**** BEGIN LOGGING AT Fri Sep 07 02:59:56 2007 Sep 07 03:01:47 * dotmef is considering going to a mate's house to drink and do builds remotely Sep 07 03:02:37 Sounds like a plan! Sep 07 03:04:33 Yep - it's looking better and better now that another package just fell over :) Sep 07 03:39:21 hows the open moko community going now? Sep 07 04:07:56 So irritating... had my Neo on for >48 hours to work with it and now I found it dead, not being able to turn it on again - and it was connected to powered USB all this time, too... Sep 07 04:09:10 openmoko: 03laforge * r2937 10/trunk/src/target/kernel/patches/pcf50633.patch: increase default LCM backlight brightness at boot to max brightness Sep 07 04:11:25 It looks like its the paralel build processes that mess up the build Sep 07 04:18:21 gcb77 i had a similar problem before.. what fails and how? Sep 07 04:21:57 openmoko: 03laforge * r2938 10/trunk/src/target/u-boot/patches/ (3 files): re-sync with u-boot mainline Sep 07 04:46:03 morning :) Sep 07 04:51:15 too late for me Sep 07 04:51:21 :D Sep 07 04:52:26 ~ugt Sep 07 04:52:28 ugt is probably Universal Greeting Time. Created in #mipslinux, it is a rule that states that whenever somebody enters an IRC channel it is always morning, and it is always late when the person leaves. The local time of any other people in the channel, including the greeter, is irrelevant. http://www.total-knowledge.com/~ilya/mips/ugt.html Sep 07 04:52:46 Still, it's 06:52 in the morning for me ;) Sep 07 05:28:56 CM, apt, so for me is too late 7:20 AM :D according to the ugt, it's time for leaving... Sep 07 05:28:58 bye Sep 07 05:34:03 * Linux_Galore waves Sep 07 05:45:34 roh: icu-native fails with: /usr/lib/libstdc++.so.6: undefined reference to `_Unwind_GetIPInfo@GCC_4.2.0' Sep 07 05:45:57 I think its because my verion of gcc... I had 3.5 Sep 07 05:46:04 I'm trying it again with 4.1 Sep 07 05:52:16 3.5? Sep 07 05:53:04 heh, I meant 3.4.6 Sep 07 05:53:30 interesting to see what price it will fetch: http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&item=260157395019 Sep 07 05:54:18 The base set that was sold a few days ago went for $401 Sep 07 06:03:13 seems like a demand when used phone sells for higher price than at store (==paying for getting it faster) Sep 07 06:13:13 I wonder if these are someday going to become collectables Sep 07 06:15:23 Well, people collect all kinds of crap so I'm sure they will be ;) Sep 07 06:17:26 does anyone know why mplayer is no longer available in the official ipkg repositories ? Sep 07 06:17:37 or am I missing something? Sep 07 06:18:53 polz: I noticed that too. I thought it was in ScaredyCat's repo, not the official. Either way, its not there anymore Sep 07 06:19:21 not having mplayer makes the phone pretty much useless for me Sep 07 06:19:51 at least it used to be a half-decent media player Sep 07 06:45:23 oh no... Sep 07 06:45:40 webkit's been updated again Sep 07 06:52:29 ScaredyCat: Build fine for me 15 minutes ago :) Sep 07 06:52:40 :) Sep 07 06:53:08 *sigh*.. Built. I always mess up d and t Sep 07 06:53:10 15 minutes is a long time in oe.dev ;) Sep 07 06:53:13 Hehe Sep 07 06:56:26 good morning Sep 07 06:58:11 hello Sep 07 06:58:37 So, does OpenMoko stream out text messages (Received), over Dbus? Sep 07 06:58:44 and phone call notifications over Dbus? Sep 07 06:58:51 or libmokocore Sep 07 06:58:53 do that? Sep 07 06:59:06 and can I send out messages/phone calls via dbus? Sep 07 06:59:23 going to sleep bye guys Sep 07 07:00:07 and how can a sniff voice? Sep 07 07:03:09 CM: canyou talk to gsmd in newer releases? Sep 07 07:04:08 like is there any formal developer documentation? Sep 07 07:04:48 roflmao Sep 07 07:04:56 * ScaredyCat gets up off the floor Sep 07 07:05:21 Sargun: unfortunately not, other than the wiki Sep 07 07:05:48 although i do remember hearing about some dev guidelines... but ... Sep 07 07:06:36 Well, I want to write push e-mail Sep 07 07:06:39 for the phone Sep 07 07:07:36 so you would have a dbus listener for text messages Sep 07 07:07:46 and when you get a "special" "signed" messagfe Sep 07 07:07:48 message Sep 07 07:08:01 it would run emailProgram.fetch() Sep 07 07:08:01 mmmm... Sep 07 07:08:10 and display a notification Sep 07 07:08:33 ok, have you looked at funambol and their clients? Sep 07 07:08:55 because iirc that was going to be a push email client... Sep 07 07:09:47 and doesn't sms have a 'flash' param that 'causes' messeages to be displayed rather than just retrieved/ Sep 07 07:09:49 ? Sep 07 07:12:35 ScaredyCat: Haven't flashed the latest image to the phone, still using yours Sep 07 07:13:04 I'll be away for a work "conference" this weekend and like to be able to show working calls at least ;) Sep 07 07:13:08 newer ones seem worse for the stability Sep 07 07:13:13 *nod* Sep 07 07:13:30 i think i'll roll back... Sep 07 07:13:41 the image from the 2nd is a good one, right? Sep 07 07:13:55 I'll check Sep 07 07:14:15 20070903 Sep 07 07:14:36 k... Sep 07 07:14:37 ta Sep 07 07:14:54 ScaredyCat: http://claesmogren.com/~cm/neo/2007.2/ (In case you can't find them in your huge list of builds :P) Sep 07 07:15:09 hey! I deleted some! Sep 07 07:15:20 CM does gsmd work in that version? Sep 07 07:15:55 gcb77: Yes, for me at least. First image where the gui dialer worked out of the box for me Sep 07 07:16:11 I got lua + gtk2 working http://scap.linuxtogo.org/files/f6c5a179ed9b98bc36ac8307398a0913.png Sep 07 07:16:33 now, do I realy want to learn lua.. :P Sep 07 07:16:57 polz: ipkg update Sep 07 07:16:58 A friend of mine really likes it for in-game scriptig Sep 07 07:17:08 polz: my repo now has mplayer Sep 07 07:17:10 s/tig/ting/ Sep 07 07:17:10 CM meant: A friend of mine really likes it for in-game scripting Sep 07 07:17:31 yeah.. not really looked at the language in detail Sep 07 07:17:58 morning Sep 07 07:18:09 mplayer on the neo? Sep 07 07:18:22 yeah... Sep 07 07:18:30 mplayer is one of my favourite pieces of software...that'd make the neo uber x 50million Sep 07 07:18:41 I've not tested it at all, I've just built it and put it in my repo Sep 07 07:21:01 just tested it, it works Sep 07 07:21:29 I have a prepaid simcard, and there is a special number that when dialed displays the balance. how does that work, with gsm and stuffs? Sep 07 07:21:30 gta01:~$ mplayer /usr/share/openmoko/sounds/ringtone_goodnight.mp3 Sep 07 07:22:18 nice one Sep 07 07:23:03 Does openmoko still login by default as root? Sep 07 07:23:36 yes Sep 07 07:23:56 but you can add a user if you want Sep 07 07:24:32 Why is it running as root? Sep 07 07:24:38 Why doesn't it add a user itself Sep 07 07:24:48 (and login as that newly created user) Sep 07 07:25:12 because it's being developed... the software side is very immature Sep 07 07:25:33 ScaredyCat: what is your repo link? Sep 07 07:26:14 http://builthost.automated.it/OM2007.2/packages/ Sep 07 07:26:22 there'sa .conf in there... Sep 07 07:26:40 download it and put it in /etc/ipgk/ on your phone Sep 07 07:26:46 then ipkg update Sep 07 07:27:01 and it'll add the available package lists Sep 07 07:27:34 http://buildhost.automated.it/OM2007.2/packages/ Sep 07 07:27:47 * ScaredyCat catches the t/d disease from CM Sep 07 07:28:27 Hehe Sep 07 07:28:42 I tiddend mean do... :P Sep 07 07:29:19 titn'd Sep 07 07:29:39 besserwisser Sep 07 07:31:50 you're only eifersüchtig Sep 07 07:34:37 CM: how are you going to keep the battery charged? Sep 07 07:34:54 By not having it on much.. Sep 07 07:34:59 lol Sep 07 07:35:07 And taking out the battery once I've powered down. ;) Sep 07 07:35:35 aha! USSD Sep 07 07:35:50 is webkit-gtk currently borked/ Sep 07 07:35:53 ? Sep 07 07:36:06 No, I built it this morning Sep 07 07:36:12 hmm. Sep 07 07:37:29 erm Sep 07 07:38:29 webit built here ok too Sep 07 07:39:21 Ist die Katze aus dem Haus, tanzen die Mäuse auf dem Tisch. Sep 07 07:41:07 Hehe.. Sep 07 07:41:10 Genau Sep 07 07:41:59 woo hoo, finally, first succesfull build in a week Sep 07 07:43:41 | /home/j/hak/openmoko/build/tmp/work/armv4t-angstrom-linux-gnueabi/webkit-gtk-0.0+svnr25410-r25406-r1/temp/run.do_compile.14178: line 228: cd: /home/j/hak/openmoko/build/tmp/work/armv4t-angstrom-linux-gnueabi/webkit-gtk-0.0+svnr25410-r25406-r1//JavaScriptCore/pcre: No such file or directory Sep 07 07:47:38 torpor: When was the last time you restarted? Sep 07 07:47:46 restarted the machine or the build? Sep 07 07:47:54 The build Sep 07 07:48:04 yesterday Sep 07 07:48:06 Meaning cleaned out build/tmp Sep 07 07:48:11 Oh.. Hmm.. Sep 07 07:48:26 i try to do make updat && make openmoko-devel-image at least once a day Sep 07 07:48:36 oh you mean i should really clean out build/tmp? Sep 07 07:48:40 that'll take me days to catch up again. Sep 07 07:48:41 Same here, clean start on sunday eve Sep 07 07:49:35 I removed my build/tmp on sunday, and it's built every day since (except the uic3.pro thing and the binutils-cross) Sep 07 07:50:29 okay so what do i do .. just rm build/tmp and start again? Sep 07 07:50:41 is there anything i can leave in there that i don't need to rebuild/ Sep 07 07:57:17 rm -rfv build/tmp && make update && make all Sep 07 07:57:23 ^^^ off i go. see you all next week. Sep 07 07:59:59 :-/ Sep 07 08:10:02 (script) wiki RecentChanges: [[Wish_List]] [[Talk:Hardware:Neo1973:Alternate_Cases:General-purpose_buttons_%28side-mounted%29]] [[Hardware:Neo1973:Alternate_Cases:Square_case]] [[Hardware:Neo1973:Alternate_Cases:General-purpose_buttons_%28side-mounted%29]] [[OpenMoko_Local_Groups:_San_Francisco]] Sep 07 08:13:36 What's the definition of "qualified developers" wrgt Oct 20 GTA02v3 release? Sep 07 08:13:36 nano-: See http://wiki.openmoko.org/wiki/FAQ (and also SH1, SH1_FAQ, Neo1973, Developer_preview and Neo1973_Hardware#GTA01Bv4 pages) (Hopefully these links answers your question. This autotrigger for new nick and question during first 15 minutes done only once.) Sep 07 08:14:51 The FAQ's didn't answer the question. Sep 07 08:14:57 aloril is a bot Sep 07 08:15:01 cool Sep 07 08:15:44 The question is if my company should buy the Neo Advanced kit now, or if we can buy the October 20 release? Sep 07 08:16:27 hp__: aloril is a real person with scripts. Sep 07 08:16:49 nano-: if you can wait, I'd say wait Sep 07 08:17:08 But my question remains, what makes a developer qualified? Sep 07 08:17:16 nano-: AFAIK it means people that get offered them, otherwise you have to wait for the release. Sep 07 08:17:21 And yes, I think it will be a P0, P1, P2 with GTA02 too Sep 07 08:17:24 hads ahh okie Sep 07 08:17:49 nano-: Once GTA02 is in P1, anyone who pays can get one Sep 07 08:18:06 Talking about that, I just made this page in the wiki: http://wiki.openmoko.org/wiki/Community_nominees_for_GTA02_P0 Sep 07 08:19:04 Thanks. Sep 07 08:19:30 CM: Nice page idea. Sep 07 08:19:30 I thought it was a good idea. mickeyl said yesterday that there will be 50 P0 for gta02, but it's not the same people who got the gta01 for free. Sep 07 08:19:58 And it's not so easy for him(them) to see wich ones are helping out the most. :) Sep 07 08:19:58 nano-: where did you read about it? Sep 07 08:20:07 aloril: Read what? Sep 07 08:20:16 Oh the bot. Sep 07 08:20:35 * CM gets some coffee Sep 07 08:20:39 aloril: here Sep 07 08:20:39 nano-: I mean "October 20 release" Sep 07 08:21:04 aloril: http://wiki.openmoko.org/wiki/Neo1973 Sep 07 08:21:16 nano-, wait for DC20 Sep 07 08:21:25 yeah, so I've gathered. Sep 07 08:23:26 Does anyone know who you'd talk to about reselling P2? We run an online phone store here and would sell them. Sep 07 08:23:38 CM: Could they not just look at CVS(or is it svn? or bit bake, or or or) commit status? Sep 07 08:24:34 CM: I suppose their could be some exceptions, but as a generally rule I would have thought that would give a general impression. Sep 07 08:25:07 CM: Though, personally I think I'd favour developers working on core aspects of either OpenMoko, or may be OE. Though I could of course be talking out of me xxxx Sep 07 08:28:50 * ScaredyCat hugs CM Sep 07 08:31:49 hads: If you check the mailing lists, there was a discussion previously about this, and I believe a member of OpenMoko responded. I believe he would be a good point of contact. Sep 07 08:32:37 Thanks Wibbit, I'll take a look. Do you remember which list offhand? Sep 07 08:32:52 community list i think Sep 07 08:32:58 I'm seeing if I can find it now, but I would assume the community list Sep 07 08:34:11 Yeah, cheers. Sep 07 08:35:49 don't see any dates by mickeyl yesterday Sep 07 08:36:50 hads: Have a look at this thread Sep 07 08:36:55 hads: http://marc.info/?l=openmoko-community&m=118780656521518&w=2 Sep 07 08:37:51 Thanks mate, I'll take a look and have a hunt around too. Sep 07 08:40:31 hmm.. talk page answer it, its unofficial guess Sep 07 08:40:48 (except for GPS parts) Sep 07 08:40:49 Wibbit: Mickeyl already know all the people who commit to OE and help out that way. What I wanted with that page was to point out other people who also do useful work. :) Sep 07 08:41:03 (script) openmoko-community: Sean Moss-Pultz speeding up the flow of information Sep 07 08:41:25 Wibbit: So if he gives 30 of the GTA02's to those guys, there's still 20 more up for grabs. Sep 07 08:41:34 CM: ahh, so you mean people not directly associated with OE or OM? Sep 07 08:41:44 Yes Sep 07 08:42:02 Maybe that should be clearer in the text on that page.. Sep 07 08:42:03 CM: Okie :o) Sep 07 08:42:33 roh: congrats on your free ticket to the pirahna tank Sep 07 08:42:51 XorA: That means? :) Sep 07 08:42:58 CM: here :-D Sep 07 08:46:20 Hmm, can't find anything on the lists, I'll wait till closer to the time. Sep 07 08:48:05 hads: I think its a worthy question to post on community lists Sep 07 08:48:22 Erm Sep 07 08:49:02 hads: The second in the thread that I sent you seemed to make it very clear who you should contact Sep 07 08:49:08 hads: http://marc.info/?l=openmoko-community&m=118784193527931&w=2 Sep 07 08:49:38 Weiss: Oh, so it does, I missed that email address. Thanks. Sep 07 08:49:59 erm, Wibbit Sep 07 08:50:28 (it's Friday beer o'clock here, that's my excuse). Sep 07 08:50:34 LOL Sep 07 08:50:43 lucky sod Sep 07 08:50:45 it's 10am here Sep 07 08:50:56 9pm here :) Sep 07 08:51:07 Where are you? Sep 07 08:51:26 New Zealand Sep 07 08:51:47 okie Sep 07 08:51:49 Wibbit: And that's too early for beer? ;) Sep 07 08:51:50 Just curious Sep 07 08:51:50 * zash_se wants to move to nz Sep 07 08:51:58 Wibbit: in the UK? Sep 07 08:51:58 CM: It is for me :\ Sep 07 08:52:03 XorA: yeppers Sep 07 08:52:12 Wibbit: cool, another UK OM person Sep 07 08:52:56 I was sitting down looking at my Zaurus CL-3000 a few days ago, and I hope I make more use of the Neo, than I did that Sep 07 08:54:28 Wibbit: I use Cl-3200 a lot Sep 07 08:54:47 Wibbit: I think Ill run out of disk space on NEo :-( unless microSD cards get bigger quickly Sep 07 08:55:12 XorA: microSD should be hitting 8GB relatively soon. Sep 07 08:55:23 Wibbit: yeah I saw the new Tosh ones Sep 07 08:55:38 Xora: 4 Gb is mainstream now Sep 07 08:55:41 zash_se: It's nice here, you should :) Sep 07 08:55:43 60Euro's Sep 07 08:56:20 XorA: I'm considering selling my CL-3000, I think I'd have used it more if the wireless/bluetooth was built in Sep 07 08:56:36 XorA: having the wifi sticking out of the edge all the time, was just REALLY anoying Sep 07 08:56:43 Wibbit: didn't toshiba get 32 Gb on MicroSD ? i guess just proto.. but.. they are getting bigger Sep 07 08:57:10 hp__: Really no idea :) Sep 07 08:57:11 hp__: I've seen 16gb micro, but 32 only on regular sd Sep 07 08:57:20 Was Kingston btw Sep 07 08:58:15 hads: ah, a fellow kiwi :) Sep 07 08:58:15 Wibbit: they seem to fetch a good price on ebay, enough to afford NEO :-) Sep 07 08:58:51 bluelightning1: Hey. There's 3 or 4 around here somewhere. Sep 07 08:59:15 CM: Kingston micro's suck ass, mine failed after only a few months Sep 07 08:59:39 Ouch.. I've just used their ram, and that has worked well for me Sep 07 08:59:40 XorA: I may think about putting it up for sale, as i've had it for about 3 years, and not actually used it once Sep 07 09:09:45 I have a feeling 8GB