**** BEGIN LOGGING AT Sat Mar 27 02:59:56 2010 Mar 27 03:05:44 rwhitby, okay, I was trying to figure this out earlier and couldn't so I tried a fresh compile. Still happening. I'm building off of the 1.4 doctor and I'm getting erros building evas, namely that it can't make the loaders/png because it can't find -lpng12. I can't figure out how to debug this at all. Mar 27 03:07:06 and under my sb2 environment for arm7, there is a libpng12.so under /usr/local/lib Mar 27 03:07:23 gaffo_home: I'd move evas to nonworking and continue on. Mar 27 03:07:36 okay, what is evas anyway? Mar 27 03:07:42 and there are a few things that depend on it Mar 27 03:07:45 enlightment widget set stuff Mar 27 03:08:49 rwhitby: can you makescript magic the dep line so i can start organizing the widk? Mar 27 03:10:47 rwhitby: what if we change "${MAKE} -C ../../packages/$$dep stage" to "${MAKE} -C ../$$dep stage" Mar 27 03:11:03 so far that moves evas, expedite, efreet and ecore into non-working Mar 27 03:11:39 rwhitby: wouldn't that just make it look up 1 directory from wherever you or? or should I make an eif to ../*/ or something? Mar 27 03:12:04 gaffo_home: -lpng12 means it's having problems finding libpng12... it might be being staged wrong. I looked at it shortly the other day.. it's probably something simple Mar 27 03:12:22 zsoc, this is with the 1.4.0 doctor btw Mar 27 03:12:33 ah Mar 27 03:12:44 i'm not sure what doctor i had when running into that Mar 27 03:12:48 i think it happened with the 1.3.5 also tho Mar 27 03:13:18 and I have libpng12.so in /usr/local/lib when under the sb2 armv7 shell Mar 27 03:13:23 hmm Mar 27 03:13:35 well, I'm just trying to get the 1.4 stuff in so that we can get the native PDK hooks Mar 27 03:13:49 for things like PDL_GetCurrentPath and the like Mar 27 03:15:24 zsoc, but that evas autoconf makefile is so inpenetrable. 90% of it is unused garbage like DNA. and I can't get it to spit out what it's actually doing, it's trying to be nice about it's output. Mar 27 03:15:28 that's a good idea Mar 27 03:15:41 gaffo_home: eh, just shove it in nonworking, no one is offended Mar 27 03:15:56 I'm sure someone will be :) but they can just fix it. Mar 27 03:15:57 gaffo_home: soon evas, X, and gst/plugins will all me in extras/ and out of packages anyway Mar 27 03:16:06 s/me/be Mar 27 03:16:14 oh, cool Mar 27 03:16:25 gaffo_home: all of that autoconf is used... fyi Mar 27 03:16:31 in fact, I'm still unsure of the _purpose_ of 'packages/' Mar 27 03:17:14 there should just be a testing/ and packages/, and only have testing/ for making sure the toolchain is working Mar 27 03:17:18 well, for evas, on the loaders/png, not all of it is used. I cleaned and re-built that part and several of the flags didn't seem to be used. I could be wrong tho :) Mar 27 03:17:24 true Mar 27 03:17:34 it'd be nice if there was a clobber that didn't take out the toolchain Mar 27 03:18:19 I've had to clobber a few times and totally start over figuring if my environmen is sane Mar 27 03:19:15 gaffo_home: they are used. they exist for a reason. the configur eoptions and checks do useful things. there is a lot of code that may be enabbled, disabled or modified in behavior by these build options Mar 27 03:19:55 aah Mar 27 03:20:04 doesn't make it easy to debug tho :) Mar 27 03:21:04 well -lpng12 ... its trying toi link to libpng Mar 27 03:21:12 yeah, I get that Mar 27 03:21:17 evas built fine for me earlier today Mar 27 03:21:27 yes, in 1.3.5 Mar 27 03:21:29 :) Mar 27 03:21:43 I just cant figure out how it's trying to link (Eg it's hiding the options and paths) Mar 27 03:21:49 so you are only moving it to nonworking in the 1.4.0 branch? Mar 27 03:21:54 somewhere it asked pngk-config or libpng12 - libpng10 or libpng packages Mar 27 03:22:07 and one of these said "link to libpng by doing -lpng" Mar 27 03:22:15 so check the libpng*.pc files inb your build env Mar 27 03:22:26 they are what are saying to use that linking mechanism Mar 27 03:22:27 raster, that's a good idea Mar 27 03:22:56 look at evas_loader.m4 for the autofoo that checks for loader deps Mar 27 03:22:59 etc. etc. Mar 27 03:24:03 okay, the libpng ones both point to /usr/local/lib/lpng12 Mar 27 03:24:06 er Mar 27 03:24:14 /usr/local/lib and -lpng12 Mar 27 03:24:22 which is in /usr/local/lib Mar 27 03:24:30 there u go Mar 27 03:24:39 so why wouldn't it be able to find it Mar 27 03:24:46 everythign is pointed correctly Mar 27 03:24:49 so your .pc files - which well.. are part of your libpng dev install are saying to use the wrong thing Mar 27 03:25:00 well does libpng12.so exist? Mar 27 03:25:03 yes Mar 27 03:25:06 in /usr/local/lib Mar 27 03:25:09 if it says -lpng12 ... then u need a libpng12.so Mar 27 03:25:13 yes Mar 27 03:25:18 and that's in /usr/local/lib Mar 27 03:25:28 so it should link Mar 27 03:25:34 and that neeeds to be in the lib link path (specified by the -Ldir options) Mar 27 03:25:48 which is point ing at /usr/local/lib Mar 27 03:26:10 but are u building for your desktop?? Mar 27 03:26:13 or cross-building? Mar 27 03:26:14 http://webos.pastebin.com/6ZeN5F2M Mar 27 03:26:15 cross Mar 27 03:26:21 this is all cross widk stuff Mar 27 03:26:25 then /usr/local/lib is your host system stuff Mar 27 03:26:32 not your target Mar 27 03:26:39 not when I'm in SB2 Mar 27 03:26:46 your problem is withiin the target libslibdirs/etc. etc.. Mar 27 03:26:51 not sure about sb2 Mar 27 03:26:53 cant help there Mar 27 03:26:56 which is how all this builds, I'm pretty sure Mar 27 03:27:20 but libpng12.so shoudl be an arm elf shared library Mar 27 03:27:23 run file on it tyo see Mar 27 03:27:27 if its z86 - no wonder it doesnt link Mar 27 03:27:42 /usr/local/lib under sb2 maps to /srv/preware/cross-compile/staging/armv7/usr/lib Mar 27 03:27:56 but that'd be a differnt error Mar 27 03:27:59 like endian Mar 27 03:28:11 no Mar 27 03:28:26 ity will simply ignore the file if its not arm Mar 27 03:28:32 hrm, didn't know that Mar 27 03:28:39 when linking Mar 27 03:28:47 hrm. so looking closer, it looks like libpng12.so is a broken symbolic link Mar 27 03:28:55 aaaaaah! Mar 27 03:29:03 you should have said that before Mar 27 03:29:04 :) Mar 27 03:29:09 (when i asked if u have it) Mar 27 03:29:12 as yes - its a symlinks Mar 27 03:29:18 I did have it :) Mar 27 03:29:21 and i assumed u'd have checked the link is valid :) Mar 27 03:29:38 I thought the linker barfed more verbosely Mar 27 03:29:45 http://webos.pastebin.com/mq4CCd7t Mar 27 03:29:56 so why do we think that that symlink is being created invalidly? Mar 27 03:30:05 and how should we fix that? Mar 27 03:30:12 can't help u there Mar 27 03:30:37 hmm. Mar 27 03:30:43 u'll have to dig into whoever/whatever manages those libpng files Mar 27 03:30:52 but evas is doing the right thing Mar 27 03:30:57 not evas's fault Mar 27 03:30:58 :) Mar 27 03:31:27 true, but it comes down to should the evas thing keep me from pushing an update to using 1.4 to the WIDK Mar 27 03:31:37 rwhitby, is there like an issues list with the WIDK so I can mark that down and push? Mar 27 03:31:52 but thanks for the help, and sorry if I got a bit