**** BEGIN LOGGING AT Sun Feb 20 02:59:57 2011 Feb 20 03:18:17 Hi there. I recently (well, quite recently) got me a N900, since it is based on Debian (well, I am a Debian Maintainer for about 3 years) and I would like to hack my packages. Feb 20 03:27:02 rbrito: it's not based on Debian really. Feb 20 03:27:16 rbrito: also, Maemo is dead, and only 45% open source Feb 20 03:28:11 really? Feb 20 03:28:18 i thought much higher percentage :/ Feb 20 03:29:20 well, unlike earlier models, at least N900 has most of the drivers open Feb 20 03:29:25 only exception is 3D accel I think Feb 20 03:29:32 so … you can run Gentoo! :P Feb 20 03:29:35 or port Debian Feb 20 03:29:38 luke-jr: http://wiki.maemo.org/Port_an_existing_Debian_package this pretty much makes it a Debian based system isnt it Feb 20 03:30:00 chx: it's a dpkg/apt-based system, but it doesn't use Debian's packages Feb 20 03:30:20 well, Debian packages can be modified to run, no? Feb 20 03:30:21 due to its closed nature, it's also unmaintainable by anyone but Nokia, who have terminated it Feb 20 03:30:40 you can take Debian package sources, and build them for it Feb 20 03:33:47 Is there any way of cross-compiling things without putting "alien" programs on my system? I would like, as much as possible, to use what the distributions have so far, and, if not available, upload them to Debian's repository... Feb 20 03:37:50 luke-jr, Well, I bought it thinking that it were. :-( Feb 20 03:38:16 rbrito: somehow N900 gets misrepresented a lot, without Nokia actually false advertising :/ Feb 20 03:38:22 well, not just N900 Feb 20 03:38:24 all the earlier models too Feb 20 03:38:55 luke-jr, well, since I already bought it, I am willing to squeeze the most that I can from it. Feb 20 03:38:56 It the closest you can get with phones you can buy, though Feb 20 03:38:58 by far Feb 20 03:39:13 rbrito: want to use Gentoo with me? :P Feb 20 03:39:24 budfive: NITs are not phones, kthx Feb 20 03:39:36 what's NIT? Feb 20 03:39:58 nokia internet tablet Feb 20 03:40:05 luke-jr, I bought it for the main purpose of being a smartphone, not an internet tablet. :-) Feb 20 03:40:13 huh Feb 20 03:40:15 haha. ok Feb 20 03:40:23 rbrito: then you're likely to be VERY disappointed :P Feb 20 03:40:24 it's an internet tablet with a phone crudely bolted on Feb 20 03:40:35 I actually don't even know what *is* an internet tablet... Feb 20 03:40:41 it's a phone :) Feb 20 03:40:47 rbrito: all the earlier models don't even support GSM ;) Feb 20 03:41:06 luke-jr, yes, that I knew... But not much else. Feb 20 03:41:39 oh well, it works if you want a combination half-decent phone, half-decent camera, half-decent GPS :p Feb 20 03:41:44 Anyway, I am interested in writing some software for this thing. What would you people recommend for someone that already packages stuff for debian proper? Feb 20 03:41:54 but not if you want a proper smartphone, or quality camera, or useful GPS Feb 20 03:42:07 rbrito: just use Qt and it should work :p Feb 20 03:42:33 luke-jr, I am interested to know about the peculiarities of the packaging... Feb 20 03:42:37 rbrito: you can build packages in much the same way you'd build them for a real Debian box Feb 20 03:42:59 rbrito: download the sdk from nokia, and you're set Feb 20 03:43:01 budfive, what about the recommended cross-compilation environment? Feb 20 03:43:06 rbrito: the real problem is that Maemo is unmaintained, and someone needs to replace it Feb 20 03:43:13 rbrito: yes. get that and it's stock enough Feb 20 03:44:15 budfive, how much of the SDK is already (approximately) available as packages? Feb 20 03:44:44 rbrito: the sdk itself isn't a package as I