is going to be skipped for 16GB Sep 07 09:29:30 Do we know if car chargers will be available for the Neo's? Sep 07 09:29:59 I hope so, or it will be useless as a gps Sep 07 09:30:08 Almost at least Sep 07 09:30:26 CM: yup, and a cradle Sep 07 09:30:51 The "snake" crade has been shown already, but I haven't seen a car charger Sep 07 09:31:15 do you have a link for it? Sep 07 09:31:20 Any car USB provider will work Sep 07 09:31:27 Any car USB charger will work Sep 07 09:31:37 as long as you get the power managment thingy setup right Sep 07 09:32:06 SpeedEvil: Exactly. I has to be able to draw 500mW from a "dumb" usb port Sep 07 09:32:09 The major problem is that the connector sucks for a cradle. Sep 07 09:32:41 There is absolutely no possible way you can make a 'just drop it in' cradle Sep 07 09:33:05 Though I'm contemplating making one using hardware mods. Sep 07 09:33:09 Wibbit: http://www.linuxdevices.com/files/misc/fic_neo1973_car_kit.jpg Sep 07 09:33:25 CM: Cheers, looks okay :) Sep 07 09:33:30 SpeedEvil: I see what you mean Sep 07 09:33:45 Well - absolutely may be a bit strong. Sep 07 09:33:55 SpeedEvil: Though I think having it horizontal would make more sense than vertical, in which case, you should be able to have a drop in. Sep 07 09:33:58 You could have a little robotic arm that plugs a USB lead in. Sep 07 09:34:02 you can't Sep 07 09:34:11 or may be as opposed to "drop in" have "slide in horizontally" :) Sep 07 09:34:14 the mini-ab connector requires something like +-0.5mm alignment Sep 07 09:34:22 or it will get broken Sep 07 09:34:39 and that can't really easily be done. Sep 07 09:34:40 Ähh.. Give it some wiggle-room.. ;) Sep 07 09:34:42 I don't understand. Sep 07 09:34:52 take the USB connector out. Sep 07 09:35:04 Now, move it horizontally or vertically by 0.5mm Sep 07 09:35:06 Wibbit: But I agree, I'd prefer to have it on the side too if used for maps Sep 07 09:35:14 nor try to push it back in without letting it move Sep 07 09:35:25 s/nor/now/ Sep 07 09:35:26 SpeedEvil meant: now try to push it back in without letting it move Sep 07 09:35:32 SpeedEvil: I think I'll have to wait till I get one :) Sep 07 09:35:35 Has anyone tried with an external antenna? Sep 07 09:35:49 you could make a mini usb to 'other' connector that was flush enough to have it there all the time Sep 07 09:36:24 ScaredyCat: indeed - that's one of the options I'm considering Sep 07 09:36:26 possibly usb + antenna tat's on the side Sep 07 09:36:31 that's Sep 07 09:36:42 is that gsm or gps ? Sep 07 09:36:51 gsm is hidden Sep 07 09:36:55 gps is on the side Sep 07 09:36:59 k Sep 07 09:37:14 If it was a custom made cradle, then surely one could have it so that it connects the appropriate USB and option of a GPS antena Sep 07 09:37:35 Wibbit: the problem is the physical connectors. Sep 07 09:38:03 You have to position the USB connector to within 0.5mm or so, otherwise when you plug it in at speed, it will snap. Sep 07 09:38:28 I do not understand how that can be the case... Sep 07 09:38:50 because normally when you plug something in, you don't force it. Sep 07 09:39:00 and this you have to force? Sep 07 09:39:09 Just don't make it in porcelain Sep 07 09:39:18 if you're not just moving the cable, but moving the whole phone, with the added mass of your arm, it will break. Sep 07 09:39:18 if you're dropping it in the cradle crooked, yes, it'll get forced Sep 07 09:39:32 what if.... Sep 07 09:39:34 hmmm Sep 07 09:39:39 That does sound good Sep 07 09:39:42 yes, it can work, if you always take 10s to put it in the cradle, really carefully. Sep 07 09:39:48 at the bottom of the cradle was a lever... Sep 07 09:39:59 sounds like a design floor Sep 07 09:40:06 when you put the phone in the cradle, it's weight pushes out the usb Sep 07 09:40:09 SpeedEvil: that said - i have a cradle for a usb bay thing, that uses a mini-B connector Sep 07 09:40:28 is the mini-AB much smaller? Sep 07 09:40:28 making the conneciton Sep 07 09:40:33 cjb: but I think you'll find that the thingy has better location features than the neo Sep 07 09:40:38 thus, drop in = connect Sep 07 09:40:41 the mini-ab is some 10mm*5mm Sep 07 09:40:47 8mm*5mm Sep 07 09:40:54 SpeedEvil: well, can the location features be improved for GTA02? Sep 07 09:41:04 Mini-AB? Mini-B? Sep 07 09:41:30 most phones have some grooves on them that are quite subtle, but when you drop them in a cradle it becomes clear they have a purpose Sep 07 09:41:38 yeah. Sep 07 09:41:48 ->cjb: Sep 07 09:41:56 cjb: and no Sep 07 09:42:03 pity. Sep 07 09:42:11 because that would involve case changes Sep 07 09:42:43 The solution that I came up with was basically to put the USB connector next to the GPS antenna. Sep 07 09:42:49 and there's umpty million of them made already? Sep 07 09:42:50 (this requries case changes) Sep 07 09:43:01 no - the case molds are expensive Sep 07 09:43:43 machine grooves into the cases :-) Sep 07 09:44:05 move the GPS antenna right by 5mm, redesign the GPS antenna board to be 5mm narrower, extend main PCB up to near end of case, and put a right-angled USB socket on it. Sep 07 09:44:41 This will slightly affect the GPS reception. Sep 07 09:44:49 but not noticably usually. Sep 07 09:45:05 s/notica/measura/ Sep 07 09:45:06 SpeedEvil meant: but not measurably usually. Sep 07 09:45:09 Honestly, this conversation I find very disturbing Sep 07 09:45:17 Why?] Sep 07 09:45:38 That the only method to charge the damn phone is as fragile as seems to be suggested? Sep 07 09:45:38 i can't remember the last phone i saw that wasn't designed for cradlig Sep 07 09:45:46 :/ Sep 07 09:45:55 Hmmm, let me think why I may find that disturbing... Sep 07 09:45:58 Is it really? Sep 07 09:46:00 * Wibbit ponders a bit Sep 07 09:46:09 I'm not saying a cradle can't be made. I'm saying it can't be really simple. Sep 07 09:46:16 I haven't noticed that it's fragile yet Sep 07 09:46:25 SpeedEvil: it's a pity the phone wasn't designed with cradling in mind. Sep 07 09:46:32 CM: That's the impression that I've been getting. Sep 07 09:46:36 rotate phone 90 degrees cradle is totally simple Sep 07 09:46:38 the connector is not fragile - if you are plugging a cable in carefully by hand. Sep 07 09:47:10 "An egg shell is not fragile, if you put it in the box carefully" Sep 07 09:47:11 Wibbit: SpeedEvil evil has a negative tone quite often, but that's just what he sounds like. ;) Sep 07 09:47:25 * CM pokes SpeedEvil jokingly Sep 07 09:47:33 heh heh double evil speedy Sep 07 09:47:34 * SpeedEvil stabs cm negatively. Sep 07 09:47:38 Hehe Sep 07 09:47:48 actually, SpeedEvil isn't really evil, he's just "Spee Devil" Sep 07 09:47:59 Negative stabbing. Hmm. Sep 07 09:48:10 All connectors have some alignment that's required so that they will safely self-align when you plug them in. Sep 07 09:48:12 CM: I had a HTC where the only way to charge the phone was via the usb port, and the charging mech snapped off the PCB, resulting in a useless piece of phone. I do not want the Neo to have a similar flaw. Sep 07 09:48:19 negative stabbing is done... with a high-powered vacuum cleaner? Sep 07 09:48:33 dimensionally. Sep 07 09:49:01 If you look on the nicer mobile phone connectors, they have features like interlocking cones, with the connectors down at the bottom Sep 07 09:49:07 Wibbit: I have no clue how well fastened the connector is to the pcb, but I hope it's welded on well Sep 07 09:49:25 the cones slide into each other and align well, before the actual connector parts meet. Sep 07 09:49:47 CM: I would like it to be welded, superglued, tied, riveted, screwed and shackled on, any thing less is sub standard :) Sep 07 09:49:55 It's just soldered on. Sep 07 09:50:18 is it surface-soldered or actually pinned? Sep 07 09:50:30 I _think_ it's just surface-mount Sep 07 09:50:51 surface soldered stuff that is manipulated with sideways force is... less than nice Sep 07 09:50:55 ideally the two halves of the case should clamp it as well for stability Sep 07 09:50:57 Indeed. Sep 07 09:51:09 They clamp it - slightly - vertically Sep 07 09:51:11 * Wibbit grumbles Sep 07 09:51:24 (I've resoldered my n-gage's power button once; to be fair, it was a drop that got it off) Sep 07 09:51:51 I'm contemplating selling a mod to give a nice external cradle. Sep 07 09:52:02 But it's annoyingly complex to do right. Sep 07 10:03:42 maybe one could mod it to charge by induction Sep 07 10:03:49 they do it on cheap toothbrushes even Sep 07 10:04:35 buz: How many watts can you push that way? Sep 07 10:05:03 a lot of the circuits are big enough Sep 07 10:05:10 i think there was even an electric car charged that way Sep 07 10:05:29 buz: I like that idea, it's snazzy Sep 07 10:05:46 buz: Here's a cooler use of that kind of charging: http://www.coroflot.com/public/individual_file.asp?sort_by=1&portfolio_id=592300&individual_id=153405 Sep 07 10:05:47 says 3.5W on my toothbrush charger Sep 07 10:05:55 Ok, nice :) Sep 07 10:06:38 * jeddy3 recieved 2.5mm 4P jackplugs today Sep 07 10:06:49 thats a neeeat lam Sep 07 10:06:50 p Sep 07 10:07:18 not sure how big a coil you need as "receiver" Sep 07 10:07:19 Yes, I want it ;) Sep 07 10:08:08 ideally you could abuse an rd antenna? Sep 07 10:08:19 CM: That's groovy Sep 07 10:08:52 can someone confirm: GTA02's will only be available to *selected* developers who qualify? Sep 07 10:09:04 coz i'm working feverishly on my neo apps, and i would love to get a GTA02 for them .. Sep 07 10:09:09 torpor: What do you mean about that? Sep 07 10:09:43 (script) wiki RecentChanges: [[Hardware:Neo1973:Alternate_Cases:Stylus_mount]] [[Community_nominees_for_GTA02_P0]] [[Hardware:Neo1973:Alternate_Cases:Camera]] [[Related_Hardware]] [[Wish_List_-_Hardware]] [[Neo1973]] [[Hardware:Neo1973:Alternate_Cases:General-purpose_buttons_%28side-mounted%29]] Sep 07 10:09:54 As far as I know, it will be the same as for GTA01. A P0 for free for a small selected group, a P1 when all devs who want an unfinished phone can buy, and P2 for public release Sep 07 10:09:55 torpor: The 0th pase of P2, after that it will be available just like GTA01 Sep 07 10:10:19 (plus from resellers too aparently). Sep 07 10:14:27 lol Sep 07 10:14:35 do you know what is a moco i spanish?= Sep 07 10:14:36 knoopx: See http://wiki.openmoko.org/wiki/FAQ (and also SH1, SH1_FAQ, Neo1973, Developer_preview and Neo1973_Hardware#GTA01Bv4 pages) (Hopefully these links answers your question. This autotrigger for new nick and question during first 15 minutes done only once.) Sep 07 10:14:36 xD Sep 07 10:15:41 well, did it answer you question? :) Sep 07 10:16:00 and yeah we all know, it's bogger or something Sep 07 10:21:12 mhm. Sep 07 10:21:13 strange. Sep 07 10:21:33 make results definitly not in an uboot image anymore... Sep 07 10:21:39 rwhitby: got an idea= Sep 07 10:21:42 ? Sep 07 10:22:42 CM: i just heard that the gta02 will only go to a select few developers at first - that we can't just buy them (like i did with the gta01 advanced kit) .. Sep 07 10:23:12 all i wanna know is when can i get a gta02 to test my apps as soon as possible .. Sep 07 10:23:39 and if there is a way to pierce the inner circle and get one sooner, say by releasing a fucking cool suite of apps (as is my plan) then i'd like to know about that too, and who,j etc. Sep 07 10:24:44 December looks like the first plausible time that you can buy one. Sep 07 10:24:55 :( Sep 07 10:24:58 ok Sep 07 10:25:10 i'll jsut do some releases of my apps then, in the meantime, for the gta01 crowd. Sep 07 10:25:15 Write something very cool in the next monthish, and you might get considered. Sep 07 10:25:22 what apps? Sep 07 10:25:34 i've got a little suite in the works Sep 07 10:25:39 (I don't decide who gets one) Sep 07 10:25:45 a game, a 'tool', a business app. Sep 07 10:25:58 3 music-making tools - sequencer, sampler, etc. Sep 07 10:26:12 and i'm currently working on some gadget_audio related stuff too .. Sep 07 10:26:27 but this build nonsense is getting in my way right now, i have to say. Sep 07 10:27:09 :/ Sep 07 10:31:37 s/bogger/booger/ Sep 07 10:31:38 jeddy3 meant: and yeah we all know, it's booger or something Sep 07 10:33:30 Write something very cool in the next monthish, and you might get considered. Sep 07 10:33:37 there should be a competition Sep 07 10:33:53 just imagine how that would get people writing apps Sep 07 10:38:38 i agree Sep 07 10:51:21 cb22: http://wiki.openmoko.org/wiki/Community_nominees_for_GTA02_P0 Sep 07 10:51:35 i saw Sep 07 10:51:39 I noticed you added SpeedEvil Sep 07 10:51:44 ya Sep 07 10:51:45 Good call :) Sep 07 10:51:54 think he deservies one :) Sep 07 10:52:06 I had added him too first, but decided that 4 was enough Sep 07 10:52:26 Otherwise it looks like it's my list only ;) Sep 07 10:52:33 hehe Sep 07 10:54:03 * CM reorders the list to be in alphabetic order Sep 07 10:55:22 * ScaredyCat slaps CM Sep 07 10:55:36 * ScaredyCat was in the mifddle of an edit Sep 07 10:55:51 :) Sep 07 10:56:00 * mmazur quit work Sep 07 10:56:05 Neo development, here I come! Sep 07 10:57:31 ScaredyCat: Oups, sorry! Sep 07 10:59:21 ScaredyCat: Btw, add yourself in "Additional votes for" instead of in "Nominated by". Or is that irrellevant, who nominated and who agreees? Sep 07 10:59:37 * ScaredyCat points at recent changes Sep 07 10:59:59 Ah, noticed :) Sep 07 11:00:01 CM: yeah, i wrote that browser a long time ago. there were a lot less standards to implement back then Sep 07 11:00:19 alp: Wow, you read a lot of backlog :D Sep 07 11:00:45 Anyway, impressive :) Sep 07 11:00:56 :-) Sep 07 11:01:11 Good slides too btw Sep 07 11:01:25 okay i'm gonna get my name on that list then.. Sep 07 11:02:05 what browser/slides? Sep 07 11:02:11 <-- doesn't read backlog Sep 07 11:02:21 torpor: ScaredyCat was in the mifddle of an edit Sep 07 11:02:28 Hmm.. Sorry Sep 07 11:02:35 torpor: http://www.atoker.com/blog/2007/09/05/webkitgtk-at-linuxconf-europe-2007/ Sep 07 11:03:11 torpor: And please get yourself nominated by us for your hard work on OpenMoko :D Sep 07 11:03:15 CM: thx cool! Sep 07 11:03:26 i'm doing some SVG hacking right now as it happens .. Sep 07 11:03:43 trying to work out the fastest route to SVG animations within a C lib similar to librsvg .. Sep 07 11:04:09 torpor: doesn't cairo have a svg backend? Sep 07 11:05:00 i don't suppose there is a link up somewhere for qemu with a newish 2007.2 build for windows? Sep 07 11:05:19 CM: cairo uses librsvg. Sep 07 11:06:12 ok :9 Sep 07 11:13:50 openmoko: 03thomas * r2939 10/trunk/src/target/OM-2007.2/applications/openmoko-dialer2/ (ChangeLog src/moko-dialer.c): Sep 07 11:13:50 openmoko: * src/moko-dialer.c: (moko_dialer_init):# Sep 07 11:13:50 openmoko: - Prevent critical warnings if the journal cannot be loaded. Sep 07 11:13:50 openmoko: - Exit if we cannot connect to gsmd Sep 07 11:14:31 alp i was loooking at the feedreader yesterday - is there a way yet to resize the fonts in webkit-gtk? all get_settings functions seem to return NULL or invalid structs... Sep 07 11:15:33 kristian-m: not available in the public api yet i suspect Sep 07 11:17:48 alp, to bad :-( Sep 07 11:18:32 alp okay, so I'll be patient then Sep 07 11:28:20 hallo Sep 07 11:30:09 PBeck: that has not been confirmed. Sep 07 11:30:21 SpeedEvil: ? Sep 07 11:30:27 * SpeedEvil was being silly. Sep 07 11:30:36 SpeedEvil: ?? :) Sep 07 11:36:36 kristian-m: will help if you file a bug, and implementing it would be no more than a dozen lines actually Sep 07 11:51:46 Do not make an ipkg upgrade at the moment.. otherwise you'll need to reflash, because X hangs and it doesn't boot. Sep 07 11:52:16 ouch Sep 07 12:11:03 (script) wiki RecentChanges: [[Community_nominees_for_GTA02_P0]] [[User:Alessandro]] [[User:Cb22]] Sep 07 12:52:54 Hi all :) Sep 07 12:53:25 hi there Sep 07 12:55:38 I can made pings to my Openmoko, but when I try to ssh on it I get: "ssh: connect to host 192.168.1.200 port 22: Connection refused", any pointer to a solution? Sep 07 12:56:18 is there an sshd running on it? :) Sep 07 12:56:20 I think the default IP is 192.168.0.202 Sep 07 12:56:43 so what have you made your local IP and subnet mask? Sep 07 12:58:26 * DPThought thinks the pohne needs either a dhcp server or zeroconf link-local ips and mdns. (Just my 2 Eurocents) Sep 07 12:58:30 Umm... Ok I will try with the default IP, but my computer address on intranet is 192.168.1.100 Sep 07 12:58:59 DHCP or static? Sep 07 12:59:08 offray_: are you connecting over usb? Sep 07 12:59:48 offray_: You need the 192.168.0.0/24 ip on your usb interface. This does not interfere with your other ip settings Sep 07 13:00:17 offray_: try. as root, "ifconfig usb0 192.168.0.200" Sep 07 13:00:31 offray_: do you get a usb0 interface? (check ifconfig -a) Sep 07 13:00:41 aeh, as dotmef said :) Sep 07 13:01:11 dotmef, DPThought, now is working, just using the default configs (I thought that I need to change it according to my network septup) Sep 07 13:01:32 w00t! :) Sep 07 13:01:41 hello everyone Sep 07 13:01:50 happy ssh'ing, offray_ Sep 07 13:01:58 dotmef: :) Sep 07 13:02:00 can any one help me to buy an openmoko Sep 07 13:02:17 thanks, now I'm inside of the machine... but no ls Sep 07 13:02:21 offray_: http://wiki.openmoko.org/wiki/USB_Networking There are the details. And some other tips you might want to try. Forwading, dns .... Sep 07 13:02:36 DPThought: Thanks :) Sep 07 13:03:25 je! Well now I will get sound working Sep 07 13:03:35 hm, at some *-native packages I get this error: does somebody know what's ld -lz ?: " /usr/bin/ld: cannot find -lz" Sep 07 13:04:07 apt-get install zlib-dev Sep 07 13:04:30 if he's using a Debian-based OS Sep 07 13:04:50 yum install zlib-devel :-) Sep 07 13:04:55 :D Sep 07 13:05:03 yum install apt Sep 07 13:05:08 i have a doubt.. can i use my debit card for on line purchase Sep 07 13:05:09 lol Sep 07 13:05:14 but my way is twp orders of magnitude faster Sep 07 13:05:41 smash: what kind of card? Sep 07 13:06:04 smash: and where? Sep 07 13:06:21 XorA, ok, I'm not debian-based, but the build-system is running in a debian chroot-environment :) Sep 07 13:06:41 thomasg_: wise choice :-) Sep 07 13:06:58 well, it was thought to avoid problems :) Sep 07 13:07:05 smash: I don't know if Neo1973 is still out of stock Sep 07 13:07:11 but in fact the problems are just others now :) Sep 07 13:07:31 just check about it, and if they're still selling just made the normal buy Sep 07 13:07:31 XorA , DP Sep 07 13:07:44 XorA it is visa card Sep 07 13:08:04 DPThought india Sep 07 13:08:09 well it should work Sep 07 13:08:47 smash: Visa Delta processes as a normal Visa Credit Sep 07 13:09:07 smash: my question was more of the sort. What are you talking about? Where or what fore does your card not work. Sep 07 13:09:21 s/fore/for/ Sep 07 13:09:21 DPThought meant: smash: my question was more of the sort. What are you talking about? Where or what for does your card not work. Sep 07 13:10:25 openmoko: 03thomas * r2940 10/trunk/src/target/OM-2007.2/libraries/libmokogsmd2/ (ChangeLog libmokogsmd/moko-gsmd-connection.h): * libmokogsmd/moko-gsmd-connection.h: Add missing Netreg type Sep 07 13:10:30 smash: i guess you talk about the openmoko.com shop? Sep 07 13:10:55 DPThought.. i am planning to buy an openmoko. i asked because they asks for a credit card number and mine is a debit card of an indian bank which supports Visa Sep 07 13:11:01 yess Sep 07 13:12:23 smash: to be fair. I have no idea. I was just woundering. Bu i had similar problems buying my Nokia770 back then. The card provider (mastercard in this case) blocked the transaction because it was a unusual amount of money for my profile which was transfered. (i was buying three of them...) Sep 07 13:12:23 smash: should work fine Sep 07 13:12:49 smash: and my incompetent bank was unable to fix that for weeks. :( Sep 07 13:12:50 smash: I have a VISA Debit in UK and never found a site that a) even knew it was a debit, b) refused it Sep 07 13:13:42 smash: debid makes no diference. Technically. But most shps don't tell you why your card was rejected. So it could be verything. Sep 07 13:14:46 smash: i don't know if someone running the moko shop is here. But they should be abel to do it manualle. Maybe you try sending them a mail asking why your card does not work. Sep 07 13:14:46 anyone know if there's a known issue with building webkit-gtk? Sep 07 13:14:53 many thanks.. XorA , DPThought i hope my card will work Sep 07 13:15:40 k.. DPThought.. i will ask them.. through mail Sep 07 13:16:23 smash: aeh, wait. did you try it already. It just sounded like you did not yet. Sep 07 13:17:42 in the second step for purchasing they asked for Credit card number Sep 07 13:18:18 smash: and the card was rejected? Well then go ahead and ask what the problem was ;) Sep 07 13:19:41 openmoko: 03thomas * r2941 10/trunk/src/target/OM-2007.2/libraries/libmokogsmd2/ (ChangeLog libmokogsmd/moko-gsmd-connection.c): Sep 07 13:19:41 openmoko: * libmokogsmd/moko-gsmd-connection.c: Sep 07 13:19:41 openmoko: (_moko_gsmd_connection_eventhandler): Remove work around for bug 454 as it has Sep 07 13:19:41 openmoko: now been fixed. Sep 07 13:19:46 not rejected. I thought i need a credit card so i quit and came here for your help Sep 07 13:19:56 dotmef: check build.webkit.org Sep 07 13:20:09 smash: debit is the same think just pre-loaded Sep 07 13:20:17 thing Sep 07 13:20:50 it is made to be a substitude that works everywhere Sep 07 13:28:09 DPThought: I'm not sure, but I've been told that only embossed cards (whether credit or debit) can be used for online shopping Sep 07 13:28:26 e.g. I was unable to get my visa electron card working with paypal Sep 07 13:29:59 mmp: Electron is a special case, shops dont like it as they are liable if they overspend your limit Sep 07 13:30:40 aha:) Sep 07 13:38:30 Hi all! I flashed rootfs 20070906. /etc/init.d/gsmd stop freeze the Neo Sep 07 13:39:42 Tyrael, http://bugzilla.openmoko.org/cgi-bin/bugzilla/show_bug.cgi?id=788 Sep 07 13:39:57 !ombug 788 Sep 07 13:39:58 * * Bug 788, Status: NEW, Created: Unknown Sep 07 13:39:59 * * mwester(AT)dls.net: Starting or stopping gsmd completely locks up the Neo Sep 07 13:40:00 * * http://bugzilla.openmoko.org/cgi-bin/bugzilla/show_bug.cgi?id=788 Sep 07 13:41:39 I add it to the wiki Sep 07 13:42:12 is it normal that the qemu openmoko is really slow? Sep 07 13:43:29 jupp Sep 07 13:43:53 ixs: as slow as the device itself? Sep 07 13:44:03 ixs: or slower? Sep 07 13:44:21 slower in general Sep 07 13:44:32 but right now, bootup on the device is not exactly fast... Sep 07 13:44:37 emulation is slow Sep 07 13:44:47 zash_se: thx Sep 07 13:45:05 ixs: I'm looking forward to the gta02, hope it will be out in october sometime Sep 07 13:45:24 * zash_se too Sep 07 13:45:47 well I think it is because you are emulation a MIPS CPU on a CISC one Sep 07 13:46:31 a RISC sorry Sep 07 13:46:49 well, I have a 2ghz X2, one would suppose that it is quite fast if it emulates 266mhz ;) Sep 07 13:47:04 Tyrael: any chance on installing openmoko on n770? Sep 07 13:49:37 edistar: yes, but RISC use small simple instruction that are computed very fast from the CPU (also more than 1 per CPU clock). RISC on the other way use bigger end complex instruction that also may require more than 1 cpu clock. (am I right?) Sep 07 13:50:42 edistar: yes, but RISC use small simple instruction that are computed very fast from the CPU (also more than 1 per CPU clock). CISC on the other way use bigger end complex instruction that also may require more than 1 cpu clock. (am I right?) Sep 07 13:50:50 ok Sep 07 13:50:56 have someone else got stuck on libsdl-net?, do_configure sanitycheck says that it probably uses some includes from host system Sep 07 13:53:54 * * OM Bug 805 has been created by thomas(AT)openedhand.com Sep 07 13:53:55 * * lgsm_get_signal_quality() defined but not implemented Sep 07 13:53:56 * * http://bugzilla.openmoko.org/cgi-bin/bugzilla/show_bug.cgi?id=805 Sep 07 13:57:04 is it possible to install openmoko on the nokia 770? Sep 07 13:57:17 See you folks Sep 07 13:57:26 edistar: probably Sep 07 13:57:40 zash_se: difficult? Sep 07 13:57:52 i dont know Sep 07 13:57:53 edistar, maybe it could be possible, but as nokias maemo is a closed plattform there might be some components that will not work Sep 07 13:58:00 thomasg_: hm Sep 07 13:58:11 thomasg_: cause I want to try out openmoko;) Sep 07 13:58:12 at least nokia uses some binary drivers Sep 07 13:58:22 edistar, get qemu :) Sep 07 13:58:27 thomasg_: very slow Sep 07 13:58:41 so you have the real openmoko experience :P Sep 07 13:58:47 thomasg_: haha Sep 07 14:00:49 why has the scrollwheel been abandoned? Sep 07 14:01:15 edistar: kinetic scrolling Sep 07 14:01:18 ala iphone Sep 07 14:01:32 zash_se: cool Sep 07 14:01:36 implemented yet? Sep 07 14:01:46 edistar: yes Sep 07 14:01:52 not bad Sep 07 14:03:32 kinetic scrolling sucks :( Sep 07 14:04:58 thomasg_: never tried it, why does it suck? ;) Sep 07 14:05:29 it's the most inefficient way of scrolling I can imagine Sep 07 14:05:33 eyecandy, but not usable (slow) Sep 07 14:05:38 inprecise? Sep 07 14:05:52 imho the only thing it ise usable for is thumb-muscle-training Sep 07 14:05:59 thomasg_: :P Sep 07 14:06:06 some people need it Sep 07 14:06:22 in the current state I'd scroll about 1 minute from a to z in my phonebook Sep 07 14:06:51 hm Sep 07 14:07:25 thomasg_, that's why you can hold your thumb at the bottom or top and it will continue scrolling Sep 07 14:07:28 we need the fast scroll are like in the iphone. (the intex at the right side) Sep 07 14:07:34 thomasg_, and also why there are filter and search options Sep 07 14:07:39 thos, so or so, it takes much time Sep 07 14:07:43 what about after scrolling from eg A to F it recognizes it and scrolls through the letters directly? Sep 07 14:07:48 s/are/area/ Sep 07 14:07:48 DPThought meant: we need the fast scroll area like in the iphone. (the intex at the right side) Sep 07 14:07:56 thos, so you mean I have to open my virtual keyboard to set a filter? Sep 07 14:08:25 thomasg_, no, the filter is based on groups, or you can use the keyboard to search for a particular name Sep 07 14:08:48 if my buildsystem would work I'd test a new image now :) Sep 07 14:09:02 * zash_se thinks openmoko + ipaq 610 would be nice Sep 07 14:09:04 apt, does the iphone have a screen area for slow and one for fast scrolling? Sep 07 14:09:11 thomasg_, what would you suggest instead of scrolling though? Sep 07 14:09:59 thos: this was my message. ;) Kind of. It has a alphabetical index at the right side like a scrollbar. If you tap a letter it jumps to it. Sep 07 14:10:22 aerg i meant thomasg_ Sep 07 14:10:42 * DPThought has too much coffee Sep 07 14:10:54 DPThought: moar coffee! Sep 07 14:10:58 * thomasg_ has to less coffee Sep 07 14:11:04 only 2 cups today Sep 07 14:11:06 less is more! Sep 07 14:11:14 or was it the other way around Sep 07 14:11:23 (script) wiki RecentChanges: [[Phase_1_Software_Testing]] [[User:Frihet]] Sep 07 14:12:42 just have a look at the apple demo videos. it's a simple and elegant solution. Sep 07 14:22:12 DPThought: true, just saw it Sep 07 14:26:03 is anyone else having trouble building openmoko-panel-gsm ? Sep 07 14:26:32 perhaps we need some 'autogrouping' alphabetical with latin letters only is not a solution for other charsets Sep 07 14:26:42 zecke: thanks Sep 07 14:26:58 dotmef: what is your build issue? Sep 07 14:27:01 like a hash table :) Sep 07 14:27:43 zecke - I think I've fixed it. someone placed a comment in the dependencies section of the bitbake recipe Sep 07 14:27:52 or, first unique chars that cover more than a screen.. Sep 07 14:28:55 polz, what's your error? Sep 07 14:29:48 | make[1]: *** No rule to make target `install'. Stop. Sep 07 14:29:54 zecko - the issue was that it was trying to package the files and couldn't find "webkit-gdklauncher-dbg.lock" Sep 07 14:30:08 s/zecko/zecke/ Sep 07 14:30:08 dotmef meant: zecke - the issue was that it was trying to package the files and couldn't find "webkit-gdklauncher-dbg.lock" Sep 07 14:31:02 zecke - but it looked like it hadn't built anything at all Sep 07 14:43:54 openmoko: 03thomas * r2942 10/trunk/src/target/OM-2007.2/applications/openmoko-dialer2/ (ChangeLog src/moko-contacts.c): Sep 07 14:43:54 openmoko: * src/moko-contacts.c: (moko_contacts_add_contact): Use fullname for Sep 07 14:43:54 openmoko: contact display name Sep 07 15:18:49 great, for some reason the image I just flashed doesn't load X Sep 07 15:19:59 cesarb - can you start it from a shell? Sep 07 15:20:43 dotmef: it drops me on a shell... and doesn't seem to load any daemons (so no ssh) Sep 07 15:21:21 ooops Sep 07 15:22:10 whoah. Sep 07 15:23:03 dotmef: I wonder if the build breakage I hand-cleaned-around yesterday wasn't the whole history... It's quite possible the whole build is borked Sep 07 15:23:31 cesarb - what runlevel does it say its at? Sep 07 15:23:43 cesarb - in which case, I'm heading for the same issue Sep 07 15:23:57 dotmef: typing runlevel is kinda hard without keyboard Sep 07 15:24:07 dotmef: let me check if I can see anything Sep 07 15:24:10 * cesarb boots again Sep 07 15:24:11 doh! >_< Sep 07 15:24:53 dotmef: doesn't even seen to have a runlevel Sep 07 15:25:09 dotmef: it's as if it couldn't find /sbin/init and tried the fallback /bin/sh Sep 07 15:25:31 dotmef: so you could, you know, fix the problem... after all, everything can be fixed if you can get a working shell ;-) Sep 07 15:25:50 * cesarb thinks both uboot and the linux kernel really need a touchscreen input method Sep 07 15:25:54 :D Sep 07 15:26:14 uboot-morse Sep 07 15:26:44 the lack of an init would explain why the "send the shutdown signal to init" gesture (hold power for some time) doesn't work Sep 07 15:28:52 * cesarb gives up and nukes the whole build Sep 07 15:29:25 * cesarb gets ready to wait two more days (losing the extra free day due to the holiday... sigh) until it finishes building Sep 07 15:30:42 * dotmef feels dread setting in that he'll have to nuke his build too Sep 07 15:34:35 anyone know how to get binutils to build on amd64? Clearly I need to understand OE better Sep 07 15:34:54 eggsy: http://bugs.openembedded.org/show_bug.cgi?id=2928 Sep 07 15:34:59 eggsy: get my patch there Sep 07 15:36:05 cesarb: thx - I'd looked on OM bugzilla not OE Sep 07 15:36:06 eggsy: but if after it you start getting errors in several packages (starting with libelf) which can be fired by cleaning the package, don't bother and nuke the whole build and start from scratch; it looks like it results in a broken image Sep 07 15:36:42 * cesarb is still at the "rm" step of nuking the build, which takes several long minutes Sep 07 15:36:54 cesarb: I saw you were having trouble Sep 07 15:37:47 cesarb: mv tmp; tmp.bak; bitbake openmoko-devel-image &; rm -rf tmp.bak Sep 07 15:37:56 missing out the excess ; Sep 07 15:38:53 OK, since it looks like it will be some time before I can build it to test, can someone with a working build which sees the same gsmd symptoms as I do (/tmp/gsm.log has back-to-back ATZ, AT+CFUN=1 and AT+COPS=0 with no response from the modem) try the patch I made at http://pastebin.ca/683328 to see if it makes any difference? Sep 07 15:39:13 It's supposed to be for bug 766, but I don't know if it will even compile Sep 07 15:39:41 cesarb: I can upload a build to website for you, built this morning Sep 07 15:40:20 XorA: just a normal build isn't enough, I would need someone who can build gsmd with that patch... Sep 07 15:40:34 XorA: (I know there are several places where I can get a "pure" build) Sep 07 15:40:37 cesarb: ah right, I only have 5 mins left at work, not enough time Sep 07 15:41:06 * cesarb is in a holiday... what a waste of a perfectly good holiday on build issues Sep 07 15:42:14 Have you considered ignoring the technical toys and looking at the countryside? Sep 07 15:42:18 :) Sep 07 15:42:45 SpeedEvil: maybe geocaching for a compromise... :) Sep 07 15:43:56 Or openstreetmapping. Sep 07 15:44:12 * spikebike perks up Sep 07 15:44:19 maemo mapper is one of my favorite apps Sep 07 15:44:31 Does anyone know of good trackers for J2ME or S40? Sep 07 15:44:33 definive Sep 07 15:44:56 * SpeedEvil has used gnuplot to show OSM. Sep 07 15:44:57 I'd like to do tracking for