**** BEGIN LOGGING AT Thu Oct 22 02:59:56 2009 Oct 22 03:28:05 going to try bytestores 2008.18 image an latest shr kernel on a fresh android sdcard install-current with android "Special" Qi installed Oct 22 03:28:13 2008.12* Oct 22 03:43:59 well nothing worked for a few boots back an forth then android booted an repaired itself:) Oct 22 04:04:36 linux on this? http://en.wikipedia.org/wiki/Motorola_Droid Oct 22 04:04:39 hardware looks nice Oct 22 04:20:07 then it appears to have installed itself to nand after i flashed the 2008.12 image to nand Oct 22 04:47:59 2008.12 Bytestor's with latest shr-u kernel ,,Qi Oct 22 05:02:15 ndnihil: hardware is def cool. more than anything, I dig the advertisement (and that's hard for me to say. I'm a commercial muter.) Oct 22 05:13:56 slaxxer: the current android releases require installation to NAND. the version of Qi will still let you boot from a uSD that has been formatted with FAT or EXT2. I'm using android and qtmoko on one uSD, then I have SHR on another. Oct 22 05:56:48 mostly_useless: Qi doesn't boot from fat Oct 22 06:07:36 morning Oct 22 07:05:31 PaulFertser: Qi does when you ask nicely. Oct 22 07:05:45 beer: morning. Oct 22 07:07:23 mostly_useless: what makes you think that? Oct 22 07:26:24 PaulFertser: because you catch more flies with honey. Oct 22 07:26:42 mostly_useless: moot argument Oct 22 07:28:21 PaulFertser: because I'm able to perform an install of the new weekly android builds off of a fat formatted card, and also able to boot qtmoko on an ext2 formatted card. Oct 22 08:03:04 http://xkcd.com/357/ Oct 22 08:16:32 molgrum: Qi ignores fat partition Oct 22 08:16:42 Fuck, he left Oct 22 08:16:50 He didn't prove anything by code Oct 22 09:54:46 Hi PaulFertser, mickey|zzZZzz Oct 22 09:54:54 is there a command to find the cell id neighboring Oct 22 09:54:56 ? Oct 22 09:56:16 panicking: hey Oct 22 09:56:26 sorry for the wheel question Oct 22 09:56:38 panicking: you left too early the previous time so i didn't manage to answer. Oct 22 09:56:50 Ok, sorry Oct 22 09:56:52 panicking: could you please repeat that question, i'm afraid i forgot it. Oct 22 09:56:53 my wife Oct 22 09:57:05 panicking: np, just please repeat the question :) Oct 22 09:57:41 #1024 issue I think Oct 22 09:58:02 Oh yes How to test it Oct 22 09:58:03 ? Oct 22 09:58:12 I have reregistration here in valencia Oct 22 09:58:29 panicking: ah Oct 22 09:58:33 where I work but I don't understand how it is so frequently Oct 22 09:58:40 it not depends on the signal strenght Oct 22 09:58:44 ? Oct 22 09:58:55 I can't remain in deep sleep mode Oct 22 09:59:16 panicking: http://wiki.openmoko.org/wiki/Neo_1973_and_Neo_FreeRunner_gsm_modem#AT.25EM Oct 22 09:59:29 panicking: (neighborhood cell info) Oct 22 09:59:51 Thanks Oct 22 09:59:52 :) Oct 22 10:00:29 panicking: it's commonly believed that all devices are affected by #1024 in certain conditions. Oct 22 10:00:40 * mjr came in, noticed "panicking: [url]" and expected something panicworthy to be there Oct 22 10:00:44 what a disappointment Oct 22 10:01:19 But when the modem go in de-registration status Oct 22 10:01:29 and you send an AT%SLEEP=2 Oct 22 10:01:53 the system must re-reregister automatically to the net Oct 22 10:01:59 or it's necessary Oct 22 10:02:04 to force re-registration Oct 22 10:02:34 panicking: no, the modem should reregister automatically Oct 22 10:03:12 But It can takes time Oct 22 10:03:16 it is must fast Oct 22 10:03:31 panicking: if you see sudden loss of registration then it's most probably due to #1024. Oct 22 10:03:31 to ask for a new registration Oct 22 10:04:20 panicking: why do not you want to set %SLEEP=2 as default? Those who did the fix will switch it to 4. Oct 22 10:04:34 panicking: the only sane approach Oct 22 10:04:51 panicking: because you can't do anything when the device is suspended. Oct 