recall. It uses scratchbox Feb 20 03:45:35 budfive, for instance, they say that I should download stuff using a script, but I see that Debian already has scratchbox. Is there anything different from Nokia's scratchbox from the Debian packaged version? Feb 20 03:45:59 I would like to avoid having as many packages as possible from places outside... Feb 20 03:46:33 rbrito: yeah, I know how you feel. As I recall using the scratchbox from debian directly didn't work for me. Let me check. It's been a little while Feb 20 03:47:00 budfive, Ah, thanks for that clarification. Feb 20 03:47:17 I see that Debian has scratchbox 2.0 packaged. Feb 20 03:47:26 (Actually, the version is 2.0-3.1). Feb 20 03:48:20 budfive, what's the basic procedure for generating a package, once I have the software written? Just prepare my package as usual and build it with scratchbox? Feb 20 03:48:30 rbrito: yes Feb 20 03:48:45 rbrito: the "sdk" is scratchbox + debs of the core system Feb 20 03:48:57 rbrito: so you build a package normally from inside scratchbox Feb 20 03:49:29 rbrito: I don't have it installed anymore, so I'm looking on the web. Can't find it for some reason Feb 20 03:49:30 budfive, great. I wish that the instructions were as clear as you... :-) Feb 20 03:51:33 rbrito: oh. I DO have it installed :) Feb 20 03:51:33 budfive, Does the SDK contain non-free software? Feb 20 03:51:42 yes Feb 20 03:51:54 I see that I would have to accept an EULA... :-( Feb 20 03:51:57 rbrito: as others have said, it uses apt, dpkg, but it's NOT debian Feb 20 03:52:00 yes Feb 20 03:52:04 it's not ideal Feb 20 03:52:06 I would like to avoid that, as much as possible. Feb 20 03:52:33 you'll have to forfeit much of the stock functionality then Feb 20 03:52:35 budfive, Oh, I thought that it was an exaggeration of not being Debian. :-) Feb 20 03:53:08 I see... I think that I will just download it, then and have a dirty system. :-( Feb 20 03:53:33 rbrito: it's not THAT dirty Feb 20 03:53:36 What do the non-free parts include? Anything essential for a basic "Hello, World" application? Feb 20 03:53:38 everything goes into /scratchbox Feb 20 03:53:51 so it bypasses the package manager, but you still know where everything is Feb 20 03:53:51 budfive, Well, less bad. Feb 20 03:53:54 yes Feb 20 03:54:00 highly unideal Feb 20 03:54:41 rbrito: the "normal" installation method of the sdk grabs a huge blob of packages from nokia. the blob contains some open and some not open things Feb 20 03:55:01 rbrito: the stuff you need for a "hello world" is open, but you'll have to get it all yourself Feb 20 03:55:19 rbrito: but you can always run stock debian or gentoo or whatever Feb 20 03:55:37 rbrito: you'll lose much of the phone-ness of the device, though Feb 20 03:56:13 budfive, I see... :-( In my general workflow, I usually grab stuff through an approx cache, so that I can install the software in many computers that I have here. Feb 20 03:56:38 The phone part for me is essential... Feb 20 03:56:56 rbrito: yeah. you really should just grab the sdk to see for yourself what's open and what isn't Feb 20 03:57:12 I guess that I will cope with things and just grab that SDK. Feb 20 03:57:32 supposedly you can run FSO,SHR on the n900 if you want. You'll have a phone, but not a very nice one Feb 20 03:57:32 Thanks for the clear answers that I could not find right from the sites, BTW. Feb 20 03:58:00 rbrito: yeah. It's not as open as could be, but it's by far the closest you can get without rolling everything yourself Feb 20 03:58:48 budfive, Yes, I saw that the other options were much worse regarding openness... :-( Feb 20 03:59:07 ever looked at the openmoko? :) Feb 20 