OSM, but couldn't find any software that would work on my mobile phone yet. Sep 07 15:45:02 well - awk + gnuplot Sep 07 15:45:16 +netcat Sep 07 15:47:05 did someone already port maemo mapper on moko? Sep 07 15:47:47 No. Sep 07 15:48:15 At the moment I'm simply using logfiles and postprocessing them. Sep 07 15:48:26 I haven't got a bluetooth thingy - I want to add that. Sep 07 15:48:51 Anyone interested have a look at http://wiki.openmoko.org/wiki/Openstreetmap Sep 07 15:49:54 i have one. trashed it only days after i bought it. I just neet do open very device i get. There is always the risk you might break a capacitor at the power supply and don't realise it. But thats the risk of beeing a nerd. Sep 07 15:50:28 re Sep 07 15:50:43 130Eur out of the window *g* Sep 07 15:51:47 what did you trash? Sep 07 15:52:07 a bluetooth gps mouse Sep 07 15:53:03 ah, i have one too many of those ;-) Sep 07 15:53:14 How did you trash it? Sep 07 15:54:21 break the case, bend a capacitor, put it all back together, insert battery... Sep 07 15:54:39 why not replace the capacitor? Sep 07 15:54:54 but the magic smoke is still in there. So maybe someone can fix it Sep 07 15:55:09 ;-) Sep 07 15:55:47 SpeedEvil: i did. either i choosed the wrong capacitance (or however this is written) or it caused secondary failures Sep 07 15:56:04 and now i lost the original... Sep 07 15:56:09 Ah Sep 07 15:56:10 DPThought: there's more to a capacitor than capacitance Sep 07 15:56:53 ESR, ESL, voltage, dielectric storage, ... Sep 07 15:58:56 SpeedEvil: well i don't think i have to match them all. Sep 07 15:59:07 True. Sep 07 15:59:10 DPThought: you don't have, except when you have to Sep 07 15:59:21 At least ESR, and voltage probably. Sep 07 15:59:30 DPThought: most capacitors you don't have; however, for some of them one or more of these parameters will be critical Sep 07 15:59:40 the guy in the shop told me, yep thats the same capacitor - sort of. :) Sep 07 16:00:56 i could not google the code on the original. Dont know what this was. But is was a simple ceramic capacitor for i belive 30V Sep 07 16:01:53 however. If someone wants to repair it, i give it to him Sep 07 16:03:19 Where are you DP? Sep 07 16:03:24 germany Sep 07 16:04:39 Hmm. I could do with a bluetooth GPS thing - to test bluetooth as much as GPS. Sep 07 16:05:11 give me your address *g* Sep 07 16:05:51 You have privmsg, thanks! Sep 07 16:10:42 (script) wiki RecentChanges: [[User:Mtfk]] [[FAQ/ru]] Sep 07 16:31:46 u_l-lap: poke Sep 07 16:33:13 cw666: morning Sep 07 16:35:39 u_l-lap: you dont happen to have a pppd debug log from your gprs session gone up w/ t-mobile handy do you? Sep 07 16:36:10 cw666: lemme see if pppd logs anywhere otyher that syslog Sep 07 16:36:39 cw666: it only logs to syslog, but I can try filtering the pppd lines out Sep 07 16:36:58 u_l-lap: you dont get output to stdout/stderr as well? Sep 07 16:37:20 i get something like: http://stupidest.org/tmp/om/ppplog1.txt Sep 07 16:37:48 so it never gives me nameservers... i wonder if you see that same? Sep 07 16:38:10 cw666: I run pon/poff from wmppp Sep 07 16:39:13 wmppp? Sep 07 16:39:55 windowmaker dock app for controlling and monitoring pppd Sep 07 16:41:07 unknownlamer.org/files/tmp/pppd.log Sep 07 16:41:18 err unknownlamer.org/files/tmp/pppd.log Sep 07 16:41:37 404 Sep 07 16:42:56 grumble my repasted was missing e too s/tmp/temp/ Sep 07 16:42:56 if it's not working replacing old default route to usb0 [192.168.0.200] seems the most suspect line Sep 07 16:43:02 I had to fiddle with my ethernet settings to have more than one connection up at once' Sep 07 16:43:47 u_l-lap: no, i have mine working, but it's not exactly what i would expect Sep 07 16:44:36 u_l-lap: ok, so you logs show you do get DNS servers pushed down, i wonder why i dont? Sep 07 16:45:15 u_l-lap: what APN are you using? Sep 07 16:47:54 An interesting difference is that u_l-lap is sending ms-dns* on the conf req, while cw666 isn't Sep 07 16:48:20 re Sep 07 16:48:36 If that isn't sent on the IPCP conf req, IIRC the other side doesn't tell Sep 07 16:49:16 So, cw666's pppd command line or default options file might be missing something... Sep 07 16:49:35 im going to use the values he has and check, though i tried a lot of different things Sep 07 16:49:42 (the log i posted is actually pretty old) Sep 07 16:50:33 cw666: you need the usepeerdns option (either on the pppd command line or in the options file) Sep 07 16:51:06 cw666: internet2.voicestream.com Sep 07 16:51:17 cesarb: no, that just frobs /etc/resolv.conf which i dont want anyhow Sep 07 16:51:27 u_l-lap: using guest w/ a blank password? Sep 07 16:51:35 cw666: yes Sep 07 16:51:57 cw666: well, blank user and password I'm assuming guest is the default when your username is blank Sep 07 16:52:50 cw666: in fact, it doesn't... it just passes the nameservers to the ip-up script, which is what actually frobs resolv.conf ;-) Sep 07 16:53:53 cesarb: oh, quite possibly... but anyhow, it shouldn't influence the IPCP layer, just act upon it Sep 07 17:01:46 cw666: the manual page says it does influence what's asked at the IPCP layer Sep 07 17:02:30 cesarb: ok, that would make me wrong, im testing some things now and ill test that shortly and let you know **** ENDING LOGGING AT Fri Sep 07 17:09:29 2007 **** BEGIN LOGGING AT Fri Sep 07 17:09:50 2007 Sep 07 17:11:16 cesarb - not unexpectedly, my latest build crashes out to shell like yours Sep 07 17:11:58 'night all Sep 07 17:30:10 cesarb: your patch seems to work. No other build errors (yet). In 4hrs hope to have a better image than you and dotmef Sep 07 17:33:37 <_diego_> hi, it's possible to run openmoko on the micro-sd, on neo1973 Sep 07 17:33:47 <_diego_> s/it's/is it Sep 07 17:34:20 _diego_: yes Sep 07 17:34:56 <_diego_> nice, have you tried it? Sep 07 17:35:17 no, but technically there is nothing to prevent it working Sep 07 17:35:57 _diego_: works fine for me Sep 07 17:36:18 <_diego_> fgau: do u have some instruction? Sep 07 17:36:23 http://wiki.openmoko.org/wiki/Booting_from_SD Sep 07 17:37:31 <_diego_> fgau: ok thanks, i'm gonna to try it Sep 07 18:10:23 (script) wiki RecentChanges: [[Manually_using_Bluetooth]] [[Community_nominees_for_GTA02_P0]] [[OpenMoko:Searching]] Sep 07 18:13:10 Question for anyone who might know: where the heck is the env variable HOME being set to / ? It's annoying me. Sep 07 18:13:37 In /etc/profile.d I have a script setting it to /home/root, but something subsequently is setting it back to / Sep 07 18:14:20 (Other things in that script ARE being picked up and stay set, so I know that it's being sourced at some point) Sep 07 18:20:37 Argh! Busybox grep has no -r! Sep 07 18:21:32 find . -type f -print0|xargs -0 grep /dev/null fish Sep 07 18:27:09 SpeedEvil: Thanks. I didn't even bother to think about "find" as my last experience with it on this platform led me to discover that busybox cpio has neither -p nor -0. Sep 07 18:32:29 ah. Sep 07 18:35:37 is anyone here using mokomakefile and having problems building openmoko-panel-gsm-0.1.0+svnr2936-r0 Sep 07 18:35:38 ? Sep 07 18:35:52 polz: yes Sep 07 18:36:20 can anyone suggest a workaround ? Sep 07 18:36:26 or even better - a fix ? Sep 07 18:36:37 it seems weird - the directory is actually empty Sep 07 18:36:42 what happened ? Sep 07 18:37:12 polz: my build crashed on it a few minutes ago. Did a make-update openmoko-dvel-image and am waiting for it to crash again :-) Sep 07 18:38:12 for me, it's been crashing for over a day now Sep 07 18:39:38 and it's not even in the bugzilla Sep 07 18:39:43 polz: Here I go task 2711 .... Sep 07 18:43:30 joao: and ? Sep 07 18:45:05 polz: and it crashed. I need to have a look at it but I'm not too good at this. The files directory is empty.I assume it should have contained a few png files Sep 07 18:45:26 pol: Maybe someone has forgotten to check them in. Sep 07 18:47:19 maybe, but how does one get around it now ? Sep 07 18:49:19 SpeedEvil: For the record, busybox xargs doesn't do -0 either. (But no matter, the filenames I'm looking at aren't weird enough to need it.) Sep 07 18:49:35 At least our busybox, that is. Sep 07 18:52:07 joao: I think there's a problem with mokomakefile Sep 07 18:52:16 not checking out the latest svn revision Sep 07 18:58:26 hmm, strange I have a mokomakefile build crasch too but it crashed in webkit-gtk-0.0+svnr25418-r25415-r1 Sep 07 18:58:54 saying that directory: /home/moko/build/tmp/work/armv4t-angstrom-linux-gnueabi/webkit-gtk-0.0+svnr25418-r25415-r1//JavaScriptCore/pcre isnt found Sep 07 18:59:07 anyone with the same prob ? Sep 07 18:59:17 ok, i've decided to get one neo, but in the case is not shipped until 15 i need to delete the order, is it possible? Sep 07 19:02:36 bedboi: post the question on the community mailing list Sep 07 19:03:00 barmeier: I think I ran into something similar a few days ago, but that just magically fixed itself :/ Sep 07 19:03:56 barmeier: a bunch of folks seem to be having that problem, myself included. Except mine was webkit-gtk-0.0+svnr25410-r25406-r1 or some such (last night). Sep 07 19:04:07 I don't have a solution. Sep 07 19:04:13 polz: I tried several days now, but it doesnt fix itself :( Sep 07 19:04:50 does anyone know if the missing directory is for output only may be I could create one Sep 07 19:05:17 hello there, does openmoko run j2me apps? Sep 07 19:05:28 I'm not sure why the ${SRCREV} variable apparently resolves as 25418-r25415 -- that seems weird to me. (If I'm reading the bb file correctly) Sep 07 19:05:35 I don't think there is a working j2me thingy Sep 07 19:05:58 quicksand: did your build problem disappeared or is it still there ? Sep 07 19:06:12 Still there, but I haven't checked yet today. Sep 07 19:06:32 SpeedEvil, kk just wondering Sep 07 19:06:42 quicksand: I have checked a few minutes ago, but it still crashes Sep 07 19:06:45 neo1973 looks cool btw Sep 07 19:07:14 polz: i've placed the order. Sep 07 19:07:18 polz: no rule to make install stopped... I think I'll look into it tomorrow bye Sep 07 19:07:28 Basically - 90% of the work is going into getting the next hardware version out. 9% is going into getting the software working, and .9% into other platforms and communication with users. (of core teams effort) Sep 07 19:08:31 yo yo yo, not an open moko question, but the neo as is stands is to be released in oct, is there a guesstimate on when? Sep 07 19:08:38 barmeier: OK. One thing I plan to try (but haven't yet) is upgrade bitbake from 1.8.8 to the latest svn head checkout. I don't know if that will make a difference. Sep 07 19:08:42 * mypapit out Sep 07 19:08:42 ColdFyre: no. Sep 07 19:10:10 quicksand: Ok, unfortunately I am not very experienced with bitbake :( Sep 07 19:11:21 barmeier: me neither! But I'm getting there. Just for comparison's sake, what distro is your build machine, and is it 32 or 64 bit? Sep 07 19:11:48 quicksand: 32Bit/Debian/testing/ Sep 07 19:12:22 Okay, I'm 64Bit Ubuntu Feisty. Not a 32/64 thing, as several other weird issues have been. Sep 07 19:13:49 quicksand: it looks like something is missing. When I create the missing folder, make complains about a missing dftables.c Sep 07 19:15:03 quicksand: when I make a make clean-package-webkit-gtk it says that webkit-gtk-0.0+svnr25419-r25415 was cleaned, but the versions are not the same ??? Sep 07 19:15:07 strange Sep 07 19:16:55 Yeah, weird. Hence my comment about ${SRCREV}. Something unusual is going on. Why are there two concatenated version numbers, just a few digits off from each other? Sep 07 19:18:58 I want to search the bug tracker for bitbake, just in case -- but http://developer.berlios.de/projects/bitbake/ seems to be unreachable for me ATM. Sep 07 19:19:02 root and brance number? Sep 07 19:19:14 s/e/h/ Sep 07 19:19:16 SpeedEvil meant: root and branch number? Sep 07 19:19:27 SpeedEvil: Could be, I guess. Sep 07 19:21:53 Ah, on the openmoko buildhost, it looks like the numbers are webcore/webkit. Sep 07 19:22:09 See URL: http://buildhost.openmoko.org/OM2007.2/tmp/work/armv4t-angstrom-linux-gnueabi/webkit-gtk-0.0+svnrwebcore-rwebkit-r1/temp/ Sep 07 19:24:54 But that prompts the subsidiary question: why does the buildhost show "webkit-gtk-0.0+svnrwebcore-rwebkit-r1" whereas barmeier and myself get "webkit-gtk-0.0+svnr#####-r#####-r1"? Sep 07 19:26:19 hi all Sep 07 19:26:42 Greets. Sep 07 19:27:39 Ghiottone: did I see in the logs yesterday you were talking about the free42 calculator package at some point? Did you get it built? Sep 07 19:28:05 yes Sep 07 19:28:09 it is in my repo Sep 07 19:28:26 Cool. It never occurred to me to put the ipk somewhere, so I'm glad you did. Sep 07 19:28:56 the url to add is in my userpage Sep 07 19:29:03 my username is Alessandro Sep 07 19:29:40 I can put in my repo also other packages that are not included in the official build. any suggestion? Sep 07 19:29:50 Firefox! Sep 07 19:29:55 http://wiki.openmoko.org/wiki/User:Alessandro Sep 07 19:30:00 I'm the one who made the recipe -- so I'm primarily interested whether you needed to tweak anything to get it to build. It was my first attempt at a bitbake recipe from scratch. Sep 07 19:30:12 oh Sep 07 19:30:30 I assembled the recipe from the file attached to the bug report in oe Sep 07 19:30:43 it has been easy Sep 07 19:31:05 Excellent. OE has a steep learning curve. Sep 07 19:33:33 I think it is a quite complex recipe Sep 07 19:33:38 very well done Sep 07 19:34:19 just the icon in the application list is not working Sep 07 19:34:29 Thanks. That represents several days of trial and error -- plus reading many .bb files and logs/scripts from failed builds! Sep 07 19:35:05 Have you tried the latest version? I updated it since the original upload to OE, and that is one of the things I (hopefully) fixed. Sep 07 19:36:24 The package name has changed -- "free42" was the original, and now it's "free42vga" because it will probably not work on non-VGA devices now. Sep 07 19:41:17 do we get any logs if the ui crashes? Sep 07 19:45:33 back Sep 07 19:45:45 just realized that Packages.gz on my repo was not correct Sep 07 19:46:19 quicksand, I did this yesterday. is there a more recent version? Sep 07 19:46:33 quicksand, anyway great app. on the neo it looks great indeed! Sep 07 19:47:26 As of yesterday, yes -- but I don't know whether earlier or later than your build. Sep 07 19:47:42 If the default skin filled less than the full screen on your Neo, then that's the old one. Sep 07 19:47:58 The new one has two VGA skins, one portrait (default) and one landscape. Sep 07 19:48:03 And the icon. ;) Sep 07 19:48:15 I just tried it again while showing it to my father Sep 07 19:48:27 and showed him both portrait and landscape version Sep 07 19:48:37 so I assume that the version is the newest :) Sep 07 19:48:51 and the icon in the application list is working as well Sep 07 19:48:55 Yup. I'll look into the icon issue. Sep 07 19:49:00 Oops, nevermind! Sep 07 19:49:09 I think it needed a reboot Sep 07 19:49:25 Awesome, glad it was easy for you. Sep 07 19:49:36 I'll be working on more cool apps soon. Sep 07 19:50:35 yeah. I started from simple bb recipe and oe manual and now I am pretty enjoying this packaging work :) Sep 07 19:50:44 (Free42 is an awesome program, and the author -- Thomas Okken, iirc -- deserves all the credit for it. I have used it for a couple years on other devices, and I'm thrilled it works so well on the Neo.) Sep 07 19:50:49 I'll be happy to help you in packaging them Sep 07 19:50:56 Right on. Thanks. Sep 07 19:51:26 I'll let you know if I get anything else together. Sep 07 19:51:37 if you add my repo to ipkg, you can enjoy the latest version of SettingsGUI (another great app made by kristian-m ) Sep 07 19:52:04 ok. just drop me an email via my userpage on the wiki Sep 07 19:52:09 i looking at the mailing lists right now and forms a picture that dialer/adress book/sms are very important to get done flawless. Is MMS also beeing developed? Sep 07 19:52:46 Will do. I haven't installed that yet, though I have been meaning to. Sep 07 19:53:13 install it to the card, or python packages will fill your root Sep 07 19:54:08 