22 10:05:54 so when the device is supend does't resume on network lost ? Oct 22 10:05:57 panicking: it seems you want to implement something like adaptive mode when you switch off deep sleep upon "sensing" the bug. Now imagine device is suspended. And i'm waiting for a call walking from somewhere. Oct 22 10:06:29 panicking: suddenly i appear in a place where #1024 is surfaced. But the device is still suspended. My call is lost. Oct 22 10:07:09 ok Oct 22 10:07:31 panicking: you can leave unsol that reports registration status enabled before suspending but then you'll have to use auto-suspend by timeout all the time. Oct 22 10:08:36 Ok, that is a big problem Oct 22 10:08:49 I must discuss on android mailing list Oct 22 10:09:20 But for me is not a big problem Oct 22 10:16:37 panicking: SHR disabled deep sleep by default Oct 22 10:18:46 ok Oct 22 10:18:52 Thanks Paul Oct 22 10:19:01 I hope that can meet you in the near feature Oct 22 10:19:10 you are a very nice man Oct 22 10:19:50 panicking: thanks a lot for you kind words Oct 22 10:20:04 panicking: i'm looking forward to see you too :) Oct 22 10:20:54 panicking: i'm glad i was able to help you at least a little bit. Oct 22 10:57:27 freesmartphone.org: 03mickey 07cornucopia * rb4c2e04416da 10/fsogsmd/src/lib/atcommands.vala: fsogsmd: atcommands: allow encoding the +CRSM payload result in "", some modems do that (palm pre, for a start) Oct 22 12:17:21 freesmartphone.org: 03mickey 07cornucopia * r20798692daab 10/fsogsmd/ (5 files in 3 dirs): fsogsmd: add 3rdparty directory for C code we 'borrow' from other projects Oct 22 12:46:37 bonjour Oct 22 13:23:28 openmoko: 03werner * r5701 10/developers/werner/cncmap/ (8 files in 2 dirs): Project a line on the Z map. Work in progress. Oct 22 13:24:25 freesmartphone.org: 03mickey 07cornucopia * rcd92e547882c 10/fsogsmd/ (12 files in 4 dirs): fsogsmd: add 3rdparty code for certain format conversions Oct 22 14:16:17 freesmartphone.org: 03mickey 07cornucopia * reef5bef22917 10/fsogsmd/src/lib/ (atcommand.vala atcommands.vala channel.vala modem.vala): fosgsmd: run through a sequence of test commands to negotiate the 'best' charset we can use. Oct 22 14:19:40 freesmartphone.org: 03mickey 07libeflvala * rf01541daddec 10/vapi/elm.vapi: elm.vapi: bind object focus stuff Oct 22 14:34:49 freesmartphone.org: 03mickey 07cornucopia * rcd71bae64222 10/libfsotransport/fsotransport/commandqueue.vala: libfsotransport: move command queue timeout into command class; this permits per-command timeouts Oct 22 14:35:22 freesmartphone.org: 03mickey 07cornucopia * r06ea09afdfd2 10/fsogsmd/src/lib/atcommand.vala: fsogsmd: catch up with changes in libfsotransport command queue Oct 22 15:20:33 PaulFertser, sorry I have other question for you Oct 22 15:20:36 do you have time? Oct 22 15:20:54 panicking: sure :) Oct 22 15:21:06 flowcontrolled Oct 22 15:21:16 Ok, this is what I think: Oct 22 15:21:28 - the rts must be change by the kernel during suspend Oct 22 15:21:41 - but deactivate by the user application Oct 22 15:22:00 panicking: not exactly Oct 22 15:22:02 Because a task does't know the suspend istant Oct 22 15:22:15 but it knows when is again run Oct 22 15:22:51 panicking: you want to disable unsolicited replies first. Then you do "echo 1 > flowcontrolled"; then you suspend; then you resume and do "echo 0 > flowcontrolled" and this way you're guaranteed to not lose anything. Oct 22 15:23:20 This is true when you know Oct 22 15:23:25 the suspend istant Oct 22 15:23:26 but Oct 22 15:23:55 Suppose that you ask to go to suspend but the system takes time to go to suspend Oct 22 15:24:11 (in my cases I dont' know if the system will be suspended) Oct 22 15:24:23 you can loose data Oct 22 15:24:25 is it correct Oct 22 15:24:25 ? Oct 22 15:24:48 Yes, for standard distrib Oct 22 15:24:56 it can be controlled