**** BEGIN LOGGING AT Wed Feb 28 03:00:01 2007 Feb 28 03:44:06 "The thing we can promise at this point is a faster CPU. We're still working on the WiFi stuff. Graphics acceleration is much later. -Sean" http://lists.openmoko.org/pipermail/community/2007-February/003857.html Feb 28 03:44:39 atleast they're thinking about graphics :) Feb 28 03:46:24 :... Feb 28 03:52:43 I don't see a pressing need for it. (accelleration) Ok - it'd be nice to have slightly faster graphics, there are things I care about more though. Feb 28 03:53:02 like WiFi. Feb 28 03:53:08 If there was to be a new chip, I'd like a MP3 chip that can talk to the SD. Feb 28 03:53:13 (apart from widi) Feb 28 03:53:25 Aren't there any miniSD to SD adapters? Feb 28 03:53:28 Many. Feb 28 03:53:32 FIC founder posting in community list confirmed: http://lists.openmoko.org/pipermail/community/2007-February/003852.html Feb 28 03:53:49 That way we could maybe use an SD wifi card on the Neo (maybe I mean SD to miniSD adapters. Feb 28 03:53:56 so we now know for sure that nice interest goes to top Feb 28 03:54:06 Yes, it's trivial to do, um_eddie. Feb 28 03:54:19 The problem then arises that you pick between wifi, and SD. Feb 28 03:54:31 sd-flash that is. Feb 28 03:55:27 Well, I'd mostly just like to use faster internet to read Google Reader, and e-mail faster... not download things. Feb 28 03:55:35 So it's not really a problem. Feb 28 03:55:56 accelleration has nothing to do with internet. Feb 28 03:55:59 Switch out the WiFi card and put in the SD card to listen to your music when your done. Feb 28 03:56:28 ? Are we talking about the same things? Feb 28 03:56:32 Having only 10M or so of storage non-volatile may not be enough for many pwople. Feb 28 03:56:51 You can use bluetooth internet too. Feb 28 03:57:27 Yeah, you know with all those bluetooth hot-spots. Feb 28 03:57:31 :) Feb 28 03:58:16 wm_eddie: yes, I have one in my home and one at work. and I read (and respond to) email and rss while riding the bus between Feb 28 03:58:25 (using a Treo650 presently) Feb 28 03:58:50 bluetooth is way fast enough for downloading stuff for offline usage. Feb 28 03:58:59 * wm_eddie uses a Treo600 and hates the slowness. Feb 28 03:59:14 do you use it online or offline? Feb 28 03:59:35 online via PCS Vision. Feb 28 03:59:46 that's the difference then. Feb 28 03:59:51 I wish I had a wifi card so I can speed up the internet. Feb 28 04:00:07 I use offline IMAP (Chatter Email) and cached web pages (Xiino) Feb 28 04:00:29 I have to use gmail because there's not a single e-mail program for the treo that can do Japanese email. Feb 28 04:00:34 so I don't care how fast the connection is - as long as it can download enough stuff to keep me occupied for a half hour bus trip. Feb 28 04:01:27 (No flat-rate GPRS here) Feb 28 04:02:11 Just use it, and see if it's less stable than usual. Feb 28 04:02:11 oops Feb 28 04:02:20 I get unlimited for $15 a month. :) It's the ONLY thing that makes Sprint worthwile. Feb 28 04:02:50 yep, the availability of flat rate wireless data makes a *big* difference in how you use a device Feb 28 04:03:15 $60/mo here (UK) for unlimited GPRS (1 pound per day purchase) Feb 28 04:03:26 Or 3 quid a meg. Feb 28 04:04:07 Sprint says it's unlimited but they actually limit it to 200MB a day. then they throttle you to 1KB/s Feb 28 04:04:24 AUD$5 per MB here - no unlimited Feb 28 04:04:33 200M/day? Feb 28 04:04:36 * SpeedEvil wishes. Feb 28 04:04:53 It was kind of a problem back in the day before I had internet and was playing Final Fantasy XI via phone->laptop-PS2 Feb 28 04:04:53 I'd love 20. Feb 28 04:05:32 It's the price one has to pay when using TDMA. Feb 28 04:14:46 (Or when one is in Japan and they charge rediculous ammounts for data dispite their 3.5gness Feb 28 04:17:16 * aloril needs to rename Wishlist:Go back ;-) (once I have implemented enough of it of course, engines are there already) Feb 28 04:32:37 * SpeedEvil is currently trying to bring some sanity to Wishlist:Hardware. Moving some to Wishlist:Unlikely Feb 