Heh, no problem there. I have a partition of my MicroSD mounted at /usr, so everything goes on the card! Sep 07 19:54:20 marsan: MMS is not on the core teams roadmap - MMS prevention is considered a feature Sep 07 19:54:31 quicksand, mmm nifty idea Sep 07 19:54:44 Writchie, ok thanks for that information. I was just checking. Sep 07 19:55:00 quicksand, but I prefer to use ipkgs features :) Sep 07 19:55:11 strange the numbers changed again now webkit-gtk-0.0+svnr25421-r25420-r1 maybe some one is working right now I will give up for today Sep 07 19:55:46 ghiottone, are you italian? Sep 07 19:55:56 yeah I have recompiled webkit-gtk hundreds of time and could never come to a full openmoko-devel-image build Sep 07 19:56:00 El_Salvador, yes Sep 07 19:56:24 El_Salvador, why? Sep 07 19:56:25 Ghiottone: Fair enough. I had installed AbiWord and Firefox, and all of a sudden my rootfs was full. So I took the easy way and just cloned /usr elsewhere. Sep 07 19:56:40 il mondo è piccolo Sep 07 19:56:44 (Firefox, by the way, doesn't work. Segfault.) Sep 07 19:56:47 hi Sep 07 19:56:48 El_Salvador, :) Sep 07 19:56:58 mickey|sports, here? Sep 07 19:57:04 :P Sep 07 19:57:06 hwo can i enable my internet connection in qemu? Sep 07 19:57:59 quicksand, I managed to compile and install firefox 1 but removed it because I filled root (at that time I was not using the card) without even trying it. Sep 07 19:58:15 quicksand, but I think I'll try again sooner or later Sep 07 19:58:49 Let me know if you can get it to work. It will probably be too slow/big to be useful, but it would be fun to try. Sep 07 19:59:02 i want to install some stuff and test, but i didn't get it working Sep 07 19:59:26 serenity: Never had any luck with that myself, sorry. Sep 07 19:59:40 hmm Sep 07 19:59:58 should be the same, than every other os in qemu, eh? Sep 07 20:01:28 serenity: You tried what's on the wiki? Sep 07 20:01:39 http://wiki.openmoko.org/wiki/OpenMoko_under_QEMU#Setting_up_USB_connection Sep 07 20:01:51 quicksand: will try, thx Sep 07 20:03:35 Hi all Sep 07 20:04:04 hi Sep 07 20:04:06 I'm trying to see how to store a small amount of data on the SIM. Sep 07 20:04:34 I've been playing with it, using libgsmd-tool in atcmd mode Sep 07 20:04:40 jseghers_: what kind of data and how much? Sep 07 20:05:07 writchie: 10-20 bytes Sep 07 20:05:30 writchie: identification data for a subscription Sep 07 20:05:37 usually you have about 100 short messages of storage. Sep 07 20:06:07 you can probably use a short message to store your arbitrary data. Sep 07 20:06:39 it may also fit in a phone book record Sep 07 20:06:41 Writchie: but wouldn't that also mean it could be seen through SMS User Interface? Sep 07 20:06:56 yes Sep 07 20:07:49 and phone book data would be seen in the phone book interface. Sep 07 20:08:23 I've been reading over the various GSM SIM documetation, and it seems like a separate DF could be created and an EF inside that... Sep 07 20:08:56 but I've been stumbling over how to navigate and/or create such Sep 07 20:09:10 http://www.google.de/trends?q=openmoko&ctab=0&geo=all&date=all&sort=0 Sep 07 20:09:46 jseghers_: i may depend a lot on the type of SIM card. Sep 07 20:10:04 i'm not sure you can create a new EF. Sep 07 20:10:23 (script) wiki RecentChanges: [[Community_nominees_for_GTA02_P0]] Sep 07 20:12:59 erm, I just did a bitbake openmoko-devel-image, but in my deploy folder is no u-boot (the uImage, zImage and rootfs are there -.-) could it be that openmoko-devel-image not longer contains u-boot? Sep 07 20:13:55 thomasg: no u-boot. u-boot build is borked at the moment and is temporarily removed from mokomakefile until it builds correctly Sep 07 20:14:10 Writchie, k, thx Sep 07 20:14:25 is the gsmd-cant-be-restartet bug fixed allrady? Sep 07 20:14:29 Hmmmm... Do you know what "class" should be used in the AT+CSIM commands? The documentation I can find suggests A0, for GSM... trying to get a Status value using that returns CME ERROR 3 which is a prohibited action. Trying 00 for the class gets me a status response... but I can't perform a SELECT... Sep 07 20:14:39 as it seems that my last u-boot update broke the connection to /dev/ttyACM0 - is there a working u-boot image somewhere? Sep 07 20:15:02 when I try to connect to ttyACM0 I get a random output without end (like cat /dev/urandom ^^) Sep 07 20:15:11 and after 2 minutes the neo turns off Sep 07 20:16:50 thomasg: I'm not sure anything better than the original u-boot image. Sep 07 20:17:05 Writchie, it's too late for this :) Sep 07 20:17:30 i think there is one later than the original which eliminated the need to erase Sep 07 20:17:46 but after that they may have gone downhill. Sep 07 20:18:02 is there a way to erase kernel and rootfs without using ACM? Sep 07 20:18:26 re Sep 07 20:18:40 hey hhf423 Sep 07 20:18:42 IMHO the u-boot images should be taken from some officially released and tested place Sep 07 20:19:07 not from random builds of on-the-fly random changes Sep 07 20:19:25 Writchie: i think it is not that critical if you own a debug board Sep 07 20:19:53 hi Ghiottone Sep 07 20:19:54 borg_: true and for as long as the debug board continues to work Sep 07 20:20:03 :D Sep 07 20:20:13 but for those without the consequences can be prettty bad Sep 07 20:20:36 i dot have one, so i will not update u-boot at all ;) Sep 07 20:20:53 borg_: good idea Sep 07 20:20:57 * hhf423 would be interested in the uboot with gsm passthrough Sep 07 20:21:09 hhf423: me too when it works Sep 07 20:21:20 in fact that's why I'm looking for u-boot in the build Sep 07 20:21:30 but without debug board ... Sep 07 20:21:58 i have a debug board and when the round of changes to u-boot is done and i have tested them i'll let you know. Sep 07 20:22:20 Writchie: ah, thanks. Sep 07 20:22:24 similar things could be set for dfutil Sep 07 20:22:36 dfu-util or whatever Sep 07 20:22:42 Writchie, so you have an idea where to get a tested and working u-boot image? Sep 07 20:22:50 why is cupsd running on the neo?? Sep 07 20:22:58 a properly tested uboot would be nice Sep 07 20:23:17 rwhitby: wil add it to mokomakefile as soon as it is building again for him Sep 07 20:23:51 as soon as its added I will test the new modem access. Sep 07 20:24:27 i'll post a working and tested image somewhere if there is not already one that matches the build from source. Sep 07 20:27:23 Writchie: excellent Sep 07 20:29:23 hhf423: i think that at least u-boot and dfu-util need to be under some kind of release control because they are critical to those without debug boards Sep 07 20:32:57 How is the openmoko working out? Sep 07 20:39:35 TheNetGoD: it is very fit Sep 07 20:39:42 up in the gym Sep 07 20:40:03 how long does the battery last when using the gps? Sep 07 20:41:47 is yaff being considered as a replacement for jffs? Sep 07 20:44:08 Has anyone had any problems getting networks to activate the openmoko? Sep 07 20:45:02 usb networking works out of the box :) Sep 07 20:45:10 TheNetGoD: networks? gsm, usb or bluetooth ? Sep 07 20:45:23 you mean GSM providers? People have had trouble with AT&T's some SIMs Sep 07 20:45:29 heh, yes Sep 07 20:45:36 with gsm I had no problem with Wind in italy Sep 07 20:45:48 no problems for me with two Finnish providers Sep 07 20:45:58 Writchie: dfuutil is only critical if you flash uboot ;-) Sep 07 20:46:10 att sucks anways. worst company evar.. What about tmobile? Sep 07 20:46:18 should work Sep 07 20:46:36 as t-mobile uses the 1900 mhz band, I have not heard of problems with them Sep 07 20:46:37 hhf423: exactly Sep 07 20:47:01 my eplus sim does not work, tmobile no issue Sep 07 20:47:16 hhf423, germany? Sep 07 20:47:20 is the openmoko ever going to come out with a cdma version? Sep 07 20:47:29 but i probably only need a new eplus sim Sep 07 20:47:44 thomasg: yes, germany Sep 07 20:47:46 hhf423, can you take a look on the serial number of your sim? there's a r03 for usims and a r02 for normal sims Sep 07 20:49:01 thomasg: there is no r on it, its a umts capable card, that i know Sep 07 20:49:04 afaik you can also see the difference on the chip itself: there are 6 big sections on normal sims, and 6 big sections in mid circled by some other sections around it for usims Sep 07 20:49:38 hhf423, oops, my fault: its n3 for usim and n2 for sim Sep 07 20:50:07 there is no n either Sep 07 20:50:23 on my t-mobile cards there are such revision numbers Sep 07 20:53:44 thomasg: interestingly enough, no issues with a brand new blau.de sim, which is also eplus, so i'll just get myself a new card Sep 07 20:55:28 is the uboot from buildhost (august 19th) safe? Sep 07 20:56:37 the august 19 is save i think. i used it, no issues Sep 07 21:00:28 * kristian-m also has one from eplus (base) - works perfectly here Sep 07 21:00:56 there is no r{2,3} though Sep 07 21:01:46 aehm n2/n3 Sep 07 21:02:13 i have one from t-mobile and one from simyo Sep 07 21:02:18 both work fine Sep 07 21:02:22 kristian-m: can it do umts? Sep 07 21:02:23 and simyo is e-plus too Sep 07 21:02:30 hhf423 yes Sep 07 21:02:42 how old ca? Sep 07 21:02:45 hhf423 bought it a months ago Sep 07 21:02:57 hm, I can't connect to dev/ttyACM0 anymore... Sep 07 21:02:58 i bought the simyo card for the neo, its 2 weeks old Sep 07 21:03:02 and it can do utms Sep 07 21:03:03 I always get this nasty output Sep 07 21:03:31 ah, mine is at least a year old, so a new one should fix it Sep 07 21:03:34 s/utms/umts/ Sep 07 21:03:35 borg_ meant: and it can do umts Sep 07 21:03:56 thomasg: nasty output? Sep 07 21:04:07 hhf423: look at the gprs pricing of simyo, it is really by far the best Sep 07 21:04:13 hhf423, do cat /dev/urandom and you'll know what I mean :) Sep 07 21:04:22 that's how it looks when I connect to the neo via ACM Sep 07 21:04:22 you should get this if you want to surf with the neo Sep 07 21:04:29 thomasg: there is no way you can tell the difference between a 2G and 3G SIM based on the contacts. Sep 07 21:04:43 eggsy, of course not Sep 07 21:05:04 erm Sep 07 21:05:07 you said contacts Sep 07 21:05:11 I meant contracts :) Sep 07 21:05:23 borg_: better than blau.de? Sep 07 21:05:25 I'm not sure about this, but my newer sims have more contacts than the old ones Sep 07 21:05:41 hhf423: its 24cent/mb Sep 07 21:06:07 24 Cent/MB Datennutzung per GPRS oder UMTS (z.B. Internet oder WAP) Sep 07 21:06:11 ah, blau.de is the same Sep 07 21:06:21 mmh.. all the prices are the same Sep 07 21:06:27 think so, yes Sep 07 21:06:27 is it the same company? Sep 07 21:06:35 dunno Sep 07 21:06:43 hmm, I'm using cdc acm in kernel 2.6.22, could this be the problem for the weird output when I connect to the neo via acm? Sep 07 21:06:44 looks like this Sep 07 21:07:00 thomasg: do you hac Sep 07 21:07:20 hac? Sep 07 21:07:24 thomasg: no they have always had 8 but until recently the bottom 2 were NC (not connected) and some module packagers didn't put contacts there. A lot or Orga SIMs are like that Sep 07 21:07:31 have a rights problem or get just crap? Sep 07 21:07:41 no, I get just crap Sep 07 21:08:07 thomasg: and you want to talk to uboot? Sep 07 21:08:08 some times (with luck) I can see the terminals first line, and I know I'm connected to the neo Sep 07 21:08:19 hhf423, I just updated u-boot, this doesn't seem to be the problem Sep 07 21:08:31 *downgraded, not updated Sep 07 21:08:56 thomasg: did you unplug the neo before booting into uboot? Sep 07 21:11:23 no, not this time Sep 07 21:11:25 I'll try :) Sep 07 21:11:49 do that Sep 07 21:11:52 ok, it works :) Sep 07 21:11:53 thanks! Sep 07 21:12:11 np, had the same issue once Sep 07 21:12:23 that's the reason why I had no problems in past: I every time had to unplug, because the neo hasn't started the u-boot menu when plugged in Sep 07 21:12:49 i thik it says in the wiki to unplug, too Sep 07 21:23:43 what does <5>jffs2_scan_dirent_node(): Node CRC failed on node at 0x03caf3f0: Read 0xffffffff, calculated 0xf8825d50 mean? Sep 07 21:28:04 Ghiottone: let me guess Sep 07 21:28:20 Ghiottone: reading 0xff is what happens when you read unwritten space Sep 07 21:28:36 Ghiottone: so, it's trying to read a partially written page Sep 07 21:30:40 cesarb, so it is probably calling for a reflashing you mean? Sep 07 21:30:58 cesarb, I think it is written while checking rootfs Sep 07 21:32:43 Ghiottone: it could mean you pulled the "plug" while the kernel was writing a block Sep 07 21:33:21 Ghiottone: yeah, I'd expect that message to be written while the kernel is scanning the rootfs to build its memory map of the filesystem's contents Sep 07 21:33:54 the starting sound just rocks :) Sep 07 21:41:19 kristian-m, gprs connection is not working :( Sep 07 21:42:50 hm, gsm is not working at all for me :( Sep 07 21:43:20 thomasg, you really like the startup sound? you obviously don't have to listen to it multiple times a day Sep 07 21:43:43 hehe Sep 07 21:43:48 it's really freaky Sep 07 21:43:50 it's like pins in my ears each time Sep 07 21:46:49 * HistoryOnTheRoad wonders what "pins in the ears" actually sound like Sep 07 21:47:24 evenin' Sep 07 21:49:52 HistoryOnTheRoad, I imagine it sounds similar to the OpenMoko startup sound ;-) Sep 07 21:50:16 lol ... did they change the sounds from .1 to .2? Sep 07 21:50:23 no Sep 07 21:50:34 it's still the one with all the random noises in it Sep 07 21:50:40 I for some reason only heard it a few times Sep 07 21:50:57 After I flashed with a self-built image, it somehow stopped playing Sep 07 21:51:08 cesarb, probably broken audio Sep 07 21:51:10 cesarb: You just murdered it. :P Sep 07 21:52:37 I have to reflash my device. I'm thinking of going the long way this time, until I have someone near me to unbrick it if I flub the uboot flash. Sep 07 21:52:50 thos: I managed to later test the audio with (IIRC) mpg123, so it isn't a broken audio... Sep 07 21:55:46 cesarb, oh, but there are various audio setup routes Sep 07 21:56:15 fwiw, I've had mpg123 work but not the audio for a call Sep 07 21:56:30 So, just because you have audio doesn't mean it's all "working" Sep 07 21:58:34 Yeah, there's mixer settings to worry about. The audio for a call doesn't take the same path as audio generated by the CPU Sep 07 21:58:51 It's directly plumbed to the mixer chip from the GSM. No CPU involvement aside from throwing switches Sep 07 21:59:12 BTW: Is there a new maintainer for gsmd? Or is it still laforge? Sep 07 21:59:55 I asked this question a few weeks back, looked around a bit but haven't had much time to play with it: Is there any way to capture audio from the mic to local storage? Sep 07 22:01:56 HistoryOnTheRoad - Yep. Sep 07 22:02:15 HistoryOnTheRoad, audio can pretty much be routed however you want Sep 07 22:02:47 you can have the SoC do some Cylon effects to both sides of the conversation if that's your thing Sep 07 22:03:09 Elrond: :D .. I know it's theoretically possible ... I had been hoping there was someone to point me in the way of "look at this piece for a magical incantation which will turn it on" Sep 07 22:03:18 HistoryOnTheRoad: alas - no. Sep 07 22:03:23 HistoryOnTheRoad: there is the WM* datasheet Sep 07 22:03:28 and cat /dev/dsp Sep 07 22:03:30 HistoryOnTheRoad - I did it a month ago. Sep 07 22:03:44 Let me boot the neo and try. Sep 07 22:03:44 Set the mixer to the right magic, and you can do it Sep 07 22:04:04 Elrond: woo hoo !! Sep 07 22:04:33 <-- has to distract himself from crappy other wanna-be-opensource-projects... Sep 07 22:05:12 I'm not an ALSA expert. I can write a script to try 1000 possibilites of switches, but that datasheet is a bit daunting to look at. Sep 07 22:05:32 It's not 1000 Sep 07 22:05:35 It's 97! Sep 07 22:05:44 SpeedEvil - 2^97 ;o) Sep 07 22:06:15 I don't mind hacking software ... groking hardware datasheets isn't impossible, but I'd need some help at this point. Sep 07 22:06:21 HistoryOnTheRoad: run alsamixer. count the number of switches Sep 07 22:06:33 Heh, even 2^97 isn't so bad. Just a matter of time :P Sep 07 22:06:43 HistoryOnTheRoad: it would be *much* more productive to describe the set of switches as a graph Sep 07 22:06:52 HistoryOnTheRoad: and search on the graph for the connection