**** BEGIN LOGGING AT Thu Nov 04 02:59:57 2010 Nov 04 05:00:53 somebody ban that dude Nov 04 05:01:47 what dude? Nov 04 05:01:53 he is gone now Nov 04 05:02:16 interesting...he is IPv6...probably thats the problem. Nov 04 05:02:39 oh, something I don't see cause I have parts and join events turned off Nov 04 05:03:09 yeah, flooding the channel with that stuff Nov 04 05:03:10 yep...a full 3 pages of k[t has quit(Excessive Flood) Nov 04 05:03:18 * rwhitby never understood why anyone wants to see those Nov 04 05:03:42 it's not a matter of want, it's just how it is on IRC Nov 04 05:03:54 not on my IRC Nov 04 05:04:01 well, you're the exception Nov 04 05:04:04 there is it again Nov 04 05:04:06 ... Nov 04 05:04:12 no, I just turn those events off in the client Nov 04 05:04:30 makes for a much more enjoyable IRC experience :) Nov 04 05:04:33 * bhuey shakes his head Nov 04 05:04:40 My irc is not on speaking terms with me Nov 04 05:05:25 really somebody should ban that dude Nov 04 05:07:59 omg Nov 04 05:08:04 yeah, really now Nov 04 05:08:35 here's a good solution to the problem Nov 04 05:12:15 * halfhalo watches ka6sox-work try to contain the flood with sugary donuts Nov 04 05:14:40 rwhitby: ok, so things I think we need to hack up the 2.0 doctor... Nov 04 05:14:40 rwhitby: installer.xml points at boot-roadrunner.bin (as well as roadrunner modem updaters which is less important since we're skippign them). recoverytool.config mentions devicetype roadrunner, which we should patch because maybe the java code is looking at it and taking different code paths based on what the doctor says rather than the device. and then obviously there's the installer boot image which we'll need to replace with Nov 04 05:15:28 (well along with changing the xml to point at it I mean) Nov 04 05:16:26 ok, so an OVERRIDE_DEVICETYPE directive and CUSTOM_BOOTLOADER and CUSTOM_KERNEL and CUSTOM_INSTALLER_UIMAGE are needed Nov 04 05:17:01 ka6sox: is it not letting you ban an ipv6 address? Nov 04 05:17:12 yep..that seems to be the problem. Nov 04 05:17:25 seriously, just turn off parts and joins in your client Nov 04 05:17:55 heh, that'll work :) Nov 04 05:17:57 yay! Nov 04 05:18:09 * halfhalo waits for a nick change Nov 04 05:18:51 * k[t has now known as k[t_ Nov 04 05:19:13 rwhitby: by that you mean those four directives need to exist and haven't been written yet? Nov 04 05:19:19 I thought the ! would fix that. Nov 04 05:19:19 correct Nov 04 05:19:41 * destinal wishes he had more skill with make :) Nov 04 05:20:11 checking out the other ones you've implemented now Nov 04 05:20:28 destinal: bootloader will just point to a .bin file, kernel probably needs to point to an .ipk (can use that for pre-flashing overclocked kernels) and UIMAGE just points to a uImage file Nov 04 05:20:43 we can point it to the stock recovery kernel Nov 04 05:21:28 destinal: here's the deal - I'll do the make code if you do the doco Nov 04 05:22:12 rwhitby: sure, you have time now? Nov 04 05:22:22 maybe Nov 04 05:22:32 you can test as I do it too Nov 04 05:22:53 rwhitby: sounds like a plan Nov 04 05:26:45 ok, CUSTOM_DEVICETYPE = castle Nov 04 05:31:19 rwhitby: so something like: Nov 04 05:31:22 make CUSTOM_BOOTLOADER=~/dr/145/resources/webOS/nova-cust-image-castle.rootfs/boot/boot.bin