. |
| 11:27.39 | magnet | :) |
| 11:36.18 | cp15 | What problems do you still have? |
| 11:39.40 | *** join/#htc-blueangel immolo (n=immolo@cpc1-cosh1-4-0-cust126.cos2.cable.ntl.com) |
| 11:53.17 | immolo | cp15- I've decided the problem could be my hacked together blueangel config |
| 11:53.41 | immolo | so I'm going to force the tslib.bb to think it is a h3900 |
| 12:06.55 | magnet | cp15: sorry I was on phone. |
| 12:07.13 | magnet | my biggest problem is my laptop ;) |
| 12:07.44 | magnet | right now I'm installing dependencies for xanadux-ba |
| 12:11.46 | immolo | magnet- I'll race you :P |
| 12:12.11 | magnet | :) |
| 13:01.04 | *** join/#htc-blueangel rob_w (n=bob@213.54.150.237) |
| 13:01.04 | *** mode/#htc-blueangel [+o rob_w] by ChanServ |
| 13:44.47 | *** join/#htc-blueangel pH5 (n=ph5@p5485FE4C.dip.t-dialin.net) |
| 13:49.05 | cp15 | Ah, hi pH5. If I remember correctly you also want to port linux to the magician |
| 13:49.35 | cp15 | magnet is working on the same project, maybe you could work together |
| 13:50.33 | magnet | hi pH5 . |
| 13:50.39 | pH5 | hi cp15, hi magnet |
| 13:51.12 | pH5 | until now i didn't have time to do anything but play with haret / wgpio and register a machine id, unfortunately |
| 13:51.35 | cp15 | Well, a machine id is a good start |
| 13:51.44 | magnet | what is it used for ? |
| 13:52.04 | cp15 | To let the kernel know on which machine it runs |
| 13:53.00 | cp15 | You could, for example, make a kernel which runs on blueangel, himalaya, wallaby and magician, and the kernel detects via the machine id what hardware is used |
| 13:53.21 | magnet | oh nice. |
| 13:54.14 | cp15 | cr2 has already done some work on magician gpio, so it shouln't be too hard to get any progress |
| 13:54.38 | magnet | I'm very glad to hear this. |
| 14:01.05 | pH5 | i'm very happy that a lot of the drivers are already there (pxafb, pxa-keys) |
| 14:01.30 | pH5 | but cr2 told me that there are different models with differend LCDs |
| 14:01.43 | magnet | oh |
| 14:01.51 | magnet | which model do you own ? |
| 14:02.33 | pH5 | mine is an o2 xda mini |
| 14:03.01 | magnet | I have a qtek s100. |
| 14:07.04 | pH5 | can you have a look if your LCCRR values are the same as mine? |
| 14:07.17 | magnet | sure. |
| 14:07.20 | pH5 | I have this: |
| 14:07.20 | pH5 | HaRET(1)# PD 0x44000000 24 |
| 14:07.20 | pH5 | 44000000 | 00000081 120928ef 0202093f 04300004 |
| 14:07.20 | pH5 | 44000010 | 00000000 3f3f3f3f |
| 14:08.36 | magnet | lemme find the usb cable for the pda. |
| 14:08.51 | *** join/#htc-blueangel rob-w_ba (n=rob-w_ba@213.54.150.237) |
| 14:09.18 | rob-w_ba | hi |
| 14:09.30 | pH5 | hi rob-w_ba |
| 14:10.38 | rob-w_ba | hi pH5 |
| 14:11.18 | magnet | pH5: can you tell me ho I look my LCCRR values ? I'm new to this. |
| 14:11.23 | magnet | I launched haret. |
| 14:12.20 | magnet | oh |
| 14:12.27 | rob-w_ba | magnet how is it going ? |
| 14:12.40 | magnet | using PD 0x44000000 24 , give me the same values as you. |
| 14:12.53 | magnet | rob-w_ba: not so bad.I solved all my toolchain troubles. |
| 14:13.12 | rob-w_ba | goody |
| 14:13.25 | magnet | it's compiling cross-gcc now. |
| 14:13.32 | magnet | I found out why it was SO slow. |
| 14:14.00 | magnet | first cache for bb don't seems to be used unless in interactive mode. |
| 14:14.07 | rob-w_ba | cool, how many boxes |
| 14:14.35 | pH5 | magnet: you're using bitbake/OE too? |
| 14:14.40 | magnet | then doing bluid a on the first time, is 10 time faster than build xanadux-xxxxxx |
| 14:14.44 | magnet | yep |
| 14:14.56 | magnet | s/bluild/build/g |
| 14:15.02 | pH5 | nice. |