you want Sep 07 22:06:58 Naah. Sep 07 22:07:23 It would be much more productive to take the WM* diagram on the audio subsystem page, and annotate it with the alsa mixer names. Sep 07 22:07:29 (a friend of mine wants to do that... with Lua.) Sep 07 22:07:36 cesarb: Ok, fair enough. I don't mind a thinking mans problem rather then the brute force method ... but I still need a doc to make up a graph, then. Sep 07 22:07:54 With bonus points for a javascript thing to let you click on mixers to change them, and output an alsamixer line Sep 07 22:08:07 HistoryOnTheRoad - ... testing... will soon tell you, how it works (IF it works) Sep 07 22:08:17 JAVASCRIPT? We don't need no stinking javascript. Sep 07 22:08:32 * HistoryOnTheRoad codes it up in Flash that only works in IE. Sep 07 22:08:34 HistoryOnTheRoad: you need, as speedevil just suggested, the datasheets for both audio chips (the WM and the LM one), annotated with alsamixer names... Sep 07 22:09:03 However, don't we have some "scenario" thing where an alsamixer switch changes a whole set of switches on the chip? Sep 07 22:09:19 Yeah - some of the settigns probably do. Sep 07 22:09:21 * cesarb saw these "scenario" stuff when looking for a way to turn off the audio amp Sep 07 22:09:43 HistoryOnTheRoad - alsactl -f /etc/alsa/capturehandset.state restore Sep 07 22:09:57 HistoryOnTheRoad - arecord -f cd -t wav -d 10 foo.wav Sep 07 22:10:04 I just installed upstart and everything worked as before... it does not seem good Sep 07 22:10:11 HistoryOnTheRoad - Make noise for 10 seconds Sep 07 22:10:24 HistoryOnTheRoad - alsactl -f /etc/alsa/stereoout.state restore Sep 07 22:10:32 HistoryOnTheRoad - aplay foo.wav Sep 07 22:10:43 (script) wiki RecentChanges: [[P1_Owners]] [[Community_nominees_for_GTA02_P0]] [[User:Alessandro]] [[User_talk:HaraldWelte]] [[Talk:Main_Page]] Sep 07 22:11:26 Elrond: does that work for you? Sep 07 22:11:40 SpeedEvil - Yes. Just tested. Sep 07 22:12:11 where should I look for some logs? /var/log is almost empty Sep 07 22:12:33 rwhitby: dead mon' ? Sep 07 22:12:37 SpeedEvil: this [[Community_nominees_for_GTA02_P0]] page is interesting... looks like you are already winning :D Sep 07 22:14:04 * SpeedEvil doesn't know. Sep 07 22:14:22 Huh, I'm on there too. Fun, didn't know. :) Sep 07 22:14:24 I'm at the moment ordering bits so GTA01 users can have accelerometers too. Sep 07 22:14:35 For a small fee. Sep 07 22:14:37 Elrond: merci buckets Sep 07 22:14:54 HistoryOnTheRoad - Does it work for you? (I have 2007.1 here...) Sep 07 22:15:37 I'm still on .1 ... let me see if the battery is not dead :) Sep 07 22:17:20 somebody knows how to reach pH5? Sep 07 22:17:40 cesarb: I also have no startup sound any more. Strange. Sep 07 22:18:13 MP3 playback works fine with the stereo ALSA state, and phone audio works fine with the handset state. Sep 07 22:18:21 But no startup sound, and no screen-tap sound. Sep 07 22:19:00 tihi... I just was about to type "w3m ...." on the neo... having it look like a normal machine is fun. :) Sep 07 22:19:49 like a normal machine? Sep 07 22:20:13 I have bootscreen :P Sep 07 22:20:16 fully connected to the internet, having _all_ crazy stuff installed. ;) Sep 07 22:20:36 ... like w3m. ;o) Sep 07 22:20:55 quicksand: oh, btw, weren't you asking earlier about $HOME? Sep 07 22:21:04 Yeah! Have a thought on that? Sep 07 22:21:12 I couldn't find anything in /etc or subdirs. Sep 07 22:21:21 quicksand: I think it's the shell which sets that, as the result of getpwent() (not sure about the name of the function) Sep 07 22:21:36 quicksand: which in turn gets (since it's glibc) the info from NSS Sep 07 22:21:56 cesarb: That's the right name, if memory serves Sep 07 22:21:56 quicksand: so, check /etc/nsswitch.conf, which should tell where it gets the "home directory" field from Sep 07 22:22:13 Probably /etc/passwd? Sep 07 22:22:23 quicksand: long history short: if /etc/nsswitch.conf says "files", it's /etc/passwd Sep 07 22:22:48 quicksand: just hope it isn't getting it from a LDAP server somewhere else in the world ;-) Sep 07 22:23:44 cesarb: http://pastebin.ca/686338 -- can you help me interpret? Sep 07 22:24:20 quicksand: "passwd: compat" also gets from /etc/passwd, but has some extra things to deal with NIS stuff AFAIK Sep 07 22:25:24 (actually, it can get from a compat: line, but there isn't one in that file) Sep 07 22:25:33 OK, now in /etc/passwd I have: root:[blahblah]:0:0:root:/home/root:/bin/sh Sep 07 22:25:45 Among other entries, of course. Sep 07 22:25:52 quicksand: and it still sets $HOME to /? Sep 07 22:25:56 Yup. Sep 07 22:26:04 quicksand: does /home/root exist? Sep 07 22:27:01 quicksand, and has +rx permission for at least user? Sep 07 22:27:12 It exists, but not right away -- it's a FAT partition on my MicroSD that gets automounted at boot. All of /home is, actually. Good point. Could that be it? Sep 07 22:27:54 quicksand: ...which is why FHS says root's home should NOT be in /home, btw... it should be /root, since /home might be on another disk Sep 07 22:28:01 Heh. Sep 07 22:28:13 I haven't looked at FHS in a long time. Sep 07 22:28:19 quicksand: could be, if what you are starting isn't a "login shell" Sep 07 22:29:05 quicksand: only the "login shells" would set the home (or would it be /bin/login? I think I'm getting rusty at these sorts of old arcane unix knowledge) Sep 07 22:29:19 oh my Sep 07 22:29:24 GOD Sep 07 22:29:25 quicksand, what if you do a su - ? Sep 07 22:29:29 I think you nailed it. Subsequent shells and SSH sessions pick up the correct value in .ashrc -- which I am forcing because it has irritated me sufficiently. Sep 07 22:29:32 Ghiottone: beat me to it ;-) Sep 07 22:29:33 <|Lynx> thtas Sep 07 22:29:37 <|Lynx> http://www.peeandpoo.com Sep 07 22:29:46 cesarb, lol :) Sep 07 22:30:06 <|Lynx> ho has such sick ideas? Sep 07 22:30:16 <|Lynx> pee and poo baby wear Sep 07 22:30:46 rofl Sep 07 22:30:57 . . . but GUI programs keep wanting me to store my documents in /. Su - puts me at /home/root, as expected, but again it's probably because of my forcing the issue. Sep 07 22:31:21 Elrond: produced a file, but no playback with that incantation. But it's promising ;) Sep 07 22:31:32 I think you set me on the right path, cesarb, so tonight I'll muck around with it further. Thanks. Sep 07 22:31:33 quicksand: the GUI is probably inheriting the home from something which gets set early on the boot proccess Sep 07 22:31:45 morning Sep 07 22:31:48 cesarb: exactly my suspicion. Sep 07 22:31:55 rwhitby: hi Sep 07 22:32:09 * rwhitby notices http://wiki.openmoko.org/wiki/Community_nominees_for_GTA02_P0 Sep 07 22:32:13 quicksand: perhaps init itself, the neo doesn't have login Sep 07 22:32:14 Wait, I can try this with the new broken terminal app that (most likely) doesn't read any .rc . . . testing . . . Sep 07 22:32:33 quicksand: the new terminal app probably also doesn't use a login shell ;-) Sep 07 22:33:33 Well, if I go into the new terminal and immediately type 'env', it tells me HOME=/ Sep 07 22:33:44 And after su - it tells me HOME=/home/root Sep 07 22:33:54 I think we got it. Sep 07 22:33:54 what about setting HOME at the start of .ashrc ? Sep 07 22:34:27 Ghiottone: that works for subsequent busybox shells, but not for the GUI, even if it's stored in / Sep 07 22:34:27 the shell should parse it even if it is not a login shell Sep 07 22:34:50 quicksand, why? where does the gui gets launched from? Sep 07 22:34:56 Beats me. Sep 07 22:35:30 mmm maybe init? as xserver script... Sep 07 22:36:24 try setting home in /etc/init.d/xserver-nodm :D Sep 07 22:42:31 quicksand, did you try? Sep 07 22:43:12 rwhitby: maybe you can save me a lot of effort, i'm trying to find the upstream repo for the kernel used on the moko Sep 07 22:45:32 cw666: the upstream repo is kernel.org itself... Sep 07 22:45:46 cw666: just look at openembedded/packages/linux/linux-openmoko_2.6.22.5.bb and openembedded/packages/linux/linux-openmoko.inc Sep 07 22:46:07 cw666: the SRC_URI lines Sep 07 22:46:12 yeah, im looking at that Sep 07 22:46:17 i wondered if that was really it Sep 07 22:46:28 HistoryOnTheRoad - You have to switch the alsa state for playback. See my list of commands for you. Sep 07 22:46:46 s/svn/http/ in the URL i can use it, svn URIs don't work for some reason Sep 07 22:47:05 cesarb: thanks! Sep 07 22:51:53 cw666: svn URIs only work with a svn client... as long as they aren't blocked by some firewall ;-) Sep 07 22:52:03 cw666: take a look at svnweb.openmoko.org, btw Sep 07 22:52:14 cw666: it might be more useful Sep 07 22:52:38 Elrond: I did the stereoout restore alsacommand. Maybe my volume is too low? Sep 07 22:53:14 cw666, the command to get the patches is svn co http://svn.openmoko.org/trunk/src/target/kernel/patches patches Sep 07 22:54:47 HistoryOnTheRoad - Hmm, no. stereoout usualy is quite loud... Sep 07 22:55:05 HistoryOnTheRoad - strange. Sep 07 22:55:13 hm, after a crash my top-menu-bar seems to be completely broken. even a "reboot" doesnt bring it back Sep 07 22:55:24 is there a way to manually start the top "tray"-bar? Sep 07 22:55:26 Ghiottone: I bet that will work. But it's odd, I think, that the init would process profile.d stuff *before* devices come up. I'll try it now . . . it will take a couple minutes to reboot. Sep 07 22:56:39 quicksand: init runs things before devices come up all the time Sep 07 22:56:58 quicksand: everything that's a module and not loaded by the initrd (which, btw, openmoko doesn't use) Sep 07 22:57:10 Elrond: Well, you've got me in the right direction. I have to go out for a bit, I'll try piddling with it some more, and when I get home, I'll try scp-ing the file over and see if I can hear the sound. If so, then we can figure out the non-neo-play issue later. Sep 07 22:57:13 quicksand: and, IIRC, the MMC/SD stuff is compiled as a module Sep 07 22:57:16 Ghiottone: yeah, i did that... before i did a cut and paste of the URL assuming http wasn't working and it failed Sep 07 22:57:36 Elrond: Thanks very much, i appreciate your help. Sep 07 22:57:37 * cworth is aghast at how slow monotone is... Sep 07 22:58:06 And even more aghast that people are using it to try to get real work done. Sep 07 22:58:25 cworth: heh.. what is your distributed version control system of choice? ;) Sep 07 22:58:59 HistoryOnTheRoad - No problem. :) Sep 07 22:59:15 HistoryOnTheRoad - That's the nice thing on this channel: Everybody is helpful. :-) Sep 07 22:59:45 hozer: just about anything is better than monotone Sep 07 22:59:50 hozer: I definitely prefer git. But either git or hg would perform a lot better than what I'm watching mtn do here I think, (I don't have any significant experience with bzr). Sep 07 23:00:07 But I can't imagine that bzr could be _this_ slow either. Sep 07 23:00:41 quicksand, have a look at /etc/init.d/mountall.sh Sep 07 23:01:05 mtn is at the "write your own version-control system while you wait" speed, (at least on the openembedded repository) Sep 07 23:01:16 Elrond: True enough. I've gotten lots of help. If I have some time, i might do up that graph, if I can decipher all of the alsacontrol switches. Sep 07 23:01:24 cworth: any way to go from mtn to HG easily? Sep 07 23:01:25 back in a bit! Sep 07 23:02:08 hozer: No idea. For a single-branch project, tailor is supposed to do a fairly good job of many-to-many conversions. Sep 07 23:02:32 Ghiottone: Waiting on a reboot now . . . I know init processes some things before devices come up -- like you say, it has to -- but I would have expected profile.d to be pretty much last. But then again, some settings in there might be prerequisites to other startups. Anyway, I'll know more in a sec. Sep 07 23:02:35 * hozer has gotten bitten by tailor svn->hg a few times Sep 07 23:03:05 maybe it's just sqlite so slow with so huge DB Sep 07 23:03:15 hozer: Yeah. I'm really just repeating hearsay, (we had no luck with tailor-based cvs->git conversions). Sep 07 23:04:45 time mtn log packages/gcc/gcc-cross_4.2.1.bb --brief --last 5 Sep 07 23:04:47 real 4m5.761s Sep 07 23:04:52 that's scary :p Sep 07 23:05:41 ynezz: It doesn't really matter the reason. It's clear that it's entirely unusable. And there are lots of alternatives. Sep 07 23:05:43 Ghiottone: and cesarb: Yup, setting HOME in /etc/init.d/xserver-nodm does the trick. Sep 07 23:06:48 Thanks for the input. It's an ugly hack, but now that I know what the problem is, perhaps I'll go 'rearrange the furniture' a bit and put root's HOME on the rootfs. Sep 07 23:06:59 ynezz: I started an update well over 10 minutes ago, and it's only about 5% of the way through its certificates. Sep 07 23:07:05 in mountall.sh there is a check _after_ mount -a if the root home dir exists. if it is not, then it is created Sep 07 23:07:27 but $HOME is taken by grep'ing /etc/password... Sep 07 23:07:36 strange Sep 07 23:07:50 cworth: i don't update so frequently so i don't care about it's speed Sep 07 23:08:42 cworth: but looking up a logs/diffs is what scm is about Sep 07 23:08:51 I might also file an openmoko bug report with a reference to the FHS. People put /home on other devices all the time. I always do. Sep 07 23:09:46 (For very small values of 'always' because I have only installed desktop Linux once.) Sep 07 23:10:00 (Still, that's 100% of the time.) Sep 07 23:10:38 Monotone's website says: "Tailor is an any-to-any version control system converter, with support for most free VCSes. Note that as of July 2007, histories are linearized on the timestamp (so don't expect it to losslessly convert between systems with DAG histories, like monotone/git/mercurial/bzr, for instance)." Sep 07 23:10:48 So that's useless. Sep 07 23:12:16 Ghiottone: you would thing "getent passwd root" followed by splitting on : would be easier... Sep 07 23:12:28 * cesarb wonders if busybox has getent Sep 07 23:13:10 cworth: I'm not here to defend monotone. The way I make it useable is to regularly run mtn pull. Sep 07 23:13:45 rwhitby: Just spread the pain out, eh? ;-) Sep 07 23:13:51 rwhitby: it makes the log command faster also? Sep 07 23:13:53 (note that mtn pull and mtn update are separate commands, specifically so you can do the time-intensive pull regularly, but still choose when to mtn update your local checkout Sep 07 23:14:07 ynezz: what version are you running? Sep 07 23:14:11 0.31 Sep 07 23:14:29 * hozer would personally like to see everything put into heirarchical hg or git repositories as raw source files Sep 07 23:14:32 ubuntu... Sep 07 23:14:32 update to the latest and you will see a huge speedup on log Sep 07 23:14:41 cesarb, yeah. I also wonder why if it does the grep thing (and thus assuming that HOME is not valid) it those not export $HOME to the right value.... Sep 07 23:14:42 ah, ok then Sep 07 23:15:53 One last observation on this subject: Xinit also sets HOME=/home/root, then checks whether it exists. If it doesn't, and /root exists, then HOME is changed to /root. Sep 07 23:16:02 over 19 minutes and only 10% of the certs. still---I guess I won't get to any openmoko hacking today... :-( Sep 07 23:16:06 I've said it before - if someone creates a stable, reliable, correct, two-way gateway between git and monotone, and hosts a production server, then I'll be happy to make the mokomakefile point to it. No dinky on-my-home-adsl-connection solutions need apply please. Sep 07 23:16:21 But if /root doesn't exist . . . then . . . nothing. Perhaps that's a good place for a patch. Sep 07 23:16:54 cworth: put the mtn pull in a cron job, so the stuff is already on your machine when you want to start hacking Sep 07 23:17:22 Thoughts? It makes more sense to create /root if neither /home/root nor /root exists, and that's the patch I have in mind. Sep 07 23:17:43 rwhitby: We've got gobs of bandwidth we can use on freedesktop.org, but I can't find even a one-way conversion tool, (looks like monotone was never popular enough for significant projects to leave it), let alone a two-way tool. Sep 07 23:17:45 quicksand, it sound reasonable for me Sep 07 23:18:20 also monotone does not support http proxies Sep 07 23:18:32 rwhitby: It's not even network transfer