by software Oct 22 15:25:05 but in android the suspend istant can be defferable Oct 22 15:25:16 by wake lock acquisition of other subsystem Oct 22 15:25:36 so I propose to create a new flag Oct 22 15:25:45 panicking: can you please elaborate more on that "wake lock"? Oct 22 15:25:54 ok, Oct 22 15:26:09 echo mem > /sys/power/state Oct 22 15:26:17 the system start to go suspend Oct 22 15:26:35 A task in the middle acquire a wake lock, the system will go to suspend after wakelock Oct 22 15:26:36 release Oct 22 15:26:44 or echo on > /sys/power/state Oct 22 15:27:00 so the final suspend can be delayed for few or more millescond Oct 22 15:27:20 Basically I don't know if I can suspend the system at t instant Oct 22 15:27:38 but for sure the display will be shutdown immidiately Oct 22 15:27:42 panicking: you won't lose any data from the modem. As soon as you do 'echo 1 > flowcontrolled' the modem will stop sending data and will keep it in its internal buffer. Oct 22 15:28:09 Yes, I know but there is any good istant to do this Oct 22 15:28:23 the only right instant is in the suspend calll Oct 22 15:28:31 of the gta02_pm_gsm call Oct 22 15:28:41 Is that true? Oct 22 15:29:13 panicking: i'm not sure i understand what problem do you have in mind, sorry Oct 22 15:29:17 This is for sure a good instat. Because user space process are freeze and the wifi Oct 22 15:29:27 sorry for wifi Oct 22 15:29:44 and the user space can't read any message Oct 22 15:29:48 panicking: the idea of "flowcontrolled" node is exactly to force RTS before userland is frozen. Oct 22 15:29:50 from the serial device Oct 22 15:30:27 ok, but why do it in the suspend path Oct 22 15:30:27 of the gsm_pm platform data Oct 22 15:30:27 ? Oct 22 15:30:27 You don't loose anything Oct 22 15:30:45 you are sure that you stop the receiveing because nobody can take data from serial Oct 22 15:30:59 I must change the kernel in this way to avoid data loose Oct 22 15:31:02 No Oct 22 15:31:09 because I don't know when this happen Oct 22 15:31:12 mwester, ? Oct 22 15:31:14 You will lose data if you make the change you are proposing. Oct 22 15:31:21 panicking: can you please read this http://trac.freesmartphone.org/ticket/31#comment:3 ? Oct 22 15:31:41 yes I'm reading Oct 22 15:31:49 mwester, can you give me an exmple Oct 22 15:31:53 of data loosing Oct 22 15:31:54 mwester: hey :) your favourite topic Oct 22 15:33:13 ok Oct 22 15:33:17 I'm reading Oct 22 15:33:23 it's in the ticket Oct 22 15:34:21 Assuming your proposed solution: user space issues suspend, but does not flow-control. Kernel begins to suspend, but it takes 1 ms because of some wake lock. During this 1 ms, data arrives from GSM -- because GSM is *not yet* flowcontrolled, the data is received and is in the serial driver buffers. The kernel suspends. At this point, the message that arrived from the GSM will be delivered when some other event finally wakes the phone. Oct 22 15:35:28 So the current approach is that user-space issues the flow-control. Then as suspend happens, the GSM will be forced to issue a serial interrupt, which will abort the suspend if data arrives while the kernel is suspending. Oct 22 15:35:29 ok Oct 22 15:35:40 it's a correct scenario Oct 22 15:36:09 And if the data was arrived after you flowcontrol the modem but before kernel suspends, it will simply refuse to suspend. Oct 22 15:36:29 s/was // Oct 22 15:36:29 PaulFertser meant: And if the data arrived after you flowcontrol the modem but before kernel suspends, it will simply refuse to suspend. Oct 22 15:38:05 It suspends but wake up immiditally for the interrupt Oct 22 15:38:30 that's what I suggested, but PaulFertser disagreed on that afaik Oct 22 15:39:06 iirc kernel has some nasty useless "abort suspend action" hooks Oct 22 15:40:00 not useless; used by the GSM flowcontrol logic Oct 22 15:40:06 