28 04:39:52 SpeedEvil: hehe Feb 28 04:46:29 Hmm. On reflection, I think maybe moving the whole thing to wishlist_hardware:neo1973 may be appropriate, and some splitting into wishlist_hardware:openmoko Feb 28 04:49:46 P1 might be April 1st week, so moved date a bit (lower limit moved to 2 days forward too), also shortened Feb 28 04:49:48 counter Feb 28 04:49:48 P0: a week, a day 06:35:31 (-0.725 - 8.275 days); P1: 24.275 - 39.275 (see faq) (421) Feb 28 04:51:03 (commenting on some past talk, sorry if you saw this already) libervisco: what makes you think they offer only EDGE and not also plain GPRS? Feb 28 04:51:42 they offer only GPRS, it says in their specs Feb 28 04:52:01 I hope EDGE could work too though Feb 28 04:52:15 libervisco: I meant your provider Feb 28 04:52:27 oh Feb 28 04:52:46 well, I am not sure.. maybe GPRS would work Feb 28 04:53:00 but I doubt it, if requirements are different Feb 28 04:53:04 3G only does exists I think, but EDGE only? Feb 28 04:53:10 No. Feb 28 04:53:11 well.. EDGE is basically just GPRS+ Feb 28 04:53:14 hmm.. Feb 28 04:53:18 neo? it does gprs only from what I understand Feb 28 04:53:20 no edge or 3g Feb 28 04:53:25 EDGE is a different modulation standard for GSM over the same frequencies. Feb 28 04:53:28 well the company does offer GPRS internet access over mobile phones Feb 28 04:53:32 yes, Neo1973 does GPRS only Feb 28 04:53:36 but they are expensive Feb 28 04:53:44 EDGE trancievers switch between EDGE and normal GSM/GPRS as needed. Feb 28 04:53:52 data trafic is 8$ / mbyte here Feb 28 04:53:55 Dynamically according to traffic. (on the tower) Feb 28 04:53:59 or you can get a flat fee for monthly use Feb 28 04:54:28 I think an EDGE only phone is impossible too - Feb 28 04:54:36 GPRS -> EDGE: some hardware and software updates in tower, but I haven't heard yet tower that would offer only EDGE and not also GPRS Feb 28 04:54:38 as it requires GSM to negotiate. Feb 28 04:54:58 As I understand it anyway. Feb 28 04:55:09 3G only is possible I think? Feb 28 04:55:14 Yeah. Feb 28 04:55:17 yeah you can't have edge without gprs, atleast no one is doing it Feb 28 04:55:20 (Japan, South Korea for example) Feb 28 04:55:20 Utterly different strandard. Feb 28 04:59:36 btw Feb 28 04:59:58 http://www.maplin.co.uk/Module.aspx?ModuleNo=99114&C=Newsletter&U=07P03-2_Office&T=-11854041 - or a neo, for cheaper! Feb 28 05:00:06 we're making some progress with that open mobile site :) Feb 28 05:00:43 SpeedEvil, nice :) Feb 28 05:01:15 Just interesting how much some 'simple' devices can be. Feb 28 05:01:29 indeed Feb 28 05:04:24 if anyone wants to make suggestions about the new site... http://www.mobiliberty.com/shaping_concept Feb 28 05:04:40 anonymous comments are open :) Feb 28 05:05:42 'written by a multiple of authors' -> many authors Feb 28 05:07:42 SpeedEvil, gonna fix :) Feb 28 05:07:57 Nothing else leapt out at me. Feb 28 05:09:04 np, thanks for reading Feb 28 05:16:40 hi all ... I was wondering if there is a harder release date than a month or so (I know I know) ... just ran across this tonight and I'm strongly considering a rething of the hardware for a m2m project I'm working on. At this price point and feature set the Neo is about perfect for what we need, but I'm on a tight timeline and would need one asap... Feb 28 05:17:34 no. Feb 28 05:17:42 nothing really solid. Feb 28 05:17:57 ok, was afraid of that Feb 28 05:18:07 If I was on a tight timeline, and needed one before the middle of april or the project would fail, I wouldn't consider it. Feb 28 05:18:23 It _should_ be available late March, but... Feb 28 05:19:02 yeah, too bad, would save me about 60 hours of board design to integrate the pile of shite I currently have sitting here to work with :) Feb 28 05:19:08 :/ Feb 28 05:19:43 What are you doing? Maybe there are other options that someone here would have a clue of. Feb 28 05:20:35 reading industrial sensors, pushing the data to a database ... needs gps and a local user interface though, so this was a fairly elegant solution Feb 