03:59:11 And given that we have git repositories of some of the things, that's great. Feb 20 03:59:29 budfive, I tried to buy a Freerunner, but I couldn't find one here where I live. Feb 20 03:59:47 Oh, a silly question: what does osso stand for? :-) Feb 20 03:59:52 no idea Feb 20 04:00:35 I reached the same conclusion. :-) Feb 20 04:00:58 Open Source ...? Feb 20 04:01:02 :-) Feb 20 04:01:20 rbrito: I gotta go. good luck Feb 20 04:01:27 OK, just installed xserver-xephyr. Feb 20 04:01:37 budfive, thank you so very much for your clear answers. Feb 20 05:13:48 can somebody explain to me the two fork() in there please http://paste.debian.net/108251/ Feb 20 05:14:03 why two? Feb 20 05:15:48 my suspicion is in direction daemon, but I don't really see what it does rsp why it's needed Feb 20 05:18:01 I understand it forks and parent waits for child to exit. child forks again, then parent of that fork exits Feb 20 05:18:29 I think the goal there is to make sure init owns the process Feb 20 05:18:44 ooh, how does that happen? Feb 20 05:19:13 if I read that correctly, it causes it to be considered orphaned which causes an assignment to init Feb 20 05:19:29 yeah, that'S probably it Feb 20 05:19:32 normally, if you fork(), you have to wait() on it or it'll be stuck in the Z state Feb 20 05:19:33 thanks Feb 20 05:32:22 I missed to paste the setsid(); as last line. This makes it clear Feb 20 05:33:48 actually setsid(); close(0); close(1); close(2); do_something() Feb 20 05:38:51 yes, that'd make more sense Feb 20 05:39:58 thanks again Feb 20 05:43:30 just one more question: which #include imports function 'event' ? Feb 20 05:44:35 sched.h ? Feb 20 05:50:13 or is that a reserved word in c? Feb 20 05:52:26 (void)event; Feb 20 05:54:53 seems I can't find it in #include #include #include #include #include #include #include #include Feb 20 05:58:40 run only the preprocessor with 'gcc -E'. then look at the output to see where it came from Feb 20 06:00:01 hehe, yeah I'd do that if I were compiling this little progie, but I'm just reading source atm Feb 20 06:00:37 grepping my /usr/include give many matches :) Feb 20 06:00:45 tooo many Feb 20 06:00:58 already did that Feb 20 06:03:37 cscope says..... Feb 20 06:03:52 lots of various event variables, but no global "event" function Feb 20 06:04:11 you suspect it's a system thing? not X or gui or anything? Feb 20 06:10:10 void quit_loop(int event) { running = 0; (void)event; } Feb 20 06:10:26 signal-handler Feb 20 06:10:50 so I guess I know what it does, but I have no idea where it comes from Feb 20 06:11:26 DocScrutinizer: (void)event is a function call? Feb 20 06:11:39 looks like to me Feb 20 06:11:42 errrr Feb 20 06:11:45 sure? Feb 20 06:12:10 c or c++? Feb 20 06:12:12 well, not exactly a function, as I'd expect event(); then Feb 20 06:12:22 c Feb 20 06:12:56 could be a #define as well Feb 20 06:12:58 :-S Feb 20 06:13:10 not a function call Feb 20 06:13:47 it's valid C by itself, but (void)event; does nothing Feb 20 06:13:54 you need () to make it a function call Feb 20 06:14:10 it's most likely a call to scheduler to tell "I'm done" without really exiting the process Feb 20 06:14:11 if it was a macro, the "int event" argument would be expanded to something weird Feb 20 06:14:35 is the full source open? can I have a look if it is? Feb 20 06:14:55 oooooh Feb 20 06:15:07 it's a parameter - darn Feb 20 06:15:14 yah Feb 20 06:15:21 I should stop and take some chill time Feb 20 06:15:40 I can imagine some trickery where it's also a global something and the compiler can resolve the overload, but more likely it doesn't do anything Feb 20 06:16:29 it's probably the standard way