| 14:15.14 | magnet | I start to enjoy it. |
| 14:17.50 | rob-w_ba | on 2 machines ? |
| 14:18.24 | magnet | just one. |
| 14:20.02 | rob-w_ba | ? |
| 14:20.46 | magnet | I use bitbake/oe on only one machine. |
| 16:05.00 | *** join/#htc-blueangel llange (n=llange@lange.nom.fr) |
| 16:07.46 | llange | Hello, I'm working on the HTC Himalaya. I'm trying to use the leds. I've insmod'ed the led modules. Now I'd like them to blink. Is there something I forgot other than echoing some bits to /sys/class/leds/[green|red|yellow] ? What would be the parameters for, say, a 1Hz blink rate on a led ? |
| 16:31.03 | Funklord | llange: how far has himalaya development progressed? |
| 16:31.34 | Funklord | llange: do you know of a "quick-start" url for himalaya? |
| 16:32.46 | llange | hard to say, I have no outside contacts (apart from this channel). To me, it seems that the development is frozen( at kernel 2.6.3 ) |
| 16:33.08 | llange | I'm trying to build a functional kernel 2.6.12, to synch it with blueangel development... |
| 16:33.20 | llange | But it's not very easy =) |
| 16:33.40 | Funklord | ok, my friend just got one, and he's got serious issues with wince, so I thought he should try it out |
| 16:33.40 | llange | So far, I boot the kernel with very little hardware support (only serial). |
| 16:34.11 | llange | He may start here : http://wiki.xda-developers.com/index.php?pagename=HimalayaProgress |
| 16:34.26 | llange | Or there : http://wiki.xda-developers.com/index.php?pagename=HimalayaLinuxBooting |
| 16:37.12 | llange | If he then wants to try with 2.6.12, he will have to get out of Xanadux's CVS http://cvs.sourceforge.net/viewcvs.py/xanadux/linux-2.6-xda/ the release that is tagged : K2-6-12-hh2-xda0-ba0 , and apply this patch: http://himalaya.lange.nom.fr/K2-6-12-hh2-xda0-ba0-himalaya.diff |
| 16:50.43 | Funklord | llange: do you know if himalaya uses the same interface cable as ba? |
| 16:51.40 | llange | I'm not sure, let me check this out. AFAIC, I modified the USB craddle to add a serial interface. |
| 16:53.05 | Funklord | llange: btw, I don't want to get your hopes up, he's a developer, but a really lazy bugger =P |
| 16:53.19 | magnet | llange: do you know if such modification is possible on qtek s100 ( magician ) ? |
| 16:54.23 | llange | funklord, magnet: check this out for the wiring on the Himalaya : http://wiki.xda-developers.com/index.php?pagename=Connectors |
| 16:55.09 | llange | magnet: It may be possible but I don't have any clue on the magician connector. |
| 16:55.44 | magnet | having serial port would a be a *great* help for installing linux. |
| 16:57.36 | llange | magnet: I agree, I couldn't do my debugging on the kernel right now without serial. )à |
| 16:57.55 | magnet | my usb connector is not like the one on the pic :\ |
| 16:58.00 | magnet | it's 'mini usb' |
| 16:58.02 | pH5 | magnet: i think that's impossible, the only pins we have on the outside are the usb ones |
| 16:58.17 | magnet | arg. |
| 16:58.19 | llange | Ah, yes, you may need to open it :-( |
| 16:58.32 | magnet | I'd better open it than fail :) |
| 16:59.27 | magnet | bitbake is still installing dependencies for xanadu-ba. :\ |
| 17:00.13 | pH5 | i somehow hoped that we could get it to work blindly. we "just" need a working pxa27x_udc, g_ether and initrd with ssh |
| 17:01.44 | magnet | I hope I won't need to open it too. |
| 17:02.53 | llange | does any one know how to "test" Leds, Keypad and touchscreen ? I only have a kernel + a shell, and just want a simple way to test that the hardware/driver is ok. for leds, I think there is something to do with /sys/class/leds/... , but need more advice =) |