28 05:21:15 current plan is a rabbit semiconductor processor and a serial gprs module ... but I like easy :) Feb 28 05:22:06 besides, the screen is sooo pretty ... and in all it is barely more expensive than what we are currently doing ... well within budget Feb 28 05:22:43 seems like this always happens ... next greatest thing is announced that will work perfectly ... it's just behind on release too far to use Feb 28 05:23:12 or I hear about it a week after we finish the current designs :) Feb 28 05:26:27 anyway ...thanks for the answer Feb 28 05:26:32 Seems always to be the way. Feb 28 05:26:34 Good luck! Feb 28 05:26:40 thanks Feb 28 06:05:03 from old devel mail: "Second, it probably won't be worth it to develop applications on the device itself (if that were even possible given the memory constraints)." Feb 28 06:05:19 well.. Neo1973 is like an old supercomputer really ;-) Feb 28 06:07:02 I dunno. Feb 28 06:07:13 It's not impossible to compile stuff on. Feb 28 06:07:17 Especially overnight. Feb 28 06:07:18 * aloril gets image of 100 text terminals being connected trough USB to Neo1973 Feb 28 06:08:32 Morning :) Feb 28 06:09:29 hall with 100 text terminals and wires and in middle of it small Neo1973 that is actually doing stuff Feb 28 06:09:33 funny ;-) Feb 28 06:10:35 What's the default max number of ttys? 255 Feb 28 06:10:42 or circle of text terminals all connected to middle of circle Feb 28 06:22:16 That would be funny. A picture of 100 terminals spread round one neo on cables leading into the middle through a tree of hubs. Feb 28 06:26:04 time travel back in time with this image when main frame meant 128MiB of memory, etc.. and lots of connected terminals Feb 28 06:26:16 see, in future these are this small Feb 28 06:26:20 Yeah. :) Feb 28 06:26:31 Somewhere I've got a 24" drive platter. Feb 28 06:26:40 oh.. btw.. here is one "takes one out of pocket" ;-) Feb 28 06:26:55 And Io should probably throw away myu stack of 64K DRAM chips. Feb 28 06:27:31 other person: "Woah" then you go to describe your desktop machine and other person faints ;-) Feb 28 06:27:45 I have version 1.0 of the system I work with now on a big tape roll in my desk drawer Feb 28 06:28:17 Paper tape? Feb 28 06:28:30 Hehe, luckily not Feb 28 06:28:46 But I don't know what it is.. Says ALPHA 800/1600/6250 on it Feb 28 06:29:05 That's 800/100/6250bpi Feb 28 06:29:12 Is it about 1/2" wide? Feb 28 06:29:19 Yes Feb 28 06:29:35 It's probably still readable in the right drive. Feb 28 06:29:38 openmoko: 03tony_guan * r1156 10/trunk/src/target/gsm/ (4 files in 3 dirs): resolved bug 108, please varify it. thks. Feb 28 06:29:38 Finding one is fun. Feb 28 06:29:53 The system is still running on an old ibm vm mainframe Feb 28 06:30:27 I've asked around, the still have a reader for it Feb 28 06:30:38 Don't want to see that old APL code though Feb 28 06:51:27 openmoko: 03zhiyong_sun * r1157 10/trunk/src/target/OM-2007/applications/openmoko-mainmenu/ (9 files in 2 dirs): add dbus message send function: when one item was selected, its name will be sent to footer Feb 28 07:22:13 aloril: 111 now on http://wiki.openmoko.org/wiki/Buying_Interest_List Feb 28 07:23:33 CM: yeah, interesting to see what people plan to do Feb 28 07:23:45 :) Feb 28 07:24:06 CM: I suspect that actual official private preorder page would have at least 10x more ;-) Feb 28 07:24:32 Heh, I think so too Feb 28 07:25:51 updated counter to now include middle of range as target day +- uncertainty Feb 28 07:26:02 counter Feb 28 07:26:02 P0: 4 days 01:59:38 (4.083 +-4.1 days); P1: 26.166 +-10.0 (see faq) (422) Feb 28 07:26:28 good point is that remaining time is less, bad point is that it moves slower for P0 ;-) Feb 28 07:26:47 (for P1 there is only good point currently ;-) Feb 28 07:34:53 * aloril guess is that FIC wants to first get P0 out before putting pre-order page up Feb 28 07:36:10 s/guess/guesses/ Feb 28 07:36:12 And P0 phones in the hands of devs to find latent bugs I suspect. Feb 28 07:36:37 P1 phones with unworkaroundable bugs that