to exit a signal-handler Feb 20 06:16:50 calback as parameter to a callback :-) Feb 20 06:16:59 (void)event doesn't do anything Feb 20 06:17:06 give that snippet to a compiler Feb 20 06:17:11 it may as well not be there Feb 20 06:17:19 ummm Feb 20 06:17:21 ah Feb 20 06:17:21 no Feb 20 06:17:23 i get it Feb 20 06:17:33 it's to prevent a "variable unused" warning Feb 20 06:17:34 ! Feb 20 06:17:42 hmm, wonder why http://maemo.org/downloads/product/Maemo5/fmms/ shows 992k downloads but adding "http://maemo.org/download-stats2/fetch.php?unixname=fmms&os=fremantle gives 1.1M Feb 20 06:17:49 wut? Feb 20 06:17:56 (void)event doesn't do anything Feb 20 06:18:11 but without it the compiler would warn you that the "event" variable (argument to the function) isn't used Feb 20 06:18:51 look at the -Wunused parameters to gcc Feb 20 06:23:18 wait a minute (/me reading man 2 signal) if ( signal(SIGTERM, quit_loop) == SIG_ERR ) Feb 20 06:23:36 http://paste.debian.net/108254/ Feb 20 06:27:52 dkogan@smart:/tmp$ gcc -Wextra -Wall -o tst.o -c tst.c Feb 20 06:27:52 dkogan@smart:/tmp$ md5sum tst.o Feb 20 06:27:52 e2659ec7813dd12c84f260f514ad06e7 tst.o Feb 20 06:27:52 dkogan@smart:/tmp$ perl -pi -e 's/$void$event//g' tst.c Feb 20 06:27:52 dkogan@smart:/tmp$ gcc -Wextra -Wall -o tst.o -c tst.c Feb 20 06:27:52 tst.c: In function ‘quit_loop’: Feb 20 06:27:52 tst.c:38: warning: unused parameter ‘event’ Feb 20 06:27:53 dkogan@smart:/tmp$ md5sum tst.o Feb 20 06:27:53 e2659ec7813dd12c84f260f514ad06e7 tst.o Feb 20 06:27:54 dkogan@smart:/tmp$ Feb 20 06:28:14 I compile your file. then I remove that line and compile it again Feb 20 06:28:26 both times the generated output is the same (identical md5 hashes) Feb 20 06:28:33 but without the (void)event, there's a warning Feb 20 06:30:47 heh, OK :-) Feb 20 06:30:53 thanks Feb 20 06:36:16 c is really friggin weird, you can write complete nonsense Feb 20 06:36:48 yep Feb 20 06:36:54 ask the compiler when in doubt :) Feb 20 06:37:26 also for this particular instance, __attribute__((unused)) is much more appropriate, and less obscure Feb 20 06:38:24 well, this seems to be a dirty hack anyway, even signal() is explicitly deprecated Feb 20 06:38:43 btw it's from bootmenu, if you wonder Feb 20 06:39:35 bootmenu itself is similarly weird, the examples make no proper use of the bootmenu.d/* logic Feb 20 06:39:40 sounds about right Feb 20 06:44:46 i.e. bootmenu.conf iterates over all the files in bootmenu.d/* to collect menu items. But the examples bootmenu.conf.N8x0.example which are supposed to go to this bootmenu.d/ directory don't use the index var $menu_count as in >>eval MENU_${menu_count}_NAME=\$ITEM_NAME<<, rather they use hardcoded MENU_2_NAME="Internal MMC card, partition 2, ext2" Feb 20 06:51:53 and backupmenu is using this bootmenu.d/backupmenu.conf to never return and instead drawing its own menu :-P Feb 20 06:52:24 an incredible mess Feb 20 06:55:17 and all this headache just to implement[2011-02-20 01:25:59] I mean, how hard can it be to quit the backup-menu process same why on pressing enter that it would when slide is closed DocScrutinizer, yeah. that would be nice Feb 20 06:57:14 at least I can understand now why robbiethe1st hasn't implemented this improvement-request yet :-D Feb 20 07:49:05 :D I love this phone. 