**** BEGIN LOGGING AT Tue Apr 06 02:59:56 2010 Apr 06 02:59:57 * dtzWill is starving and snaping on poor zsoc. time to make some food, bbiab Apr 06 03:00:06 it's nearly done downloading, I'll see what happens Apr 06 03:00:07 enjoy Apr 06 03:00:25 pwd Apr 06 03:00:28 oops Apr 06 03:00:33 ha Apr 06 03:01:19 * dtzWill is back. "make a meal now, or grab a beer instead and make dinner later?" wasn't a very difficult decision Apr 06 03:02:12 Jack87: haven't seen Jeff in here, but I gave him the URL to join Apr 06 03:03:29 rwhitby: thanks... he is asking me to write up a little step by step guide but only way i can contact him is via precentral PM which i hate. I suppose i can email him but not sure if thats any better. Apr 06 03:03:36 echo $CROSS_COMPILE Apr 06 03:03:39 oh ffs Apr 06 03:03:47 ha Apr 06 03:06:26 *sigh* Apr 06 03:06:28 tis too late Apr 06 03:06:32 I must be off to bed Apr 06 03:06:54 would anyone be interested in a kernel tree of the palm kernel with the realtime patches applied that at least compiles? Apr 06 03:07:02 goodnight sbromwich Apr 06 03:07:12 sbromwich: put it somewhere, certainly Apr 06 03:07:59 sbromwich: definitely Apr 06 03:08:01 k Apr 06 03:08:22 I was really hoping to get the -rt kernel going with this :-( Apr 06 03:08:25 what's the command i can use to run novacom again w/o restarting my pc Apr 06 03:10:05 How does palm-install work? Apr 06 03:17:35 zsoc: novacom put, then executes a luna-send to call the application install service Apr 06 03:18:21 so.. i can just novacom put it.. and then ipkg install it? Apr 06 03:18:36 rwhitby: or do i need to use this magical install service since the service and app are packaged together? Apr 06 03:20:08 zsoc: oh, you need BLDFLAGS = -p Apr 06 03:20:20 i have no idea what that means Apr 06 03:20:25 the palm-install can execute the postinst if the postinst is written correctly Apr 06 03:20:30 i see Apr 06 03:20:34 it's in the packaging makefile Apr 06 03:20:55 tells it to build a palm-format package with PmPostInstall.script Apr 06 03:21:11 hmm.. i think i had that Apr 06 03:21:19 save/restore uses it? Apr 06 03:21:22 yep Apr 06 03:21:34 then probably Apr 06 03:21:54 ipkgservice (Java version) 0.9.45 in the testing feed for Preware to Preware Alpha upgrade testing Apr 06 03:24:32 rwhitby: hi Apr 06 03:24:42 Bmyers: g'day Apr 06 03:24:59 still looking for help on that date thing? Apr 06 03:25:12 Bmyers: sure Apr 06 03:26:52 what its the save and restore git that i should look work with? Apr 06 03:30:49 rwhitby: its the save and restore git that i should look work with? Apr 06 03:32:33 Bmyers: yep Apr 06 03:36:59 So, who thinks we should include the webos-internals-testing feed in Preware by default, but disabled? Apr 06 03:37:34 http://www.fop.ns.ca/pre/linux_palm_rtpatch_tar.bz2 (in about 20 seconds) Apr 06 03:37:36 not me Apr 06 03:37:54 and it's copied Apr 06 03:38:22 that was patches with the rt patch and then the palm patch and then the overclock patch, if I recall correctly Apr 06 03:38:49 if needs be I can probably get one of the RT guys from novell to come talk to anyone who's genuinely interested in implementing it Apr 06 03:38:59 but for now I must go... laters :-) Apr 06 03:40:58 is there a .. uh.. palm-/uninstall/? Apr 06 03:54:24 * jacques just registered for the dev days Apr 06 03:56:58 hi JayCanuck Apr 06 03:57:15 hi jack87 Apr 06 03:58:10 jacques, kewl Apr 06 03:58:29 ka6sox, :-D Apr 06 03:59:33 JayCanuck: someone has a request for you..... i just noticed Apr 06 04:00:07 ? Apr 06 04:00:25 JayCanuck: haha i know there are a million requests... but do what you please.. here is the link... http://forums.precentral.net/palm-pre-tips-information-resources/238141-b-optimized-webos-1-4-kernel-video-wosqi-720mhz-800-mhz-easy-install-b-124.html#post2361929 Apr 06 04:04:16 hmm Apr 06 04:04:39 Is character counter working for anybody else? I've seen a few threads of people saying the patch is working for them with 1.4.1 installed. Apr 06 04:04:49 But for me it's just installed as a placeholder. Apr 06 04:05:01 dtzWill, I know you're not around but I just wanted to let you know that I made a *ton* of progress on the X window manger front. I've got a custom fvwm setup that works beautifully with virtual desktops, I'll post a vid tomorrow so look for it Apr 06 04:05:20 JayCanuck: seems like it might have to be an app all in it self rather then a patch? Apr 06 04:05:33 jack87, maybe at 55*C I can give a warning that if it stays above 55*C, in 5minutes, the device will shut down. sound good? Apr 06 04:06:16 definitely do-able as a patch :) Apr 06 04:06:24 JayCanuck: ya i think that sounds good.. haha not my request :-) i just figured id relay you since i ran into you here. Apr 06 04:06:24 phil_bw: congrats! Apr 06 04:06:35 Jack87, thanks Apr 06 04:06:54 I have just about made X on the Pre a "user friendly" experience Apr 06 04:07:25 phil_bw: thats awesome Apr 06 04:08:10 with very little effort can load it up, open up a document in openoffice, switch virtual desktops if something goes off the screen, all with very little memory impact Apr 06 04:08:12 JayCanuck: could it be a patch in the screen settings to add it as an option? Apr 06 04:08:40 phil_bw: does it run in SDL? Apr 06 04:08:56 would be significant more work as those are 2 separe applicatinons and would need to communicate together Apr 06 04:09:33 Jack87, yes Apr 06 04:09:46 so as my wife texts me I just switch between X and the text Apr 06 04:09:52 JayCanuck: roger that.. probably not worth the effort at least for now. be good to get input after you try your first idea Apr 06 04:10:17 zsoc|away: palm-install -r, or orange+tap Apr 06 04:10:22 phil_bw: i cant wait to see the video Apr 06 04:10:32 JayCanuck: ping Apr 06 04:10:37 hi rod Apr 06 04:10:50 Jack87, I was going to make the vid tonight but wound up screwin around thw fvwm for way longer than I had anticipated Apr 06 04:10:52 JayCanuck: hi, did all your patch updates go through correctly? Apr 06 04:11:24 the updates I sumbitted, I believe, yea Apr 06 04:11:54 phil_bw: what do we need to do to start packaging that work up and releasing it? Apr 06 04:12:21 JayCanuck: so that was me doing them last night, still no sign of dBsooner Apr 06 04:12:36 JayCanuck: but I don't get notifications of new ones, so ping me if you upload any Apr 06 04:13:23 haven't checked yet (@ a friend's place at the moment) though I did get the approval notifcationd Apr 06 04:15:34 rwhitby, well since the server part is all dtzWill (and it is already in the testing feed) I'd say not much on that end Apr 06 04:16:17 phil_bw: but we need to set up a debian chroot infrastructure, right? Apr 06 04:16:22 rwhitby, packaging up a whole debian environment and configuring it to work right via chroot is a completely different task that I'm sure only you and jason could answer Apr 06 04:17:01 rwhitby, correct, my hope is to eventually build an img file that is already configured with fvwm and some minimal apps Apr 06 04:17:23 so if all was done right you'd just install the server from preware then drop the img file on the usb drive and load it up Apr 06 04:17:46 then just use apt-get (or a visual alternative if you wanted) for the apps you want Apr 06 04:18:14 install the server? Apr 06 04:18:55 phil_bw: what about webOS customisations of the apps you apt-get install ? Apr 06 04:19:07 (presumably most would need some?) Apr 06 04:19:17 the tough part is that the img file would be already be just about max capacity when it's made lean and mean Apr 06 04:19:35 and most users don't have access to a linux box to expand it (mine's at 1.5GB and i'm pushing needing to size it up again) Apr 06 04:20:00 rwhitby, I'm not sure I get what you mean by customisations Apr 06 04:20:07 tmzt_, what about the server? Apr 06 04:20:42 what is it? Apr 06 04:21:10 a debootstrapped image is pretty small. Apr 06 04:21:16 tmzt_: X Server in webos-internals testing feed - install it via preware Apr 06 04:21:49 just use the filelist from that initally Apr 06 04:21:50 ok Apr 06 04:22:17 tmzt_, it's just an X server with nothing attached to it, you can do with it what you want, I just happen to use a debian environment through chroot where I have X apps connecting to it Apr 06 04:22:37 phil_bw: customisations like geometry and such, or file paths etc Apr 06 04:23:00 ka6sox, it's easy to get a small debian image, it's the resizing later that'll be tough for users Apr 06 04:23:20 phil_bw, kk Apr 06 04:23:27 does anybody have a .c program to mount and chroot Apr 06 04:23:28 rwhitby, I'm not sure about geometry, I've found that every app I've tested sets itself to the max screen width by default anyway Apr 06 04:24:06 maybe even as a service Apr 06 04:24:32 the other option, which I've thought about but haven't experiment with would be to basically bind the /var back out to the /media portion Apr 06 04:25:14 you mean X apps? Apr 06 04:25:26 yeah, if there's no wm or geom they usually run full screen Apr 06 04:25:32 except xterm and a few other programs Apr 06 04:26:24 what's the irc username of the guy who runs prethemer? Apr 06 04:26:41 and fvwm has a virtual desktop so if windows go off the screen you can easily go find them Apr 06 04:27:16 phil_bw: we currently bind mount /var/opt over to /opt for optware Apr 06 04:28:07 phil_bw: but /var is quite small unless you've meta-doctored it to 2GB Apr 06 04:28:27 2GiB /var FTW Apr 06 04:29:47 rwhitby, I'm thinking more of binding /var *in* the debian portion back *out* to the /media portion Apr 06 04:30:20 phil_bw: oh, ok. well, note that nothing under /media has symlinks or permissions - it's vfat Apr 06 04:30:32 rwhitby, crap... that's right Apr 06 04:30:48 so back to figuring out how to size the img Apr 06 04:31:28 I've looked but it doesn't appear you can get resize2fs as a windows app Apr 06 04:31:40 phil_bw: note that Palm have said in the PDK video that they are going to have some sort of chroot security solution for PDK apps by mid-year Apr 06 04:32:05 rwhitby, interesting Apr 06 04:32:46 jail or someting more like dsl/puppy packages? Apr 06 04:33:03 tmzt_: we have no info Apr 06 04:33:31 that's usual Apr 06 04:34:58 well I better get going Apr 06 04:35:04 talk you you all tomorrow Apr 06 04:39:00 Finally, there’s the issue of the App Store. I’m on the fence about it. My hunch is that Apple should follow Palm’s lead and allow users to install applications from the web, Apr 06 04:39:06 is that correct? Apr 06 04:39:13 I think Android had that first Apr 06 04:41:53 phil_bw: great news! :D be sure to ping me if you get that video--I've been swamped with work and haven't had a chance to put the final polish on it to make an announcement, etc Apr 06 04:42:10 phil_bw: and i'm very glad more than just me is pondering how to make a good experience with this :) Apr 06 04:42:37 JayCanuck: are there other temperature warning patches to be submitted? Apr 06 04:42:48 phil_bw: interesting idea re: binding back /out/ :) Apr 06 04:42:57 there's only 1 warning patch Apr 06 04:43:20 2 device info app patches (C and F) and 2 device menu patches (C and F) Apr 06 04:43:39 JayCanuck: ok, I think I got those last 4, but not the warning patch Apr 06 04:43:54 warning one was uploaded by dbsooner a long while ago Apr 06 04:43:57 * rwhitby loves how Canada uses the correct temperature scale ;-) Apr 06 04:44:02 :p Apr 06 04:44:16 C is the general standard for computer compents anyway Apr 06 04:44:21 phil_bw: fvwm sounds good, we could probably package that up separate from any debian work Apr 06 04:44:44 I had a surprrisingly large # of people who wanted F, so i obligued ;) Apr 06 04:45:10 dtzWill: we need some sort of system when different installed apps can put files in a central location and they are automatically added to the menus and stuff Apr 06 04:46:17 rwhitby: Im working on sending 10or so updated patches as we speak Apr 06 04:46:28 rwhitby: we absolutely do, I agree. as to *how* that's not as clear. if we support a particular wm (or two, _maybe_) we might be able to hack that together, similar to the way some commercial linux applications install themselves onto the gnome and kde menus Apr 06 04:46:34 this 5 min wait in between is killing me Apr 06 04:47:04 ok, I'm back for a sec Apr 06 04:47:21 phil_bw: wb :) Apr 06 04:47:24 thnx Apr 06 04:47:45 dtzWill, so I think fvwm is the answer to X usability on the Pre Apr 06 04:48:06 it's *incredibly* customizable and it's virtual desktops make it work so well Apr 06 04:48:21 phil_bw: that's great news. I was hoping some wm would fit the bill :) Apr 06 04:48:37 phil_bw: j/w, how much customization did it take? :) Apr 06 04:49:12 if we decide that's the wm we want to support it might be worth to just stage that and build it so people have a working wm. Apr 06 04:49:21 dtzWill, well first I had to make it *not* look like windows 95 which is how it defaults, killed the task bar and this big ugly button list that had buttons for apps that weren't even installed Apr 06 04:49:29 it's hard to draw the line where we should make it preware-happy and where we should just say "build a debian image" Apr 06 04:49:38 phil_bw: haha Apr 06 04:50:17 dtzWill, I also tweaked the "desker" (virtual desktop) module so it was more fat finger friendly Apr 06 04:50:29 tomorrow I will mod the menus to make them fat finger friendly as well Apr 06 04:50:31 dtzWill: I'd say a windows manager fits the bill of being worth packaging separately Apr 06 04:50:36 phil_bw: wonderful! were there any source changes, or just configuration files, etc? Apr 06 04:51:22 dtzWill, just config files, one shell script could easily convert it over, infact it comes completely stripped down and you rin the "95 script" to make it look like windows 95 Apr 06 04:51:49 rwhitby: agreed, but at some point we can't exactly launch our own complete distribution--so how much do we build (and as separate packages, etc?) ourselves and how much do we say "use a debian image" (that perhaps we provide and automate and make more usable) Apr 06 04:51:50 I could easily take the "95 script" and modify it to do what we want, perhaps change up the icons to Pre-looking ones (instead of direct MS rips) Apr 06 04:52:27 dtzWill: yeah, it's a fine line. Similar to the one I have to draw with optware - which packages do I add to preware, and which do I leave as command line install onlyl Apr 06 04:53:27 phil_bw: that sounds great. we should definitely get on that, that sounds fantastic Apr 06 04:54:06 rwhitby: and of course there's the issue you mentioned of wanting to make these applications more pre-friendly (the ones we leave to debian chroot to provide) Apr 06 04:54:17 rwhitby: maybe we SHOULD fork our own debian distro xD Apr 06 04:54:41 * phil_bw sees his little remaining free time making it's way down the toilet Apr 06 04:54:47 what about XPde? Apr 06 04:55:01 