need shipped back would be _horrid_ for them. Feb 28 07:36:37 yes and also so that however pre-orders has had chance to read P0 impressions Feb 28 07:37:01 s/however/who/ Feb 28 07:37:03 aloril meant: yes and also so that who pre-orders has had chance to read P0 impressions Feb 28 07:37:55 can't they have uncertain dates in pre-order? if that is not good idea, then that also will push pre-order date close to actual P1 release data Feb 28 07:38:19 s/data/date Feb 28 07:45:05 openmoko: 03caowai_song * r1158 10/trunk/src/target/OM-2007/applications/openmoko-keyboard/ (8 files in 2 dirs): Feb 28 07:45:08 openmoko: 1. Add background picture to the keyboard Feb 28 07:45:09 openmoko: 2. Display normal picture or pushed picture for each keys Feb 28 07:47:37 openmoko: 03caowai_song * r1159 10/trunk/src/target/OM-2007/applications/openmoko-keyboard/layouts/ (Makefile.am spacekey.png testkey.png): Remove unnecessary pictures Feb 28 07:52:31 The thing we can promise at this point is a faster CPU. We're still Feb 28 07:52:31 working on the WiFi stuff. Graphics acceleration is much later. Feb 28 07:52:36 (thats from sean himself) Feb 28 07:55:23 what faster cpu has no graphic acceleration at all? Feb 28 07:56:31 most arm CPUs Feb 28 07:57:07 well but which ones are likely candidates? Feb 28 07:57:25 my guess would be other samsung s3c chips Feb 28 08:01:01 oh i thought the s3c2440 had graphics accelerator Feb 28 08:01:03 but it has not Feb 28 08:08:06 morning Feb 28 08:11:17 easy mail, just copy faq entry ;-) (community ML) Feb 28 08:15:29 buz: btw, most 'graphic acceleration' is either useless or require closed drivers Feb 28 08:20:17 But it is cool to get information now about the v2 neo Feb 28 08:20:33 just a few but that is ok Feb 28 08:29:43 Really, really trivial accelleration can help quite a lot in many cases. Just blitters, that can do 'copy rectangle from here to here'. Feb 28 08:31:46 SpeedEvil: most 'accel' is either YUV overlay or *dct Feb 28 08:32:02 then it jumps to "3d" Feb 28 08:32:36 Yeah. Feb 28 08:32:51 having a bit of fast vram would be nice Feb 28 08:33:11 that way you can do offscreen rendering that isn't dog slow Feb 28 08:33:59 You mean a seperate accellerator core? Feb 28 08:34:12 no, just some dedicated videoram Feb 28 08:34:41 We need to copy from the Sinclair ZX81. Feb 28 08:35:12 'FAST' mode - that blanks the display and does no RAM access. And SLOW mode that does video output. Feb 28 08:35:13 there is one reason for beefy FPGA: develop stuff first for it and then make ASIC for actual usage (kind of P1 vs P2) Feb 28 08:35:13 :) Feb 28 08:35:29 Yeah. Feb 28 08:35:38 but then.. not many want to have enough batteries with them to do above ;-) Feb 28 08:36:30 Optimising for ASIC is hard though. Feb 28 08:36:51 As not all optimised for FPGA designs translate cleanly AIUI. Feb 28 08:36:51 going from FPGA to ASIC is big step? Feb 28 08:37:16 Generally not, but the optimisations are different, as I understand it. Feb 28 08:37:36 Plus, the number of VHDL coders is much smaller than C/.. Feb 28 08:39:02 For example, if you need a 14 input AND gate, that may not be a standard block on the FPGA, but it can be easily added to the ASIC. Feb 28 08:39:07 meh Feb 28 08:39:14 openmoko stuff doesn't build again Feb 28 08:39:31 :( Feb 28 08:39:33 can those people at least try to compile stuff after a change? Feb 28 08:42:07 ~lart tony guan Feb 28 08:42:08 * apt takes large quantities of Krispy Kream donuts and stuffs them one after another down tony guan's throat until tony guan puts on 150lbs Feb 28 08:42:30 http://svnweb.openmoko.org/trunk/src/target/OM-2007/applications/openmoko-dialer/src/contacts.c?rev=1147&view=diff&r1=1147&r2=1146&p1=trunk/src/target/OM-2007/applications/openmoko-dialer/src/contacts.c&p2=/trunk/src/target/OM-2007/applications/openmoko-dialer/src/contacts.c Feb 28 08:42:35 that looks bogus to me Feb 28 08:42:53 var = 1 ; if (var==1).... Feb 28 08:48:40 I wonder if the FIC developers have been educated about the friendly response they will get