**** BEGIN LOGGING AT Fri Oct 11 02:59:58 2013 Oct 11 03:21:51 I am looking to speed up my computer and get all the crude off of it. What channel do I need for that Oct 11 03:22:21 #ubuntu? Oct 11 03:22:29 *win7? Oct 11 03:22:39 #ssd Oct 11 03:22:41 #debian? Oct 11 03:23:37 Hurrian: you missed the idea to speed it up by deleting stuff. However futile Oct 11 03:24:25 too tired to optimize my system nowadays, threw everything on a 128GB SSD and it all 'just works' now. Oct 11 03:25:10 optimize my system... lemme think! maybe get rid of friggin polypaudio? Oct 11 03:25:18 avahid Oct 11 03:25:52 any damn tracker, however it's called (akonadi prolly, nowadays) Oct 11 03:27:09 but honestly with an i5 and a 250GB SSD and a 2TB HD I don't feel like I need much "optimization" either Oct 11 03:30:54 http://privatepaste.com/64cad1e071 ;-P Oct 11 03:31:29 fwiw even a 32GB SSD to keep / on is way, way more than enough, just keep /home on a rust drive Oct 11 03:32:50 saturn:/ # df -h ~jr Oct 11 03:32:51 Filesystem Size Used Avail Use% Mounted on Oct 11 03:32:53 /dev/sdc3 867G 193G 673G 23% /home Oct 11 03:33:43 saturn:/ # df -h / Oct 11 03:33:45 Filesystem Size Used Avail Use% Mounted on Oct 11 03:33:46 /dev/sda2 79G 17G 59G 23% / Oct 11 03:36:24 saturn:/ # time du -xsh / Oct 11 03:36:26 16G / real 0m1.305s user 0m0.132s sys 0m0.633s Oct 11 03:37:12 the latter one is quite impressive, eh? ;-) Oct 11 03:40:16 btw I don't think you can find any good fast 32GB SSD nowadays Oct 11 03:40:24 prolly starts at 160 Oct 11 03:41:11 ooh, I missed read Oct 11 03:42:11 saturn:/ # time dd of=/dev/null if=/ssd-aux/x bs=50M count=100 Oct 11 03:42:12 100+0 records in Oct 11 03:42:14 100+0 records out Oct 11 03:42:15 5242880000 bytes (5.2 GB) copied, 12.6627 s, 414 MB/s Oct 11 03:44:01 funny how write is faster than read Oct 11 03:45:32 might have one GB of write buffer Oct 11 03:50:55 how much NAND would you want on Neo900? Oct 11 03:59:42 DocScrutinizer05: ideally? 1x32/64GB PCIe SSD Oct 11 04:01:25 realistically? probably 32GB of SATA MicroSSD Oct 11 04:01:47 if this could be done by interleaving 2x16GB chips, better. Oct 11 04:04:41 DocScrutinizer05: what display connector is Neo900 planning to use, by the way? Oct 11 04:05:13 same as on N900 or is the design free to use something new? Oct 11 04:07:30 Since the intent is to use the same display then one would be using the same display connector I would have thought Oct 11 04:12:07 sorry, I asked for NAND, not for eMMC Oct 11 04:13:33 how much does the N900 have? Oct 11 04:13:43 err 256MB, no? Oct 11 04:13:46 DocScrutinizer05: ah, nand, as in what comes with a OneNAND? Oct 11 04:13:58 well, guess we could get the cheapest stuff you can get your hands on Oct 11 04:14:19 and dump fremantle into the emmc/sd/usb otg/etc Oct 11 04:15:34 RAM would be the big elephant in the room - put savings from lower NAND storage to buy a OneNAND unit with more memory Oct 11 04:16:02 we can't get OneNAND with >512MB RAM Oct 11 04:16:40 So 512MB is the most the Neo900 can possibly have? Oct 11 04:16:40 so I guess we'll use a 1GB of RAM-only PoP chip, and place a discrete NAND chip next to SoC Oct 11 04:16:47 oh ok Oct 11 04:16:54 sounds good Oct 11 04:17:41 DocScrutinizer05: does the planned SoC support DDR3/DDR3L memory? Oct 11 04:17:44 thus I asked "how much NAND would you consider OK?" Oct 11 04:17:51 nope Oct 11 04:17:55 aww Oct 11 04:18:56 well, I dunno if it *works* with DDr3, but it *supports* LPDDR1 only Oct 11 04:19:14 DocScrutinizer05: 32MB sounds healthy for NAND Oct 11 04:19:27 meh, nope Oct 11 04:19:31 kernel/bootloader + tiny command line system? Oct 11 04:19:37 that's a regression from N900 Oct 11 04:20:54 512MB (N9/50 size)? Oct 11 04:21:07 sounds better Oct 11 04:21:38 too bad that adding more RAM is so hard :/ 2GB like the blackberry q10 would be awesome Oct 11 04:22:06 omap37xx has an addr space for RAM of only 1GB Oct 11 04:22:50 by no means