mostly variants of kde crystal I think Apr 06 04:55:48 also, if someone was able to build a desktop utilty to customize the img on the fly we could go that route Apr 06 04:56:00 pretty much just have to resize it Apr 06 04:56:01 * dtzWill thinks of german windows XP :P Apr 06 04:56:17 there's multiple versions of that, jigdo-image, mic2, etc. Apr 06 04:56:34 ...do we want the wm tolook like xp necessarily? Apr 06 04:56:40 tmzt_, windows versions? Apr 06 04:56:43 bah i don't wanna tell anyone what wm to use :(. Apr 06 04:57:09 hmm, I thought you were going for Windows 95 look Apr 06 04:57:26 no, those are for building debian/ubuntu images with custom packages Apr 06 04:57:26 tmzt_, trying to do away with the win95 look that fvwm defaults to Apr 06 04:57:38 oh Apr 06 04:57:51 they'd even run on the Pre Apr 06 04:58:15 i tried running resize2fs on my pre once Apr 06 04:58:19 ....I don't recommend it Apr 06 04:58:24 haha Apr 06 04:58:29 dtzWill, so did I Apr 06 04:58:34 also don't recommend it Apr 06 04:58:52 moving a lot of data? Apr 06 04:59:09 resizing on flash must be ... slow Apr 06 04:59:25 jacques: it is, and made the entire system absolutely unusable Apr 06 04:59:33 ended pulling the battery after hours :/ Apr 06 04:59:43 I kept meaning to bench the flash access on the pre - it seems like < 5MiB/s Apr 06 04:59:59 and just copying it properly. which, btw, scp -C rather helps when what you're copying is mostly 0's :) Apr 06 05:02:47 I suppose just supporting one wm officially and letting tinkerers ...tinker (and perhaps help them with some tools) isn't so bad a way to go Apr 06 05:03:32 I'm probably going to lose my sat connection soon - boat is moving back under cover Apr 06 05:03:54 jacques: man you seem to travel an awful lot :) Apr 06 05:03:55 jacques: it's a tough life ... Apr 06 05:04:41 jacques, but somebody's gotta do it. Apr 06 05:04:46 dtzWill, only recently :-) Apr 06 05:05:11 hm Apr 06 05:05:38 the guys in #linux say try mingw or cygwin Apr 06 05:06:05 tomorrow they load the ROVs onboard Apr 06 05:07:17 phil_bw: try those for what? resizing/customizing the debian chroot? Apr 06 05:07:35 i think requiring people to instal cygwin is..something to be considered carefully Apr 06 05:07:55 dtzWill, well if you could compile resize2fs to windows you could build a tool who's sole purpose is to size the img before they put it on their device Apr 06 05:08:00 no no Apr 06 05:08:12 oh you mean an env you can build windows binaries from Apr 06 05:08:12 okay Apr 06 05:08:17 right Apr 06 05:08:34 then someone could build a tool that contains the small img (that the user doesn't have to see) Apr 06 05:08:50 they open it and it simply asks how big they want it, then it spits out the new img file in that size Apr 06 05:09:01 and if they ever want to size up their current one they could use the same tool for that as well Apr 06 05:10:09 no telling if it'll actually work Apr 06 05:10:46 I suppose. i mean it's certainly moving forward. maybe I'm just stuck on finding the uber solution :). seems undesirable to have people dealing with a chroot, and sizing them, etc Apr 06 05:11:02 but that does get us a lot of apps, quickly, and possibly comparatively easily Apr 06 05:11:21 right, good luck cross compiling openoffice to run natively... ya know Apr 06 05:11:43 phil_bw: well first I'll grab the way the debian guys do it.... :) Apr 06 05:11:45 if all the user had to do was download one desktop utility and push a button it would be fine Apr 06 05:11:46 heading to the mess, biab (or back tomorrow if I lose the satellite) Apr 06 05:12:06 in fact the desktop utility could install the webos package as well (similar to preware) Apr 06 05:13:33 anyhow, I haven't seen my wife all day and she just got home so I am going to take off, I'll talk to you guys later, again if you need anything and I'm not around just drop me a line at phil@boraware.com Apr 06 05:14:01 phil_bw: alright, thanks for help and thoughts. have a good one Apr 06 05:14:07 later Apr 06 05:14:10 * phil_bw wanders off Apr 06 05:49:56 looks like they're testing the thrusters, so I still have sat :-) Apr 06 05:50:34 jacques: testing the thrusters? are you headed to the ISS or something? Apr 06 05:53:44 rwhitby, I'm on a DP boat Apr 06 05:54:37 dynamic positioning - computer controlled thrusters all around which can hold station very accurately Apr 06 05:54:46 bah, iPhone multitasking Apr 06 05:55:00 different levels for holding position in various sea conditions Apr 06 05:56:55 they were working on the boat under a huge shed which blocked sat. now they seem to be sailing around the port Apr 06 05:57:08 we never know what the captain's going to do :-) Apr 06 06:03:20 anyone online know much about the topic of software licenses? Apr 06 07:17:31 sup folks? Apr 06 07:27:20 question: has anyone else noticed that certain websites are the primary culprit in the infamous "too many cards" error? Apr 06 07:28:50 the full Facebook site, and speedtv.com both bring it on inevitably if I browse around them enough, even if nothing else is open Apr 06 10:49:13 Who's testing the new versions of Preware Alpha, Preware and Package Manager Service? Apr 06 10:49:21 ~testing-feed Apr 06 10:49:22 [testing-feed] a means of testing new WebOS Internals products, documented at http://www.webos-internals.org/wiki/Testing_Feeds Apr 06 10:54:24 has anyone else seen this msg on their device? "LunaSysMgr: {LunaSysMgr}: void CpuShareManager::writeEntry(const char*, int): Failed to open cpu share file: /dev/cgroup/game/cpu.shares Apr 06 10:54:25 " Apr 06 10:54:48 happened to see it while grepping through /var/log/messages for something else Apr 06 11:46:37 whoever tried to HEAD http://www.fop.ns.ca/pre/linux_palm_rtpatch.tar.bz2 it probably won't work as commands are restricted on my web server Apr 06 11:46:46 it's 320 meg Apr 06 11:50:37 I think I finally worked out why Preware is sometimes hanging when scanning unknown files. Reading a big file into a buffer that is too small. Apr 06 11:50:57 (seems that some little ipkg control files are actually larger than 4K) Apr 06 11:58:20 rwhitby: ccing you on a message about noop vs cfq Apr 06 11:59:44 hiya, does anyone know if the qOperatorShortName or qOperatorLongName in the CarrierNetworkSettings.db3 also controls what operator name to display on the screen? Apr 06 12:00:19 ie the text that tells you waht operator you are on at the moment. Apr 06 12:03:10 rwhitby: what's the difference between build.git/hardware and patches.git/overclocking Apr 06 12:05:01 Reason im asking about the qOperatorShortName is that there is a possibly typo on the latest carriernetworksettings.db3 where french operator SFR is called SRR. =) Apr 06 12:10:02 jhojho: the latter stores the patches, the former builds the kernel Apr 06 12:10:18 ok Apr 06 12:10:34 do I have write perms for patches.git/overclocking? Apr 06 12:10:41 the kernel build may pull patches from a number of places, including sbromwich's web site ;-) Apr 06 12:10:44 jhojho: yes Apr 06 12:12:42 oh is that what's doing the pull? Apr 06 12:12:52 I can extract out the tarball if it would help Apr 06 12:13:59 I've also asked the VP of SW dev at novell if he's aware of anyone else porting -rt to phones before and what their experience was Apr 06 12:14:13 sbromwich: so why are you trying to get rt patches working on the pre? Apr 06 12:14:29 because pulseaudio is very latency sensitive Apr 06 12:14:43 ah Apr 06 12:15:02 -rt makes it try harder to keep in sync so the sounds happen when things happen on screen instead of a perceptible few milliseconds later Apr 06 12:15:19 I will hold back on my rant about pulseaudio as I fear it would be inappropriate. Apr 06 12:15:35 lol. no. I'm well aware of it from other forums Apr 06 12:15:55 yes, it's... *cough* well known. Apr 06 12:16:04 not personally invested in audio/rt so doesnt really bother me Apr 06 12:16:29 what I would like are the unreleased patches for omap from ti =) Apr 06 12:16:48 it gives a big performance boost for other stuff as well, particularly for video, so I'm thinking that might fix the problem with artifacts people are moaning about Apr 06 12:16:54 yeah, wouldn't that be nice? Apr 06 12:17:11 well as for the jittering Apr 06 12:17:20 i think 125mhz is too low from my testing Apr 06 12:17:36 it's a matter of tweaking the settings Apr 06 12:17:36 mornign all Apr 06 12:17:48 rwhitby or zsoc: thanks for adding news room :) Apr 06 12:17:59 250mhz seems to work for me but have not done comprehensive testing with the powersave_bias param Apr 06 12:18:02 eg I posted a list on the forums of my settings and explicitly said they were extremely conservative because I have a "bad" cpu that won't stick at 800 Apr 06 12:18:12 I saw the post Apr 06 12:18:15 and then people complained my settings were too conservative and caused artifacts Apr 06 12:18:19 *roll eyes* Apr 06 12:18:47 for me upthreshold of 20 and powersave_bias (cant remember) works Apr 06 12:19:16 really need a group of people that are willing to test out various params using different governors Apr 06 12:19:18 *nod* Apr 06 12:19:39 the other thing is people setting retardedly high values for the sensor so the cpu has the wake up repeatedly to check the load Apr 06 12:19:48 have you confirmed that modifying LunaSysMgr does not matter when you do not use the userspace governor? Apr 06 12:19:53 which thereby drains the battery chop chop fasta benj Apr 06 12:20:16 have I tried both userspace and ondemand, do you mean? Apr 06 12:20:52 meaning that modifying LunaSysMgr to 600000 or higher is not needed if using the ondemand or conservative governor Apr 06 12:20:58 if I use userspace the palm daemon will frig with the frequencies (as expected) Apr 06 12:21:08 correct Apr 06 12:21:43 if non-userspace daemon is selected the palm stuff doesn't touch any of the cpufreq that I've noticed Apr 06 12:21:54 sorry, non-userspace governor Apr 06 12:23:57 actually I've noticed that lunasysmgr does not try to much with things when I use clipcarl's binary too Apr 06 12:24:08 i.e. 600mhz lcd on, 250 lcd off Apr 06 12:24:36 I experimented with my own userspace daemon but found ultimately the userspace one did just as well without needing 6 meg of ram for perl Apr 06 12:24:49 http://www.fop.ns.ca/pt.pl is where I put it, I think Apr 06 12:24:56 little more than a state machine Apr 06 12:25:17 what's clipcarl's binary? Apr 06 12:25:25 you have to search the forums Apr 06 12:25:33 k Apr 06 12:25:48 it's called cpuspeed along with a little event.d script call cet-cpuspeed Apr 06 12:25:50 precentral forums? Apr 06 12:25:52 yes Apr 06 12:25:58 oh, that one, o0k Apr 06 12:26:25 also, testers needed to see if noop/deadline performs better than cfq on the pre Apr 06 12:26:30 AnOutsider: yw Apr 06 12:26:56 sbromwich: no, we're not pulling anything from your website, it was just an example of a possibilty Apr 06 12:27:15 oh, have you seen my comments passim about /var and /var/log in /etc/fstab, and logging in /etc/PmLog.conf? Apr 06 12:27:23 it's in the current kernel patch but no reason given as of yet why the change was made and I have not seen any actual test results. Apr 06 12:27:54 rwhitby: It's a complete compiled tree so it might be easier if someone just wants to run a diff against it rather than downloading all the object code as well Apr 06 12:28:20 sbromwich: are you asking me re: /var ? Apr 06 12:28:23 my suspicion would be cfq was getting overwhelmed by the amount of data being written to /var/log Apr 06 12:28:27 sorry, yes jhojho Apr 06 12:28:34 yes I have. Apr 06 12:28:44 k Apr 06 12:28:45 agree that /var should be =ordered Apr 06 12:29:04 not sure why it's not but that's a change I would rather palm engineering weigh in on Apr 06 12:29:12 I found setting /var/log data=writeback not /var to be a major, major performance increase for me in certain areas where I saw it Apr 06 12:29:14 mmm Apr 06 12:29:34 their thought process Apr 06 12:29:53 sbromwich: really? measured or just subjective? Apr 06 12:30:01 it looks to me like a thinko Apr 06 12:30:02 sbromwich: at the moment I'm focused on just getting a stock kernel with overclocking ability correctly packaged, so I won't be getting diverted into -rt kernels Apr 06 12:30:41 jhojho: simple test: I have a 4 digit PIN, if /var/log is not writeback and syslog is enabled it lags entering the pin, otherwise it's responsive Apr 06 12:30:46 ok np Apr 06 12:31:01 rwhitby: I was hoping that unixpsycho would show up to make sure he's okay with others modifying his patch Apr 06 12:31:04 (and right at the moment I'm testing Preware to Preware Alpha migration corner cases ...) Apr 06 12:31:07 I shall leave it up there in case anyone else is interested or I'm stuck on another con call and can look at it Apr 06 12:31:09 heh Apr 06 12:31:44 sbromwich: would be interesting if you could do more testing with the writeback changes as well as cfq vs noop if you are interested Apr 06 12:32:03 jhojho: we can put multiple candidate patches in that directory, suitably named Apr 06 12:32:19 i have a suspicion that palm was pressed for time and left quite a few of the options stock. Apr 06 12:32:20 I'm interested but massively lacking in time Apr 06 12:32:30 sbromwich: well, as you have time =) Apr 06 12:32:32 jhojho: and if you want to rename the current one to indicate things like cfq, feel free Apr 06 12:32:44 ok. Apr 06 12:32:59 (just update it in hardware/uber-kernel-pre/Makefile at the same time to match) Apr 06 12:33:16 ok Apr 06 12:34:07 all I can say now is WFM Apr 06 12:34:38 heh. Apr 06 12:34:45 also I can scroll the launcher window up and down very quickly, so perhaps in my CFT I should make yet another youtube video to show that. Apr 06 12:35:09 yeah, we probably need more data points than that to put it in widespread distribution =) Apr 06 12:35:17 sbromwich: did you make a comment about whether or not to have the testing feed in Preware but disabled? Apr 06 12:36:13 rwhitby: btw were you involved at all in caj2008's kernel ipks or was that all him? Apr 06 12:36:26 jhojho: Jason created them Apr 06 12:36:30 rwhitby: yeah, it should be a minimum IQ test to gain access Apr 06 12:36:36 along the lines of alt.sysadmin.recovery Apr 06 12:36:41 rwhitby: ah okay. Apr 06 12:36:48 sbromwich: very much so Apr 06 12:38:32 sbromwich: another thing for you to play with is that rwhitby has added powertop to the optware feed so you can see the power consumption on the pre Apr 06 12:38:49 heh Apr 06 12:38:56 I installed powertop second day I got the pre Apr 06 12:40:00 I've also got a killawatt meter and did some testing... standard draw is 2 watts when turned on, 3 watts when working (resolution of the killawatt is 1w) Apr 06 12:40:39 and I think the reason people are burning out chargers is because the palm one provides a full amp not 500mA, so my guess is the palm is