on IRC here :-) Feb 28 08:49:24 thank god they're not japanese Feb 28 08:58:15 well, 'svn up -r1146' fixes the problem Feb 28 08:59:00 revision 1147 does what it intends to do: you can't segfault an app that doesn't build Feb 28 09:11:53 openmoko: 03rwhitby * r25 10mokomakefile/trunk/patches/openmoko-1146/: Tag patches at 1146 Feb 28 09:13:33 openmoko: 03rwhitby * r26 10mokomakefile/trunk/Makefile: Freeze MokoMakefile at r1146 until head is fixed Feb 28 09:16:24 ~lart Excel Feb 28 09:16:25 * apt calls Excel on the phone ... the lights are on but nobody's home Feb 28 09:33:29 * rwhitby realises it might not be as easy as that to get a specific version of a moko app to build in mokomakefile framework due to the strange openmoko-base.bbclass SRC_URI stuff Feb 28 09:35:27 * koen finally figures out how to add applets to matchbox-panel-2 Feb 28 09:36:03 it didn't help that "matchbox-panel --help" gives you the wrong syntax Feb 28 09:39:15 hmm - there is no way to set individual revs for openmoko apps in the current build framework. Feb 28 09:39:40 koen, update wiki :-) Feb 28 09:40:24 rwhitby: which is why I don't use all that SRCDATE="now" nonsense the moko people seem to love Feb 28 09:40:39 koen: your 'svn up -r1146' was a manual set in sources/svn, right? Feb 28 09:40:46 s/set/step/ Feb 28 09:40:46 rwhitby meant: koen: your 'svn up -r1146' was a manual step in sources/svn, right? Feb 28 09:41:04 bitbake -b openmoko-dialer_svn.bb -c devshell Feb 28 09:41:10 svn up -r1146 Feb 28 09:41:26 right. not easy to automate :-) Feb 28 09:41:39 but SRCDATE_openmoko-dialer = "daybeforetonybrokeit" should work as well Feb 28 09:41:52 ah, yes. thx. Feb 28 09:42:20 Is there a SRCREV equivalent for the SVN fetcher? Feb 28 09:43:06 * rwhitby reads svn.py ... Feb 28 09:43:17 only SRC_URI = "svn://host/blah;rev=1146" Feb 28 09:44:02 in bitbake 1.7.x, that is Feb 28 09:44:09 1.6.x might not even have that Feb 28 09:44:17 yeah, difficult to do on an app by app basis. looks like 1.6.x has it. Feb 28 09:44:37 (difficult to do in the current openmoko-base.bbclass situation) Feb 28 09:44:43 it would be nice to have an SCMREV or SRCREV, though Feb 28 09:45:18 like 'revision = Fetch.getSRCRevision(d) Feb 28 09:45:18 ' ... Feb 28 09:46:13 rwhitby: it is funny that openmoko is so conservative with the OE metadata, but don't have any safeguards against their internal breakages Feb 28 09:46:29 which you solve in your makefile and I in OE proper Feb 28 09:46:53 koen: I suspect that mickeyl is buffering the FIC developers from OE completely. Feb 28 09:47:30 and sadly mickeyl only has 24 hours in his day Feb 28 09:47:39 koen: 1.6.4/1.6.6 has it Feb 28 09:47:43 (hence the target/OM-2007 stuff and how the .bb files grab it from there) Feb 28 09:48:24 the build instructions should really say to co openmoko/trunk/oe, not openmoko, cause the other stuff under openmoko is never used (as the svn fetch pulls it in under sources/svn ...) Feb 28 09:49:11 and then the preferred-versions should have SRCREV variables in it, and should say HEAD, or a known good svn rev ... Feb 28 09:49:15 i'm just surfing the wiki for graphic design information for openmoko. does anyone know where i can find some or probably anyone knows the name of the font that's used in the logo? Feb 28 09:49:20 people that followed the official build instructions haven't been able to build stuff for the past weeks Feb 28 09:49:35 not without mokomakefile, no. Feb 28 09:49:42 right Feb 28 09:50:36 so you need to use the mokomakefile or not use the fic bbcollection to get stuff done Feb 28 09:50:39 although there are really only two downloads and two patches at the moment (one to omit uboot cause it doesn't build, and the other to mainmenu for the -lmb thingo) Feb 28 09:50:47 koen: yup Feb 28 09:51:16 which is why I suspect that the FIC developers are using neither, and are just working directly in the SVN without OE at all Feb 28 09:51:31 that, or they don't test at all Feb 28 09:51:51 Helmi: Do you mean for applications or just the logo? Feb 