we can get any higher than that Oct 11 04:23:07 unless we go for omap5 Oct 11 04:23:18 which we prolly won't do Oct 11 04:23:26 and going for omap5 would defeat the purpose of the Neo900 Oct 11 04:23:31 yep Oct 11 04:23:38 largely Oct 11 04:23:45 the purpose is to be backwards compatible with the N900 as much as possible Oct 11 04:24:24 otherwise I wouldn't even stabd up in the morning to work on this Oct 11 04:24:32 stand up* Oct 11 04:25:13 since when we lose fremantle compatibility, we are competing with every other crappy android/whatever device out there Oct 11 04:26:24 and *I* for one would refuse to spend _any_ money on such a device, when it doesn't come with a promise of a proper fremantle Oct 11 04:27:26 porting debian could've done on shamgung galaxy9 since years Oct 11 04:28:29 fremantle binary compatibility is the single unique selling point of this platform Oct 11 04:28:59 I wont consider Neo900 "done" until the stock Fremantle dialer app and messaging app function 100% on Neo900. Oct 11 04:29:09 yep Oct 11 04:29:26 as well as any other app fresh from original extras repo Oct 11 04:29:37 aka "HAM working as well" Oct 11 04:29:42 ;-P Oct 11 04:34:44 ok, so regarding Neo900 audio, do we know if we can even get the TLV320AIC34 (same audio chip as N900) for use in Neo900? Oct 11 04:44:43 Nikolaus is still pretty reluctant, but I'll convince him since we need that mixer/codec. Or did I get your question wrong? Oct 11 04:45:34 what I meant is, is there a supplier that we could buy it from Oct 11 04:45:55 I know for example that we wont be able to get the WL12xx wifi chip because we are too small Oct 11 04:45:56 sure, check digikey: ~12EUR Oct 11 04:46:04 ok, thats good to know Oct 11 04:46:28 Seems like the hardest chips to get hold of tend to be wireless chips (cellular, bluetooth, WiFi etc) Oct 11 04:46:44 and 1GB PoP Oct 11 04:48:23 in the end we might not be able to source anything better than 512/512 Oct 11 04:48:57 compatibility of 156p PoP 1GB RAM not yet checked Oct 11 04:49:57 unless we find the magic supplier of N9 spare parts ;-) Oct 11 05:10:36 hmmm Oct 11 05:10:38 IroN900:~# du -hs /opt Oct 11 05:10:40 575M /opt Oct 11 05:28:14 jonwil: Nikolaus lists WL1271-TiWi as option Oct 11 05:28:46 ( "LS Research" ?) Oct 11 05:29:23 WL1271-typeTN would look beeter though ( "Murata" ) Oct 11 05:29:29 better even Oct 11 05:29:58 BT4.0BLR Oct 11 05:30:24 BLE, dang Oct 11 05:30:56 TiWi has just BT2.1EDR Oct 11 05:33:29 well, but it says "software upgrade to BLE" Oct 11 05:37:22 UGH typeTN is HUUUGE Oct 11 05:39:07 hmm, actually Nikolaus says "TiWi5 like in pandaboard" - don't see how that relates to twl1271-TiWi Oct 11 05:41:43 we could use W2CBW0015 instead of W2CBW003, comes with b/g/n and BT3.0, significantly smaller than 003 (9.5*9.5*1.3 instead 12*12*1.5), based on 8787 Oct 11 05:54:20 * DocScrutinizer05 idly wonders: 802.11bgn+BT3.0 = 9.5*9.5*1.3mm; LTE modem: 30*37*2.1mm Oct 11 05:59:18 AAAH TiWi5: http://www.lsr.com/wireless-products/tiwi5 Oct 11 05:59:22 nice Oct 11 06:00:15 k, 13*18*1.9 Oct 11 06:02:10 a(!)bgn,2.1EDR/4.0BLE Oct 11 06:10:15 * DocScrutinizer05 ifdly wonders how the 2.4GHz WiFi antenna will perform on 5GHz Oct 11 06:10:48 5.8(!)GHz Oct 11 06:38:51 seems like getting a better WiFi chip than the GTA04 isn't out of the question then Oct 11 06:39:39 better WiFi and bluetooth Oct 11 06:41:32 yep Oct 11 06:48:06 real hard sourcing is modem module Oct 11 06:48:37 the Option with 2.1mm is already most flat module you can find Oct 11 06:49:01 and it's still 0.6mm too high Oct 11 06:49:07 Even finding Option modem would have been nearly impossible if it wasn't for that one shop Oct 11 07:12:25 how can I access stored SMSes in SIM card ? Oct 11 07:14:04 on n900? no normal way Oct 11 07:14:46 yes, on n900. then what is the difficult method? Oct 11 07:15:59 RE isi/phonet library Oct 11 07:16:24 I do not know about any way to access it now Oct 11 