seeing a "dumb" charger and trying to draw a full amp Apr 06 12:41:25 eg that muppet that tried to power his pre off of a dinky solar charger and complained when shit burnt out Apr 06 12:43:24 Preware Alpha 0.9.58 is in the testing feed, with a couple of good bug fixes. Apr 06 12:46:41 what's the upgrade procedure, rwhitby? Apr 06 12:46:45 installed, luna is taking its own sweet time to restart Apr 06 12:46:48 just install it and it removes package manager? Apr 06 12:47:30 hmmm... preware terminate but it doesn't look like luna is restart Apr 06 12:47:31 ing Apr 06 12:47:56 VincentLaw: this is the version which puts the upgrade procedure in place, so it's probably a case of whatever works for this version Apr 06 12:48:07 ok Apr 06 12:49:00 sbromwich: luna should never need to restart for preware Apr 06 12:49:23 (since preware doesn't modify any UI files) Apr 06 12:49:24 yeah, package manager service needs a java restart, not luna Apr 06 12:49:29 I got a "luna needs to restart" message Apr 06 12:49:38 luna or java? Apr 06 12:49:42 ah, maybe I wasn't paying close enough attention Apr 06 12:49:44 errr Apr 06 12:49:56 gonned now Apr 06 12:50:03 ipkgservice will ask for a java restart. preware alpha shouldn't Apr 06 12:50:07 java will restart and you won't see anything happening besides the phone going offline Apr 06 12:50:21 nope, not going offline Apr 06 12:50:31 restarted preware and it says 0.9.58 Apr 06 12:50:34 preware alpha is replacing ipkgservice specifically to get rid of the java restart Apr 06 12:50:55 if it says 0.9.58, then you started Preware Alpha, not Preware. Apr 06 12:51:26 btw caj2008 just emailed and mentioned he'll be joining the fun at webos-internals. Apr 06 12:51:29 sorry, yes, I should be more specific Apr 06 12:51:45 perhaps name it pralpha or something to make it easier for the hard of thinking like me? ;-) Apr 06 12:52:22 I expect it will only last for a couple more weeks Apr 06 12:52:28 (maybe 3 or 4) Apr 06 12:52:28 package feeds needed refreshing but otherwise everything looks good, just looking for sometime useful (or at least not totally useless) to install Apr 06 12:52:30 holy CRAP Apr 06 12:52:31 k Apr 06 12:52:36 preware alpha installs patches fast Apr 06 12:52:56 VincentLaw: that's interesting, cause it shouldn't Apr 06 12:53:08 sbromwich: install the temp in device menu patch by jaycanuk =) Apr 06 12:53:21 already done Apr 06 12:53:32 well. game over then Apr 06 12:54:13 just managed to lock up the device though poking randomly through Apr 06 12:54:14 rwhitby: yeah, I almost wonder if it installed correctly Apr 06 12:54:28 hmm - those temp in device patches should really be in the topbar category, not other Apr 06 12:54:34 still, this wasn't a particularly important conference call, they can wait while it reboots Apr 06 12:54:40 rwhitby: it installed the "Share Larger Videos Via Email" patch in literally 0.5 seconds. No exaggeration. I couldn't even read what it was doing it went so fast. Apr 06 12:55:22 and I just found removing the battery while the usb cable is plugged in and charging will do a full reset, handy Apr 06 12:55:33 rwhitby: agreed. (where the patches should be). has dbsooner resurfaced or are you packaging patches now? Apr 06 12:55:33 was there a problem with the 4x4 v4 plus more patch? I just was told by preware to update it for the second time in two days, and now I'm back to stock Apr 06 12:55:52 sbromwich: yeah. that was a surprise. the treo never did that Apr 06 12:55:56 jhojho: I did the last round of updates, but either dBsooner or egaudet did a round after that Apr 06 12:56:18 idontwan2know: dunno, have you asked the author? Apr 06 12:56:20 ah Apr 06 12:56:29 edit any video's metadata went pretty fast as well, not as fast as the previous, but still much faster than standard Preware. Apr 06 12:56:41 VincentLaw: did they actually install? Apr 06 12:56:51 I'm restarting luna to find out right now Apr 06 12:57:13 time to restart FF. sucking up 1.32GB of mem is just so wrong Apr 06 12:57:22 jaysus the palm forums are even worse than the precentral forums Apr 06 12:57:41 Jailbreak, patch install is it safe? Apr 06 12:57:42 i've never been to the palm forums Apr 06 12:57:42 ?! Apr 06 12:57:48 you're not missing much Apr 06 12:57:50 yep, the patches did seem to install! Apr 06 12:58:12 I'm trying to figure out where the best place is to report my /var / /var/log writeback findings Apr 06 12:58:17 sbromwich: palm developer forums or public forums Apr 06 12:58:27 whatever caused it I'm not complaining. Patches used to take ~5-8 seconds to install, now they take ~1. haha Apr 06 12:58:58 rwhitby: are they on forums.palm.com or other? Apr 06 12:58:58 VincentLaw: as long as it's writing all the info required to remove them safely too ... Apr 06 12:59:07 sbromwich: developer.palm.com Apr 06 12:59:10 rwhitby: I'll test removal for you :p Apr 06 12:59:15 ta Apr 06 13:00:16 rwhitby: he's not here to ask and there isn't any contact info other than a nick...I'll just find another version to install Apr 06 13:00:19 said "SUCCESS", restarting luna to check. Apr 06 13:01:56 rwhitby: uninstalled patch successfully! Apr 06 13:03:45 VincentLaw: well, I don't have a theory to put behind your empirical evidence Apr 06 13:04:08 me either, but I'm not complaining Apr 06 13:04:34 I /am/ running the 800mhz patch, so maybe it benefits more from the extra power, I don't know. Apr 06 13:06:36 oh fuck I have to go some idiot rolled out a new client package for patient registration to 1500 users "accidentally" Apr 06 13:07:07 ROLLBACK Apr 06 13:07:37 sbromwich: shouldn't be a problem, all the code you write is 100% bug free at all times, right? RIGHT? :p Apr 06 13:09:25 not my code Apr 06 13:09:40 written by a healthcare vendor so it's shoddy crap Apr 06 13:09:49 :p Apr 06 13:10:31 ask slashdot had a thread on healthcare software a few years ago and cerner and mckesson both got slammed for their piss-poor code Apr 06 13:10:43 ok, now that the known bugs in Preware Alpha are fixed, it's time to fix the bugs in the palm catalog conversion script Apr 06 13:11:09 Apr 6 10:04:54 h9328n1 auth|security:info sshd[1757932]: Accepted password for rlm003 from 172.24.57.202 port 3303 ssh2 Apr 06 13:11:14 sorry, wrong window Apr 06 13:11:36 * rwhitby is glad he's not in rlm003's shoes right now Apr 06 13:11:46 lol Apr 06 13:12:56 just trying to demo ssh is working for login and it's mckesson at fault Apr 06 13:13:00 "ssh isn't working!" Apr 06 13:13:11 yes it bloody well is, how do you think I'm logged in to kill your sessions? Apr 06 13:13:22 etc. Apr 06 13:17:39 sbromwich: magic! Apr 06 13:17:41 * jhojho offers sbromwich a LART Apr 06 13:22:52 it's going to take more than one at this rate, I'm going to need an arsenal Apr 06 13:23:08 3 of our hospitals have effectively stopped working and are unable to accept patients Apr 06 13:40:11 ok, palm feeds on ipkg.preware.org should additionally show packages which are not released in english Apr 06 13:43:48 jhojho: ping Apr 06 13:43:57 pong Apr 06 13:44:11 you have questions for me about schedulers? Apr 06 13:44:54 oh yeah Apr 06 13:45:44 I was asking if there was a reason for switching to noop vs cfq Apr 06 13:46:38 i havent seen any testing talk so was wondering... Apr 06 13:46:49 in a flash based system it is pointless to add latency of an IO scheduler since there are no elevators to tune. :) Apr 06 13:47:08 true. Apr 06 13:47:28 hello Apr 06 13:47:52 thanks for all your help!!! Apr 06 13:48:56 just wanted to know Apr 06 13:49:58 hi caj2008 Apr 06 13:50:15 hey Apr 06 13:50:35 learning quite a bit here Apr 06 13:50:54 jhojho: so is there an actual benefit to changing to noop, or is it just that there is no benefit to using cfq? Apr 06 13:51:06 rwhitby: see http://www.alphatek.info/2009/02/02/ssd-performance-vs-linux-kernel-io-scheduler-in-fedora-10/ Apr 06 13:51:25 and http://www.alphatek.info/2009/02/02/io-scheduler-and-ssd-part-2/ Apr 06 13:52:18 so in general the thought is that noop/deadline should help but it's a good idea to test to verify Apr 06 13:52:23 jhojho: oh, the change is to the IO scheduler, not the CPU scheduler? Apr 06 13:52:28 yes Apr 06 13:52:31 AND Apr 06 13:52:43 the change can be made by modifying a /proc entry Apr 06 13:52:46 nice article! Apr 06 13:52:51 which is why I wanted to pull it out of the patch Apr 06 13:52:51 jhojho: excellent Apr 06 13:52:59 anything that is scriptable should be Apr 06 13:53:09 jhojho: yeah, enable both but default to stock and have a script change it Apr 06 13:53:09 that would keep the kernel patch as small as possible Apr 06 13:53:16 exactly Apr 06 13:53:20 go for it Apr 06 13:53:27 same thing with the upthreshold Apr 06 13:53:58 unixpsycho logged off before i could ask him if he was cool with others modifying the code but i have no problem modding it. Apr 06 13:53:59 as such Apr 06 13:54:08 go for it Apr 06 13:54:32 alright. I'll try to find some time to make some mods. Apr 06 13:54:41 I will test it out Apr 06 13:55:40 okay Apr 06 13:55:44 jhojho: the patch was put in that place intentionally as a common area. Apr 06 13:55:53 maybe you can help test the deadline scheduler too Apr 06 13:55:56 (by unixpsycho) Apr 06 13:56:02 ok Apr 06 13:56:12 deadline won't help, seek latency is too low on flash Apr 06 13:56:17 I agree with the noop argument myself Apr 06 13:56:24 jhojho: he will still maintain his own site for his bleeding edge kernels Apr 06 13:56:36 he has a site? Apr 06 13:56:45 yes Apr 06 13:56:49 http://unixpsycho.com/palm-pre/ Apr 06 13:57:16 sbromwich: now that I realise it's IO schedulers and not CPU schedulers, I agree too. Apr 06 13:57:22 ah Apr 06 13:58:02 nice! Apr 06 13:58:10 do kernels still use .config? its been ages since I worked on the kernel =) Apr 06 13:58:30 ages=over 10 years ago Apr 06 13:58:32 jhojho: yep. I included the defconfig changes in the patch Apr 06 13:58:49 for the sirloin defconfig that is Apr 06 14:00:09 rwhitby: do you want to maintain config changes as part of that patch or as a separate piece. imo it should be broken out Apr 06 14:00:23 with a diff specifically for .config changes Apr 06 14:00:33 makes it easier to read that way. Apr 06 14:00:55 jhojho: you saw that the patch is a diff of the .config changes, right? Apr 06 14:00:58 and to track too. as none of the .config changes are as a result of a kernel source change Apr 06 14:01:23 rwhitby: right, just wanting input if that should be a separate patch Apr 06 14:01:34 because it could be applied to a stock kernel too Apr 06 14:01:36 jhojho: but yes, it does make sense to separate, especially since a single config will need to cover multiple patches from different sources. Apr 06 14:01:53 all we are doing in there is flipping more bits on Apr 06 14:02:00 okay Apr 06 14:02:09 (when different functionality patches are merged into a single unified modular kernel) Apr 06 14:02:33 awesome. looks like we have agreement there. Apr 06 14:08:42 jhojho: over the next couple of days (with egaudet), I'm going to add prologue and epilogue functionality to the aupt-4 postinst and prerm to allow for the creating the device node required Apr 06 14:08:54 (and to allow for remounting rw of /boot) Apr 06 14:09:01 cool Apr 06 14:09:12 nice touch Apr 06 14:09:24 yeah. kinda need that for installing a new kernel =) Apr 06 14:09:55 jhojho: currently the Makefile builds the kernel and then puts it in the additional_files location for aupt to handle Apr 06 14:10:18 I need to get with egaudet to make sure aupt-4 handles those correctly md5sum-wise Apr 06 14:10:33 ok Apr 06 14:11:49 May I ask a question on an issue with the current ipk (dont mean to change topic) Apr 06 14:11:57 go ahead Apr 06 14:13:04 With a minority of users, they stll get 600 MHz for either 720 and 800 kernel. I have found that install/unistall .sh file/install ipk fixes this and I have no idea why. I can send you the ipk to look at Apr 06 14:13:53 any idea?? Apr 06 14:14:50 double check scaling_max_freq? Apr 06 14:15:12 my suspicion is half the people on the forums are just blindly cut and pasting and missing bits out Apr 06 14:15:24 sounds plausible Apr 06 14:15:34 Script is the same. The problem has been minimized but not entirely eliminated (we put in safegurads, I will send you the new ipk's) Apr 06 14:15:38 sbromwich: you have to remember that over 50% of the population are below average intelligence Apr 06 14:15:45 lol Apr 06 14:16:06 I think the average intelligence has diminished over the past decade too Apr 06 14:16:11 rwhitby: per my cousin, it's not an even bell curve distribution Apr 06 14:16:14 that or september really *is* never going to end Apr 06 14:16:17 it's skewed to the left Apr 06 14:16:18 I cant believe how much time is wasted on stuff explained clearly on the first page Apr 06 14:17:05 I added wget to make .sh file download easier Apr 06 14:17:16 well, I have a rule that I never answer the same question twice in the same thread in a 24 hour period Apr 06 14:17:35 you are smarter than me... Apr 06 14:17:45 is perl part of the default install? Apr 06 14:17:52 sbromwich: nope Apr 06 14:17:56 bum Apr 06 14:17:58 sbromwich: but it's in optware Apr 06 14:18:20 I was toying with writing a "run this YFI" utility to tell them exactly what their system is set up for Apr 06 14:18:27 Sending the new ipk's by email in 5 min Apr 06 14:23:12 * rwhitby bbt - night all Apr 06 14:23:27 greetings earhtlings Apr 06 14:23:50 hey AnOutsider, you just caught me. did you see the Saved Package List stuff? Apr 06 14:24:38 AnOutsider: where did we get to with themes and aupt-4? I was just chatting with Jason that we should get theme support into aupt-4 before we push it out Apr 06 14:27:20 saved package list stuff? Apr 06 14:27:31 also, nothing has changed in the themes packaging in ages Apr 06 14:27:41 do we have a spec for aupt and themes now or? Apr 06 14:27:56 yep Apr 06 14:27:57 :) Apr 06 14:28:33 however I'd like to request egaudet to specify a different, less partisan, file location for the theme identifier Apr 06 14:29:39 JayCanuck: agreed, no need for org names to be in a common non-partisan standard Apr 06 14:30:08 AnOutsider: http://bit.ly/holy-grail-of-homebrew Apr 06 14:30:09 okie dok Apr 06 14:30:15 is this info in the wiki or? Apr 06 14:30:28 ah yes, I did see that the other day Rod Apr 06 14:30:47 I actually isntalled the Alpha this morning and played a bit Apr 06 14:31:13 AnOutsider: http://git.webos-internals.org/?p=preware/build.git;a=blob;f=autopatch/postinst.testing and http://git.webos-internals.org/?p=preware/build.git;a=blob;f=autopatch/prerm.testing Apr 06 14:31:40 AnOutsider: 0.9.58 has some important bug fixes - a buffer overrun error causing the hangs when scanning unknown packages Apr 06 14:32:09 AnOutsider: and 0.9.58 should correctly uninstall Package Manager Service without removing all the feeds Apr 06 