PaulFertser: excuse me if my memory is wrong on that, another time I haven't stored to source with the info Oct 22 15:40:08 not nice, not elegant, but useful. Oct 22 15:40:19 s/to/the/ Oct 22 15:40:19 DocScrutinizer meant: PaulFertser: excuse me if my memory is wrong on that, another time I haven't sthered to source with the info Oct 22 15:40:33 bah Oct 22 15:40:58 mwester: how? Oct 22 15:41:06 how what? Oct 22 15:41:27 how could such a nonsense be useful? Oct 22 15:41:49 * mwester smacks closed-minded people over the head with a dead fish, and refuses to continue the discussion. Oct 22 15:41:59 hehehe Oct 22 15:42:02 I explained it. Read above. Oct 22 15:42:08 :p Oct 22 15:42:46 mwester: come on, I really don't see the use of speeding up a rare race condition by a few ms Oct 22 15:43:56 isn't it completely useless to abort suspend rather than simply resuming in zero time Oct 22 15:43:58 ? Oct 22 15:44:40 DocScrutinizer, where can I find an example of such hooks Oct 22 15:44:41 ? Oct 22 15:45:24 it was potentially 10s and nearly 100 ms at one point (glamo and SD card suspend/resume), and the window from user-space issuing flow-control to the begin of suspend is actually quite long -- up to a full second in my observations. So the possibility is reasonably high, and the time-saved used to be (assuming someone fixed the glamo suspend/resume) significant, and the total code added was a couple of lines. Seems quite simple, DocScrutinizer -- Oct 22 15:45:24 but you don't have to agree. Oct 22 15:45:29 GSM is forced to issue a irq anyway, and we should resume on level IRQ, not edge Oct 22 15:45:29 DocScrutinizer: it's useful only if the GSM interrupt is edge. If the device is configured to wakeup on level it's not useful i think. Oct 22 15:45:50 mwester: nope, DocScrutinizer is talking about another issue. Oct 22 15:45:53 ok Oct 22 15:46:09 i woul like to design a my own gsm mobile phone,specs of NeoFreerunner is over my requirement ,my requirement would just (TFT 640x480,gsm+gprs module,and a bluetooth) ,is there any way to assemble one any good hw providers ? Oct 22 15:46:14 mwester: there's some very late in suspend check if there was interrupt or not. And if there was, the kernel refuses to suspend altogether. Oct 22 15:46:19 Until the irq is cleared. Oct 22 15:46:42 s/woul/would/ Oct 22 15:46:43 solo|moko meant: i would like to design a my own gsm mobile phone,specs of NeoFreerunner is over my requirement ,my requirement would just (TFT 640x480,gsm+gprs module,and a bluetooth) ,is there any way to assemble one any good hw providers ? Oct 22 15:46:53 I added a check in there for the GSM... isn't that what we speak of? That was added because we were missing interrupts, or so we thought. Oct 22 15:46:55 mwester: DocScrutinizer thinks this measure is useless because the device should resume immediately anyway in this scenario. Oct 22 15:47:15 And if it doesn't then probably irq or wakeup is not configured properly. Oct 22 15:47:34 exactly Oct 22 15:47:41 I don't know that anyone ever really did figure out if we have a window where we can miss the IRQ from the GSM as we suspend. Oct 22 15:47:52 The suspicion was that such a window existed. Oct 22 15:47:59 mwester: level irq can't be missed Oct 22 15:48:24 as there isn't such a thing like "window" Oct 22 15:48:38 Right - by hardware. the suspcision was that the interrupt was disabled or ignored at some point during the suspend/wake process. Oct 22 15:48:51 solo|moko: i think lindi- planned something a bit like that, you can ask him. Oct 22 15:48:54 So the window would have been introduced by software, not hardware. Oct 22 15:48:59 then it should latch up Oct 22 15:49:09 software can clear the latch. Oct 22 15:49:16 but we argue theories. Oct 22 15:49:20 that's a clear bug then Oct 22 15:49:23 PaulFertser: thank you,btw is he