CUSTOM_INSTALLER_UIMAGE=~/dr/145/resources/webOS/nova-installer-image-castle.uImage REMOVE_CARRIER_CHECK=1 REMOVE_MODEL_CHECK=1 REMOVE_DEVICETYPE_CHECK=1 DISABLE_MODEM_UPDATE=1 DISABLE_UPLOAD_DAEMON=1 BYPASS_FIRST_USE_APP=1 ENABLE_DEVELOPER_MODE=1 all-wr Nov 04 05:32:05 hmm the start got truncated Nov 04 05:32:06 sec Nov 04 05:33:24 destinal: REMOVE_DEVICETYPE_CHECK is replaced by CUSTOM_DEVICETYPE Nov 04 05:34:43 rwhitby: so maybe Nov 04 05:34:45 make VERSION=2.0.0 DEVICE=pre2 OVERRIDE_DEVICETYPE=castle CUSTOM_KERNEL=org.webosinternals.kernels.palm-kernel-pre_1.4.5-9_arm.ipk CUSTOM_BOOTLOADER=~/dr/145/resources/webOS/nova-cust-image-castle.rootfs/boot/boot.bin CUSTOM_INSTALLER_UIMAGE=~/dr/145/resources/webOS/nova-installer-image-castle.uImage REMOVE_CARRIER_CHECK=1 REMOVE_MODEL_CHECK=1 DISABLE_MODEM_UPDATE=1 DISABLE_UPLOAD_DAEMON=1 BYPASS_FIRST_USE_APP=1 ENABLE_DEVELOPER_MODE= Nov 04 05:36:03 or do you want to imply remove model check also? Nov 04 05:36:23 CUSTOM_DEVICETYPE Nov 04 05:36:43 REMOVE_MODEL_CHECK stays Nov 04 05:36:48 adding CUSTOM_BOOTLOADER now Nov 04 05:36:53 k Nov 04 05:42:31 CUSTOM_BOOTLOADER pushed (untested) Nov 04 05:43:46 wait, bug. Nov 04 05:44:28 tar: boot-roadrunner.bin: Not found in archive Nov 04 05:44:28 tar: Exiting with failure status due to previous errors Nov 04 05:45:23 fix pushed Nov 04 05:47:05 you should be able to test that on a device, knowing that it won't boot, but able to test the flashing and bootloader replacement Nov 04 05:47:33 ah, we need CUSTOM_INSTALLER done first Nov 04 05:50:14 ah.. yeah, we need Installer target="castle" and the ramdisk and bootfile in there Nov 04 05:50:53 bootfile? Nov 04 05:51:09 Nov 04 05:51:10 vs Nov 04 05:51:10 Nov 04 05:51:34 ah, installer.xml Nov 04 05:51:51 right Nov 04 05:52:09 and unfortunately I think we're going to have to rename a bunch of files to match the target name Nov 04 05:52:40 image value, "nova-cust-image" in this case, gets appended with castle or roadrunner Nov 04 05:52:56 to generate the filename nova-cust-image-castle.rootfs.tar.gz , doesn't it? Nov 04 05:54:16 still don't know whether we technically need to change target name but it seems likely Nov 04 05:56:36 CUSTOM_INSTALLER added, no filename changes yet Nov 04 06:00:13 that's for specifying the the uimage? Nov 04 06:00:56 yup, ok Nov 04 06:03:36 rwhitby: looks good so far, copying the right files and not exploding anyway Nov 04 06:08:13 rwhitby: my guess is we'll need to rename nova-installer-image-roadrunner.uImage nova-installer-image-castle.uImage for instance if it's also a generated name (it's not in the xml) Nov 04 06:08:30 yeah, doing that Nov 04 06:11:04 oh we'll need a CUSTOM_INSTALLER_IMAGE or CUSTOM_INSTALLER_RAMDISK , however you want to call it Nov 04 06:11:24 just CUSTOM_INSTALLER Nov 04 06:13:07 oh right nevermind, and to rename the rootfs, and maybe that's it for the target name Nov 04 06:15:18 besides a sed on the xml. I'm sure you're already on all this, just trying to see all the pieces fit.. Nov 04 06:18:20 interestingly the doctor's classes to refer to castle and roadrunner, but one each in two files. castle is in com/palm/nova/installer/core/IcssUtilities.class and roadrunner in com/palm/nova/installer/core/FlasherThread.class Nov 04 06:21:21 the former cust-castle and