heated :) Mar 27 03:31:56 well evas found a broken part of your build env Mar 27 03:32:06 *our* :) Mar 27 03:32:27 so yes - it should stop it as it says "yoou have a broken env - this could be more pervasive in more subtle ways that have not turned up as errors yet and should be investiigated" Mar 27 03:32:29 like all debugging Mar 27 03:32:50 u can try fix the symptom Mar 27 03:32:57 or simply try and ignore it (by not buidling evas) Mar 27 03:33:13 u can anually fix the libpng symlink- that will patch the symptom up Mar 27 03:33:20 but u still dont knwo HOW it became like that Mar 27 03:33:28 will some future update.action re-create the same issue Mar 27 03:33:33 willl it keeep biting you? Mar 27 03:33:42 you don't know as u dont know how it happened :) Mar 27 03:35:10 it's clear that there is nothing here which is the fault of one particular package or another. any errors are the fault of the WIDK cross-compilation scratchbox2 build environment, cause the upstream packages obviously build perfectly in their native build environments. Mar 27 03:36:37 it's all everyones fault! Mar 27 03:37:14 gaffo_home: we should categorise the packages into groups (e.g. base, e17, x11, sdl, etc) and then have each group buildable separately. Mar 27 03:37:25 kk Mar 27 03:37:45 is that what zsoc is working on? Mar 27 03:38:28 no, but i've made the suggestion many times :) and may undertake it if it doesn't happen Mar 27 03:38:45 but we need to fix how deps are handled first Mar 27 03:40:49 what's the problem there? Mar 27 03:41:03 that things are always rebuilt? Mar 27 03:41:12 or never? Mar 27 03:42:41 gaffo_home: it checks a static location at the moment (packages/$$dep) Mar 27 03:42:53 obviously now we're going to have to have deps relative to where the package is building Mar 27 03:43:08 true. we also do multiple passes through Mar 27 03:43:18 well, not really relative, because if it's 1 level up then you can't share deps Mar 27 03:43:22 which also makes thigns difficult Mar 27 03:43:28 * zsoc shrugs Mar 27 03:43:33 i'm no makescript magician Mar 27 03:43:39 nor am I:) Mar 27 03:45:30 dang, now libsdl-mixer is failing, can't find avformat Mar 27 03:46:54 looks like libavformat.so.52 is in there but no symlink from libavformat.so to it Mar 27 03:47:50 zsoc: we could list deps as relative to the top-level, e.g. packages/foo, untested/bar, nonworking/baz, and then prepend ../.. in the for loop. Mar 27 03:48:02 might it be that since I'm pulling the SDL 1.3.5 source and patch, and I now need to pull the 1.4 ones? Mar 27 03:48:24 rwhitby: we'd have to update every makefile... Mar 27 03:48:38 gaffo_home: yes, all sources and patches should also be updated the 1.4, and the stage target in each makefile should create any missing symlinks in the staging area Mar 27 03:48:41 zsoc: yup Mar 27 03:49:01 rwhitby: i guess it's not the worst option. I mean right now i'm ../'ing anyway Mar 27 03:58:29 hrm, it looks like the libpng build has a patch and has had for some time, but we're not applying it. Mar 27 04:01:42 Cool, just got my first donation from an application developer who wants their app supported by Save/Restore Mar 27 04:03:37 800mhz ty Mar 27 04:04:34 rwhitby: hopefully it's a sign of things to come Mar 27 04:04:40 castle`: hello wircer Mar 27 04:04:54 word Mar 27 04:04:59 love wirc Mar 27 04:06:27 word up Mar 27 04:09:44 word Mar 27 04:09:47 lol Mar 27 04:12:32 oil: BTW, Save/Restore is an example of how the single package Preware can be done. Mar 27 04:12:42 yeah, thats good fun Mar 27 04:17:05 hopefully