online now ? Oct 22 15:49:29 solo|moko: he always is :) Oct 22 15:49:34 o_O Oct 22 15:49:40 lindi-: ping! Oct 22 15:49:57 solo|moko: but his plan was a bit less ambitious, he wanted just "simple" uC, minimal display and usb 3g stick. Oct 22 15:50:10 The check was simple to add, and it ensured that there we no more missed IRQs so people counldn't point fingers at that, and had to fix the driver or the user-space problems. :) Oct 22 15:50:23 solo|moko: as to developing a real phone, OM project proved it's damn hard even for a real company. Oct 22 15:50:25 oh! well he might be able to help me out i guess Oct 22 15:50:47 solo|moko: possibly the Gizmo FLOW DIY is close to what you want? Oct 22 15:50:55 Someone can remove it if they find that code offensive. I don't care anymore; I'll just add it back in on my kernels, since I don't care to debug the low-level suspend and u-boot code. Oct 22 15:51:04 * solo|moko googling gizmo flow Oct 22 15:51:14 mwester: ok, I'm fine with htat. Just thought it's useless Oct 22 15:51:24 PaulFertser: well, it would be difficult for me too Oct 22 15:52:06 anyway, as long as it works... not my problem Oct 22 15:52:34 cya Oct 22 15:52:48 solo|moko, gizmo start to design flow2 Oct 22 15:53:00 if you need some info I can give you Oct 22 15:53:31 panicking: sure :) i would love to,atm i'm trying to figure out the details of gizmo Oct 22 15:53:44 flow^ Oct 22 15:54:03 Ok, the idea is to design a new device based on the imx327 Oct 22 15:54:05 cpu Oct 22 15:54:48 hm,cool,do you have specs details / Oct 22 15:55:39 not yet, we are discussing about what we want Oct 22 15:56:32 wow gizmo flow is exactly wht i wanted ty Weiss :) Oct 22 15:59:09 no problem :) Oct 22 15:59:48 solo|moko, what are the requirement of your device Oct 22 16:00:55 mwester: it seems I remeber a really braindamaged code for IRQ-handlers though, that does edge trigger on both edges, then POLLS(!!) the IRQ line, and latches up state internally. So I'd rather guess we see a lot of misconfig and races there. So probably your patch wrt that suspend-abort is last resort unless someone cleans up the elemetary IRQ-handler code eventually Oct 22 16:02:15 DocScrutinizer: probably that was code for "buttons" :) Oct 22 16:02:24 might be Oct 22 16:03:50 anyway, I have to care about fuel for my brain. running really dry Oct 22 16:03:55 bbl Oct 22 16:04:03 cu Oct 22 16:20:39 mwester, so it's clear the we need to do it in userspace because there is a window Oct 22 16:21:32 Yes, that part is clear. Oct 22 16:21:55 You won't lose data; it just may be deferred until something else wakes the phone. Oct 22 16:33:37 openmoko: 03werner * r5702 10/developers/werner/cncmap/zmap/ (Makefile zline.c zmap.c): (log message trimmed) Oct 22 16:33:37 openmoko: Fixed some of the algorithms (in progress). Oct 22 16:33:37 openmoko: - zmap/zmap.c (comp): sorting works so much better if we don't just compare Oct 22 16:33:37 openmoko: things with themselves Oct 22 16:33:37 openmoko: - zmap/zline.c (next_closest): renormalize the "t" coordinate Oct 22 16:33:41 openmoko: - zmap/zline.c (next_closest): it's okay to go "back", since we may have Oct 22 16:33:43 openmoko: multiple boundaries at the same place Oct 22 17:31:42 freesmartphone.org: 03mickey 07cornucopia * r8ace361d7b1b 10/fsogsmd/src/lib/atcommands.vala: fsogsmd: +COPS=?: apparantly both the long name and the short name can be missing Oct 22 18:44:18 anybody knows, which patches from openembedded against gcc are important for vanila gcc to build for openmoko? Oct 22 18:44:59 hey i just came to see what gps application could be used with an openmoko running on QT Oct 22 18:46:39 anybody can help? Oct 22 18:47:07 qtextended? Oct 22 18:50:04 xyzz: yes Oct 22 18:51:11 you can use tangogps, but first, you need to start X server ... Oct 22 18:51:33 I think tehre should