| 17:36.17 | cp15 | llange: If you modified the blueangel_leds.c then yes, echo 1 >/sys/class/leds/green/brightness should turn on the green led |
| 17:36.38 | llange | Ok. It doesn't, so there is a bug somewhere =) |
| 17:37.30 | llange | It does write to 4 location in memory - asic3 but I need to debug. |
| 17:38.44 | cp15 | It should be quite similar to the blueangel, since I have taken the himalaya code |
| 17:38.57 | cp15 | But I didn't like the ipaq connection |
| 17:39.29 | llange | Yes, but I think the problem is in the asic3 initialization : I had to modify some values because it was freezing. |
| 17:41.01 | cp15 | Yes, the asic3 values are very critical. Best is to dump them under windows |
| 17:41.26 | llange | With Haret ? How do I do ? |
| 17:41.36 | cp15 | Just a moment |
| 17:41.39 | llange | thx |
| 17:43.26 | cp15 | According to http://wiki.xda-developers.com/index.php?pagename=HimalayaMemoryMap your asic3 is at 0x0d800000 |
| 17:43.51 | llange | yes |
| 17:44.02 | cp15 | Since the himalaya asic3 is word aligned, gpioa will be at +0x000, b at +0x100, c at +0x200 d at +0x300 |
| 17:44.16 | llange | ok |
| 17:45.22 | cp15 | Take a look at include/asm/hardware/ipaq-asci3.h for the register offsets within the gpio block |
| 17:46.17 | llange | I will. But then with all the addr info, should I chekc with Haret ? isn't there a virtual / physical translation to do ? |
| 17:46.46 | cp15 | Then do a pd <register address> 2 |
| 17:48.09 | cp15 | You can also dump the complete block IMO. Do a pd 0x0d800000 0x34 |
| 17:48.33 | cp15 | You have to skip 2 bytes every 2 bytes |
| 17:50.06 | cp15 | BTW: May I have your email-Adress? JuNuKN (sometimes here in the channel) has now also a himalaya and wants to run linux |
| 17:50.19 | cp15 | Maybe you can work together |
| 17:51.06 | llange | yes of course. use <my irc name> @ users dot sourceforge dot net |
| 17:51.21 | cp15 | Ok, thank you very much |
| 17:51.33 | llange | you're welcome. |
| 17:53.50 | cp15 | As for the hardware flashing there should be 2 files within the led directory: hw_cycle and hw_duty |
| 17:54.16 | cp15 | With hw_cycle you can set the cycle time (time between two flashes) 128=1 second, 256=2 seconds... |
| 17:54.36 | cp15 | With hw_duty you can set the on time (same scale as above) |
| 17:54.47 | cp15 | Maybe we should normalize it to milliseconds |
| 17:54.54 | llange | The files are there, but nothing happens. I'm trying to setup Haret connection but it's not done yet =) |
| 18:05.50 | cp15 | llange: I have seen you are a synce developer |
| 18:06.18 | cp15 | Can you tell me something about the synce protocol? Is it useful? Would it be hard to write a linux client for it? |
| 18:06.22 | llange | yes I created the project a few years ago |
| 18:06.48 | cp15 | Because you can wait for people crying: "I want my device to sync with outlook" |
| 18:07.35 | llange | It should be easy to do it, but then it depends on the linux-side applications. |
| 18:08.07 | cp15 | But I guess if it works with synce it will also work with windows? |
| 18:08.47 | llange | there are some layers. One for the connection ( PPP ) then there are network protocols on the top of it that are dealing with contacts, mail, etc... |
| 18:09.32 | llange | yes, synce is an ActiveSync clone for Linux, so you can expect that a device working with linux synce will also work with windows ActiveSync |
| 18:09.45 | cp15 | Ok, the ppp connection should be pretty easy. But we probably need a special serial usb driver, since the server side also needs a special driver |