we can work out any kinks with it before we transition preware Mar 27 04:17:30 oil: any critical apps that you want supported by Save/Restore ? Mar 27 04:22:11 notes app? Mar 27 04:30:34 the built-in one? Mar 27 04:30:42 it doesn't save as part of your profile? Mar 27 04:32:28 rwhitby, pushed new updates to the doctor_14 branch. it's still not there and I need to start over with a recompile unfortunately. Mar 27 04:32:50 oh, idk rwhitby Mar 27 04:42:10 Hi Mar 27 04:42:41 I just did the 800 mhz patch and was wondering if a temp of 39C was too hot for the pre to be running at? Mar 27 04:43:26 don't know the answer to your question, but curious -- seeing much of an improvement? Mar 27 04:43:56 definetlly Mar 27 04:44:23 * rwhitby bbl Mar 27 04:44:48 So one thing I was noticing about the WIDK build is that the download cache is not based upon the palm version but the lib version, even if there are possibly differences. Mar 27 04:47:11 how do I unpack an ipk? Mar 27 04:48:25 manderton, what are you trying to do with it? Mar 27 04:48:58 learn from someone else's code Mar 27 04:49:08 make a few changes to a homebrew app Mar 27 04:49:10 7zip will take them apart Mar 27 04:49:10 etc Mar 27 04:49:37 how 'bout on linux? Mar 27 04:49:50 ^ Mar 27 04:50:04 they're not really zips are they? Mar 27 04:50:29 if by zip you mean the format, no Mar 27 04:50:40 they're ipx files Mar 27 04:51:02 just extract it with 7zip Mar 27 04:51:19 apt-get install p7zip-full Mar 27 04:52:32 perfect thanks Mar 27 04:53:17 np Mar 27 05:56:32 rwhitby, have you seen: http://developer.palm.com/distribution/viewtopic.php?f=70&t=5964 Mar 27 06:07:14 ugh, bout time for bed Mar 27 08:53:40 hi, can someone help me with the git commandline to clon doctor_14 instaed of master? Mar 27 09:07:01 How do I clone the webos 1.4 toolchain insted of 1.3.5? I serach the web for the right command but do not find them Mar 27 09:07:30 * hape Is used th cvs not to git Mar 27 09:20:10 hape: when you 'git clone' you actually grab the /entire/ repo incl all the branches--so then you just checkout the branch you want (git checkout doctor_14) Mar 27 09:28:06 ok, will try. thx Mar 27 09:38:26 where do i need to do that? when i'm in the cross-complie dir after the clone, then I ony get an error Mar 27 09:41:07 hape: should work, just tried it to verify too. what's your error? Mar 27 09:50:53 dtzWill: hjochman@tp-W500:/srv/preware/cross-compile$ git checkout doctor_14 Mar 27 09:50:56 error: pathspec 'doctor_14' did not match any file(s) known to git. Mar 27 09:51:29 It's new fresh cloend git Mar 27 09:54:34 hape: oh that command was kinda lazy. 'git checkout -b doctor_14 origin/doctor_14'? Mar 27 09:56:16 hape: i think earlier versions of git make you be more explicit, not sure Mar 27 09:58:12 dtzWill: worked, lets see if the rest now works. Mar 27 10:06:21 let's see if i now get compiled a working navit Mar 27 10:15:07 I added the info to the wiki, thanks Mar 27 10:36:26 hape: good to hear, ty for wiki edit :) Mar 27 10:53:36 hm, running on an error. dtzWill could you take a look please, i have no idea http://pastebin.com/PQE9N1Nd Mar 27 10:54:23 happen while 'make stage' Mar 27 11:03:03 hape: i'm about to crash sorry :). afaik the 1.4 stuff is a WIP, and I haven't been a part of that.. i think gaffo is leading the effort on that front Mar 27 11:03:49 ok, no problem. will wait if someone shows up how know. Mar 27 11:04:07 Happy to be able to help a bit to bring things forword. Mar 27 11:04:10 but mostly looks like the glib in 1.4 doesn't have that