be an option in all taht menus for that Oct 22 18:51:55 xyzz: okay. checking.. Oct 22 18:52:07 xyzz: what could be the tangogps .opk files? Oct 22 18:58:12 xyzz: i just cant spot how to start the x server. could someone help? Oct 22 18:58:31 xyzz: i can say that debian's gcc works ifne Oct 22 19:00:11 sjd I dont have qtextended anymore around, sorry Oct 22 19:00:27 xyzz: okay :( Oct 22 19:00:28 xyzz: what is important when you build for openmoko is to use armv4t and not armv5t. Many precompiled toolchains have libgcc compiled for v5t so these toolchains are suitable only for kernel or u-boot (that do not need _any_ function from libgcc). Oct 22 19:01:40 PaulFertser that may be ... Im using fedora native gcc on armv5, building armv4{t,l,tl} binaries, and all of them at some point and with illegal instruction error Oct 22 19:01:52 xyzz: looks exactly what i describe Oct 22 19:02:31 xyzz: if you download vanilla gcc and compile it with correct options it will work fine. Oct 22 19:03:34 what am I trying to do is rebuild base fedora arm (they have only armv5) port for openmoko Oct 22 19:04:45 but good to know, that there are no extension for openmoko over vanilla gcc Oct 22 19:07:57 xyzz: can you explain what advantages does fedora have in your opinion? Oct 22 19:08:04 freesmartphone.org: 03mok 07libframeworkd-glib * rb9cf016a126a 10/src/ (5 files in 3 dirs): Oct 22 19:08:04 freesmartphone.org: opimd: add handler for Contacts.NewContact signal Oct 22 19:08:04 freesmartphone.org: Signed-off-by: Klaus Kurzmann Oct 22 19:08:05 freesmartphone.org: 03mok 07libframeworkd-glib * rc5ead9c20f19 10/src/opimd/ (3 files in 2 dirs): Oct 22 19:08:05 freesmartphone.org: opimd: add Message.Update method Oct 22 19:08:07 freesmartphone.org: Signed-off-by: Klaus Kurzmann Oct 22 19:09:09 PaulFertser solid packaging and buildsystem Oct 22 19:17:21 xyzz: do not forget that you really want to use softfloat Oct 22 19:17:30 PaulFertser but I dont want to end this up in some endless religious flamewar ... Oct 22 19:17:39 good to know Oct 22 19:18:52 freesmartphone.org: 03mok 07libframeworkd-glib * rcefa825fc7ee 10/src/opimd/ (2 files): Oct 22 19:18:52 freesmartphone.org: opimd: give Message.Update the needed field_data hashtable Oct 22 19:18:52 freesmartphone.org: Signed-off-by: Klaus Kurzmann Oct 22 19:32:52 hey. i have another moko with SHR. tango is installed. was trying to install gllin but unfortunately there was no space to install gllin. does anyone suggest a way out? im looking for the gps to be run Oct 22 19:33:22 sjd: i hope by moko you mean gta01? Oct 22 19:33:36 PaulFertser: yes Oct 22 19:33:44 sjd: do opkg list-installed and find some unnecessary package, remove it to free space Oct 22 19:34:16 trying Oct 22 19:34:19 sjd: alternatively, you can try to install gllin to your SD and alter PATH accordingly. Oct 22 19:35:12 PaulFertser: you mean i should modify which path? Oct 22 19:35:25 sjd: environment variabl Oct 22 19:36:00 PaulFertser: i dont get you. \ Oct 22 19:36:20 sjd: hm, are you *nix user? Oct 22 19:36:44 PaulFertser: im new to openmoko. so not familiar with the terms. Oct 22 19:36:56 PaulFertser: yes im a gnu/linux user Oct 22 19:38:07 sjd: http://en.wikipedia.org/wiki/PATH_(variable) Oct 22 19:40:33 sjd: /media/card/gllin instead of mere gllin Oct 22 19:41:23 maybe even PATH=/media/card:$PATH gllin Oct 22 19:42:23 how do I install gllin into the card? Oct 22 19:42:54 sjd: i guess there was some option to opkg. Oct 22 19:43:08 sjd: at least ipkg had it, it was something related to ipkg-link or whatever. Oct 22 19:43:29 sjd: if you get desperate, just "ar x" the package and untar data.tar.gz manually to whatever you want. Oct 22 19:43:29 PaulFertser: checking.. Oct 22 19:43:52 PaulFertser: then? Oct 22 19:44:13 PaulFertser: what about the dependencies in that case? Oct 22 19:44:24 