13 web pages, 15 windows, and it's still fairly snappy. Feb 20 07:53:03 android? Feb 20 07:53:09 or wp7? Feb 20 07:54:07 Maemo, of course Feb 20 07:54:11 silly. Feb 20 08:27:30 hi Feb 20 08:27:46 hi Feb 20 08:28:14 i need some help with my nokia n900 Feb 20 08:28:20 who doesn't Feb 20 08:28:41 lol Feb 20 08:34:04 Nokia is publishing linux driver for moorestown: https://lkml.org/lkml/2011/1/5/31 --- Meego device not based on OMAP? Feb 20 08:34:34 Pr0n: what's up? Feb 20 08:34:50 trying to get bluemaemo working Feb 20 08:34:57 says i need some packages Feb 20 08:35:22 Devel's enabled, right? Feb 20 08:35:28 whats that Feb 20 08:35:37 probally not Feb 20 08:37:20 efl-core , python2.5-efl-core , python2.5-ecore , python2.5-edbus , python2.5-evas , python2.5-edje Feb 20 08:37:37 thats what it says are missing packages Feb 20 08:38:14 Doc: this may not be related to the upcoming device in any way Feb 20 08:38:31 Doc: Just to the work on Meego that they are supposedly doing with Intel Feb 20 08:44:00 wait... the upcoming Meego device is NOT based on an Intel chip? I thought that was Intel's whole angle Feb 20 08:44:26 upcoming meego device from whom? nokia? Feb 20 08:44:54 is there some other? Feb 20 08:45:07 I haven't been paying very close attention Feb 20 08:46:11 X-Fade: https://garage.maemo.org/pipermail/extras-cauldron-builds/2011-February/ there are some troubles with curl :/ i think it needs to be removed Feb 20 08:46:25 intel released their own tablet which is intel atom based, fujitsu will be releasing their netbook with meego, that's going to be atom based. nokia killed their own n9 which may run on atom processor but the devel work was ported to n900 as initial releases of n9 were arm based Feb 20 08:46:37 X-Fade: and no, i didn't reupload it Feb 20 08:46:38 err nokia killed their n9 hardware :p Feb 20 08:47:33 they killed the n9? The main announcement they made said they will be releasing a meego something this year. That's not assumed to be the n9? Feb 20 08:48:27 they do intend to release a meego device but nobody knows the ramifications, nokia signed a deal to be microsoft's bitch Feb 20 08:48:50 that's old news Feb 20 08:49:58 yeah. sure. Feb 20 08:53:00 https://garage.maemo.org/builder/fremantle/libnet-smtp-ssl-perl_1.01-2maemo1/i386.build.log.FAILED.txt any ideas what to do with it? Feb 20 09:53:37 Guys, which WebOS games are perfectly playable on n900? I.e. doesn't require multitouch, etc. NFSU is the first example/ Feb 20 10:26:32 divan: it's all in wiki Feb 20 10:28:16 Is it possible to buy the games without a WebOS device? Feb 20 10:28:39 yeah, torrents Feb 20 10:29:53 that's not the answer that I wanted to hear ;) Feb 20 10:30:54 well you asked for it :p Feb 20 10:33:21 that's true. Feb 20 10:36:16 I accidently removed the package mp-fremantle-pr, can someone please download it for me and send it? I will give you free google searches for the rest of your life Feb 20 10:37:01 free google search? lolwut Feb 20 10:37:17 :) Feb 20 10:38:53 http://maemo.org/packages/view/mp-fremantle-generic-pr/ Feb 20 10:40:02 how could I miss that? thank you! Feb 20 10:40:14 you can just apt-get it Feb 20 10:41:05 jacekowski: it's not that simple Feb 20 10:41:47 it actually is pretty simple Feb 20 10:42:15 so now where my free google search for my entire lifetime? ;P Feb 20 10:42:38 s/where/wheres/ Feb 20 10:42:38 psycho_oreos: here's your free search: htpp://www.gooogle.com Feb 20 10:42:38 psycho_oreos meant: so now wheres my free google search for my entire lifetime? ;P Feb 20 10:42:49 demute_, lol fail Feb 20 10:43:52 * psycho_oreos is so gonna bash.org that :p Feb 20 10:44:30 http://www.lmgtfy.com/?q=mp-fremantle-pr Feb 20 10:48:17 where the hell is the download button? ;) Feb 20 10:49:33 * psycho_oreos isn't going to bother answering that because he got ripped off ;p Feb 20 10:50:44 * demute_ feels stupid but hey come on, has anyone else tried to remove that package? (dont) Feb 20 10:52:07 * psycho_oreos has removed that package in favour of CSSU, hence