install target, not sure if that's not needed or if we should be doing something else Mar 27 11:04:11 thanks for your help Mar 27 11:04:19 hape: sounds good, you're welcome, and have a good one :) Mar 27 11:05:01 ALSO, grumbles @ PDK having directories like "$_OUTDIR" lol Mar 27 13:58:24 For anyone working on the Save/Restore GUI, the initial list service method now returns a "timestamp" field with a numeric seconds since the epoch value for the timestamp of the most recent save. It would be good to display that in the Restore Application Data list somewhere for each entry (different entries might have different timestamps) Mar 27 13:58:51 * rwhitby bbt - night all Mar 27 13:58:55 starting tomorrow, the pixi is free on an upgrade at R Shack. Mar 27 13:58:57 later Mar 27 15:09:58 hello Mar 27 15:25:17 hello Mar 27 15:25:36 rwhitby: i haven't forgot ;-) Mar 27 15:26:26 does anyone have some suggestions of some preware "taglines" I could add to my patch? Mar 27 15:26:51 * mjkjr thinks it's kind of lame and wants to update it Mar 27 15:27:53 * mjkjr will finally submit it to preware after that =) Mar 27 15:32:34 is it possible to prevent the pre from suspending? Mar 27 16:26:34 rwhitby, heyo, did you see that link I posted to you last night? That actually works just fine under linux, you just have to extract the PDK with 7zip and then install the linux version of the toolchain. I'm going to try it with some non-sample code. I keep getting all sorts of compile issues with the 1.4 doctor from the branch. Mar 27 16:31:38 probably a silly question. How do I "back swipe" in the emulator or minimize an app? Mar 27 16:33:41 Python run on a Palm Pre Plus? Mar 27 16:36:00 is it even possible to back swipe or minimize a window in the emulator Mar 27 16:36:05 mjkjr, esc I think Mar 27 16:36:10 s/window/card/ Mar 27 16:36:12 mjkjr meant: is it even possible to back swipe or minimize a card in the emulator Mar 27 16:36:35 gaffo_home: yep, thank you very much Mar 27 16:37:08 np Mar 27 16:44:26 i searched on p|c but didn't find anser so i have another dumb question, am i able to install preware or connect to webos qi using the emulator? Mar 27 16:47:09 i'm more than happy to read if someone could point me to some useful info. I don't need the answers handed to me, I just need to know where I can find the answers on my own. Mar 27 16:47:35 mjkjr, the emulator is all compiled for i686, so some of the stuff will work (webos tweaks) but other stuff (native) won't Mar 27 16:47:44 so like ssh servers wouldn't work. Mar 27 16:48:56 Is there a list of packages available for a hacked WebOS? Mar 27 16:51:00 hacked webos? Mar 27 16:51:04 does anyone hack webOS around here? Mar 27 16:51:06 lol Mar 27 16:51:09 I've never had to hack my webOS Mar 27 16:51:31 ;) Mar 27 16:51:37 patch? Mar 27 16:51:47 thank you gaffo_home Mar 27 16:52:03 that's right. Mar 27 16:52:33 np Mar 27 16:52:47 mischko: so what is a "hacked WebOS"? Mar 27 16:53:09 a WebOS where you can install your own apps on it? Mar 27 16:53:16 I'm brand new to the web-os community. Mar 27 16:53:30 Have a Centro now. Considering upgrading to a Pre Plus so I can do Linux on the phone. Mar 27 16:53:33 that's the regular WebOS, lol Mar 27 16:53:42 later all Mar 27 16:53:44 All the better. Mar 27 16:53:51 no "hacking" involved Mar 27 16:54:02 * chrisa twitches Mar 27 16:54:03 So it's not like a Linksys router's LInux. Mar 27 16:54:08 Linux Mar 27 16:54:31 More like a real linux? Mar 27 16:54:46 just upgrade from the centro and you'll see =) I did and I'm happy Mar 27 