28 09:52:19 CM: primarly i'm interested in the logo font - didn't see any application shots yet Feb 28 09:52:38 yeah, logo font would be nice for related web sites :-) Feb 28 09:52:55 * rwhitby wants to write "mokomakefile" in that font ... Feb 28 09:53:02 Hehe Feb 28 09:53:06 Helmi: http://wiki.openmoko.org/wiki/Look_%26_Feel Feb 28 09:53:56 rwhitby: iirc chinese new year end right around now, so lets hope we'll see some improvement Feb 28 09:54:10 CM: thanks for the link - should i find something related to the logo font there or just the shots? Feb 28 09:54:10 Helmi: See http://wiki.openmoko.org/wiki/FAQ (this autotrigger for new nick and 2 question marks during 15min done only once) Feb 28 09:54:30 Helmi: Nothing about the logo font afaict Feb 28 09:54:39 CM: ok thanks anyway Feb 28 10:06:37 counter Feb 28 10:06:37 P0: 4 days 00:39:20 (4.027 +-4.0 days); P1: 26.055 +-10.0 (see faq) (423) Feb 28 10:09:49 this time faq autotrigger wasn't useful Feb 28 10:11:50 aloril - Where is the +-10 coming from? Feb 28 10:12:06 Helmi: I guess source is your friend currently Feb 28 10:12:33 Elrond: just guess: 2007-02-25 seems to be target date, but it can go to April too Feb 28 10:12:45 probably - part is not correct ;-) Feb 28 10:13:09 aloril: sorry can't follow you!? Feb 28 10:13:17 aloril - s/02/03 you mean. ;) Feb 28 10:13:46 Elrond: yeah Feb 28 10:16:03 Helmi: http://svnweb.openmoko.org/trunk/src/target/OM-2007/artwork/themes/openmoko-standard/gtk-2.0/gtkrc Feb 28 10:16:11 koen: I wonder if we can convince mickeyl to put a SRCREV in openmoko-base.bbclass and use SRCREV=HEAD in preferred-versions instead of SRCDATE ... Feb 28 10:16:37 aloril: should this help finding the logo font from openmoko? Feb 28 10:17:07 Helmi: from that file: font_name = "Vera Sans 16" Feb 28 10:17:31 hmm no - that's not the logo font (you know the one on the webpage with the nice signet next to it) Feb 28 10:18:04 I guess only the designer knows that one. I've only seen it in images and press stuff, not in any screenshot Feb 28 10:18:05 Helmi: duh, I read it "font used in openmoko apps" Feb 28 10:18:44 aloril: CM asked if it's just about the logo font or the apps to and sent me a link to see some screenshots of apps - but i'm primarly loooking for the font used in the logo - thanks anyway# Feb 28 10:26:02 morning :-) Feb 28 10:29:39 aha. SRC_URI from openmoko-base.bbclass conveniently doesn't have a space at the end, so you can append ";rev=1146" to it in the .bb file ... Feb 28 10:32:36 hi all! Feb 28 10:32:46 koen: any idea why 'SRC_URI .= "rev=1146"' would cause "ERROR: too many values to unpack while parsing /home/moko/oe/packages/openmoko-apps/openmoko-dial Feb 28 10:32:46 er_svn.bb" ? Feb 28 10:33:20 wouldn't SRC_URI_append work better? Feb 28 10:33:43 that doesn't add a space? Feb 28 10:33:56 (oh, only += adds a space?) Feb 28 10:35:44 nope, SRC_URI_append gives the same error. I wonder if this is a bitbake 1.6.x problem ... Feb 28 10:36:22 I bet SRCDATE = "now" and ;rev=1146 don't go well together Feb 28 10:36:53 no, I looked at that specifically in svn.py, and rev overrides date. Feb 28 10:37:53 (and I commented out all the svnnow stuff anyway) Feb 28 10:42:03 hmm - scratch that idea then. No idea how to make that work :-( Feb 28 10:42:51 call FIC, have someone throw a bucket of water on tony to wake him up and have him fix it Feb 28 10:43:00 or just inject an antipatch Feb 28 10:43:02 oh, hang on, I see my error. Feb 28 10:44:36 that's better. I just forgot the leading ';' Feb 28 10:46:44 openmoko: 03rwhitby * r27 10mokomakefile/trunk/ (5 files in 3 dirs): Freeze openmoko-dialer at r1146 until compile error is fixed. Unfreeze everything else to HEAD again. Commit a checkpoint set of patches at 1146. Feb 28 10:47:19 ok, that's a general recipe for freezing individual apps at a certain rev ... Feb 28 11:01:45 openmoko: 03laforge * r1160 10/trunk/src/target/gsm/src/libgsmd/libgsmd_device.c: add missing file from Commit rev. 1155 (tony guan) Feb 28 11:20:27 openmoko: 03laforge * r1161 