| 18:11.21 | llange | yes, perharps a user-level driver could do. there is a full remote API ( RAPI ) dealing with browsing, registry, etc... that should be easy to implement. I'll probably look into this when my himalay device is up and running =)))) |
| 18:12.56 | cp15 | That would be good. I am currently trying opensync, but it isn't working too well |
| 18:13.35 | cp15 | If the sync process gets interrupted (happens about every 5th time), a full sync will be done next time, which duplicates every entry :-( |
| 18:14.42 | llange | =) You're further than I am =) |
| 18:16.56 | cp15 | I see another problem with nearly every sync software. Suppose the following case: I have a fully-featured addressbook (maybe outlook, evolution or what ever) which I want to sync to a handheld or phone |
| 18:17.25 | cp15 | So far, so good. But the problem is: The fully-featured address book has much more fields than the phone |
| 18:17.27 | llange | It' my situation. Outlook (not express) + HTC Typhoon. |
| 18:17.59 | llange | Yes, you're losing data. The 'master' DB is on the desktop, not the handheld / phone. |
| 18:18.34 | cp15 | No problem, as long as you change only data on one side. But if I, for example, change the phone number of a entry in the phone, the (limited) entry of the phone will get into the fully-featured address book, removing all entries the phone didn't support |
| 18:19.42 | cp15 | How does outlook solve this problem? |
| 18:20.14 | llange | it merges, not replaces. |
| 18:20.52 | cp15 | So it can make a difference between "field removed by user" and "field not supported"? |
| 18:20.54 | llange | it detects which fields are changes. There is an unique ID for each record. Then there is a timestamp (better check your clocks have the same times) =) |
| 18:21.31 | llange | Yes because, on the phone, the synchronization application exports the information : I only support this and this field(s). |
| 18:21.46 | llange | so the desktop can make the difference. |
| 18:21.57 | cp15 | Ok, this functionality is probably missing in opensync |
| 18:28.06 | cp15 | Do you know whether active sync supports three way sync? Following situation: I have an addressbook in the office, one in the car and one in the phone. I want to be able to sync each device with each other. Is this possible? |
| 18:30.31 | llange | The desktop application only supports 2 distinct devices (it's hardcoded that way but I don't see a reason apart from disk space). Idem, the device only support being synchronized with 2 desktops. |
| 18:30.31 | llange | On the desktop, you can choose wether your device or your desktop is the 'master' (in case of conflict after comparing timestamps), or even will ask you what to do. |
| 18:30.31 | llange | So I think this could be doable with some configuration (but may be risky =) ) |
| 18:30.36 | bhima | wince PDAs have an owner, or can be a guest. |
| 18:31.26 | bhima | ahh, llange knows much more. |
| 18:32.04 | *** join/#htc-blueangel oho (n=phm4@myrtle.ukc.ac.uk) |
| 18:32.49 | llange | however the best thing to do IMHO is over-the-air synchronization, with a central server. most smartphone allow it, however only with microsoft products (exchange) |
| 18:34.11 | cp15 | With the current gprs traffic costs, this isn't that desireable |
| 18:34.44 | cp15 | Does wince have a flag for a deleted entry? |
| 18:35.45 | cp15 | So the server can ask the mobile: Show me everything which changed since xxx and the mobile responds: entry 4 deleted, entry 5 changed, entry 6 new? |
| 18:36.05 | cp15 | This would make it quite efficient for synchronisation |