14:32:29 (it leaves Preware there) Apr 06 14:32:29 just had the crashed phone/email/calendar/IM issue again Apr 06 14:32:35 I really wish Palm could figure out that bug Apr 06 14:32:37 .58 of the base preware or C preware Apr 06 14:32:56 AnOutsider: base preware is 0.9.35 - C preware is 0.9.50+ Apr 06 14:33:02 Ah yeah, it did have all my feeds when I opened the alpha Apr 06 14:33:37 AnOutsider: if you've got ipkgservice < 0.9.45 installed, it will remove the feeds when it is uninstalled Apr 06 14:34:22 so I'll be pushing an update to the java ipkgservice too, so that it no longer does that (all feed setup is now in the postinst instead of being static files that ipkg uninstalls automatically) Apr 06 14:34:33 nope, updated the package service this morning, so I think I'm good Apr 06 14:34:47 where are the primary speed improvements in the alpha? Apr 06 14:34:51 package installation? Apr 06 14:35:06 AnOutsider: so a good test is to install preware 0.9.35, ipkgservice 0.9.45, and then install preware-alpha 0.9.58 over the top and make sure it cleanly removes ipkgservice. Apr 06 14:35:51 AnOutsider: the logic is basically the same, except it downloads the package first using curl (hence the progress stats) and then ipkg installs from the local file Apr 06 14:36:17 whereas the rpevious version called ipkg on the remote url? Apr 06 14:36:29 AnOutsider: apart from that, just the speed of C with staticly allocated buffers vs Java strings ;-) Apr 06 14:37:00 AnOutsider: Java version allows ipkg to choose the download URL based on the feeds Apr 06 14:37:11 AnOutsider: C version has that URL passed down from Preware Apr 06 14:37:52 AnOutsider: that allows for things like installing something which is not the latest version of a something Apr 06 14:38:13 roger that Apr 06 14:38:24 but is mainly for the download stats Apr 06 14:38:24 at the very least, it's really good having everything in one package now Apr 06 14:38:40 yeah, that should cut out at least 50% of support issues Apr 06 14:38:41 updating preware and PMS individually was a bit of a pain Apr 06 14:39:06 btw, don't remember if I said it this mornign, but thank you or zsoc for adding newsroom Apr 06 14:39:20 AnOutsider: no probs - I did that today over lunch Apr 06 14:39:47 pity it doesn't sync with google reader Apr 06 14:40:10 i know, manually entering urls can be a pain. and the app is prone to flipping out and deleting my feeds sometimes Apr 06 14:40:12 AnOutsider: yes, one might say I convinced rod to do it Apr 06 14:40:15 so this is invaluable Apr 06 14:40:23 hehe, either way, thanks guys Apr 06 14:40:25 AnOutsider: someone else might say he read my mind Apr 06 14:40:53 AnOutsider: still, other's might say he did it on his own accord Apr 06 14:41:07 hmm why "VERSION=aupt-1" currently the aupt scripts on the git? Apr 06 14:41:28 nvm, my mistake Apr 06 14:41:52 so, real quick, did anything ever come of moving themes to css patches? Apr 06 14:43:21 I think they had some ideas, but I believe the css idea had some big issues, like not affects several key images folders, and the frustrating "!important" css keyword Apr 06 14:43:54 hmm, the !important bit could likely be solved with some css trickery Apr 06 14:44:07 what that about the image folders though? Apr 06 14:44:42 like image folders used by lunasysmgr Apr 06 14:46:57 I have an interesting idea for themes Apr 06 14:47:07 do tell Apr 06 14:48:50 well, as you may or may not be aware, recently both rod's ipkg scripts and my jpkg-webos library were updated to include support for palm's pmPostInstall.script and pmPreRemove.script Apr 06 14:49:54 basically allowing for installation/uninstallation via palm's ApplicationInstallUtility Apr 06 14:50:02 * zsoc_wirc wonders if themes could just go away... Apr 06 14:50:18 zsoc_wirc: themes are paying for AnOutsider's retirement ;-) Apr 06 14:50:27 haha hardly Apr 06 14:50:44 me and rod recently came to an agreement on a how services should function under this system Apr 06 14:50:47 what doesnt go to server bills goes to WOI donations Apr 06 14:51:06 which we then spend on server hardware :) Apr 06 14:51:08 * AnOutsider somewhat agrees with zsoc Apr 06 14:51:12 dont some of themes mess with aupt4 compatibility of some of the patches?? Apr 06 14:51:25 nope Apr 06 14:51:42 not if themes are installed separate to patches in themes Apr 06 14:52:00 themes are currently completely outside of the scope of AUPT, and are considered to be manual uncontrolled changes if they touch the same files as patches Apr 06 14:52:03 ok Apr 06 14:52:32 me and rod agreed on having a /media/cryptofs/apps/usr/palm/applications/whatever/ folder with an appinfo and icon Apr 06 14:52:36 makes sense Apr 06 14:52:46 (for services) Apr 06 14:52:46 but all the right stakeholders are in agreement to remedy that in the fullness of time Apr 06 14:53:11 (mainly so you can orange+tap remove them) Apr 06 14:53:23 orange tap themes? Apr 06 14:53:28 so themes would be installed as apps Apr 06 14:54:00 with the appinfo.json and the icon.png, under the new postinst/prerm format, you can keep the launcher icon hidden, but allow the app to remain listed in the app list and be user-removable without any special software Apr 06 14:54:13 AnOutsider: I was referring to orange tapping services. I'm not up to speed on what Jason is describing yet Apr 06 14:54:28 gotcha to both Apr 06 14:54:57 orange-tap themes would be so cool. Apr 06 14:55:07 my suggestion would be to include the same format on themes, for user-conventient uninstalls :) Apr 06 14:55:11 my goal is to have themes as self-sufficient as possible. it becomes a pain managing it all (especially between OS updates), so that's why I was open to the CSS thing Apr 06 14:55:48 So, having post install scripts officially from palm, think homebrew will make it to the catalog? Apr 06 14:55:51 well, aupt is a huge step forward with that respect Apr 06 14:55:56 nope Apr 06 14:56:05 I know themes wouldnt (or shouldnt)-- too much can go wrong to be officially distributed Apr 06 14:56:49 AnOutsider: nope, I've already asked, and the word is that all scripts are stripped at submission time. only Palm can officially use them. (and only we can unofficially use them ;-) Apr 06 14:57:03 heh, fair enough Apr 06 14:57:21 it would be a nightmare in any case Apr 06 14:57:27 random post install scripts would cross the bright line of app catalog security Apr 06 14:58:06 I am surprised they didnt excecute a cease and desist order on some of this but they know secretly waht is best to maintian user satisfaction Apr 06 14:58:15 can I use reize2fs and just resize my already working debian chroot on the pre to be however much bigger I need it to be? or is other steps involved Apr 06 14:59:07 caj2008: uh... for what? Apr 06 14:59:09 caj2008: WebOS Internals, from day 1, has always done things in a manner which complies with all copyrights and carrier terms of service. There is nothing that can be the target of a cease and desist. Apr 06 14:59:27 They have very very clearly articulated their policy on tht caj2008 they know people will push the limits. Apr 06 14:59:59 anything that indirectlt manipulates Palm source material?? Apr 06 15:00:11 well I'm trying to make the debian chroot bigger Apr 06 15:00:27 caj2008: as long as you don't redistribute the Palm material, you are not making yourself a cease and desist target. Apr 06 15:00:33 caj2008: how is that grounds for C&D? Apr 06 15:00:37 yep!! Apr 06 15:00:52 but I'm wondering if I can just do resize2fs or I need to start over with the chroot Apr 06 15:00:56 preware and anything in it doesn't redistribute palm materials ... Apr 06 15:01:17 gelato: a resize is going to be slow on the Pre - best to copy off, resize on desktop, then copy back Apr 06 15:01:52 cool, and the file structure will remain the same? Apr 06 15:02:07 gelato: so slow, it tends to crash, I've tried Apr 06 15:02:20 gelato: yeah, just do md5 checksums in and out Apr 06 15:02:33 nice thanks all Apr 06 15:03:35 I'm turning the pre into a nice little pen tester and I need more space Apr 06 15:03:38 gelato: and keep a spare copy of the image before resizing Apr 06 15:03:57 hmm good idea rwhitby Apr 06 15:04:02 gelato: no wifi master mode though, right? Apr 06 15:04:05 also e2fsck has saved my neck a few time with bad resizes Apr 06 15:04:36 gelato: and we learnt yesterday the Pre doesn't have the required voltage regulator for OTG host mode Apr 06 15:04:45 zsoc_wirc, you have to run an e2fsck to do a resize first. Apr 06 15:04:51 so you can't do a usb wifi stick Apr 06 15:05:15 rwhitby: yeah but I got some other ideas Apr 06 15:06:04 ka6sox: er... yes certain options for it I seemed to remember helped after... Apr 06 15:06:13 was awhile ago tho Apr 06 15:08:28 gelato: I did it by editing /etc/fstab, commenting out /media/internal, rebooted, delete the LV, recreate smaller, mkdosfs it, re-enable in /etc/fstab, reboot Apr 06 15:08:56 palm-webos-device cpufreq # vgdisplay | egrep 'Alloc|Free' Apr 06 15:08:56 Alloc PE / Size 807 / 6.30 GB Apr 06 15:08:56 Free PE / Size 168 / 1.31 GB Apr 06 15:08:56 palm-webos-device cpufreq # Apr 06 15:09:06 WFM. Apr 06 15:09:53 also an extra swap LV helped me for what I was doing Apr 06 15:12:50 * nt4cats yawns Apr 06 15:13:09 sbromwich: thanks that might help with some other plans..i want to repartition /media/internal Apr 06 15:13:21 easiest is just to blow it away Apr 06 15:13:43 (allow me to not state the obvious re: backups since you know what you're doing) Apr 06 15:15:18 sbromwich: haha yeah Apr 06 15:16:34 lets hope so Apr 06 15:16:48 certainly if you're pen testing ;-) Apr 06 15:17:25 i got a usb wifi adaptor to work with rfmon using an externally powered usb hub Apr 06 15:17:25 VincentLaw: are you sure those patches were installing correctly? Apr 06 15:17:55 clunky to carry around with you and requires external power Apr 06 15:17:57 loot-: connected to a Pre? Apr 06 15:18:00 yea Apr 06 15:18:12 i dont use it though because its easier to just use my laptop honestly Apr 06 15:18:14 loot-: and you got the USB hub to enumerate? Apr 06 15:18:35 rwhitby: this was with 1.3x and a custom kernel but ya Apr 06 15:18:44 it was weird though Apr 06 15:18:46 loot-: are you able to share the details with destinal? Apr 06 15:18:48 it would only work with one device at a time Apr 06 15:19:04 i can probably recreate yes Apr 06 15:19:55 i just doctored my phone a couple weeks ago and have been extremely busy lately so i havent had time to hack up my phone again Apr 06 15:20:15 enumerating usb would be sweeeeet Apr 06 15:20:25 loot-: we are *very* keen to know how you got enumeration going Apr 06 15:21:11 rwhitby: ill pull out the hub and try it again tomorrow night Apr 06 15:21:20 loot-: thx Apr 06 15:21:21 i have dinner plans tonight otherwise i'd start today Apr 06 15:21:41 it was a huge pain though i do remember that Apr 06 15:21:43 hey just a quick ? does anyone no if you can use the chat for friendsbook ?? Apr 06 15:22:05 loot-: no rush, but we do want the info ;-) Apr 06 15:22:29 loot-: we're building a replacement kernel, so if that can be incorporated we want to do it Apr 06 15:22:35 unfortunately i do remember having to do some things to the kernel Apr 06 15:22:37 ah ya Apr 06 15:22:40 that would work well then Apr 06 15:22:56 i had my own pre patch going for awhile Apr 06 15:23:06 back ported grsecurity patch to the pre's kernel version Apr 06 15:23:09 loot-: the replacement kernel will be installable via Preware Apr 06 15:23:18 excellent Apr 06 15:23:36 now we just need palm to fix the "everything runs as root" issue and i will be completely satisfied with my pre Apr 06 15:23:38 loot-: transparently built from source and patches on the autobuilder Apr 06 15:23:56 rwhitby: how would the kernel ipk support uninstall/rollback? Apr 06 15:24:07 nt4cats: aupt Apr 06 15:24:23 rwhitby: hmmm, that sounds remarkably well thought-out. Apr 06 15:24:39 it sounds excellent IMO Apr 06 15:24:56 imagine having a grsecurity hardened kernel available to just click install on the pre Apr 06 15:25:00 anyone have a idea ? Apr 06 15:25:06 with pre-generated RBAC policy for luna and such Apr 06 15:25:13 loot-: http://git.webos-internals.org/?p=kernels/patches.git;a=tree Apr 06 15:25:15 DrFunk: I didn't understand your question, actually Apr 06 15:25:32 loot-: and http://git.webos-internals.org/?p=preware/build.git;a=tree;f=hardware/uber-kernel-pre Apr 06 15:25:40 ah speaking of Apr 06 15:25:58 how well does the 800mhz OC patch work with seido extended batteries? Apr 06 15:26:10 so glyder 2 and wirc run at the same time fine with 800mhz Apr 06 15:26:13 i noticed gaming on the regular stock palm battery it got hot enough to cook an egg Apr 06 15:26:19 loot-: I'll point you to caj2008 for that one Apr 06 15:27:04 and im annoyed with my touchstone... it left a big ugly ring stain in the wood of my night stand Apr 06 15:27:13 so the kernel is the same one as we had for 1.4.0? Apr 06 15:27:23 nt4cats: chat with friendsbook it's not facebook but the other Apr 06 15:27:38 nt4cats: g'day Apr 06 15:27:41 well en0x with embedded gear it is just much easier to backport important patches and keep consistent dependency versions Apr 06 15:27:46 en0x: yep, but that one I pasted is not released - I'll be waiting for official source code Apr 06 15:27:51 especially when everything relies on kernel headers for building and what not Apr 06 15:27:56 * nt4cats waves at zsoc_wirc Apr 06 15:28:09 DrFunk: are you asking if the frindsbook supports a chat feature? Apr 06 15:28:16 s/supports/app supports/ Apr 06 15:28:16 nt4cats meant: DrFunk: are you asking if the frindsbook app supports a chat feature? Apr 06 15:28:22 yes Apr 06 15:28:47 or is it just facebook Apr 06 15:29:09 oh another negative btw Apr 06 15:29:12 while im thinking about it Apr 06 15:29:28 getting the phone to enumerate usb also meant that it broke novacom entirely Apr 06 15:29:39 and refused to charge over USB Apr 06 15:30:04 im sure that can be fixed eventually but that was one of the problems i was having when i last worked on it Apr 06 15:30:07 well so till the official kernel source+patches won't be released by palm i'm not gonna overclock it with the old kernel Apr 06 15:30:56 luckily i was using the terminal app and openssh over wifi while testing Apr 06 15:31:05 DrFunk: I don't use that app, but I don't see a mention of chat. Apr 06 15:31:14 and touchstone charging still worked IIRC Apr 06 15:31:39 en0x: we're not aware of any incompatibilities Apr 06 15:31:48 ack so nt4cats it's just for facebook than Apr 06 15:31:52 DrFunk: I did see instructions for 'tricking' your Pre's messaging application into connecting to Facebook chat Apr 06 15:32:14 DrFunk: want me to see if I can find those instructions for you? Apr 06 15:32:32 sure if you don't mine Apr 06 15:32:36 mind* Apr 06 15:33:50 