16:54:47 It's a standard OE type linux install Mar 27 16:54:56 Don't try to compare to centro, mainly Mar 27 16:55:21 I've been doing Linux for a long time. To be able to run things like Python on my phone will be very nice. Mar 27 16:56:10 I'm also interested in Adobe Flex programming. Anyone know what status is of Flash and Air on WebOS? Mar 27 16:58:12 I'd also love to run Firefox on it if it has enough horsepower. Mar 27 16:58:37 You're not going to see air on an embedded device anytime soon Mar 27 16:58:51 (if ever) Mar 27 16:59:19 chrisa, that's what I was thinking. I can live without that. Mar 27 16:59:41 Firefox would be very nice, especially if I can use the adblocking extension, etc. Mar 27 16:59:59 I'm trying to figure out how far I can go in replacing my laptop with one of these phoens. Mar 27 17:00:03 firefox would probably run, but you'd need to kill the luna ui to free up the framebuffer and have some sort of directfb-esc ff port Mar 27 17:00:06 there is an adblocking patch for WebOS Mar 27 17:01:23 mjkjr, cool. What web browser does it run? Opera? Mar 27 17:01:56 WebOS browser? Mar 27 17:02:09 It's a custom web browser? Mar 27 17:02:11 I get this feeling you've never looked at the OS or apps on a pre/pixi Mar 27 17:02:27 chrisa, right. I'm just now considering upgrading. Have not explored it yet. Mar 27 17:05:15 I'd suggest you just go into a store and play around, it's not like Maemo (which runs X) Mar 27 17:06:44 chrisa, I will. Sorry for the "newbie" questions. Mar 27 18:43:39 * mjkjr still needs some more taglines for his patch, no creativity flowing today... Mar 27 19:56:38 wow eh Mar 27 20:00:04 wheres that kernel dev wiki Mar 27 20:19:12 hi Mar 27 20:55:07 hi MetaView Mar 27 20:56:27 hi Mar 27 21:13:50 get an segmentation fault when I try to load a png image with IMG_Load() Mar 27 21:43:23 hello Mar 27 21:43:34 is there a way to power the pre on without the power button? Mar 27 21:43:37 my button is broken Mar 27 21:43:45 [; Mar 27 21:43:47 lol Mar 27 21:44:20 u should install the wake on lan patch beforeu had pre running Mar 27 21:44:21 :) Mar 27 21:44:29 hmmmm Mar 27 21:44:41 I think it'll wake when you put the usb in Mar 27 21:44:43 there is a shiny ftp-get app on my pre Mar 27 21:44:46 en0x: haha.. i should have :) Mar 27 21:44:47 spacerbar also Mar 27 21:45:14 datafirm, sprint just replaced mine for an issue like that Mar 27 21:45:15 gaffo_home: word. that did it. thanks Mar 27 21:45:19 np Mar 27 21:45:24 gaffo_home: good to know.. Mar 27 21:45:36 gaffo_home: mine has been dropped many times.. so. we'll see :) Mar 27 21:45:38 I used space for a while, then the keyboard went out Mar 27 21:45:48 mine had chunks out of it Mar 27 21:45:50 gaffo_home: space without the usb plugged in? Mar 27 21:46:51 yeah Mar 27 21:47:04 they did that cause it's hard to hit the power with the keyboard exposed Mar 27 21:47:22 cool. Mar 27 21:54:18 gaffo_home: I supped the only way to hard reset would be to pull battery then? Mar 27 21:54:26 since the power button is broke. Mar 27 21:54:39 sym+orange+r is just a restart? Mar 27 21:55:43 yap Mar 27 22:00:25 broken power buttons are Palm's all-time-favorit Mar 27 22:11:16 hi Mar 27 22:11:35 is here someone who have the carrier o2? Mar 27 22:11:58 jep, me in germany Mar 27 22:13:06 kk Mar 27 22:13:21 karotte_exee: why you ask? Mar 27 22:14:52 using IMG_Load() with SDL_INIT_NOPARACHUTE will crash Mar 27 22:15:17 o2 will release a Wifi-Hotspot app Mar 27 22:17:44 MetaView: hey, are you by o2? Mar 27 22:23:41 yes Mar 27 22:23:50 well, what do