sjd: i guess it has none Oct 22 19:44:50 sjd: that proprietary shit needs a separate OABI-chroot anyway, so it has everything needed inside the package. Oct 22 19:45:11 PaulFertser: it says something like libstdc++6 Oct 22 19:46:31 sjd: i'd just try and solve problems when i face them, not trying to predict too much in this case. Oct 22 19:46:32 PaulFertser: lol, that's why I suggested to actually change the path Oct 22 19:47:11 DocScrutinizer: i suggested that too ;) Oct 22 19:47:16 yep Oct 22 19:47:27 but not me in first proposal Oct 22 19:48:27 DocScrutinizer: btw, have you already vaccinated against seasonal flu? Oct 22 19:48:53 yep, seems I got a cold since then :-P Oct 22 19:49:27 so? any hope to get glinn installed and working?\ Oct 22 19:49:35 sjd: yes, just do that Oct 22 19:50:08 sjd: btw, i think the latest SHR image is small enough to fit NAND and it has enough space for gllin i hope. Oct 22 19:50:17 sjd: so you might want to refresh. Oct 22 19:50:40 sjd: alternatively, just make your rootfs on uSD, then you'll have no problems with free space. Oct 22 19:51:33 PaulFertser: I am having the latest SHR. after installing dependency in the internal memory, there is no space for gllin Oct 22 19:51:39 -d : use for opkg root directory Oct 22 19:51:46 sjd: aha, thanks for reporting a bug then Oct 22 19:52:56 for installation, removal, updating... Oct 22 19:52:57 so can I have opkg install -d /media/mmc blablabla.ipk Oct 22 19:53:15 will that install all files to the mmc, if its mounted? Oct 22 19:55:08 sjd: no idea, never used that opkg option Oct 22 19:55:44 but probably you'll want /media/card anyway Oct 22 19:56:26 or better even, chack what your partitions on uSD are, select one that's ext2/3, and install to that Oct 22 19:56:34 check Oct 22 20:01:03 DocScrutinizer: hav an ext2 Oct 22 20:01:30 so, does the command I typed earlier work? Oct 22 20:07:19 sjd: i'd try to just uninstall the libraries that gllin required as dependencies. I'm not sure it needs them. Oct 22 20:08:44 opkg can't install it without the dependencies Oct 22 20:10:16 opkg --help Oct 22 20:12:50 freesmartphone.org: 03mok 07libframeworkd-glib * r0f6f9d3865e3 10/src/opimd/dbus/messages.h: Oct 22 20:12:50 freesmartphone.org: regenerate opimd/dbus/messages.h from the xml spec Oct 22 20:12:50 freesmartphone.org: Signed-off-by: Klaus Kurzmann Oct 22 20:40:55 sjd: glling pakage installs on /home/root/gillin/ see http://wiki.openmoko.org/wiki/Gllin Oct 22 20:49:15 so you could simply mount or link that dir from uSD, prior to installing, I guess Oct 22 21:10:51 im looking for Oct 22 21:11:27 sorry to hit enter so fast. i just got few more dependency issues like libc6 and libgcc1 Oct 22 21:12:11 what should i do now? Oct 22 21:16:25 http://cgi.ebay.co.uk/openmoko-neo-freerunner-neo-1973-neo1973-lcd-screen_W0QQitemZ280407990325QQcmdZViewItemQQptZPDA_Accessories?hash=item41499e6035 Oct 22 21:16:40 If you've broken your screen, or know someone who has Oct 22 21:19:36 :) Oct 22 21:20:00 Or you plan on breaking it in the future. Oct 22 21:20:14 :| Oct 22 21:23:09 openmoko: 03werner * r5703 10/developers/werner/cncmap/zmap/ (zline.c zmap.c): Oct 22 21:23:09 openmoko: Make zline work as designed. Now, improve the design ... Oct 22 21:23:09 openmoko: - zmap/zline.c (zline): if selecting the last point, make sure we don't swap Oct 22 21:23:09 openmoko: back the point we're trying to get rid of Oct 22 21:23:09 openmoko: - zmap: cleaned up debugging output Oct 22 21:28:32 SpeedEvil: please compare to a real LCM! This ebay thing looks somewhat strange. Missing FPC on bottom left. And there's a weird black patch (or censor bar) where the other end of the FPC should connect to ts-panel on upper surface Oct 22 21:29:50 though it actually *seems* the ts panel is mounted on top of the LCM Oct 22 21:32:56 sorry - I'm