10/trunk/src/target/gsm/ (5 files in 3 dirs): revert 'bug108' patch from tony guan. Feb 28 11:25:31 heh Feb 28 11:25:48 it seems tony is going to have a bad day Feb 28 11:28:46 hmm.. laforge edited [[Getting_Started_with_your_Neo1973]], I wonder if that means anything about shipping going to be soon? Feb 28 11:28:48 counter Feb 28 11:28:49 P0: 3 days 23:58:14 (3.999 +-4.0 days); P1: 25.998 +-10.0 (see faq) (424) Feb 28 11:30:04 aloril: shipping should start after chinese new year Feb 28 11:30:08 e.g. now Feb 28 11:30:11 openmoko: 03thomas * r1162 10/trunk/src/target/OM-2007/applications/openmoko-dialer/ (configure.ac src/Makefile.am src/contacts.c): * Clean up configure.ac and add a check for EContactPhotoType. Fixes bug 219. Feb 28 11:30:26 yay! Feb 28 11:37:33 hmmm Feb 28 11:37:46 according to wikipedia there are still ~5 days left in chinese new year Feb 28 11:39:33 koen: actual vacation was only 1 week of it? Feb 28 11:39:46 that's what I heard Feb 28 11:39:49 but which week? Feb 28 11:39:59 * aloril wonders if actual order would be 100x more than bying_interest count Feb 28 11:40:11 last week? Feb 28 11:40:19 * koen hopes so Feb 28 11:41:24 dunno how much additional time they need for cable + other problems, if week enough, then I guess shipping soon Feb 28 11:42:27 * aloril also notes .org has not yet been /.:ed or digged Feb 28 11:45:21 <_buz> http://feeds.feedburner.com/~r/weblogsinc/engadgetmobile/~3/97284837/ for all of those who think 2.8inch is too smal Feb 28 11:45:26 <_buz> (talk about geeky) Feb 28 11:59:12 koen: The vacation week ended last sunday Feb 28 11:59:20 ah, cool Feb 28 11:59:30 At least the cinese chef here is back, and that's a sure sign Feb 28 11:59:37 :) Feb 28 12:00:16 Also, read this, less good: http://www.theregister.co.uk/2007/02/28/chinese_gaming_fatality/ Feb 28 12:00:33 * CoreDump|afk personally hope he will get an email or another notice before shipping starts Feb 28 12:13:43 openmoko: 03rwhitby * r28 10mokomakefile/trunk/patches/openmoko-HEAD/series: Disable openmoko-dialer patch, since upstream has fixed the bug. Feb 28 12:18:53 morning Feb 28 12:22:03 * rwhitby considers commenting out the SRCDATE_* lines in preferred-openmoko-versions.inc, and adding a makefile target which blows away the corresponding stamp files instead. Feb 28 12:23:25 hi greentux Feb 28 12:25:21 hi hrw Feb 28 12:34:29 openmoko: 03rwhitby * r29 10mokomakefile/trunk/ (Makefile patches/openmoko-HEAD/series): Feb 28 12:34:29 openmoko: Enable mokomakefile patches which disable the 'svnnow' paradigm, and provide a Feb 28 12:34:29 openmoko: 'force-rebuild' makefile target to allow developers to have a build environment Feb 28 12:34:29 openmoko: that doesn't rebuild everything every time it is executed, yet also allows them Feb 28 12:34:29 openmoko: to easily force the rebuild all the openmoko apps that would otherwise be Feb 28 12:34:31 openmoko: tracking svnnow. Feb 28 12:35:15 rwhitby: hey, around? Feb 28 12:35:19 yup Feb 28 12:35:29 rwhitby: I was wondering what a good way to skip/revert openmoko-dialer is. Feb 28 12:35:41 cjb: it's fixed Feb 28 12:35:48 cool, I'll just rerun, then. Feb 28 12:35:57 make update-patches first Feb 28 12:36:32 and enjoy the new removal of the svnnow stuff, replaced by a 'force-rebuild' makefile target for when you want to rebuild the svnnow marked apps. Feb 28 12:36:35 whoa. I just did make openmoko-devel-image again and it started at gta01-kernel-2.6.. does it not go back to the last package that didn't build? Feb 28 12:36:39 :) Feb 28 12:36:46 it didn't before. now it will. Feb 28 12:37:11 well okay then :) Feb 28 12:37:52 took me a while to work out the right perl->xargs->find->xargs->rm to force a rebuild of the precise apps that openmoko sets to svnnow. Feb 28 12:38:00 neat. Feb 28 12:38:11 thanks! Feb 28 12:38:32 so I have a patch which comments out the svnnow stuff, and a correponding makefile target which forces the same behaviour on the next build only. Feb 28 12:39:01 much more civilised (in my opinion at least) Feb 28 