| 18:36.38 | llange | I don't know about the flag. In fact I'm not very active on the synce front since my iPaq broke a few years ago =) But it may be possible they do it that way. |
| 18:36.55 | cp15 | :-) Ok |
| 18:37.47 | cp15 | The current gpe synchronisation doesn't do it. The server asks "Show me all ids of all entries and ther modification date" and then compares them with the previous run |
| 18:38.24 | cp15 | It is already slow for local sync, since the device takes some time to get all ids and produces way too much traffic for gprs |
| 18:38.33 | llange | however I'll probably try to implement a synce personality for linux once my Himalaya is able to phone call... |
| 18:38.33 | llange | gpe is ? What about the trolltech synchronisation ? |
| 18:39.26 | cp15 | Haven't take a look at it recently, but I didn't like what I have seen from old zaurus days |
| 18:39.58 | cp15 | It is a xml-based file and synchronisation uses parts of rsync to get the changes |
| 18:40.40 | llange | Ah, ok. |
| 18:41.36 | cp15 | One reason why I prefer gpe: It is based on sqlite. It is quite easy to add/remove/modify/query data from other applications |
| 18:42.52 | cp15 | With qpe this was very hard. The application buffered the changes, so if you had your addressbook open and another application modified the xml file, the modification was most likely lost |
| 18:43.07 | llange | but then your applications on the device deal with the SQLite DB ? |
| 18:43.21 | *** join/#htc-blueangel magnet (n=magnet@vsg94-1-82-234-118-20.fbx.proxad.net) |
| 18:43.44 | cp15 | Yes, for example, i run a query on the addressbook database when a call comes in |
| 18:44.34 | llange | ok. but that means the existing applications must have their backend rewritten to be qpe-compatible ? |
| 18:45.14 | cp15 | gpe-compatible? gpe and qpe are very different |
| 18:45.48 | cp15 | qpe is based on framebuffer and a special qt library, gpe is based on X11 and gtk |
| 18:46.44 | llange | gpe sorry =) |
| 18:47.46 | cp15 | The only applications I currently use which interact with the addressbook are gomunicator (the phone...) and gpesyncd (the synchronisation application) |
| 18:50.55 | llange | ok |
| 19:02.15 | *** join/#htc-blueangel llang1 (n=llange@lange.nom.fr) |
| 19:07.41 | Funklord | cp15: do you know how different qt4 and qt-embedded are nowadays? |
| 19:08.20 | Funklord | I use qt4 regularly, haven't tried qt-emb yet |
| 19:10.04 | cp15 | Don't know. I have played with embedded qt a bit, but haven't done anything with qt4 |
| 19:10.22 | cp15 | Anyway I am not that much of a c++ fan |
| 19:11.39 | *** part/#htc-blueangel pH5 (n=ph5@p5485FE4C.dip.t-dialin.net) |
| 19:12.55 | Funklord | cp15: me neither, but I'm very fond of the qt ui style, and how fast it is |
| 19:13.21 | cp15 | Yeah, qt seems faster and leaner than gtk |
| 19:13.51 | Funklord | it's also easy to make use of in "corporate" environments, since there're those pretty pages, installers for windows and licensing |
| 19:14.25 | Funklord | I'd have a hard time talking my co-workers into using gtk+ |
| 19:14.52 | cp15 | :-) |
| 20:36.57 | *** join/#htc-blueangel rob_w (n=bob@p213.54.151.83.tisdip.tiscali.de) |
| 20:36.57 | *** mode/#htc-blueangel [+o rob_w] by ChanServ |
| 20:41.11 | rob_w | hi |
| 20:41.19 | magnet | hi. |
| 22:36.18 | *** join/#htc-blueangel llange (n=llange@lange.nom.fr) |
| 22:43.09 | cp15 | Will go to bed now |
| 22:43.11 | cp15 | Good night |
| 22:43.25 | llange | good night cp15 =) |
| 22:43.33 | llange | where do you live ? |
| 22:44.40 | cp15 | Germany, not too far from munich |
| 22:45.20 | llange | ok. I'm from France, near Paris. =) it's the beginning of the night there too =) |