mp-fremantle-community-pr Feb 20 11:15:53 hi there Feb 20 11:17:59 /home/user/MyDocs gets mounted readonly, this is set in /etc/fstab (-rodir). I tried deleting that part and rebooting, but the system restores the old fstab... Feb 20 11:19:13 phryk, perhaps it need an fsck Feb 20 11:20:42 phryk: look in /etc/event.d/rcS-late, that file overwrites fstab Feb 20 11:21:04 anybody here knows how wifi-switcher's popup is changed from interactive to passive ? Feb 20 11:21:38 i'm trying to do the same to ssh-switcher but it's the same command in both scripts wifi.sh and ssh-status.sh Feb 20 11:22:03 print_msg "SSH started" "OK" & print_msg "WLAN DISABLED" "OK" Feb 20 11:22:12 i thought it was here where it's modified Feb 20 11:29:16 Hi is the Community SSU stable to install ? Feb 20 11:29:27 i checked the Wiki , i was a bit confused Feb 20 11:30:11 xDaReaperx, I think the official answer is no, but I've installed it and have had no problems (that haven't been fixed). Feb 20 11:31:00 The initial version messed up some keyboard shortcuts like cntrl+shift+P = screengrab, but that was fixed a white ago. Feb 20 11:52:47 happy bday lcuk! \o/ Feb 20 12:14:21 * jonwil has run out of things worth reverse engineering Feb 20 12:16:52 thanks frals \o/ Feb 20 12:17:13 jonwil, then forward engineer some fixed and functionality into liqbase for me please :) Feb 20 12:17:30 what do you need? :P Feb 20 12:18:06 jonwil: What about the first level - simply replacing the little wifi widget with one that shows signal leve? Feb 20 12:18:16 jonwil, I need most of the stuff to work Feb 20 12:18:20 * lcuk has poor fingers Feb 20 12:18:35 I tried to reverse engineer the relavent bits of the wifi widget and got nowhere Feb 20 12:18:42 due to all the cruft that is libconnui Feb 20 12:19:07 jonwil, on MeeGo, there is a component called connman Feb 20 12:19:15 it does not support ad-hoc wifi Feb 20 12:19:20 but the lower levels do Feb 20 12:19:32 there has been an open bug about it since moblin days Feb 20 12:19:38 Thats on MeeGo, not Maemo. Not going touching MeeGo at all :) Feb 20 12:19:38 and brought into meego tracker Feb 20 12:19:54 thats a shame, lots of r&d happening there :) Feb 20 12:20:13 jonwil: ah Feb 20 12:21:13 I was planning to reverse engineer the dialer app (at least anything external it talks to that isn't a GUI call or a basic system function) but I never got anywhere because its far too complex Feb 20 12:21:33 even the Notification app was too complex Feb 20 12:21:46 :/ Feb 20 12:22:43 the connectivity UI is too hard to reverse engineer (i.e. the status bar widgets, control panels etc etc etc) Feb 20 12:23:12 Annoying. Feb 20 12:23:19 * lcuk does not like reverse engineering and will just try to build stuff Feb 20 12:23:49 I wonder what'd be involved in ripping out the connectivity stuff. Feb 20 12:23:50 Even the SMS libraries (libsms etc) proved too hard Feb 20 12:24:04 do you mean ripping out the CSD daemon and the cell stack Feb 20 12:24:09 or the ICD daemon and the internet bits Feb 20 12:24:10 The IP connectivity, and just using 'internet is up' and 'internet is down' hooks Feb 20 12:24:11 or both? Feb 20 12:24:21 I suppose also too involved? Feb 20 12:24:26 so you mean replacing the ICD daemon? Feb 20 12:24:31 That might be possible Feb 20 12:24:31 yes Feb 20 12:24:38 figuring out what talks to it Feb 20 12:25:10 So you can do stuff like have it connect through a VPN, and know it's connected, or tether backards, or ... Feb 20 12:25:33 Some of that could be done by writing new ICD network plugins AFAIK Feb 20 12:25:40 hmm Feb 20 12:27:30 The hardest part is going to be identifying what the icd daemon and its plugins