you mean? Mar 27 22:23:56 I'm not working for them Mar 27 22:24:05 it's just my carrier for the Palm Pre Mar 27 22:24:17 i have a strange problem with tethering Mar 27 22:24:45 I'm not using it. No need ot be honest. Mar 27 22:25:11 :( Mar 27 22:26:06 du know, when (approximate) o 2 will release the hotspot? Mar 27 22:26:20 *do Mar 27 22:26:26 it was a news on www.heise.de Mar 27 22:26:29 *do you Mar 27 22:26:35 danke Mar 27 22:26:38 bitte Mar 27 22:29:00 vielleicht kommts am 31. märz mitm vollem appcatalog Mar 27 22:29:17 vielleicht Mar 27 22:31:16 bb, i'm tired! Mar 27 22:41:57 chrisa: http://wdtz.org/pre_x/ice2.png in X running on the pre, in a card. no eed to kill luna or use directfb :) Mar 27 22:43:43 just saying :D Mar 27 22:44:35 fine :) Mar 27 22:45:59 zsoc: ! 'ello! :) Mar 27 22:46:18 dtzWill: ello gov Mar 27 22:49:27 every time l think there's nothing left to patch... 5 new patches come out Mar 27 23:04:10 Has anyone been able to get "Mode Switcher" to work? Mar 27 23:04:24 GreedyB: i'm imagine it's developer Mar 27 23:04:29 s/i'm/i'd/ Mar 27 23:04:30 zsoc meant: GreedyB: i'd imagine it's developer Mar 27 23:45:38 hi all Mar 27 23:46:21 Too bad webchat.freenode.net doesn't work correctly on my Pre. Mar 27 23:47:02 you don't use webos's irc client? Mar 27 23:48:26 no I don't, I tried and it didn't work correctly and I don't want to mess with it. Mar 27 23:48:59 It's about to cut out here in a second Mar 27 23:49:13 Just need one more line to pass Mar 27 23:49:41 ok, great. I have a question. Is it possible to have a verbose boot screen? Mar 27 23:53:17 I think it would be neat to see it start up like a regular linuxbox Mar 27 23:56:16 what's the startup manager for webOS called? Mar 27 23:56:38 upstart? Mar 27 23:57:20 simple as that ltm? Mar 27 23:59:58 a quick research sprint, there init also. So is it init or upstart? Mar 28 00:03:58 yes it's upstart. So it should be possible to enable a verbose startup instead of a boot screen Mar 28 00:04:41 i don't see why not, i'm no boot expert Mar 28 00:05:18 hmm, I'm going to check out the upstart IRC room. Mar 28 00:05:39 ha, I'm no expert at anything really Mar 28 00:13:39 gaffo_home, mjkjr: most packages in preware will work on both device and emulator. especially optware stuff. Mar 28 00:25:42 myearthhour.org Mar 28 00:49:32 Hi all! Question regarding the 720 and 800 MHz patches: can anyone explain what these do under the hood to a layman? I haven't been able to find that kind of info on PreCentral. Is there any frequency scaling or voltage scaling going on? It all sounds too good to be true so far. Mar 28 00:50:06 jbg7474: the kernel is rewritten, in a sense. That's the easiest way to explain it. Mar 28 00:50:16 jbg7474: a little of both. I think it essentially uses the profiles of higher end/newer OMAP CPUs Mar 28 00:51:21 It is rewritten and recompiled with some additional goodies that Palm left out? Mar 28 00:56:24 not exactly no Mar 28 00:56:36 it wasn't so much left out as it was not intended for the device in the first place. Mar 28 00:56:52 (overclocking isn't exactly to-spec for most hardware :P ) Mar 28 00:58:42 Okay, so instead of the Pre running at 500MHz all the time, it would run at 720 or 800 MHz all the time? The thing that's really bugging me is the battery life reports that say there's essentially no hit. Mar 28 00:58:58 Sounds like a free lunch, and I tend not to believe in them. Mar 28 00:59:36 jbg7474: no, there's battery life hit Mar 28 00:59:43 it's just less than 10% difference so i'm on board Mar 28 01:01:41 