07:17:00 well, I'm going to find my Nokia 3310 Oct 11 07:26:00 I dont know if the cellmo even supports accessing SMS on the SIM Oct 11 07:26:47 Oh wait it does Oct 11 07:26:48 or appears to Oct 11 07:26:51 com.nokia.csd.SMS.SIM exists Oct 11 07:27:17 Let me grep the codebase and see if anyone actually uses it Oct 11 07:27:22 grep the filesystem I mean Oct 11 07:27:50 jonwil: maybe creating program for access sim sms could be usefull! Oct 11 07:29:19 maybe Oct 11 07:41:01 ok so yes it looks like the cellular services daemon does support reading SIM SMS Oct 11 07:41:09 but nothing else on the system uses that functionalty Oct 11 07:41:33 And I dont know how to decode the SMSs as returned by the relavent dbus call Oct 11 07:41:43 why would any Oct 11 07:41:46 Binary noise? Oct 11 07:42:04 jonwil: check smscb Oct 11 07:42:28 jonwil pastebin a few, see if any patterns drop out. Oct 11 07:42:53 http://maemo.cloud-7.de/maemo5/usr/local/bin/smscb.py Oct 11 07:43:02 Maybe its a standard SMS PDU Oct 11 07:43:12 there won't be any patterns, it's 7/8 Oct 11 07:43:23 yes, it is Oct 11 07:43:58 at least afaik Oct 11 07:48:23 there's a standard SMS "codec" on ~phonecontrol Oct 11 07:48:32 ~phonecontrol Oct 11 07:48:32 somebody said phonecontrol was http://wiki.maemo.org/Phone_control Oct 11 07:49:49 modem can buffer SMS locally, store them to SIM, or directly and synchronously deliver them to APE Oct 11 07:50:12 theoretically Oct 11 07:50:36 and - also theoretically - you can read the local buffer and the SIM SMS storage Oct 11 07:51:01 and I guess that's exactly what CSD does. All three of them Oct 11 07:51:54 probably from SIM it will read only SMS marked as "unread" Oct 11 07:54:20 CSD has specific dbus calls for SIM SMS Oct 11 07:54:38 and I dont think it feeds SIM SMSs to anything else in the AP stack, just regular incoming SMSs Oct 11 07:55:07 i.e. the only way to read SIM SMSs is by calling manually com.nokia.csd.sms.sim Oct 11 07:57:35 DocScrutinizer05: that was my reasoning: packed 7-bits is simple to decide Oct 11 07:57:39 decode, even Oct 11 08:33:19 3gpp AT command lists give some inspiration how a modem *might* handle that stuff Oct 11 08:34:02 see http://talk.maemo.org/showpost.php?p=548948&postcount=52 Oct 11 08:40:47 I bet if I wanted to I could probably write something and use some code from ofono to decode the SMS (same as I did to decode the cell broadcast SMSs) Oct 11 08:52:10 jonwil: SMs from modem are always same encoding, no matter if realtime or read from SIM Oct 11 08:52:37 yeah Oct 11 09:36:18 hey Pali is 0xffff able to flash emmc yet? Oct 11 09:36:43 noticed some activity in on the repo Oct 11 09:36:49 no, 0xFFFF has no eMMC support Oct 11 09:36:49 seems to be progressing nicely Oct 11 09:38:19 good to know, thanks for the info :) Oct 11 13:06:41 Still getting nowhere with pulseaudio stuff Oct 11 13:14:46 I still cant figure out what tells pulseaudio-nokia-voice "hey, feed this audio through the cellular audio pathway and libcmtspeech" Oct 11 13:17:26 I also cant figure out just what processing is applied to VoIP (or to cellular) audio and whether we will need that same processing on the Neo900 or not Oct 11 14:13:33 hi all Oct 11 14:14:37 for some resone cssu crashed in install, I will need to reflush meamo or Is it possible to reinstall it? Oct 11 14:24:37 uhm..this is bad Oct 11 14:25:26 What is wrong when my n900 does not charge from AC-10E anymore? Tried a new charger => same problem. It still charges from USB Oct 11 14:28:33 DrCode: I would hope that the install is just a sequence of apt commands, so should be relatively stateless, and capable of being rerun. Emphasis on "hope". Oct 11 14:39:59 hi faari-ddr Oct 11 14:40:10 hi FatPhil Oct 11 14:40:28 I will reflaush it again Oct 11 14:41:35 I am waitting till my battery recharge Oct 11 15:13:35 uhm does nokia still repair n900s? Oct 11 15:13:48 unlikely Oct 11 15:13:55 damn Oct 11 15:13:55 since