expose to the upper layers Feb 20 12:27:36 jonwil: what's the status of cbsms? Feb 20 12:27:45 going nowhere Feb 20 12:28:01 see bug 8347 for the details of why Feb 20 12:28:02 Bug https://bugs.maemo.org/8347 Cell Broadcast Feature not available Feb 20 12:28:05 :/ Feb 20 12:29:18 Also see bug 10870 Feb 20 12:29:19 Bug https://bugs.maemo.org/10870 Open source libsms and libsms-utils Feb 20 12:31:32 so yeah replacing the ICD daemon is not at all easy Feb 20 12:32:15 The fact that the icd2-doc package is so great does help though Feb 20 12:32:55 I suspect the GPRS plugin would be one of the hard parts Feb 20 12:32:59 how can i use lxp injection driver on boot-up. I placed the module load script in /etc/event.d but lost the list of APs in the applet Feb 20 12:33:10 figuring out what it talks to in the CSD GPRS plugin Feb 20 12:33:16 and figuring out what it exposes to higher levels Feb 20 12:33:39 atm i have to manually load the wlan drivers after bootup Feb 20 12:34:24 and lxp drivers seem to give less choppiness for sip Feb 20 12:35:50 Perhaps i should load the drivers after the applet which enumerates the AP has loaded Feb 20 12:38:20 so yeah all the things I wanted to reverse engineer have hit brick walls Feb 20 12:38:38 Why? Feb 20 12:38:50 too complex to reverse engineer Feb 20 12:38:52 at least for me Feb 20 12:39:31 maybe an ARM Linux ASM guru might have more luck Feb 20 12:39:39 but I have given up on most of these things for now Feb 20 12:40:10 If http://wiki.maemo.org/index.php?title=Open_development/Why_the_closed_packages#List_of_outstanding_requests_that_are_still_relevant leads anywhere then maybe I will use anything comming out of that as a way forward Feb 20 12:40:18 to try and reverse engineer more stuff Feb 20 12:41:56 jonwil, you sure are optimistic... Feb 20 12:42:18 Someone was going to push that through to the right people at Nokia Feb 20 12:42:26 I forget who off the top of my head Feb 20 12:42:43 considering the great success this had in the past... Feb 20 12:42:49 jonwil: I thought you had it working Feb 20 12:43:04 nope, I dont have Cell Broadcast working Feb 20 12:43:13 I am able to dump the low level packets comming from the cellular modem Feb 20 12:43:22 and observe proper Cell Broadcast messages in those packets Feb 20 12:43:39 but there is a bug inside libsms/csd-sms that prevents things from getting through the cell stack Feb 20 12:43:57 jonwil: so what stops us from creating a cell broadcast displayer desktop widget for example? Feb 20 12:43:59 well getting through the stack properly Feb 20 12:44:23 the issue is that talking to the cell modem by bypassing the cell stack wont work Feb 20 12:45:53 so how did it work for you? Feb 20 12:46:06 I was dumping the data flow manually with tcpdump Feb 20 12:46:12 and getting a raw dump of data comming from the cell modem Feb 20 12:46:19 not an option for real-world use Feb 20 12:46:45 mhm Feb 20 12:46:48 Our only option is to work WITH the CSD daemon Feb 20 12:46:59 which means we either need Nokia to help Feb 20 12:47:11 or we need someone with guru ARM ASM skills to step forward and help reverse engineer stuff Feb 20 12:47:23 especially all the functions in libisi and elsewhere that need to be reverse engineered Feb 20 12:48:29 i.e. figuring out how to call them Feb 20 12:49:16 unfortunatly, ARM ASM gurus seem to be thin on the ground around here :P Feb 20 12:51:23 Also note that I do NOT intend to use any hacks to get this working such as binary patching the closed source library (or in-memory patching it) Feb 20 12:51:42 Those are hacks and are not the Right Way to do things :) Feb 20 12:54:31 jonwil: so find a Nokia