bleh... now i want to mess with my pre again and i have to continue bug fixing some crappy PHP someone desperately needs Apr 06 15:34:27 DrFunk: I am not specifically endorsing this technique -- I just noticed an article about it. It could be the best thing ever (or maybe not). I haven't tried it ... http://www.precentral.net/facebook-chat-comes-messaging-patch Apr 06 15:35:11 nt4cats: got it that's for facebook Apr 06 15:35:16 thanks Apr 06 15:38:23 So the alternate machine we were using for autobuilding didn't have lsdiff installed, and so was building incorrect patch packages for those updated since the autobuilder broke. They're all rebuilding correctly now and should upload soon. Apr 06 15:38:39 YIPPEE!!! Apr 06 15:39:22 oh i thought my pre died ... while installing the 720mhz kernel i got cannot mount / ... and it shut down ;) Apr 06 15:39:26 for loot, I found use of the 2600 Seidio back minimizes heating and that for reasons unknown an uninstall/reinstall process can improve on unusual activity seen like excessive loss of batter and heating Apr 06 15:39:53 i almost sh... myself Apr 06 15:40:06 I dont know why the latter seems to clear up issues with some of the users..just something we learned Apr 06 15:41:32 My seidio 2600 goes over a day with constant use and on touchstone temp never exceeds 41C Apr 06 15:43:00 did I answer your questions, loot? Apr 06 15:46:39 anyone know how to paste in the terminal app Apr 06 15:47:17 or use symbols I need | Apr 06 15:48:22 anyone? Apr 06 15:48:37 nfoxTc: :http://www.webos-internals.org/wiki/Application:Terminal Apr 06 15:48:51 thanks Apr 06 15:50:28 ah that sucks no pasting Apr 06 15:50:42 nfoxTc: the app also has a "non-obvious keys" help page in its menu -- but I wouldn't recommend you open it up because doing so just crashed Luna on my Pre :) Apr 06 15:50:49 guess I'll have to install my tether from my pc Apr 06 15:51:21 Does anybody find their phone prone to crashing if you tap the power button to wake/sleep your phone? Apr 06 15:51:40 I've had 3 restarts yesterday... And this was trying to wake the phone... Apr 06 15:52:03 same here Apr 06 15:52:04 Mitalis: I had a problem with that in 1.4 often. With 1.4.1 it happend a few times, but it hasn't happened to me since I updated to the newer Facebook beta. Apr 06 15:52:12 i had one right after the update and that's all Apr 06 15:52:16 havent seen one since Apr 06 15:52:44 Mitalis: either it is a complete coincidence (possible) -- or all of the orphaned background tasks spawned by the 1.1.4 Facebook beta were making things unstable Apr 06 15:52:47 anyone know why I would get ipgk errors with freetether? Apr 06 15:52:50 nfoxTc: you can cat /tmp/webkit-paste Apr 06 15:53:14 I don't know what that means Apr 06 15:53:30 to substitute paste Apr 06 15:53:41 even if I could just get a | in there I'd just type the wget line Apr 06 15:54:19 cat /tmp/webkit-clipboard it is Apr 06 15:54:20 pipe is Orange + . Apr 06 15:54:27 orange key and a period Apr 06 15:54:36 what exactly is your clipboard content? Apr 06 15:54:55 ah that's what pipe means lol Apr 06 15:55:16 wget -qO- http://mytether.net/install.php | /bin/sh Apr 06 15:55:25 that's what it is dthough Apr 06 15:56:14 ah Apr 06 15:56:31 you'd need to type $(cat /tmp/webkit-clipboard) Apr 06 15:56:39 but i don't get the $ .... Apr 06 15:58:21 nah I got it...i just didn't feel like installing from my pc Apr 06 15:58:34 anyone else try the paid version yet? Apr 06 15:59:28 hold on switchin to pc Apr 06 16:01:30 nfoxTc ah!. sh /tmp/webkit-clipboard should have worked for your case; too Apr 06 16:54:04 stupid sprint Apr 06 16:54:17 they redirecting all my phone calls to sprint debit department Apr 06 16:54:25 but I don't have any outstanding balance or past due Apr 06 16:54:27 eh Apr 06 16:57:00 lady from sprint "please disregard this message" Apr 06 16:57:18 me: I wish I could but I can't because it's redirecting me to sprint debit department Apr 06 17:01:39 what version of Terminal doesn't make my luna reboot? Apr 06 17:01:47 ...and where do i get it? Apr 06 17:02:10 just don't attempt to swipe backwards Apr 06 17:02:16 oh Apr 06 17:02:17 uh Apr 06 17:02:22 you're psychic Apr 06 17:02:49 but of course :D Apr 06 17:03:06 how does one get out of the preferences screen in that case? Apr 06 17:04:12 you can't. If you go into pref's you're there forever. Apr 06 17:04:40 is it ok to remove the kernel ipk by using preware? Apr 06 17:04:41 there is the official answer Apr 06 17:04:58 hmm i think so Apr 06 17:05:01 nvm the question Apr 06 17:23:52 Mousey: if you go into the preferences screen you just have to go buy a new Pre, that's all. Apr 06 17:25:31 apparently =P Apr 06 17:25:47 grr Apr 06 17:25:51 same problem tho Apr 06 17:25:58 makes terminus more attractive again! Apr 06 17:27:23 oh Apr 06 17:27:24 i see Apr 06 17:27:25 nm Apr 06 17:27:31 ^_^ Apr 06 17:34:42 isn't xterm in the beta feeds? Apr 06 18:14:18 Mousey: use xterm! :D Apr 06 18:14:24 Rick_work: yes xterm is in testing feeds :) Apr 06 18:14:26 k! Apr 06 18:14:30 i saw it there ^_^ Apr 06 18:14:35 but what wm? Apr 06 18:15:12 Mousey: but really the normal terminal application works fine other than the preferences.... which the other terminals don't offer, so shrug. as much as i'd like everyone using xterm.. :) Apr 06 18:15:24 Mousey: the xterm package has no wm--just grab the 'xterm', and click it and away you go. Apr 06 18:15:35 doesn't it need the xserver? Apr 06 18:15:39 (which i have ^_^) Apr 06 18:15:40 Mousey: it'll grab it Apr 06 18:15:51 Mousey: in that case it'll use it :) Apr 06 18:16:00 Mousey: http://www.webos-internals.org/wiki/Application:X as a reference. Apr 06 18:16:02 * Mousey would grab it, but phone is busy pretending it's a USB drive atm Apr 06 18:16:08 oh right you have it so maybe you've seen it Apr 06 18:16:09 hehe kk Apr 06 18:16:27 i saw xterm this morning, yha Apr 06 18:16:29 yah, even Apr 06 18:20:02 hello! Apr 06 18:20:27 * Jack87 has an app idea but needs input on how to start or support for it Apr 06 18:21:28 * Jack87 wants to be able to wirelessly upload to palm pre as well as download from (while pre runs a webserver) Apr 06 18:22:22 Any idea how to implement such a thing. is it possible via ngix? Apr 06 18:29:04 wirc has suddently decided to reorder the natual order of my windows, it's quite odd Apr 06 18:32:47 I have an alpha perl script that applies all the go faster settings I've found so far Apr 06 18:32:53 http://www.fop.ns.ca/pre/tnb.pl Apr 06 18:32:55 zsoc_wirc: you might be just the man to help with my idea hre Apr 06 18:33:05 take a copy of /etc/fstab before you run it just in case Apr 06 18:33:14 [12:20] * Jack87 has an app idea but needs input on how to start or support for it [12:21] * Jack87 wants to be able to wirelessly upload to palm pre as well as download from (while pre runs a webserver) Apr 06 18:33:24 (amazing how much work one can get done sat in a worthless conference call) Apr 06 18:33:48 Jack87: sftp Apr 06 18:33:57 Jack87: there's already a wiki page on WebDav which does that Apr 06 18:34:14 Jack87: there's actually many ways to do that. Apr 06 18:34:44 write an app to simulate a front panel and toggle the binary opcodes in by hand Apr 06 18:35:01 so basically will use IP to connect to the pre webserver (DDNS) Apr 06 18:37:08 how would on use web browser on computer to upload files Apr 06 18:38:20 rwhitby: can you point me to wiki page? Apr 06 18:41:25 Jack87: a web browser would involve something like lighty Apr 06 18:41:30 on a non-standard port Apr 06 18:41:39 or maybe 443 (https) Apr 06 18:42:06 and lighty has already been ported to the pre? Apr 06 18:42:47 now what would be needed is a MOJO app that can control the web server for an end user and tell them the https://IP:443 Apr 06 18:44:08 Jack87: right, it would have to assist setting up ddns, and allow you to enter user/pass (so you can have my.pre.cx or similar instead of ip) Apr 06 18:44:10 Jack87: http://forums.precentral.net/palm-pre-tips-information-resources/219094-howto-sync-your-palm-pre-files-without-external-software-using-webdav.html Apr 06 18:44:15 and a user/pass for things Apr 06 18:45:18 zsoc_wirc: right... ddns can be an option. user/pass can be an option as well. (it should have controls to turn server on and off via app) Apr 06 18:57:19 it be a perfect way to send files to another pre in a sense... Apr 06 18:57:36 other pre goes to your pre's web address via browser and downloads Apr 06 18:58:45 thanks for the linke rwhitby Apr 06 18:59:35 any 1.4.1.1 overclocking patch available yet ? Apr 06 19:05:09 jerms: http://tinyurl.com/ye4hqsy Apr 06 19:05:23 I'm running the original overclocked kernel on 1.4.1 FWIW Apr 06 19:05:30 i just installed it Apr 06 19:05:34 the 720Mhz kernel Apr 06 19:05:38 hi all Apr 06 19:06:20 i have an offer, if anyone wants to make a google wave app, i will give them an invite to google wave Apr 06 19:09:02 is anyone interested? Apr 06 19:10:10 a monetary donation to the webos team might get a better response Apr 06 19:10:36 google wave sux Apr 06 19:10:47 all software sucks Apr 06 19:10:57 the trick is finding the least sucky software Apr 06 19:10:58 yap google wave especially:D Apr 06 19:11:09 en0x they have improved it ALOT! Apr 06 19:11:24 lemme check Apr 06 19:12:11 yapo i still see no use for it Apr 06 19:12:28 and the gui is a suxorz Apr 06 19:12:28 :P Apr 06 19:12:50 if there are alot of people you know that have accounts it is amazing. Apr 06 19:14:52 yeah, that's called the "chicken and egg" problem in the software world Apr 06 19:21:57 howdy Apr 06 19:22:40 sbromwich: actually donations to webos internals often have little effect on motivation if no-one on the team already has an itch to scratch which matches ... Apr 06 19:23:00 (donations expect a certain app to be created that is) Apr 06 19:23:07 s/expect/expecting/ Apr 06 19:23:08 rwhitby meant: (donations expecting a certain app to be created that is) Apr 06 19:23:53 * rwhitby tries to go back to sleep now that the patches autobuilder fiasco seems to be resolved Apr 06 19:26:02 rwhitby: slightly more likely than just offering invites to some web service I would think, though ;-) Apr 06 19:28:51 sbromwich: what happened to not being back till august? Apr 06 19:29:17 What I really want at the moment is an app that watches running javascript processes (like say jstop), and restarts the background phone app whenever it dies. Apr 06 19:29:34 What, the TIL? Apr 06 19:29:52 TIL? Apr 06 19:30:13 is it possible to install service via palm app catalog? Apr 06 19:32:06 i thought luna already had a configuration thing to keep certain applications in memory no matter what Apr 06 19:32:09 and phone was one fo them Apr 06 19:34:19 I've noticed that phone, messaging and the browser always load way faster after the first time, suggesting that they are cached to some degree. Apr 06 19:35:18 that would make sense idontwan2know Apr 06 19:36:12 s/would/would totaly Apr 06 19:36:50 preware on 1.4.1.1: I used bootstrap method, but when I launch preware I get: Service Method Exception java.lang.NullPointerException undefined. How to fix? Apr 06 19:37:25 or can an ipk package and install a service and app in one install Apr 06 19:38:20 Jack87, ^ is this for me? Apr 06 19:38:47 timothy1jones: no jsut a general question whoever can answer :) Apr 06 19:39:54 Jack87, or sorry, I just joined and saw your line after my Apr 06 19:41:41 after rebooting preware now say The Package Manager Service is not running... Maybe error is boot-strap process? Apr 06 19:50:42 the install goes no problesm: http://tinyurl.com/yhn2xdv Apr 06 19:51:34 but I notice that ipkgservice version is 0.9.44, preware is 0.9.33 Apr 06 19:51:39 is problem for this? Apr 06 19:58:21 Okay...done worrying about battery life. Just bought a couple of Centro batteries and a wall charger on ebay for a grand total of $14 Apr 06 20:04:18 zsoc_wirc: I can write dinky little perl scripts while I'm in meetings doing SFA, but kernel compiles are a bit much Apr 06 20:06:15 I might possibly have time to make a youtube video showing how fast I can scroll the launcher up and down as well. Apr 06 20:08:12 Maybe these related? http://pastebin.ca/1857319 Apr 06 20:09:15 any preware devs here? Or am I speaking into black hole? :) Apr 06 20:09:39 patience, grasshopper...its all volunteers here... Apr 06 20:10:56 ka6sox, thank you brother. I wasn't sure maybe nobody here. I wait Apr 06 20:13:56 timothy1jones, I can't answer your question but about 1-2hrs from now things will start hopping. Apr 06 20:14:55 ka6sox, ah I see. thank you again my friend. I come back later and ask again :) Apr 06 20:16:38 that works! Apr 06 20:21:39 so... I try to avoid windows so I may be out of the loop... Apr 06 20:21:52 does novacom not work on windows 7 x64? Apr 06 20:22:05 considering I'm using it... Apr 06 20:22:09 like, at this second Apr 06 20:22:24 novaterm is not installed by default though Apr 06 20:22:48 halfhalo_T400: the version in both the sdk and the pdk don't seem to work Apr 06 20:22:52 is there a process? Apr 06 20:23:05 how are you trying to use it Apr 06 20:23:22 in any way Apr 06 20:23:33 well, I mean how are you trying to run it Apr 06 20:23:34 novaterm, novacom commandline, or wosqi Apr 06 20:23:51 from command line? or with installer in wosqi Apr 06 20:24:12 try from the command prompt Apr 06 20:24:13 novacom -t open tty:// Apr 06 20:24:45 .. don't you think I know how it works? lol Apr 06 20:24:50 'unable to find device' Apr 06 20:24:51 no I don't Apr 06 20:25:08 the device isn't recognized Apr 06 20:25:24 does it connect to the emu? Apr 06 20:25:56 ... I'm not using the emulator. I only have access to a win7 machine at the moment. this is a one time fix thing for a friend. Apr 06 20:26:38 well, I was just wondering if its a novacom <-> pre issue vs a novacom <-> all issue Apr 06 20:27:17 I'll figure it out. thanks Apr 06 20:27:37 oh, seriously? you can have novacom wait for a device? Apr 06 20:27:42 thats freaking cool Apr 06 20:27:53 -w Apr 06 20:28:42 zsoc_wirc: It CAN be touchy sometimes. But a reboot of both pre and comp usually fixes that Apr 06 20:30:48 I should never have to reboot a computer. Apr 06 20:30:58 this is windows here.... Apr 06 20:31:23 well rebooting a webos phone is just silly imo. Apr 06 20:32:38 unless you are enabling usbnet ;) Apr 06 20:33:53 reinstalling the sdk worked Apr 06 20:34:31 heh Apr 06 20:34:39 sounds like the novacom service died Apr 06 20:52:10 [22:31] well rebooting a webos phone is just silly imo. <- how to fix dns issues then? ;) Apr 06 20:52:29 hucksy: what dns issues? Apr 06 20:52:40 there is no spoon Apr 06 20:53:00 there's almost no reason to reboot a linux box unless you are changing the kernel in some way, or for some new kernel modules. Apr 06 20:53:03 sometimes my pre wont resolve any names, but the 