though. It's just CPU wasting, (19 minutes of CPU time now). Sep 07 23:19:12 Thanks. Unless I hear an objection, I'll file a bug report and submit a patch later tonight. I'll check the IRC log before I do. Sep 07 23:19:30 cworth: in that case, the only way forward is to convince the #oe core team (of which at least mickey|sports and RP in this channel are members) to change. Venting about monotone in this channel is only useful for releasing your frustrations - it won't actually help anything change. Sep 07 23:19:42 guys what's the status of power managment? Sep 07 23:19:52 Not there yet Sep 07 23:19:52 quicksand: what would make more sense would be to create /root on the rootfs creation step and set the home to it on /etc/passwd, and then use a _non-root_ user for the GUI! Sep 07 23:19:56 not started yet Sep 07 23:20:05 bedboi: Minimal. Screen blanking works. Sep 07 23:20:18 rwhitby: Right. So a one-way conversion would be helpful for demonstrating the current problem. That's why I'm looking for one. Sep 07 23:20:21 bedboi: screen blanking and manual power saving works. Sep 07 23:20:27 bedboi: well, it's better than what was before: we now _know_ all the magical incantations you need to type before powering off so it won't drain your battery! Sep 07 23:20:29 But it's basically not usable yet Sep 07 23:20:40 cworth: yep, in fact a one-way is probably good enough for mokomakefile. Sep 07 23:21:05 cworth: what is actually *in* oe? it's not the actually source, just bb files and patches, right? Sep 07 23:21:20 hozer: correct Sep 07 23:21:35 rwhitby: You're kidding! Sep 07 23:21:46 cesarb: Yup, agreed. There are a few sub-issues here, and I'll try to hit them all in the bug report. Fixing /home/root is just a minor bandage, but it should be done anyway IMHO. Sep 07 23:21:54 rwhitby: It's taking this much time for that little amount of data? That's insane! Sep 07 23:22:09 hozer: but it's 190M :) Sep 07 23:22:12 cworth: as I said, I'm not here to defend monotone. Sep 07 23:22:20 cworth: There are a LOT of packages, and some of the patches are quite large. Sep 07 23:22:23 rwhitby: Yeah, I know. Sep 07 23:22:33 rwhitby: tailor would probably work, if enough people care i guess i can host a git repo Sep 07 23:22:45 damn! webkit has changed svn revision between my do_compile and do_install and now it is looking in another dir! Sep 07 23:22:46 anyway kernel is few times bigger, but with git few times faster also :) Sep 07 23:22:49 cworth: I think a better one-way tool would be to check out mtn once a day, extract the BB files, then extract source and *apply* the patches, then check the whole thing into git Sep 07 23:22:58 Ghiottone: doesn't it do that all the time? Sep 07 23:23:10 Ghiottone: the trick is to start your build just after 0:00 GMT Sep 07 23:23:10 cw666: I would love that! Sep 07 23:23:29 someone puts a git mirror in production, I change mokomakefile to use it, that's the way to show that it's a viable replacmement. Sep 07 23:23:35 cesarb, that worked until few days ago Sep 07 23:23:51 build it and they will come. do something other than just vent frustrations :-) Sep 07 23:23:52 cesarb, now the packages are named after svn HEAD revision Sep 07 23:24:11 cedric, so a commit of Joe Developer is enough Sep 07 23:24:20 s/cedric/cesarb/ Sep 07 23:24:20 Ghiottone meant: cesarb, so a commit of Joe Developer is enough Sep 07 23:25:03 rwhitby: I'm trying. It appears that tailor has a useless command-line and man page, (gotta fetch the README and construct a configuration file...). Sep 07 23:25:17 Ghiottone: great. so we spend at least 90% of our neo-hacking time fighting the damn build system. Sep 07 23:25:28 cesarb, could be worse Sep 07 23:25:45 you chould be spending 99% of your time fighting it Sep 07 23:25:47 :) Sep 07 23:26:52 cb22: you know, it's been like _two days_ with me with a patch for some gsmd timing issues, and I don't even know if it compiles Sep 07 23:26:53 cworth: i have to run off now, if someone can make taylor behave i can host a git server Sep 07 23:27:19 cw666: I'll keep trying. If I can get it to work, I could host it too. Sep 07 23:27:22 cb22: I think this patch could be enough to make the gsmd actually work correctly on boot... but I cannot test Sep 07 23:27:22 ouch Sep 07 23:27:28 big ouch Sep 07 23:27:48 cworth: www.terminus.org/hg/hgforest Sep 07 23:27:49 what's the need? to convert from monotone to git? Sep 07 23:27:55 cesarb: where's the patch? Sep 07 23:28:01 Ghiottone: reliably Sep 07 23:28:10 Ghiottone: yes Sep 07 23:28:11 and incrementally Sep 07 23:28:15 cw666, eh. got it Sep 07 23:28:21 cb22: all because a binutils-cross issue stopped my build in the middle, and while I was trying to fix it something in some .bbclass or recipe changed and completely broke the build (no init AFAIK) Sep 07 23:28:28 oh, wait... if it's just a single branch it wouldn't be hard to do anyhow Sep 07 23:28:33 rwhitby: http://pastebin.ca/683328 Sep 07 23:28:46 cw666: Well, incremental would let us workaround the bug. Non-incremental would hopefully help us convince the right people to just fix the bug. Sep 07 23:28:59 rwhitby: it's more of a theory, "I think this is the cause for the gsmd issues", but I cannot test :( Sep 07 23:29:17 cesarb: how do you know which commands can't be pipelined? does the spec make this clear? or experience? Sep 07 23:29:24 cesarb: Cool, I'll be glad to test, (if I can ever get a build environment up). Sep 07 23:29:29 cesarb: certainly some commands back to back confuse the modem terribly Sep 07 23:29:34 cesarb: looks like a very good theory to me Sep 07 23:29:44 cw666: I know at least ATZ, AT+CFUN=1, AT+COPS=0, and AT+COPS=? don't like it Sep 07 23:29:52 * cworth is still only 18% of the way through the update Sep 07 23:30:10 cesarb: i need to debug this, but for me gsmd fails when first started Sep 07 23:30:11 cw666: in fact, I think I heard around here that sending *any* command will cancel an AT+COPS=? Sep 07 23:30:16 it has to be killed and restarted Sep 07 23:30:37 i sorta wondered if the calypso firmware was hosed and might get an update or fixes Sep 07 23:30:52 cw666: the symptom which would be fixed with that patch is when you have in your /tmp/gsm.log a line with ATZ, a line with AT+CFUN=1, and a line with AT+COPS=0, with no modem responses between then, and stuck after these three Sep 07 23:31:09 !ombug 766 Sep 07 23:31:12 cesarb: i actually have another workaround Sep 07 23:31:26 cesarb: i should try what you have there, the idea seems sane enough Sep 07 23:31:33 hm, where's cdbot? Sep 07 23:31:48 cesarb: i'm on that bug already thanks :-) Sep 07 23:32:21 cw666: I intent to attach that patch to that bug once I can check it actually works and make any difference Sep 07 23:32:25 s/intent/intend/ Sep 07 23:32:26 cesarb meant: cw666: I intend to attach that patch to that bug once I can check it actually works and make any difference Sep 07 23:34:50 foo Sep 07 23:34:52 /s/foo/bar/ Sep 07 23:34:55 s/foo/bar/ Sep 07 23:34:56 cw666 meant: /s/bar/bar/ Sep 07 23:34:56 doi Sep 07 23:35:06 heh, apt is a bot! i didn't realize Sep 07 23:35:07 Most of my best moments have something to do with Susa, he thought. With Susan, or with a computer. Never with Susan and computers. These where the worst moments. Sep 07 23:35:20 lol Sep 07 23:35:20 Douglas Adams is a genious Sep 07 23:35:35 it's The Truth Sep 07 23:36:12 Ghiottone hi Sep 07 23:36:20 Ghiottone whats the problem with gprs? Sep 07 23:36:24 lol cw666 Sep 07 23:36:40 cesarb: http://stupidest.org/tmp/om/gsm-log.txt Sep 07 23:37:36 cw666: that with the patch, or without it? Sep 07 23:37:36 cesarb: it's also far from clear how much can be pipelined, i think it might make sense to have a runtime settable limit there (so when i doubt you can turn if off) Sep 07 23:37:39 cesarb: without Sep 07 23:37:54 cesarb: im rebuilding now, it'll probably take, umm, not sure actually Sep 07 23:38:02 I would expect you'd get the thing stable with *no* pipelining first Sep 07 23:38:14 cw666: hm, a bit different than mine, but the CFUN/COPS part is the same Sep 07 23:38:30 cw666: mine doesn't even get the OK from ATZ Sep 07 23:38:34 hey kristian-m Sep 07 23:38:54 kristian-m, it does not seem to launch a ppp process Sep 07 23:39:01 Meanwhile, looking past the gsmd problems---what's the word on echo issues while talking on the phone? (Bad echos on the other end.) Sep 07 23:39:05 cesarb: well, it used to be you could frob the gsm power (at processor power?) to get it back end it's hung Sep 07 23:39:12 kristian-m, also, I would like to change the parameters it sents Sep 07 23:39:20 Ghiottone oh, btw, we have a dependency on pppd now Sep 07 23:39:25 cesarb: i've def had cases where the AT interp went away Sep 07 23:39:31 cw666: I don't even have to do it, just killing gsmd and restarting it is enough, so the modem chip isn't wedged Sep 07 23:39:32 Ghiottone have you tryed it manually? Sep 07 23:39:48 cesarb: yeah, that works too usually Sep 07 23:39:59 Ghiottone maybe the modem is not initialized - it has to be registered with a network Sep 07 23:40:02 cw666: I'm just trying to see if my theory on bug 766 is the cause for gsmd not working correctly on boot Sep 07 23:40:03 cesarb: oddly it seems only you and me see this ... not others, i wonder why? Sep 07 23:40:26 kristian-m, oh ops! did not realize it. no I did not try manually Sep 07 23:40:29 cw666: other people seem to have PINs on their SIMs Sep 07 23:40:33 cw666: I don't Sep 07 23:40:38 oh, perhaps yeah Sep 07 23:40:41 i dont bother either Sep 07 23:40:45 kristian-m, the gsm page reported good signal strenght. and icon in the panel bar was changed Sep 07 23:40:47 im not even sure if i did how i would deal :-) Sep 07 23:40:51 cesarb i have the same here, gsmd sometimes just ignored the modem for ever or something Sep 07 23:41:02 ok, time for the kids Sep 07 23:41:18 Ghiottone hmm, then try killing the gsmd before connecting to gprs Sep 07 23:41:31 kristian-m, is the package ppp? Sep 07 23:42:03 Ghiottone i guess Sep 07 23:42:17 ERROR: Exception:exceptions.OSError Message:[Errno 2] No such file or directory: '/home/polz/openmoko/build/tmp/work/armv4t-angstrom-linux-gnueabi/libelf-0.8.6-r0/install/libelf-locale' Sep 07 23:42:37 does anyone have any ideas why I might get this error ? Sep 07 23:42:46 polz: ...your build is hosed Sep 07 23:43:00 polz: you can go ahead by cleaning libelf ("make clean-package-libelf") Sep 07 23:43:09 polz: but you'll hit that with another package Sep 07 23:43:18 polz: and another, and another, and another... (more than 20) Sep 07 23:43:22 Ghiottone yap, ppp includes the pppd binary Sep 07 23:43:35 fsck Sep 07 23:43:35 kristian-m, could you add a switch to write a pppd log? Sep 07 23:43:38 polz: then, when you actually finish building and try to flash, you'll find out you have no init Sep 07 23:43:49 argh. Sep 07 23:43:50 kristian-m, ok I'll release -r3 asap Sep 07 23:43:52 polz: just nuke the build (rm -rf build/tmp) and start over Sep 07 23:44:00 greeeeeeeat Sep 07 23:44:02 Ghiottone try starting it from a terminal - its in the debug messages Sep 07 23:44:15 polz: you had binutils-cross problems, right? ;-) Sep 07 23:44:19 that's just going to take... oh, well, I don't know, 2 days or so Sep 07 23:44:23 kristian-m, ok Sep 07 23:44:25 Ghiottone great, wait, i'm working at an icon Sep 07 23:44:29 polz: welcome to the club :D Sep 07 23:44:42 kristian-m, oh ok :) Sep 07 23:44:59 rwhitby: So I'm looking for the openembedded mtn repository to feed to tailor, (looking in MokoMakefile). Could you help me out? Sep 07 23:45:02 cesarb: actually, I had problems with the build system constantly using some older revision for openmoko-panel-gsm .... Sep 07 23:45:11 and complaining about missing files Sep 07 23:45:33 so I deleted some .sqlite3 file which seemed to be connected with bitbake Sep 07 23:45:34 cworth: monotone.openembedded.org, org.openembedded.dev branch Sep 07 23:46:02 rwhitby: OK. I'll try feeding that to tailor and see what happens. Sep 07 23:46:06 and the thing finally started building the latest svn... which doesn't really help me that much if it fails at a different point Sep 07 23:46:07 Ghiottone http://mput.de/~kristian/.openmoko/Icon/ - but I'm not sure if it is okay yet Sep 07 23:46:10 cworth: crossing fingers :) Sep 07 23:46:33 kristian-m, in the top box (that one with the name of the ppp profile, I think) nothing is showing until I try to select some text Sep 07 23:47:46 Ghiottone ah, right thats a known bug - see #786 in bugzilla - they have yet to apply the patch Sep 07 23:48:02 cworth: I think you should go read openbemedded.org and figure out how they lay stuff out Sep 07 23:48:39 kristian-m, ah ok. I launched it from the shell. loaded edited and save my profile. clicking on connect to gprs does not seem to do anything Sep 07 23:48:59 hozer: tailor is only asking me for --repository and --module so far. Sep 07 23:49:02 kristian-m, on the terminal I can see "executes subprocess" and "Done" Sep 07 23:49:39 cworth: I know.. but I think there are some inherent problems in the way they build things Sep 07 23:49:59 Ghiottone try to start manually: pppd debug call latest_gprs Sep 07 23:50:11 and using a different VCS is just sorta papering over other problems Sep 07 23:51:03 kristian-m, it seems to hang my ssh connection and then close Sep 07 23:51:11 Ghiottone i guess the subprocess is showing an error and quits... - should have caught this. :-( Sep 07 23:51:45 it shows atz for a few seconds Sep 07 23:51:56 I type but nothing shows or happens Sep 07 23:51:56 Ghiottone huh? - very strange. - try from the terminal Sep 07 23:53:06 hozer: I'm not sure what you're suggesting I look for. Sep 07 23:53:27 kristian-m, done. same results Sep 07 23:53:29 cworth: openembedded checks a bunch of patches into a distributed version control system Sep 07 23:53:45 hozer: yes... Sep 07 23:53:56 Ghiottone i suspect '\d' from the chat script is somehow executed - can't emagine why... - so the terminal hangs or is it getting closed? Sep 07 23:54:18 damn, i just read on the community ml that (http://lists.openmoko.org/pipermail/community/2007-September/009944.html) estimated time for new shipping is 20 sept? Sep 07 23:54:19 which seems to me like a hack because the VCS doesn't really do what you'd watn Sep 07 23:55:01 I have this idea you should be able to clone the GCC HG repo, apply a patch, check those changes into your local GCC repo, and use that directly Sep 07 23:55:16 hozer: I'm not following. But even assuming openembedded is totally broken, it's what the openmoko code expects to work with now, and monotone is definitely getting in my way. Sep 07 23:55:24 same for the kernel.. clone the hg or git kernel repos Sep 07 23:55:35 hozer: Yes, you should definitely be able to do that. Sep 07 23:55:42 apply patches, check the changes in as changesets, *not* patches Sep 07 23:55:42 (script) openmoko-community: "Michael 'Mickey' Lauer" Re: Apple is going to beat all competitors Sep 07 23:55:50 hozer: Git has a notion of "submodules" which would be perfect for this kind of setup. Sep 07 23:56:05 hozer: Oh, I see what you mean. Yeah. Sep 07 23:56:21 yeah.. mercurial has an 'hgforest' extension that seems to be similiar Sep 07 23:56:24 kristian-m, terminal hangs for a bit and then returns to prompt Sep 07 23:56:30 * rwhitby still wonders why people are talking here about SCM Sep 07 23:56:38 when #oe is where the decision will be made Sep 07 23:56:40 hozer: Fedora does the same totally broken thing of adding the actual patch files to the VCS. And yes, I think this is really stupid. Sep 07 23:57:05 (Though, maybe it made sense given the tools at the time. But definitely not taking advantage of distributed VCS.) Sep 07 23:57:13 Ghiottone hmm to stange no idea, can you pastebin a log? Sep 07 23:57:21 rwhitby: I think the real issue can't be addressed by OE without really starting over Sep 07 23:57:33 kristian-m, yeah I'd like to. but how do I create one? Sep 07 23:57:36 talk about changing SCMs is just hot air here. talk in #oe - set up an alternative - lobby the OM core team to change on