they're probably all out of warranty by now Oct 11 15:19:19 looks like I have to try fixing the usb myself then? :( Oct 11 17:26:36 I guess in 35 min there's another council meeting? I hope this time all council members ahve a chance to join in, we got urgent need to discuss and coordinate elections Oct 11 17:27:27 thedead1440: sixwheeledbeast: chem|st: merlin1991: qwazix: ^^^ Oct 11 17:27:58 ~seen woody14619 Oct 11 17:28:01 woody14619 <~Woody@66-162-186-66.static.twtelecom.net> was last seen on IRC in channel #maemo, 28d 23h 34m 33s ago, saying: 'odd... Time to reboot. :)'. Oct 11 17:28:16 FYI on holiday for a week starting now, lift just arrived. TTFN o/ Laters, DocScritinizer05 Oct 11 17:31:04 DocScrutinizer05, present! Oct 11 17:40:17 present! Oct 11 17:49:21 good Oct 11 17:51:09 warfare: kolp: FatPhil: freemangordon: we'll need you there too, since I hope for discussion of karma as a mandatory prerequisite for elections Oct 11 17:52:21 I'm (t)here, but can only pay attention tothe chat with half an eye Oct 11 17:53:57 are processes suspended when switching tasks? Oct 11 17:55:03 DocScrutinizer05: meeting where? Oct 11 17:55:06 sono: In maemo? No. Oct 11 17:55:22 #maemo-meeting Oct 11 17:55:22 heaven forfend! Oct 11 17:56:20 must have been a freak incident Oct 11 17:56:29 i am downloading the debian image and switched to a second running xterm Oct 11 17:56:39 and when i switched back, the download was in a stall Oct 11 18:00:55 a nice device for its time Oct 11 18:00:59 * sono never owned one before Oct 11 18:01:14 i hope to make it my primary phone, perhaps Oct 11 18:03:48 sorry, i am talking about an n900 of course. Oct 11 18:09:08 hi all Oct 11 18:09:14 I try to flash again Oct 11 18:10:48 n.m Oct 11 18:10:50 It worked Oct 11 18:47:59 hi all Oct 11 18:48:05 I have strange problem in n900 Oct 11 18:48:40 when I am using 3g data, I can ping to wan address, but it can't resolve names Oct 11 18:48:41 any idea? Oct 11 18:48:51 dns not working? Oct 11 18:49:05 I tryid to had goole dns , nameserver 8.8.8.8 Oct 11 18:49:09 and it dosn';t work Oct 11 18:49:13 put ping work ok Oct 11 18:49:29 yes Oct 11 18:49:39 like somthing blocking udp port 53 Oct 11 18:50:13 Wizzup, any idea? Oct 11 18:51:05 not really, beyond no name server being set Oct 11 18:51:36 ok Oct 11 18:51:39 I will check it again Oct 11 18:52:17 I use those flash: RX-51_2009SE_10.2010.13-2.VANILLA_PR_EMMC_MR0_ARM.bin Oct 11 18:52:18 RX-51_2009SE_20.2010.36-2_PR_COMBINED_MR0_ARM.bin Oct 11 18:52:18 is it ok? Oct 11 20:48:49 hmm it appears that xcb has an endianness bug resulting in a failed assertion at xcb_io.c:249 Oct 11 20:49:09 * sono doesn't have a build toolchain up yet Oct 11 20:59:52 endianess bug?? Oct 11 21:00:17 I think that endianness can appear only on big endian arch Oct 11 21:00:49 because sw is probably written for pc/i386 little endian Oct 11 21:08:39 on armel (it is low endian) there should be no endianness problem in sw written for i386 Oct 11 21:10:19 right, silly me Oct 11 21:10:48 probably those (long)'s, the remote is 64bit. Oct 11 21:10:53 anyway, i'll look at it tomorrow Oct 11 21:29:42 use int64_t or int32_t Oct 11 21:29:54 yea i have no build toolchain Oct 11 21:29:59 and no problem with longs... Oct 11 21:30:03 yea.. Oct 11 21:31:01 but I heard that there are platforms which char is not 8 bit... Oct 11 21:31:10 some say so Oct 11 21:31:36 iirc, the pdp-11 was one Oct 11 21:31:45 i'm way to young for that stuff, however Oct 11 21:32:03 * sono only knows the old days from lore Oct 11 21:58:56 ~bootmenu Oct 11 23:23:12 DrCode: COMBINED is OK, not completely sure about VAANILLA, but looks right too Oct 12 02:26:34 heya, since the nmap 5.59 package for maemo5 is broken, would it be possible to revert to 5.50 that actually worked? Oct 12 02:27:13 in other words, is there a quick fix with dpkg or do I have to compile the tarball myself? :/ **** ENDING LOGGING AT Sat Oct 12 02:59:59 2013