3g connection is fine Apr 06 20:53:29 hucksy: that doesn't make sense, but you could probably just bring the interface down and backup Apr 06 20:54:27 nah, looks like the resolv-conf is broken in this case but i dont know where webos saves the nameserver ips Apr 06 20:54:53 I have literally never heard of such a thing. how did you break it? :D Apr 06 20:55:07 i dont know :p Apr 06 20:59:46 but yes, rebooting linux sucks Apr 06 21:19:58 zsoc_wirc: btw, do you have some "lost lines" in wirc? Apr 06 21:20:17 (thats not just my problem ;) Apr 06 21:22:34 lost.. lines? Apr 06 21:22:47 * zsoc pokes zsoc_wirc Apr 06 21:22:55 ow Apr 06 21:23:12 * zsoc pokes zsoc_wirc harder Apr 06 21:23:20 careful there buddy Apr 06 21:23:46 I'm a professional. Apr 06 21:23:50 hehe Apr 06 21:24:32 I my service would work. Instead of just exploding. Apr 06 21:24:33 yeah, sometimes there a lines that wirc just not show Apr 06 21:24:52 hucksy: i don't run into that.. unless you mean the ones that come up all white... they show once i scroll down a bit Apr 06 21:26:11 mh no, they are really ... "lost" ... ;) i have no idea why, that has nothing to do with unstable connections etc. Apr 06 21:26:49 odd Apr 06 21:29:07 yep Apr 06 21:29:38 must be you :> Apr 06 21:30:04 nonono :P Apr 06 21:30:13 i'm not the only one Apr 06 21:35:42 give them a break, this IRC client is 100 times better than the old ptpChat clients for PalmOS Apr 06 21:41:50 phil_bw: and it was developed in 4 weeks by 2 people Apr 06 21:41:55 JayCanuck: fancy meetin' you here Apr 06 21:42:08 nah, wirc rocks. just that weird problem... :) Apr 06 21:42:09 * halfhalo goes on stabbing spree Apr 06 21:42:23 yes, ptpChat took me years to write and still sucked Apr 06 21:43:17 ptp2 was the only client available that would connect to multiple servers *and* run in the background. A feat not easily achieved on PalmOS Apr 06 21:44:36 anyhow, I'd take wirc over my own client any day :) Apr 06 21:44:52 hehe Apr 06 21:45:02 pre is my first palm device Apr 06 21:48:02 zsoc: stop poking zsoc_wirc... thats not nice! look what you did he left Apr 06 21:48:23 Jack87: pretty sure he fell off a cliff Apr 06 21:51:34 zsoc: ouch! think he might be okay? Apr 06 21:52:50 probably Apr 06 21:53:13 nn Apr 06 21:59:04 hucksy: grep -v ^\# /etc/dnsmasq.palm.conf | grep -v ^$ will point you to where the dns info goes Apr 06 21:59:38 oh, nice, tnx! Apr 06 21:59:47 YAMW Apr 06 22:01:43 ok, so I'm going to switch editors on my Pre, am I better off with vi or vim? Apr 06 22:01:54 define "better" Apr 06 22:02:11 better for use with the Pre's kb Apr 06 22:02:35 they both use the same command set Apr 06 22:02:49 vim has a superset but for basic editing there's not much between them Apr 06 22:03:10 I use joe myself Apr 06 22:03:48 doesn't come on the Pre Apr 06 22:03:50 ed ftw! Apr 06 22:04:23 palm-webos-device /etc # ipkg-opt list | grep joe Apr 06 22:04:23 joe - 3.6-1 - Joe's own editor. A text editor with wordstar-like and emacs-like keybindings. Apr 06 22:04:23 palm-webos-device /etc # Apr 06 22:06:26 which is easiest to start with? Apr 06 22:06:52 what editor are you used to using? Apr 06 22:07:02 just nano Apr 06 22:07:14 um Apr 06 22:07:20 then I would recommend installing nano Apr 06 22:07:31 I have it Apr 06 22:07:55 I want to switch to something that copies and pastes Apr 06 22:08:24 just use vi Apr 06 22:08:27 use echo and sed ;) Apr 06 22:08:39 well, the nice thing about joe is it uses wordstar keybindings by default, so CTRL-K B and CTRL-K K to mark a block of text are relatively easy Apr 06 22:08:59 if you've never used wordstar they won't make a huge amount of sense to you Apr 06 22:09:08 never have Apr 06 22:09:25 and again, most editing is done *on pre* in the terminal Apr 06 22:09:47 so I'm without the F keys, etc Apr 06 22:09:57 then I would recommend trying all the editors from ipkg-opt list | grep editor and see which one works best for your workflow Apr 06 22:12:10 sbromwich: you get your kernel compiled? Apr 06 22:12:36 compiled, but I don't have the time to sit down and try getting it to boot Apr 06 22:13:20 I have the whole compiled tree if you are in the mood to download it and give it a whirl yourself Apr 06 22:13:47 you have the uImage? i can try that for you Apr 06 22:13:49 http://www.fop.ns.ca/pre/linux_palm_rtpatch.tar.bz2 Apr 06 22:13:57 sure, let me dig it out Apr 06 22:14:27 i can boot it seperately incase it doesnt work :) Apr 06 22:15:40 http://www.fop.ns.ca/pre/uImage Apr 06 22:16:04 wow. its huge Apr 06 22:16:13 want me to gzip it? Apr 06 22:16:35 naw. i got it. gonna give it a whirl. Apr 06 22:16:50 I just used the .config rwhitby sent me (which I think came for you) and enabled -rt Apr 06 22:16:58 ^for^from Apr 06 22:17:17 odd. i dont remember my builds being so big. Apr 06 22:17:36 I make no promises that I built it correctly ;-) Apr 06 22:17:55 I was in a rush trying to get it all done while I had some free time last night Apr 06 22:17:56 well if you got a uImage it built something :) Apr 06 22:18:04 that was my thought Apr 06 22:18:15 I was trying to get the emulator going and just ran out of time Apr 06 22:19:58 cd ~opis/patient_reg Apr 06 22:20:03 sod Apr 06 22:20:18 some day I will pay attention to what window I'm in before typing Apr 06 22:23:07 sbromwich: i'm updated/commited/pushed entire repos with the wrong info from the wrong shell Apr 06 22:23:39 I've at least managed to avoid shutdown -r now in the wrong window Apr 06 22:23:54 sbromwich: i'm done that a few good times too, heh Apr 06 22:24:15 I did once run killall httpd on solaris instead of the linux terminal I thought I was in, which led to quite some mirth and woe Apr 06 22:24:50 sbromwich: well i can say i've never seen my phone not boot with a bad kernel :) Apr 06 22:25:25 *grin* Apr 06 22:25:46 who needs the phone functionality *anyway*. Apr 06 22:26:05 but this one hangs like i've never seen before. it usually just reboots. Apr 06 22:26:52 do you get any output at all? Apr 06 22:27:12 nope :( Apr 06 22:27:30 OK I think I might know what that is then Apr 06 22:30:02 I don't suppose there's any known way to get serial console? Apr 06 22:33:30 enumerate usb :P Apr 06 22:33:53 you could plug in a usb to serial device Apr 06 22:34:15 hmm, good idea Apr 06 22:34:18 palm might not be too happy about doing it though :P Apr 06 22:34:23 I'll see if I can grab one tomorrow Apr 06 22:34:26 *shrug* Apr 06 22:34:28 it's my phone Apr 06 22:35:09 ya at least they have a die hard fan base Apr 06 22:35:12 why would palm care? Apr 06 22:35:25 people need to stop assuming palm get's mad about things for no reason Apr 06 22:35:36 because they intentionally try to make it so you cant just plug any old usb device into it Apr 06 22:35:49 what Palm /does/ get mad about is my constant bickering, which I probably won't stop Apr 06 22:35:58 haha Apr 06 22:36:00 loot-: give an evidence of them intentionally trying to do that Apr 06 22:36:31 The current reason you can't "plug any old usb device into it" is a hardware limitation. Apr 06 22:36:33 try using usb devices with it like Apr 06 22:37:00 now you're going to make me dig it up and i have to leave in a minute haha Apr 06 22:37:10 haha n/m it's cool Apr 06 22:37:12 i'm just arguing Apr 06 22:37:25 hardware is not the entire problem though Apr 06 22:38:00 a stock palm kernel/userland wont do it i remember getting it to see a usb stick over a usb hub being a huge pain in the ass Apr 06 22:39:12 it did work with some crappy powered mini usb hub though Apr 06 22:40:34 pre tries to draw an amp from dumb chargers Apr 06 22:43:59 getting the usb to work broke a ton of must-have stuff Apr 06 22:44:31 I'm stumped on the kernel compile, which worked fine yesterday; when I do make I get /lib/ld-linux.so.3: No such file or directory Apr 06 22:44:54 I must have missed setting something with LIBPATH... anyone recognise that off-hand? Apr 06 22:46:19 i dont know why it would want to see that lib Apr 06 22:54:37 nova thingy isn't working when i try to run the preware Apr 06 22:56:30 timothy1jones: are you using wosqi 3.02 or later to install preware and package manager service? Apr 06 22:57:24 i have no clue Apr 06 23:02:58 hmmm, I can't even get make menuconfig working... something's quite, quite awry Apr 06 23:14:42 ncurses just built quite happily so something must be cocked up in the kernel dir Apr 06 23:20:48 NFI what happened but things seem to be trundling along more happily after make mrproper Apr 06 23:28:52 so would I be able to mount my debian chroot (copied from the pre) on my linux desktop and compile software for it? or would the cpu architecture screw things up Apr 06 23:29:14 latta: the latter Apr 06 23:29:28 yeah thats why i need to get scratchbox going right? Apr 06 23:29:39 that's the fnu I'm going through right now Apr 06 23:29:40 latta: you could create a cross compile environment however. options include WIDK (with scratchbox) or gentoo and crossdev Apr 06 23:29:46 WIDK is the most supported method by this channel. Apr 06 23:30:05 There are literally tens of popular methods of setting up CC environments tho Apr 06 23:30:23 if not twenties Apr 06 23:30:33 yeah..one of those things i figured out as soon as i hit enter :) Apr 06 23:30:58 sbromwich: what are you, Egyptian? Apr 06 23:31:10 polish Apr 06 23:31:22 ... it was a number counting method jest... nevermind Apr 06 23:31:29 Shadowhawk: hello webchatter Apr 06 23:31:35 I was expecting an rpn joke in return ;-) Apr 06 23:32:35 lssrc -a | grep -i rsct Apr 06 23:32:45 *sigh* Apr 06 23:33:13 haha Apr 06 23:33:41 multihead widescreen is great... most of the time. Apr 06 23:33:44 Sometime's I'm amazed I can even manage to cross compile anything in any environment. Apr 06 23:34:07 I have a suspicion the PIII I was using was a bit shonky Apr 06 23:34:42 compiling again on an atom netbook without the rcu call I was dubious about, cross fingers it'll work now Apr 06 23:36:33 Testing unstable services on a production device is actually quite invigorating. Apr 06 23:37:24 not in my world it isn't ;-) Apr 06 23:37:51 happened this morning and took out patient registration for a bunch of hospitals, that's always brown trousers time Apr 06 23:43:54 ? Apr 06 23:44:17 ?? Apr 06 23:44:18 Shadowhawk: hi Apr 06 23:44:20 ??? Apr 06 23:44:29 ! Apr 06 23:44:44 hmmmm....been a loooong time since I've been in IRC...thought maybe the ? would give me a list of commands or something. Apr 06 23:44:57 Shadowhawk: use /help for that Apr 06 23:44:59 try /help Apr 06 23:45:10 Shadowhawk: can you get a linux command line on your Pre ? Apr 06 23:45:53 Rwhitby'd suggested I should come here to debug a problem. Not sure how to go about that. Should I try to contact him directly via here? I don't even know whether that's possible or not. Apr 06 23:46:12 Shadowhawk: I'm here Apr 06 23:47:16 Shadowhawk: can you get a linux command line on your Pre ? Apr 06 23:48:13 rod take your invisible cloak off Apr 06 23:48:36 Shadowhawk: can you get a linux command line on your Pre ? Apr 06 23:49:05 I think his up arrow's working just fine, zsoc Apr 06 23:50:46 sbromwich: haha good one Apr 06 23:52:46 rwhitby, sorry, I stepped out. I just did boot strap process via novaterm Apr 06 23:53:53 timothy1jones: remind me what the problem is? Apr 06 23:55:16 rwhitby, I installed preware using boot strap and get Service Method Exception java.lang.NullPointerException undefined Apr 06 23:55:58 here is install process: http://pastebin.ca/1857310 Apr 06 23:56:37 I think could be following related: http://pastebin.ca/1857319 Apr 06 23:56:54 timothy1jones: for some reason those pastebin urls are not working for me Apr 06 23:57:00 timothy1jones: can you use webos.pastebin.com ? Apr 06 23:57:21 rwhitby, sure. please hold on Apr 06 23:57:48 from http://www.webos-internals.org/wiki/Research_Pre_Boot_Process if I boot a kernel with cat uImage-kexecboot-2.6.24-r50-palmpre.bin | novaterm boot mem:// is that a one-shot deal or is it saved? Apr 06 23:58:05 timothy1jones: since you're using the bootstrap script, we *will* be able to solve this. Apr 06 23:58:14 sbromwich: one shot deal Apr 06 23:58:24 :) Apr 06 23:58:31 ah, that sounds like just what I'm looking for, thanks Apr 06 23:58:42 I shall be a good little doobie and update the page to clarify Apr 06 23:59:08 rwhitby, this is install proccess: http://webos.pastebin.com/EgzxKXFZ Apr 06 23:59:14 everything looks ok Apr 07 00:00:41 timothy1jones: can you pastebin /tmp/preware-bootstrap.log ? Apr 07 00:00:55 comme ca Apr 07 00:03:42 rwhitby, please see http://webos.pastebin.com/N7nfABvh Apr 07 00:04:19 timothy1jones: ok, that installation looks correct. what is the exact error again? Apr 07 00:04:24 (and when does it happen?) Apr 07 00:05:53 rwhitby, I launch preware and right away dialog at bottom: Preware \n Service Method Exception: java.lang.NullPointerException undefined Apr 07 00:06:24 timothy1jones: ok, is there anything else in /var/log/messages? Apr 07 00:06:30 latta:hmm, i'm getting some errors on the 'make stage' part of sb2, perhaps reinstall Apr 07 00:06:53 openssl bash error? Apr 07 00:08:06 no, seems like a PATH issue im gonna check my steps Apr 07 00:08:20 okie Apr 07 00:08:22 rwhitby, http://webos.pastebin.com/yaU8qPxn Apr 07 00:08:32 but i thought i heard someone here saying something about 'make stage' unneccessary? is that true? Apr 07 00:08:40 just curious if you were seeing the same error I did, latta Apr 07 00:09:09 rwhitby, stuff about cgroup is suspicios Apr 07 00:09:52 /dev/cgroup/game/cpu.shares doesn't exist, but I have /dev/cgroup/cpu.shares Apr 07 00:10:05 not sure if relevant Apr 07 00:10:21 latta: http://logs.nslu2-linux.org/livelogs/webos-internals/webos-internals.20100406.txt Apr 07 00:10:26 Apr 05 23:37:43 sbromwich: skip the last "make stage" command, you'll be there all day and it's just building things you aren't using for this. You can try make stage'ing a single package in packages/ somewhere to test the toolchain Apr 07 00:10:56 timothy1jones: not relevant Apr 07 00:11:47 timothy1jones: what are you using to access the command line on the Pre? Apr 07 00:12:05 rwhitby, novaterm Apr 07 00:12:26 timothy1jones: what host OS ? Apr 07 00:12:39 rwhitby, gentoo linux Apr 07 00:13:14 timothy1jones: excellent Apr 07 00:13:32 timothy1jones: open a window and run "dbus-util --capture ipkgservice" then reproduce the problem Apr 07 00:13:42 (run that on the Pre) Apr 07 00:14:27 is it necessary to run this? cuz i get a fatal error when doing so: git checkout -b doctor_14 origin/doctor_14 Apr 07 00:15:01 no Apr 07 00:15:04 latta: it is quite unecessary to run that, but you can but an image of the 141 doctor in it's place if you want Apr 07 00:15:36 latta: that command is used to get the branch of the widk that is being prepared for compiling stuff against 1.4 instead of 1.3.5 Apr 07 