it does however, lower the chip's overall lifetime Mar 28 01:02:29 I remember back when 600 MHz experiments were being run, battery life was a major issue. Late summer, early fall timeframe. I'm wondering what has changed. Mar 28 01:02:52 Lifetime calculation would seem to depend heavily on whether there is any frequency scaling going on or not, no? Mar 28 01:03:48 yes, which there isn't, by default Mar 28 01:04:08 battery life was a big /concern/ back then with 600mhz, but it wasn't an /issue/ Mar 28 01:08:00 And Palm is already upping the frequency to 600 MHz for 3D games, is that correct? Mar 28 01:09:04 If so, that says to me that Palm is not too worried about running the silicon at 600 MHz, which means that another 15-20% probably is also okay. Is that pretty much the logic everyone is using in figuring that this is pretty safe? Mar 28 01:13:28 jbg7474: no, the logic is it's been thoroughly tested with heat monitoring Mar 28 01:21:08 also jbg7474, the chip was built for 600mhz Mar 28 01:21:18 palm downclocked it Mar 28 01:21:39 I think the reasons being passed around included battery life, and TI recommending it due to a bug in the chip? dunno Mar 28 01:22:02 Yeah, good point. Man, why am I feeling so mamby-pamby about this? :) Mar 28 01:26:34 zsoc: the heat monitoring is at the battery, "miles away" from the OMAP3430 junction temperature. Mar 28 01:27:01 rwhitby: that's good to know Mar 28 01:27:19 jbg7474: the kernel defines new OPP's based on the kernel patches for the OMAP3440, which is an 800MHz device. Mar 28 01:27:57 See http://twitter.com/webosinternals/status/11126494433 for the facts of the matter. Mar 28 01:28:59 as long as you're not going to do a fraudulent warranty or insurance claim if things go wrong, then go for it. Mar 28 01:29:30 rwhitby: please blame Puff that I have yet to restructure the widk tree Mar 28 01:30:31 rwhitby: the new OPPs is the key thing, then. And switching OPPs is done on the Linux side of things? Mar 28 01:31:43 jbg7474: yep, the patch is at http://git.webos-internals.org/?p=kernels/patches.git;a=blob;f=overclocking/overclock-patch - there's no magic, it's just defining new OPPs based on the 800MHz device voltage settings (which are well above the overdrive settings for the 600MHz part) Mar 28 01:33:05 OPP=operating program point or something like that? Set of voltages and frequencies? Mar 28 01:33:46 jbg7474: read the datasheet linked from my post Mar 28 01:34:27 OPP = combination of freq, voltage and PLL settings. device lifetime degradation is determined by time at various OPPs. Mar 28 01:35:23 Alright, thanks all. I'll do a bit more reading and come back if I have more questions (likely). Thanks zsoc, rwhitby, VincentLaw Mar 28 01:35:35 btw you might want to note that's the published numbers for the chip itself Mar 28 01:35:55 there is no public info that I know about what happens to surrounding circuitry Mar 28 01:36:26 OPP6 (720MHz) reduces device life from 100K hours to 44K hours. The 800MHz OPP7 would likely reduce it further. No-one knows by how much. Mar 28 01:37:18 jhojho: indeed, who knows what the limits are for the other chips on the PCB, although the freq and voltage is not affecting them, it's only an increase in heat which should have any effect there. Mar 28 01:38:07 the OPP modifications only affect certain domains inside the OMAP, not the surrounding circuitry outside the SoC (apart from thermal conduction effects) Mar 28 01:38:11 imo there will be an effect but we wont know without measurement if its negligible Mar 28 01:39:00