employee who wants to leave the company because of "black friday" and worked on Maemo and has access to the code Feb 20 12:56:14 I dont particularly want to break the law either (and using code released by a Nokia employee without permission from Nokia breaks the law, copyright and probably also trade secrets also) Feb 20 12:59:15 truth be told that in cases like these, nokia needs a big slap across its face for pissing off the maemo community. Granted, they may file up complaints/lawsuits if their rights were demolished Feb 20 13:00:09 I think really all the arguments for not open sourcing a fair whack more of Maemo have vanished (except for the argument about the time taken to do so) Feb 20 13:00:13 * psycho_oreos thinks qwerty12 reverse engineered that fmtx thingy Feb 20 13:00:32 There are still legitimate reasons not to open source some parts of the system Feb 20 13:00:34 including BME Feb 20 13:01:02 the irony is that many owners are facing their devices soon to be out of warranty Feb 20 13:01:34 that and when things break without much hope, they are left inevitably to either sell or throw it away Feb 20 13:01:40 but all the reasons for not open sourcing, say, the cellular stack or the internet connectivity daemon or any number of other bits of middle-ware have all gone away now IMO Feb 20 13:02:49 Is there any way to restore deleted files or folders in maemo? Feb 20 13:03:15 there are testdisk and the like Feb 20 13:03:49 but i don't fine that package in devel Feb 20 13:03:54 any idea? Feb 20 13:04:38 if you haven't got it installed and the files you want to recover sits inside the rootfs or /home/user, your chances may become very minimal Feb 20 13:05:10 especially rootfs Feb 20 13:05:24 i deleted the folder espeak inside /opt which was being used by mstardict Feb 20 13:06:20 though i have now reinstalled all these mstardict,qstardict,espeak and espeak caller,i still cant use the mstardict Feb 20 13:06:57 The error being shown is failed to load espeak-data Feb 20 13:07:32 unfortunately /opt is bind to /home/user. You might be able to umount /home/user in theory and then dd that partition. Feb 20 13:08:00 so you've accidentally deleted the files and directories and you've tried to reinstall those? Feb 20 13:08:35 yeah,absolutely.. Feb 20 13:08:42 * psycho_oreos facepalms Feb 20 13:08:55 :P Feb 20 13:09:34 the thing is before i deleted the directory,the espeak installation was removed.. Feb 20 13:10:02 But still i cud see the direcftory inside opt so i tried to free up some space.. Feb 20 13:11:05 consider the chance of recovering deleted files very very slim Feb 20 13:11:11 :( Feb 20 13:11:25 read what file recovery is supposed to do Feb 20 13:11:37 but now i re installed all the utilities.. but no luck.. Feb 20 13:11:57 indeed, now go read what file recovery actually does Feb 20 13:12:28 i could see the folder /opt/espeak/data been created newly. but there is not phontab file/folder in it.. Feb 20 13:13:01 yeah will do that.. Feb 20 13:16:48 where can i find the working version of advanced power monitor? Feb 20 13:17:36 I find the one in repo is not working and it just is of 14kb size which seems to be weird.. Feb 20 13:19:21 its supposed to be working, if not then you can try reinstalling it which should bring the battery icon display up Feb 20 13:53:16 hello Feb 20 13:53:37 I need a little help to create a .deb package Feb 20 13:54:19 I've generate tar.gz using py2deb Feb 20 13:54:37 and installed sdk on my station (debian) Feb 20 13:55:41 what have I to do now to generate on my station a .deb from the files generated on my n900 ? Feb 20 13:59:00 Well. Moo. Feb 20 14:00:21 mm Feb 20 14:00:31