00:15:36 I think you're following the same path I did last night, you might want to follow through the irc log link I posted and search for that timestamp Apr 07 00:17:09 oh cool thanks guys, that link was helpful, sbromwich Apr 07 00:17:55 YAMW Apr 07 00:18:00 rwhitby, here is: http://webos.pastebin.com/s5RjpzaX Apr 07 00:18:53 oh wait when I click OK, a lot more stuff shows up... Apr 07 00:20:15 timothy1jones: ok, that means something in the update method is causing a java null pointer exception Apr 07 00:20:44 timothy1jones: what version of ipkgservice and what version of webOS ? Apr 07 00:22:02 rwhitby, webos 1.4.1.1 when install preware it said ipkgservice 0.9.44 Apr 07 00:22:15 timothy1jones: and what version of wosqi did you use to install it? Apr 07 00:22:27 timothy1jones: ls -lR /media/cryptofs/apps/usr/lib/ipkg Apr 07 00:24:24 rwhitby, not wosqi using bootstrap method with novaterm http://tinyurl.com/yfclgpt Apr 07 00:24:34 timothy1jones: oh, right, sorry. Apr 07 00:24:56 timothy1jones: good, that eliminates lots of sources of errors Apr 07 00:25:09 here is ls command: http://webos.pastebin.com/TK3tzJbU Apr 07 00:25:41 rwhitby, maybe related: when I try to install with app catalog, give error insufficient space Apr 07 00:26:22 timothy1jones: df -k Apr 07 00:26:59 rwhitby, df -k : http://webos.pastebin.com/TADpnt7N Apr 07 00:27:26 timothy1jones: the code being executed should be line 85 of http://git.webos-internals.org/?p=services/ipkgservice.git;a=blob;f=org/webosinternals/ipkgservice/IPKGService.java Apr 07 00:27:49 timothy1jones: that code should create a tmp and cache directory under /media/cryptofs/apps/usr/lib/ipkg/ Apr 07 00:28:22 timothy1jones: for some reason you don't have those Apr 07 00:28:30 timothy1jones: and I suspect that is causing the problem Apr 07 00:28:40 timothy1jones: do you know Java at all? Apr 07 00:29:12 rwhitby, I know c/c++ and hate java, but I can read it :) Apr 07 00:29:22 timothy1jones: output of mount please Apr 07 00:29:41 timothy1jones: I'm the same. I'm currently converting this java service to C Apr 07 00:30:02 timothy1jones: one thing you can try is to just run Preware Alpha Apr 07 00:30:08 (it uses the new C service) Apr 07 00:30:26 then if you have a problem you can run gdb on the Pre to help me debug it Apr 07 00:30:44 rwhitby, sorry you are too fast for me :) here is mount : http://webos.pastebin.com/3cF0SFdu Apr 07 00:32:20 rwhitby, let's do preware alpha :) Apr 07 00:32:30 timothy1jones: ok, you have a serious problem. /media/cryptofs is not mounted Apr 07 00:32:56 rwhitby, hehe, that's nice 8o Apr 07 00:33:17 timothy1jones: so it's writing into the / partition instead of into a cryptofs loop mounted from /media/internal/.palm Apr 07 00:33:37 timothy1jones: so that's gonna need to be fixed before *anything* will work on your device Apr 07 00:35:30 rwhitby, that's weird. maybe this is relevant: before upgrading to 1.4.1.1 I ran webos doctor and installed I think it was 1.3.5 and wiped everything... Apr 07 00:35:50 I had a few issues which I later found out were bugs in 1.4.1 :) Apr 07 00:36:29 I upgraded to 1.4.1.1 from 1.3.5 and maybe that cause it? Apr 07 00:36:52 timothy1jones: are you sure it was 1.3.5 and not something earlier? Apr 07 00:37:16 rwhitby, could have been earlier. I didn't actually look at the version Apr 07 00:37:30 timothy1jones: at this point, I would suggest to backup your data, complete erase, webosdoctor to 1.4.1(.1) Apr 07 00:37:43 timothy1jones: cause your phone is screwed at the moment and won't get better Apr 07 00:40:12 rwhitby, I didn't realize there were diff versions of webosdoctor. I am looking inside .jar is there a file like /etc/issue that indicate webos version? Apr 07 00:41:10 timothy1jones: check the md5sum from our wiki page Apr 07 00:41:51 Did that messaging backup go anywhere? Apr 07 00:42:51 chrisa: it backs up, it doesn't restore Apr 07 00:43:06 That is marginally more useful than just trashing my fs! Apr 07 00:43:07 ;) Apr 07 00:46:30 i agree Apr 07 00:47:02 anyone who wana help a guy with sprint pre and mhs? Apr 07 00:55:58 hello Apr 07 00:56:27 anyone there? Apr 07 01:00:09 i don't have a staging directory in cross-compile and i get errors when i run this sb2 -M /srv/preware/cross-compile/staging/mapping-armv7 Apr 07 01:01:02 uNiXpSyChO: http://www.fop.ns.ca/pre/uImage Apr 07 01:01:20 is there something I'm missing? Apr 07 01:01:25 I'm leery of booting it myself and that boot mem:// command gives unknown command Apr 07 01:01:31 what's the errors? Apr 07 01:02:24 Messages from sb2: Apr 07 01:02:24 2010-04-06 21:02:18.343 (ERROR) bash[7560] ERROR: Fatal: Rule file interface version check failed: No version information in /tmp/sb2-latta-20100406-210217.yKKOW0/rules/Default.lua Apr 07 01:02:24 # exit 99 (99) Apr 07 01:03:20 rwhitby, while I am reloading my pre, how do I install preware alpha? Apr 07 01:03:39 latta: what are you trying to do? Apr 07 01:03:40 timothy1jones: can you run palm-install ? Apr 07 01:04:26 well im just following the instructions on building doom w/ sb2 cuz i've never cross compiled b4 Apr 07 01:04:31 rwhitby, not sure. I used to just run good ole bootstrap script Apr 07 01:04:39 chrisa: save/restore backs up IMs and SMSs to a HTML table Apr 07 01:05:04 I should find who works on the messenging app and pick their brain Apr 07 01:05:16 I still want a true IM/SMS import so it's like I never wiped my device Apr 07 01:05:50 burrp Apr 07 01:06:04 chrisa: issue is it's kept in the PalmDatabase with everything else on the planet Apr 07 01:06:11 right Apr 07 01:06:15 it's difficult/impossible to reinject those tables safely Apr 07 01:06:18 I explored that once for the same reason Apr 07 01:06:20 and saw dragons Apr 07 01:06:26 ah it's not easy... I'm trying to compile the kernel and I think I ended up having to manually massage it into the build tree Apr 07 01:06:46 zsoc: zs! Apr 07 01:08:18 raster: sup Apr 07 01:10:20 would i skip calling ...staging/mapping-armv7 if i never did a "make stage" Apr 07 01:10:46 sbromwich: nope. still hangs Apr 07 01:11:17 rwhitby, what's the version of preware alpha? Apr 07 01:13:05 zsoc: pants are up. thanks god Apr 07 01:13:09 thank god Apr 07 01:13:10 yeah, I figured out how to get it to boot Apr 07 01:13:21 I'm assuming the "no flashy palm logo" means hang? ;-) Apr 07 01:13:24 raster: i like pants Apr 07 01:13:36 yep Apr 07 01:13:57 ok... then this seems like a good time to give up for the evening ;-) Apr 07 01:14:08 laters :-) Apr 07 01:14:25 I love how he promised yesterday he wouldn't be back till august lol Apr 07 01:15:49 timothy1jones: 0.9.58 Apr 07 01:16:29 zsoc: I hooked him successfully Apr 07 01:18:54 Anyone having problems with the App Launcher patches that change how many icons are displayed? Apr 07 01:19:14 I got an update to the 4x4 icons v4 today and it completely stopped working and am unable to load any other ones Apr 07 01:23:00 something did break today, i'm not sure Apr 07 01:23:56 something with preware itself? Apr 07 01:24:20 patch feed Apr 07 01:24:22 oh ok Apr 07 01:24:23 i think it was fixed Apr 07 01:24:39 i just got the same error trying to load it through preware on my pre Apr 07 01:24:50 trying it through the webos quick install now and i think it may have taken care of it, as well as removed a "stuck" patch Apr 07 01:24:56 restarting it now Apr 07 01:28:10 zsoc: is there a difference between installing a patch or using the Tweaks part of preware on my computer? Apr 07 01:28:34 preware does not have a tweaks part Apr 07 01:28:41 nor does it run on your computer Apr 07 01:28:46 well, outside of the emulator Apr 07 01:29:12 rwhitby: so I've been trying out the justinput IME, it's decent so far(although lacks a certain input that's been reported) but it's activated the same way the Virtual Keyboard is activated. =/ Apr 07 01:29:24 AMFS: do you realise that Preware and WebOS Quick Install are two different things written be two completely separate developer groups? Apr 07 01:29:30 yes Apr 07 01:29:43 egaudet: !!!! Apr 07 01:29:52 zsoc!! Apr 07 01:29:52 im asking if theres a difference, since i use webos quick install to remove patches that get stuck on my pre, as well as fix other problems with it Apr 07 01:29:54 RUN Apr 07 01:30:01 AMFS: so where is tweaks part of preware? Apr 07 01:30:13 egaudet: hi Apr 07 01:30:17 hello Apr 07 01:30:23 i mis-spoke, sue me...of webos quick install Apr 07 01:30:25 feel better now? Apr 07 01:30:33 slightly Apr 07 01:30:42 No, you'll be hearing from my lawyer. Apr 07 01:30:44 JayCanuck, rwhitby: can you clarify what should change in aupt for themes? Apr 07 01:30:54 AMFS: chill out dude, just trying to work out what you're having problems with. Apr 07 01:31:05 hi rwhitby how goes it Apr 07 01:31:21 egaudet, all that need to change afaik, is the location of the theme identifier Apr 07 01:31:43 what's a good generic name/location for that? Apr 07 01:31:50 JayCanuck, my initial thought on that is what is the need for a theme identifier Apr 07 01:31:51 my problems are resolved, i took care of it through the webos quick install...im asking if theres a difference since i dont ever use the webos quick install for anything other than removing problematic patches Apr 07 01:32:20 AMFS: honestly, it's not a great idea to use wosqi to remove 'stuck' patches, because it removes the 'patch' without undoing it (assuming it was done at all) Apr 07 01:32:27 if it was never applied however, that's probably ok. Apr 07 01:32:27 AMFS: webos quick install uses the same AUPT scripts as Preware for patches. for tweaks, you need to ask JayCanuck Apr 07 01:32:32 they should work interchangably, don't they? Apr 07 01:32:37 zsoc: I don't believe that is accurate any more Apr 07 01:33:00 rwhitby: If that wasn't true anymore, then it would be no different then using 'uninstall' in preware, they would both work or not work Apr 07 01:33:07 zsoc: no clue if it was applied anymore, it was showing up as it had an update but wouldn't update, then i got an error trying to remove it Apr 07 01:33:08 egaudet, wouldn't it be preferred? it'a definitely not required, but woud be nice for quick identification Apr 07 01:33:10 Otherwise, we should be using whatever method he is using. Apr 07 01:33:16 ah true those special tweaks in WOSQI Apr 07 01:33:44 why would the installed patch for VKB say its not available for 1.4.1 and is a place holder? is it because i updated out of the testing feed, it's working fine. Apr 07 01:34:01 JayCanuck, the way I was thinking when I last worked on the aupt script was that any non text file (i.e. images) being replaced HAD to be stock which resolved the multi-theme install issues Apr 07 01:34:16 test http://www.webos-internals.org/ Apr 07 01:34:28 and it let's there be hybrid patch/image replacements without labelling them theme or patch etc... Apr 07 01:34:43 the package metadata can still be used to categorize though i.e. Type: Patch and Type: Theme Apr 07 01:35:29 egaudet, yea, but that poses the same speed issue I had with patches Apr 07 01:35:49 I'd need to parse every package's control file to determine if it were a theme Apr 07 01:36:21 If you want to treat them specially yeah Apr 07 01:36:36 alternative, aupt supporta additional extra postinst/prerm eh? Apr 07 01:36:43 I see no need to treat things specially aside from emergency recoveries etc... Apr 07 01:37:00 aupt used to but does not anymore support extra scripts Apr 07 01:37:02 special cases are where the bugs hid Apr 07 01:37:02 k, well, 'm referring to off-feed installs Apr 07 01:37:04 hide Apr 07 01:37:23 off-feed installs? Apr 07 01:37:31 as in .ipk files not on feeds Apr 07 01:37:35 off-feed installs have the power of a desktop Apr 07 01:37:42 It seems to me an off-feed install should just be "package and then install like normal" Apr 07 01:38:09 yea, but if it's a theme, I want wosqi to identify the installed theme Apr 07 01:38:12 and the package creation (which is where the parsing is done) is done well before then Apr 07 01:38:44 JayCanuck, for what purpose? Apr 07 01:38:52 well, WOSQI themer Apr 07 01:39:08 it is 1 part repository and 1 part current theme status Apr 07 01:39:16 not following? Apr 07 01:39:40 1 have of the themer window shows the currently installed theme and its info Apr 07 01:39:49 I understand the need for special theme support in the old mechanisms but if using the aupt-4 way it should be seemless as any other advanced ipk Apr 07 01:39:55 the other have of the window show the combo prethemer/precentral repo Apr 07 01:40:18 JayCanuck: if themes have appinfo files, then you should be able to read from there, no? Apr 07 01:40:37 appinfo doesn't have desscriptions Apr 07 01:40:51 description will be in control file. Apr 07 01:40:55 exactly Apr 07 01:40:56 so all you need is the appid, right? Apr 07 01:41:12 my issue is needing a quick way to id the current theme Apr 07 01:41:14 ]yea Apr 07 01:41:28 so what we need to work out is a generic "singleton install identifier" mechanism Apr 07 01:41:42 and use that for themes and kernels and anything else where you can only have one of a type of thing installed Apr 07 01:41:53 I don't like that because it goes outside of the standards for an unnecessary special case Apr 07 01:42:09 not a special case for themes, but a new generic class for singletons of all types (themes, kernels, binary patches, etc) Apr 07 01:42:37 but the aupt-4 does not prevent two "themes" from being installed Apr 07 01:42:45 and the only thing that makes it a "theme" is the packager's metadata Apr 07 01:42:50 currently I'm using a "org.webosinternals.patches.*"/"ca.canucksoftware.pateches.*" filter for patches Apr 07 01:43:25 Preware for example just organizes all packages by metadata category and works fine without need for special cases/knowledge Apr 07 01:43:43 it reads all off-feed control files Apr 07 01:43:56 egaudet: I think we need to add singleton capabilities to preware too Apr 07 01:44:28 I could do a similar filter for "com.prethemer.theme.*" and "com.precentral.themes.