assuming hte auction description is correct Oct 22 21:34:41 well, you *could* retrofit the old FPC to a new LCM. You even *could* move the ts panel from old broken screen to a new LCD. The auction doesn't say anything about that Oct 22 21:35:38 they are just mentioning LCD (not LCM!). And I found those strange differences between the photo an a real display in my left hand Oct 22 21:36:07 where do i get few lib files for solving dependencies for my moko Oct 22 21:52:23 hi.. is there any other OS that works well in neo1973? Oct 22 21:52:29 other than TSR Oct 22 21:52:52 sorry SHR Oct 22 21:54:01 I guess debian. Ask our debian wizard PaulFertser Oct 22 21:55:01 will gps work in it? Oct 22 21:55:09 DocScrutinizer: our Debian wizard is lindi- and other guys. I'm just a user. Oct 22 21:55:13 Guest46328: sure Oct 22 21:56:14 does it need any extra daemon to work? Oct 22 21:57:39 Guest46328: if you use full FSO, no Oct 22 21:58:51 will full fso run from the internal memory? Oct 22 22:01:47 Guest46328: i'm afraid it's not easy to fit debian to nand. Oct 22 22:02:04 Guest46328: i guess almost everybody runs from uSD Oct 22 22:04:01 where can i get the files from? its for neo1973 Oct 22 22:07:30 Guest46328: i think debian supports both gta01 and gta02 Oct 22 22:17:31 freesmartphone.org: 03mickey 07cornucopia * r3c294459912d 10/fsogsmd/src/lib/ (atmediators.vala channel.vala modem.vala): fsogsmd: support multiple modem state init sequences, all customizable via configuration Oct 22 22:23:54 after removing android reinstalling uboot an shr im getting "not enough space left on the device" Oct 22 22:33:01 freesmartphone.org: 03mickey 07cornucopia * r1a969d896c12 10/fsogsmd/src/lib/consts.vala: fsogsmd: name unknown phonebooks as unsupported, but allow to query them Oct 22 23:27:59 openmoko: 03werner * r5704 10/developers/werner/cncmap/zmap/ (zline.c zmap.c): Oct 22 23:27:59 openmoko: More algorithm fixes. We should now be good, except for some corner cases. Oct 22 23:27:59 openmoko: - zmap/zmap.c (zmap_point): treat this as a linear equation with two unknowns. Oct 22 23:27:59 openmoko: The previous algorithm only worked if the points formed a triangle with a Oct 22 23:28:00 openmoko: right angle and the first point was at that angle. Oct 22 23:28:01 openmoko: - zmap/zline.c (zline): we still have to consider "old" points, as they may Oct 22 23:28:03 openmoko: replace different points as we crawl the line Oct 22 23:52:28 openmoko: 03werner * r5705 10/developers/werner/cncmap/zmap/zmap.c: Oct 22 23:52:28 openmoko: - zmap/zmap.c (zmap_point): handle the special case of all the closest points Oct 22 23:52:28 openmoko: being on a single line, so we don't get a plane Oct 22 23:52:35 freesmartphone.org: 03lukpank 07libeflvala * r54ebe67bd02f 10/vapi/ (edje.vapi evas.vapi): fix free_function of Evas.Object and Edje.Object Oct 22 23:52:36 freesmartphone.org: 03mickey 07libeflvala * r5121eb74467b 10/vapi/elm.vapi: elm.vapi: add middle_bring_in for genlist Oct 22 23:52:37 freesmartphone.org: 03lukpank 07libeflvala * r0f3a23c0cbd7 10/vapi/elm.vapi: elm.vapi: bind Scroller.policy_set Oct 22 23:52:38 freesmartphone.org: 03lukpank 07libeflvala * r2663d17f8872 10/vapi/elm.vapi: elm.vapi: ListItem.label_get and Entry.entry_get return unowned string Oct 23 01:02:42 opkg upgrading:S Oct 23 01:25:28 How usable is the Freerunner as a phone *really* ? Oct 23 01:31:06 That's rather an open question. There are some who didn't consider the iPhone to be usable until recently, when AT&T finally enabled MMS. Oct 23 01:34:30 still nearly useless Oct 23 01:34:46 its a dev. toy Oct 23 01:36:03 we really need to know how many freerunners are left Oct 23 01:36:10 unsold Oct 23 01:37:07 then maybe a plan can be formulated sooner to build new better hardware for this case **** ENDING LOGGING AT Fri Oct 23 02:59:57 2009