the openmoko-devel mailing list in public, giving a real working alternative Sep 07 23:57:58 rwhitby: I don't follow. If #oe is fundamentally broken, why would openmoko necessarilly have to convince it to fix itself before doing something better. Sep 07 23:58:18 rwhitby: I'm all for setting up an alternative and talking about in on the mailing list. Sep 07 23:58:34 And that's why I'm here getting help on setting up the alternative. Sep 07 23:58:45 cworth: cause the lead software architect of OM is also on the core team of #oe and was part of the decision (which was the *right* decision back *then* when it was made) Sep 07 23:59:24 cworth: yep - you are doing the right thing Sep 07 23:59:34 rwhitby: I agree.. the decision was the right one at the time Sep 07 23:59:35 rwhitby: Either way, we're not trying to convince anyone of anything right now. Sep 07 23:59:43 Ghiottone you don't have the ssh terminal open anymore? hmm - so it happens after atz - that means "Defining Packet Data Protocol" has already shown up? - I still wonder why the subprocess from within the gui is not printing that output... Sep 08 00:00:15 cworth: I disagree. I want to move to something more "mainline" than monotone. Sep 08 00:00:40 rwhitby: And yeah, the "core team" of openmoko does the right thing by adopting something that exists. It's for casual-time hobbyists to do experiments that reinvent everything. Sep 08 00:00:46 rwhitby: people are talking here about SCM because someone mentioned the "mtn" word Sep 08 00:00:50 rwhitby: Oh, definitely. Sep 08 00:00:59 rwhitby: it's one of those "trigger" words for this group of people Sep 08 00:01:00 kristian-m, well it prints just atz. and when terminal is back again, it overwrites even the atz Sep 08 00:01:03 Ghiottone btw "excecutes subprocess" does not show the system-calles comandline? Sep 08 00:01:09 kristian-m, no Sep 08 00:01:12 rwhitby: I _will_ be working on convincing people to change eventually. Just that it's the "get something working" stage right now. Sep 08 00:01:56 Ghiottone but you have got /etc/ppp/peers/latest_... ? Sep 08 00:02:10 cesarb: I'm talking about SCM here at least until my "mtn update" completes, (since otherwise I'm just waiting for it). About 33% done now(!). Sep 08 00:02:17 cworth: excellent - that's the right (and "Linux") way to do it. "show me the working code" ;-) Sep 08 00:02:19 kristian-m, yes Sep 08 00:02:52 * cworth can't get tailor to do anything at all... :-( Sep 08 00:02:55 I pledge that MokoMakefile will be modified to work with any viable alternative, as part of showing the viability of that alternative. Sep 08 00:03:00 kristian-m, http://pastebin.com/d40a4e7e4 Sep 08 00:03:09 Ghiottone i whish i could reconstruct it here... I'll try a clean install Sep 08 00:03:42 rwhitby: Awesome. So nobody else I need to convince for now. ;-) Sep 08 00:04:18 kristian-m, this is with contents http://pastebin.com/d6b27469f Sep 08 00:04:21 cworth: I need an excuse to use git, cause I eventually want to replace svn on two continents at work with git. Sep 08 00:04:45 Ghiottone can you also pastebin a cat /etc/ppp/peers/* please? Sep 08 00:04:53 But yeah, if we _really_ want to throw away the entire approach that openembedded currently uses, (patches inside of version control), then something like a tailor conversion might not be what we want anyway. Sep 08 00:05:25 kristian-m, did it. do you want the full dir or is that enough? Sep 08 00:05:25 cworth: I'm going to suggest you start by pulling down the GCC mercurial repo Sep 08 00:05:38 cworth: that's a whole different ball game. I think people should wait to see what OM comes up with for the "binary SDK and toolchain" that mickey|sports has been talking about. Sep 08 00:05:47 Might be better to take the whole source tree that results currently, shove that all into git, and then just go forward by applying new openembedded patches from there. Sep 08 00:05:47 and then check the patches that OE/Openmoko/mokomakefile end up applying in Sep 08 00:05:53 Ghiottone ahkay, thanks Sep 08 00:06:08 cworth: the openembedded approach is a proven approach used by every single linux distribution Sep 08 00:06:18 hozer: Yeah. And that might be a faster way to get me to where I want to be. Sep 08 00:06:26 cworth: all linux distributions I've seen use a "upstream source + patches + control file" model Sep 08 00:06:39 cesarb: which is exactly what OE is Sep 08 00:06:43 cesarb: I _work_ for a large distribution, and I've been pushing internally to totally revamp this similarly. Sep 08 00:07:03 cesarb: "Always been done that way" doesn't mean it's right at all. Sep 08 00:07:07 * Aria shakes her head. That model works, but I'll sure say that OE is the most insane system I've worked with Sep 08 00:07:42 cworth: good luck, one of the better things of linux is that people can do try different things, and if they succeed (sp?), everyone else can eventually switch Sep 08 00:07:48 cesarb: Things like "git bisect" are really powerful, and I'd _love_ to be able to apply them at the distribution level. Sep 08 00:08:39 cesarb: Yeah. This seems like something that OpenMoko really "gets", (and Nokia's maemo work, for example, totally misses out on by not being 100% free). Sep 08 00:08:58 Ghiottone the only thing i see is that you have no number (atdt) there - there should only be one '#' thats a bug - but it should show the at commands before that, i think - try entering *123# or something as number if your provider doen not need one... - hmm still no idea Sep 08 00:09:19 hozer: Since where I want to be is "able to write some neo-targeted games while on a long plane ride tomorrow". Sep 08 00:09:31 heeh yep Sep 08 00:09:52 * kristian-m can't start his neo due to two really empty batteries.... Sep 08 00:10:16 kristian-m, ok I'm trying Sep 08 00:11:23 (script) wiki RecentChanges: [[Talk:IPhone]] [[User:Historybuff/notes]] Sep 08 00:12:26 kristian-m, no difference. http://pastebin.com/d2dfadbb5 Sep 08 00:14:09 Ghiottone I guess *99***1# is still the best choice - try: "mv /etc/ppp /etc/ppp_" and restart settingsgui Sep 08 00:15:03 Ghiottone you've got 4 other peers there, maybe its best to create a new one from the default scripts within settingsgui Sep 08 00:15:11 kristian-m, ok what exactly should I do? fill in data and click connect? do I have to save anything? which data are mandatory? Sep 08 00:16:05 Ghiottone you should fill in the apn and the user - maybe the password - i guess *99***1# is mostly okay Sep 08 00:16:39 Ghiottone you dont have to save - this is why i generate the latest_* scripts - so the original configuration does not have to be changed Sep 08 00:17:33 kristian-m, same results Sep 08 00:17:59 http://pastebin.com/d5d1c9744 Sep 08 00:19:07 Ghiottone hmm, mom Sep 08 00:19:31 kristian-m, lol Sep 08 00:19:48 Ghiottone got it Sep 08 00:19:50 kristian-m, what should I do to try it manually? what is the command to send AT codes? Sep 08 00:20:13 Ghiottone are you sure the .py files are the new once? Sep 08 00:20:39 Ghiottone it seems as if you still have an old ProcessInterface.py Sep 08 00:21:06 kristian-m, mmmm the package has been upgraded Sep 08 00:21:19 kristian-m, I'll try a remove/install Sep 08 00:21:21 Ghiottone it should print the subprocess commandline... hmmm Sep 08 00:22:38 Ghiottone have you had the mute option in the audio panel? Sep 08 00:23:25 kristian-m, no mute option Sep 08 00:23:37 kristian-m, mmmmmm Sep 08 00:23:43 kristian-m, that should be 0.6 Sep 08 00:24:01 Ghiottone that you are still using the files from 0.5 - maybe there are in your current directory Sep 08 00:25:29 Ghiottone 0.5 had a hardcoded start of libgsmd-tool in the ProcessInterface - this is why it ignores its parameters - still strange that you've had no ther issues Sep 08 00:25:52 kristian-m, there were some old files in /usr/lib/python ! Sep 08 00:26:00 kristian-m, I'm checking again Sep 08 00:27:55 * kristian-m keeps his fingers crossed... ;-) Sep 08 00:28:16 kristian-m, same results :( Sep 08 00:28:54 Ghiottone but the output should be different now anyway Sep 08 00:29:54 same as before Sep 08 00:30:03 should I check manually what is in that file? Sep 08 00:31:37 Ghiottone you can activate the test at the buttom of the file (comment the last 2 lines in) - then you should at least see if it's used Sep 08 00:33:00 test = test() ? Sep 08 00:33:22 right Sep 08 00:34:48 Ghiottone and the next one behind - that sould activate a test as soon as the file is included into any other python script Sep 08 00:35:08 no change :( Sep 08 00:35:35 I'm doing a find / Sep 08 00:36:46 there is the .pyo Sep 08 00:36:52 should I remove it? Sep 08 00:37:18 Ghiottone oh, yes - that could be it. Sep 08 00:37:49 Ghiottone btw, you could do a release with only th .pyo files - should start up faster Sep 08 00:37:57 i think Sep 08 00:38:05 kristian-m, that changed Sep 08 00:38:21 kristian-m, but now the interface does not show up... Sep 08 00:40:17 Ghiottone the test is still active Sep 08 00:40:43 Test interface Sep 08 00:40:48 Executes subprocess [/bin/sh] Sep 08 00:40:55 is the test blocking? Sep 08 00:41:13 I think I'll restore the # and clean all the pyo Sep 08 00:41:16 brb Sep 08 00:41:36 Ghiottone actually it should not, but who knows what else is happening there... ;-) - cu Sep 08 00:44:11 kristian-m, argh. I have to use my sim. bbl Sep 08 00:45:04 ah kristian-m , is the right icon the -48 one? Sep 08 00:47:04 Ghiottone i think so - btw. the testinterface hangs you're rigth - haven't used it for a while Sep 08 01:02:30 DOES ANYONE KNOW WHY MY SAMSUNG CANT RECOGNISE MY 2GB STORAGE CARD? Sep 08 01:02:31 Guest798: See http://wiki.openmoko.org/wiki/FAQ (and also SH1, SH1_FAQ, Neo1973, Developer_preview and Neo1973_Hardware#GTA01Bv4 pages) (Hopefully these links answers your question. This autotrigger for new nick and question during first 15 minutes done only once.) Sep 08 01:03:16 .oO(propably he spoke uppercase to the samsung) Sep 08 01:03:37 Wow, that was incredibly rude. Asking on a unrelated channel, and yelling... Sep 08 01:03:46 Not to mention the "guest" nick Sep 08 01:08:55 webkit-gtk-0.0+svnr25426-r25427 Sep 08 01:09:14 someone seems to have commited something _in the middle of my do_fetch_ Sep 08 01:09:21 * cesarb hopes that doesn't break the build... again Sep 08 01:10:02 * cesarb wonders how long before webkit has a proprer versioned release, so we can stop having to track a moving target Sep 08 01:10:35 I also wonder why do we have a feedreader when we can't even call out-of-the-box yet, but that's another issue Sep 08 01:12:17 cesarb because its a totally different field of working. Sep 08 01:12:48 the webkit stuff/feedreader is done by zecke, who got a google summer of code for it. Sep 08 01:13:11 * mwester agrees with cesarb's last comment, and points out that it is very simple to remove the feedreader -- which has the side effect of removing a great deal of frustration along with removing a great deal of unnecessary (as of yet) compilation time. Sep 08 01:13:14 roh: the problem is that webkit is slow to compile, and wants to compile _every single time_ I do a build Sep 08 01:13:28 roh: when I'm eager to try out my patch for the gsmd issues Sep 08 01:14:05 why don't you build gsmd instead and just install the ipkg? Sep 08 01:14:17 Oh, btw, when Guest-whatever comes back, tell him that we all know why his samsung won't work, but we refuse to tell him. Sep 08 01:14:18 * kristian-m actually uses the feedreader - you can add rss urls, remove the search bar and recompile it... Sep 08 01:14:43 ... and then what? Sep 08 01:14:47 mwester: yeah, but last time I used my local overlay to change the contents of the rootfs, the way the image was done changed, and I ended breaking my build and having to recompile everything Sep 08 01:15:16 roh: because I don't have a working rootfs and its matching build environment Sep 08 01:15:26 I didn't do that in an overlay. I edited the file in openembedded/packages; it won't get overwritten. Sep 08 01:15:33 hm.. i rebuilt clean yesterday and that was through in about 300min. Sep 08 01:15:54 roh: your computer is probably faster than mine Sep 08 01:16:09 * mwester guesses roh has a dual-core @ about 2.8GHz Sep 08 01:16:24 * cesarb has a single-core @ about 2.0GHz Sep 08 01:16:45 * mwester thinks that cesarb needs an upgrade Sep 08 01:16:47 mwester nope.. c2d t7600 at 2.3ghz with 2g ram.. thats enough if you fetched the sources before Sep 08 01:17:47 Hmm. interesting. Takes me 340 minutes to build, with all sources in the download directory, on a dual-core 2.4GHz system, also 2GB RAM Sep 08 01:17:54 but i mostly recompile from scratch every few days to find hidden bugs with the buildsys or to replicate reports i get Sep 08 01:18:08 with rm_work it also doesnt take too much space Sep 08 01:18:13 mwester: I upgraded like the end of last year (from a "slot A" Ahtlon 750 with 256M RAM) Sep 08 01:18:18 mwester: it's way too soon to upgrade again Sep 08 01:18:19 hehe - I *attempt* to recompile every other day - usually it suceeds about once a week. Sep 08 01:19:08 mwester try checking out after europe went to sleep and before taiwan awakes ;) Sep 08 01:19:28 about 2-4am CEST Sep 08 01:20:13 back Sep 08 01:20:17 kristian-m, much better now :) Sep 08 01:21:35 kristian-m, where do you add rss feed to the rss reader? Sep 08 01:23:07 mwester ah.. and something which i used to make the compile always compareable long.. the bb cache is already built.. (the thing that counts up tasks at start) Sep 08 01:23:44 i used bitbake -g openmoko-devel-image for that... because i need the .dot files anyways from time to time Sep 08 01:24:13 Yeah, that's gotten especialliy bad, as it now connects to SVN and CVS repos as it processes the bb files!!! Sep 08 01:25:12 (as I discovered the other evening when I tried to build a much smaller OE project from a hotel, on their crappy internet connection) Sep 08 01:25:36 mwester: it did that also on 2007.1, so it's not that new Sep 08 01:26:21 Ok, then perhaps I've been lucky for two years building with OE... ? Sep 08 01:26:29 * mwester is skeptical Sep 08 01:26:32 Ghiottone ./openmoko-feedreader2/src/feed-configuration.c Sep 08 01:26:55 Ghiottone nice to hear :-) - are you online with gprs now? Sep 08 01:27:14 mwester: on 2007.1 it was even more interesting... it _printed_ the svn/etc revisions while it was parsing the files Sep 08 01:27:19 kristian-m, not yete Sep 08 01:27:56 http://pastebin.com/d60bdcff Sep 08 01:28:26 thre tries gradually removing the telefone numbere. program seems to ignore that value Sep 08 01:29:31 Ghiottone maybe you'll have to reconnect to the gsm network Sep 08 01:30:41 Ghiottone, with my old phone (a P800) I used to send at+cgdcont=1 Sep 08 01:30:52 s/=1/=5/ Sep 08 01:30:53 Ghiottone meant: Ghiottone, with my old phone (a P800) I used to send at+cgdcont=5 Sep 08 01:31:50 I am really too tired now Sep 08 01:31:58 thanks for the help kristian-m Sep 08 01:32:11 I'll keep trying tomorro Sep 08 01:32:18 tomorrow Sep 08 01:32:22 nite all Sep 08 01:32:42 Ghiottone 'night i thank you. Sep 08 01:33:11 kristian-m, I think I'll do new package tomorrow, as now I'm too tired and I would probably mistake something Sep 08 01:33:33 i know what you mean ;-) Sep 08 01:33:40 * kristian-m also has to sleep, still flashing though Sep 08 01:33:51 nite alle Sep 08 02:39:25 hi there. Sep 08 02:47:27 cesarb - how that scratch rebuild go? Sep 08 02:47:40 s/how/how'd/ Sep 08 02:47:40 dotmef meant: cesarb - how'd that scratch rebuild go? Sep 08 02:47:49 still 2556 of 3331 Sep 08 02:48:31 cesarb - fingers crossed Sep 08 02:48:47 dotmef: I don't see why it would not work Sep 08 02:49:10 dotmef: other than a different build breakage Sep 08 02:49:30 cesarb - yup Sep 08 02:50:23 cesarb - did you do an "rm -rf build/tmp" as before? Sep 08 02:50:43 dotmef: yeah, that's the quickest way to clean the build Sep 08 02:50:55 dotmef: "make clobber" clobbers too much Sep 08 02:51:13 * dotmef thinks he should just bite the bullet Sep 08 02:54:17 * dotmef bites the bullet **** ENDING LOGGING AT Sat Sep 08 02:59:57 2007