*" Apr 07 01:44:29 i.e. from this set of packages, only one of them can be installed at any one time, and you can have multiple independent sets Apr 07 01:45:19 I think something like this belongs in the packaging standards, not in postinst/prerm scripts Apr 07 01:45:22 though my worry would be people using my theme builder creating themes with ids different Apr 07 01:45:36 the file replacements are protected though Apr 07 01:45:53 egaudet: I agree the singleton support should be in the standards, and preware and wosqi and the scripts should support it Apr 07 01:45:54 for example image A.png can only be replaced ONCE, if it is not palm stock the script will not allow you to overwrite it Apr 07 01:46:10 exactly Apr 07 01:46:22 you'd need to remove existing theme to install new one Apr 07 01:46:45 right, but there will be no harm done Apr 07 01:47:10 If it is ideal for this type of singleton support I don't see the scripts as the place for it though Apr 07 01:47:19 I think that should be handled in the packaging metadata and installers Apr 07 01:47:32 i.e. make that decision/error message report prior to attempting to run the postinst Apr 07 01:47:39 k Apr 07 01:47:47 egaudet: I think it should be in both. the installers trap it at top level, and if something gets through the file replacement last resort catches it Apr 07 01:47:55 So I see where you guys are coming from just think that the scripts are the wrong place for it Apr 07 01:48:13 egaudet, aupt allows for running extra scripts for postinst/prerm, right? Apr 07 01:48:19 well the scripts will catch it now as-is via the file replacement md5sum check Apr 07 01:48:26 JayCanuck, not anymore no Apr 07 01:48:30 O.o Apr 07 01:48:50 It supports file replacement via additional-files.tar.gz, text file patching and binary patching Apr 07 01:48:52 egaudet: for the singleton support, we need a common area where the info of what member of each singleton group is installed is captured, so that the different installers can communicate through that common area Apr 07 01:49:07 hmm, I guess as a temporary solution I could just go with my filter idea until a full standard is established Apr 07 01:49:13 rwhitby, I think it should go in the existing ipk database area Apr 07 01:49:36 egaudet: yep, that would work. now, how to implement it Apr 07 01:49:51 similar to postinst flag? Apr 07 01:49:54 metadata SINGELTON: type field Apr 07 01:50:14 egaudet: doesn't get passed through to control file Apr 07 01:50:31 oh, unless in the Source JSON field Apr 07 01:50:43 does Source: get into the control file? Apr 07 01:50:50 what about using app type field? Apr 07 01:51:07 actually what JayCanuck said the flags Apr 07 01:51:10 which field is that? Apr 07 01:51:13 if it's a theme, then write package id to ipkg area Apr 07 01:51:18 Type Apr 07 01:51:23 in source json Apr 07 01:51:25 why not have a "singleton" flag that indicates this "Type:" can only have one package installed Apr 07 01:51:59 rwhitby, ok my pre is back to normal and preware works now. I am ready for preware alpha :) Apr 07 01:52:07 timothy1jones: awesome Apr 07 01:52:11 ~testing-feed Apr 07 01:52:12 i guess testing-feed is a means of testing new WebOS Internals products, documented at http://www.webos-internals.org/wiki/Testing_Feeds Apr 07 01:52:25 timothy1jones: do that and then just install it in Preware Apr 07 01:52:45 rwhitby, hehe, thank you my friend :) Apr 07 01:53:05 egaudet, JayCanuck: I'm thinking more general than just Theme: - we may be able to have more granular singletons Apr 07 01:53:26 e.g. Theme which touches certain files, Theme which touches certain other files, etc Apr 07 01:53:42 maybe even use it for groups of conflicting patches Apr 07 01:54:04 CONFLICTS: should be used for that Apr 07 01:54:21 egaudet: it's different from Conflicts Apr 07 01:54:28 lol, long like of conflicts if you want to have every precentral/prethemer package conflict :p Apr 07 01:54:45 *long line Apr 07 01:54:50 egaudet: and scales better - you don't need to change existing package metadata to add a new member to the conflicting set Apr 07 01:55:06 so add a GROUP field to the sources.json Apr 07 01:55:17 and the rule is only ONE package of a certain GROUP id can be installed Apr 07 01:55:24 egaudet: yeah, SingletonGroup: in Sources JSON would do it. Apr 07 01:55:27 rwhitby: people love alpha's little dashboard notification backup thingy, makes them feel warm and fuzzy Apr 07 01:55:48 o, I like SingletonGroup idea Apr 07 01:56:19 and then for speed-wise, the installers would have to agree to a file to touch in a SingletonGroup/ directory? Apr 07 01:56:24 what would be the default action though; pass error when SingletonGroup is already in use, or remove package of current using particular SingletonGroup? Apr 07 01:56:38 JayCanuck: first do no harm Apr 07 01:56:58 k Apr 07 01:56:58 egaudet: yeah, you'd record the current appid for each group somewhere Apr 07 01:57:31 So does SingletonGroup with the currentappid file and md5um checking solve all problems? Apr 07 01:58:27 and singleton groups need a namespace that has both common standards and an extension mechanism Apr 07 01:59:50 or are singleton groups managed by the corresponding Palm package name? Apr 07 02:00:01 (if they touch palm package files) Apr 07 02:00:41 that's probably too restrictive, but may be the basis of a singleton group name prefix Apr 07 02:02:58 in the info/ dir there is .list, .md5sum etc... Apr 07 02:03:16 egaudet: I think we need a /groups dir Apr 07 02:03:17 make .sg ? Apr 07 02:04:11 use /media/cryptofs/apps/usr/lib/ipkg/groups/ and put a reverse DNS file for each installed group in there and the contents of the file is the packagid of the installed member of that group Apr 07 02:04:18 $(IPKG_OFFLINE_ROOT)/usr/lib/ipkg/info/.sg could contain the group name Apr 07 02:04:31 egaudet: but you want to search by group, not by appid Apr 07 02:04:41 you need a reverse lookup Apr 07 02:04:48 it will be very quick to parse a small handful of *.sg files though no? Apr 07 02:05:14 parse/iterate Apr 07 02:05:15 egaudet: a separate dir makes it clear that it is not handled by ipkg Apr 07 02:05:30 md5sums files aren't handled by ipkg are they? Apr 07 02:05:36 yes Apr 07 02:05:41 k Apr 07 02:06:34 groups dir sounds good to me Apr 07 02:07:09 g2g, ttyl Apr 07 02:07:19 so then we just need to decide on the standard for the SingletonGroup namespace Apr 07 02:07:39 and agree on common names for shared items Apr 07 02:08:09 do like patches but substitute patches for groups Apr 07 02:08:15 egaudet: ah, the reason why the postinst has to do it is for palm-install and orange+tap support Apr 07 02:09:09 :/ Apr 07 02:09:09 but it's a general postinst facility now, rather than a special case for themes Apr 07 02:10:39 egaudet: I wonder if a package can require multiple singleton groups, and refuse to install unless all are satisfied Apr 07 02:11:12 palm-install doesn't support dependencies Apr 07 02:11:20 or other advanced homebrew concepts Apr 07 02:11:29 then themes and kernels could just list the palm rootfs packages they touch Apr 07 02:11:51 why though Apr 07 02:12:02 egaudet: agreed, but we want to make sure that a palm-install can't corrupt the singleton property Apr 07 02:12:05 All of this JUST to get an error message to the user Apr 07 02:12:28 no, so you can theme separate apps with separate theme packages, and have multiple independent ones installed Apr 07 02:12:37 you can as-is Apr 07 02:13:52 the ONLY thing missing in the current setup is telling the user "Oh i Failed BECAUSE you are trying to install a theme with one already installed" Apr 07 02:13:57 same issue as incompatible patches Apr 07 02:16:17 egaudet: this can also be used for situations where the conflict is not at the file level but at the conflicting functionality level but the files do not conflict Apr 07 02:17:13 too bad you cant institute an electrical shock when users try to intsall two themes...whole new means of support...lol Apr 07 02:17:13 But when does that go away from being a specific issue that is package dependent and can be resolved with CONFLICTS: Apr 07 02:18:36 egaudet: when there are more than two packages involved, and you want to add a third without touching the other two Apr 07 02:18:58 are there any current cases? Apr 07 02:19:21 I mean it just seems like it's a lot for a rare outside case Apr 07 02:19:28 egaudet: overclocking will have a case I think Apr 07 02:19:51 what will overclocking change that is not conflicting file replacement / patch? Apr 07 02:20:06 e.g. scaling scripts vs governer scripts vs new stuff coming. no file conflicts but you don't want more than one of a group installed at a time Apr 07 02:20:29 and the group grows and you don't want to update the earlier members Apr 07 02:21:03 I still don't think we should add that to the scripts Apr 07 02:21:11 BTW, that one error message to the user will pay back the development time 10x Apr 07 02:21:55 in theory, I highly doubt that for example patches are going to correctly utilize this new stuff Apr 07 02:22:00 egaudet: I think the scripts should have it as a fail-safe, but the user friendly stuff is done in the installer Apr 07 02:22:44 you could argue that scripts should have dependencies support too then Apr 07 02:22:58 egaudet: a set of patches that change the battery to different colours will benefit from a better error message than the last-resort file conflict message Apr 07 02:23:41 egaudet: as in refuse to install until the known dependencies are in place? Apr 07 02:23:47 yeah well that requires the individual patch authors to identify the conflicting other patches Apr 07 02:23:49 (but not actually try and install dependencies) Apr 07 02:23:52 something I know most are not willing to care about Apr 07 02:24:08 and then for them to submit it and dBsooner to verify that they really are conflicting groups Apr 07 02:24:32 egaudet: most are not, but some are, and we can give them the tools to indicate it and make thier and our lives easier for those best-practice patches. Apr 07 02:24:51 yeah, I think if we are going to go down the road of making advanced hombrew palm-install and orange-tap compatible it should go all the way or not at all Apr 07 02:25:16 egaudet: I think we should go as far as fail-safe, for palm-install and orange+tap, but no further Apr 07 02:25:38 we're not going to implement an installer in a postinst, but we can implement a fail-safe Apr 07 02:26:11 I dont' like it because the scripts get bigger and bigger and bigger Apr 07 02:26:22 and it gets increasingly more difficult to keep up with them Apr 07 02:26:32 agreed. how can we modularise them? Apr 07 02:26:41 C-service Apr 07 02:27:21 The amount of discussion that has taken place tonight has permanently damaged parts of my brain. Apr 07 02:27:51 only the unnecessary parts hopefully zsoc :P Apr 07 02:28:22 rwhitby, so how opposed would you be to postinst using luna-send to call a service method to do some of these package wide specific tasks Apr 07 02:28:25 I shall bite my tongue Apr 07 02:28:50 zsoc, I wouldn't do that it hurts Apr 07 02:28:59 egaudet: hmm - so palm-install fails if ipkgservice is not installed Apr 07 02:29:23 for these advanced packages yeah Apr 07 02:29:30 (not all will need to call a service method) Apr 07 02:29:56 egaudet: what about a common location for a functions.sh which contains functions the script can call? Apr 07 02:30:09 we can depend all packages on preware then and put preware install in ALL package postinsts Apr 07 02:30:11 (rather than going through dbus) Apr 07 02:30:13 and take over the world! Apr 07 02:30:43 rwhitby, the thing is that modularizing in scripts is still very ugly and doesn't make things that much easier Apr 07 02:31:33 egaudet: I would have thought calling a function in a script which can return a value would be easier than formatting a json string and parsing a return json string Apr 07 02:31:42 rwhitby, also it could be set up so that the installers call tehse methods normally and the postinst only calls it for palm-install Apr 07 02:32:31 rwhitby, it is in that singular case. But as my experience with aupt scripts tell me, it's when you keep adding and adding (which will happen inevitably) and then you start sacrificing things because it's just ugly to do in scripting Apr 07 02:33:33 egaudet: it's the error checking on the return values of luna-send that I'm concerned about Apr 07 02:33:42 and then it's so much more forward compatible for example if the methodology to install a patch was to call some service method, the future of changing to some type of aufs methodolgy would not require massive pacakge changes and only require the service change Apr 07 02:34:25 egaudet: ok, that is a good point Apr 07 02:34:53 but in the same way, the common function.sh which is not part of each individual package could just change in the same way Apr 07 02:35:03 rwhitby, yes that is the area for most error is return value of luna-send, we'd need a solid awk script Apr 07 02:35:35 rwhitby, yes but can you imagine implementing a change like aufs in a script ;-) Apr 07 02:35:53 egaudet: given that the functions.sh is common (and could even itself call luna-send), I don't see it as different Apr 07 02:36:03 egaudet: in some cases, a script is easier than C Apr 07 02:36:21 but I see your point Apr 07 02:37:06 but at least now we're talking about implementation decisions for a common agreed conceptual framework :-) Apr 07 02:37:51 yeah I've had some of these ideas in the back of my mind for a while too Apr 07 02:38:08 indeed Apr 07 02:38:36 egaudet: BTW, the P in AUPT need to change from "Patch" to "Package" ;-) Apr 07 02:38:47 done :P Apr 07 02:38:57 ~aupt Apr 07 02:38:58 i heard aupt is the best thing for patching webos since .. well.. patching: http://forums.precentral.net/webos-patches/227163-aupt-auto-update-patch-technology.html Apr 07 02:39:20 egaudet: can you take a look at the uber-kernel-pre Makefile in build.git/hardware to see if I am using additional_files correctly for the kernel and modules? Apr 07 02:39:26 1 Apr 07 02:39:38 oil you missed an ol Apr 07 02:39:48 or is that a 1 nevermind Apr 07 02:39:53 controller fell on keyboard Apr 07 02:40:01 hah Apr 07 02:40:13 happens sometimes Apr 07 02:40:47 oil: so the reason why saved package list needed a separate screen is so people can swipe to delete packages from it, separately from swiping to uninstall packages in the installed packages list Apr 07 02:41:03 (neither of those is implemented yet) Apr 07 02:41:26 i just thought it was going to be done completely different Apr 07 02:41:33 whoa saved package list? swipe to delete? Apr 07 02:41:35 I'm so far behind Apr 07 02:42:00 that the save script was simply going to store a list of package ids in a text file or whatever Apr 07 02:42:01 egaudet: http://bit.ly/holy-grail-of-homebrew Apr 07 02:42:13 and for restore, launch preware with the list of packages Apr 07 02:42:28 oil: it has to store a full package model entry so that it can display the info when then feed is not selected Apr 07 02:42:38 hence the depot Apr 07 02:42:46 if its not there, it simply ignores it Apr 07 02:43:00 but it would also copy in and out the feeds Apr 07 02:43:05 so it wouldnt be all that much of an issue Apr 07 02:43:23 it just seems like it was overcomplicated Apr 07 02:43:39 time for dinner :) Apr 07 02:43:40 I just haven't explained the future bits well enough Apr 07 02:46:59 nice save/restore :) Apr 07 02:57:12 rwhitby, 1.4.0 doctor not available via palm? Apr 07 02:57:33 egaudet: correct, all pre-1.4.1 doctors have been removed by Palm it seems **** ENDING LOGGING AT Wed Apr 07 02:59:57 2010