12:39:32 forcing SRCDATE to 'now' should never be needed Feb 28 12:39:51 koen: FIC devs and OpenMoko devs are not touching PR - that's why it's being used. Feb 28 12:40:03 rwhitby: actually, now Feb 28 12:40:23 s/now/no/ Feb 28 12:40:24 koen meant: rwhitby: actually, no Feb 28 12:40:43 that was the reason mickeyl gave me directly. Feb 28 12:41:14 it's because developers wanted to checkout revs they just committed Feb 28 12:41:19 which can't be done with PR Feb 28 12:41:47 nod Feb 28 12:42:12 koen: how would you recommend that goal be achieved? Feb 28 12:42:23 I wouldn't Feb 28 12:42:27 but if people insist: Feb 28 12:42:40 do_configure_prepend() { svn up } Feb 28 12:43:22 koen: wouldn't you need to combine that with removing stamps or incrementing PRs ? Feb 28 12:43:36 not really Feb 28 12:43:51 just set no_stamp[] appropriately Feb 28 12:44:28 it will break the assumption that SRCDATE = "" will always build the same rev (midnight GMT) Feb 28 12:45:08 if you want to build stuff you are working on, I'd recommend devshell Feb 28 12:45:49 yeah, but mickey wants that to happen across multiple people and sites Feb 28 12:46:10 (with people who don't know how to use devshell) Feb 28 12:46:10 it's fundamentally flawed Feb 28 12:46:24 OE isnt a development system Feb 28 12:46:30 even if we want it to be Feb 28 12:46:53 I'm not arguing that point, but given the OpenMoko core devs goal, I'm looking for the best way to meet that goal (rather than arguing against their decision) Feb 28 12:46:55 but I can see the trouble trying to explain that to non open-source developers in asia Feb 28 12:47:24 and meet it in a way which doesn't force a rebuild of all openmoko packages every time you type bitbake ... Feb 28 12:47:40 rwhitby: I don't think non-FIC employees would need/want srcdate-"now" behaviour Feb 28 12:48:12 especially seeing that commits to svn are rare Feb 28 12:48:25 koen: agreed - that's why I've turned it off in mokomakefile now Feb 28 12:48:32 (and that half of the commits get reverted anyway) Feb 28 12:49:13 people advanced enough to want rebuild can figure out -c rebuild themselves, right? Feb 28 12:49:49 ah, but one may only want to rebuild the apps which are marked as svnnow, which is what I've just added with the force-rebuild target. Feb 28 12:50:06 (assuming that all others apps are having PR updated properly) Feb 28 12:50:20 I really suspect the 'now' tag is arbritrary Feb 28 12:50:48 in the sense of "we need to finish this before going public" Feb 28 12:54:40 Anyway, I believe what i just committed to mokomakefile will give external devs the ability to not have everything rebuild every time, but be able to comply with the goal of the openmoko core devs (svnnow for a specific set of packages only) on demand Feb 28 12:55:16 irrespective of the sensibility of the goal, or whether that's the best way to use openembedded or not. Feb 28 12:57:20 * rwhitby notes that it has just gone March where he is ... Feb 28 12:57:38 it's early march - where's my Neo :-) Feb 28 12:59:17 * rwhitby kicks off a build, and goes to bed. Feb 28 13:07:54 * koen files some more bugreports Feb 28 13:25:09 mickeyl: i'm gonna start writing a patch to get udev and hal startup times down. could you do me a favour and do /sbin/udevd --daemon; time /sbin/udevsynthesize ? Feb 28 13:25:19 on the neo hardware? Feb 28 13:25:40 (or in fact anyone with some neo hardware..) Feb 28 13:26:54 * SpeedEvil wishes... Feb 28 13:27:25 We need a 'scripts for people with neos to run' on the wiki. Feb 28 13:29:20 * LuitvD might be braindead after tomorrow... Feb 28 13:30:00 but if I'm sane after tomorrow I'll try to do whatever I can do for openmoko :) Feb 28 13:30:21 after tomorrow's presentations, I'll finally have some free time :) Feb 28 13:30:27 lo Feb 28 13:30:38 hey... Feb 28 13:33:41 mickeyl: also can you tell me wether you have a /sys/subsystem with your current kernel Feb 28 13:35:01 robtaylor: try mailing mickeyl, better response time :) Feb 28 13:35:24 koen:well, its not a huge rush Feb 28 13:35:41