sprint-castle, just the customization packages Nov 04 06:21:52 ooh that's interesting in flasherthread Nov 04 06:21:54 String machineType = this.device.getDeviceInfo().getMachineName().toLowerCase(); Nov 04 06:21:54 if ((machineType.contentEquals("broadway")) || (machineType.contentEquals("windsor")) || (machineType.contentEquals("mantaray")) || (machineType.contentEquals("stingray")) || (machineType.contentEquals("roadrunner"))) { Nov 04 06:21:54 this.flasherLogger.logPrintln("temporarily not restoring logs for newer devices"); Nov 04 06:21:54 } Nov 04 06:22:31 broadway, windsor, mantaray, stingray, and roadrunner Nov 04 06:26:04 broadway, windsor, and mantaray get special mention in A6UpdateStage due to having an A6 Nov 04 06:26:04 if ((!machineType.contentEquals("broadway")) && (!machineType.contentEquals("windsor")) && (!machineType.contentEquals("mantaray"))) { reporter.commentOnJob(this.jobId, "Not updating A6 firmware: no A6 on the device"); Nov 04 06:28:24 testing before push Nov 04 06:28:32 is that like apple's A4 but two better? or something else.. Nov 04 06:28:39 heh Nov 04 06:29:04 rwhitby: has anyone heard of these codenames before? Nov 04 06:30:23 http://www.webosroundup.com/2010/10/webos-device-broadway-found-in-app-logs/ Nov 04 06:31:34 if you have an A6 you need /usr/bin/PmA6Updater to update it. whatever an A6 is.. Nov 04 06:32:36 destinal: pushed Makefile.new - it's too raw to push as Makefile Nov 04 06:33:08 k, pulling Nov 04 06:33:11 destinal: I need to run - any fixes, make them in Makefile.new and push back or email Nov 04 06:33:26 are we missing anything to try it? Nov 04 06:33:39 if so I can hack somehting and see what happens :) Nov 04 06:33:59 we're missing kernel, but the rest may work to flash it Nov 04 06:35:00 k Nov 04 06:36:24 A6 is an ARM Core type Nov 04 06:36:46 the OMAP is A8 Nov 04 06:36:53 and the 44Xx series is A9 Nov 04 06:37:30 make that OMAP 3430 is A8 Nov 04 06:37:59 ah, the N900 has an A6, right? Nov 04 06:38:26 kind of odd that they'd go that way for new devices though Nov 04 06:38:26 dunno..will look Nov 04 06:38:46 there are like 5 ARM devices in each phone Nov 04 06:39:01 so it could be a Modem or something processor. Nov 04 06:39:07 good point Nov 04 06:39:25 and I was wrong, N900 is A8 like pre Nov 04 06:39:56 the A8 is the first one that really was able to handle any sort of graphics correctly. Nov 04 06:40:04 (with co=processors) Nov 04 06:40:21 IIRC the N900 has a graphics co-processor. Nov 04 06:45:07 destinal: on bus Nov 04 06:46:45 destinal: the other option is we say CUSTOM_DEVICETYPE and point CUSTOM_DOCTOR to an unpacked doctor directory Nov 04 06:47:10 then we grab everything from that CUSTOM_DOCTOR_SOURCE directory Nov 04 06:47:56 rwhitby: that's not a bad idea, except the partition layout for one is different in 2.0 it seems with mojodb Nov 04 06:48:26 I wonder how they're going to manage the update if I'm right Nov 04 06:48:37 lvm resize I expect Nov 04 06:54:53 destinal: /usr/share/ota-scripts/post-update Nov 04 06:55:42 rwhitby: do you think this might do the trick for now? Nov 04 06:55:57 make -f Makefile.new VERSION=2.0.0 DEVICE=pre2 unpack Nov 04 06:55:57 rm -rf build/pre2-p103ueu-wr-2.0.0/rootfs/boot; cp -rp ~/dr/145/resources/webOS/nova-cust-image-castle.rootfs/boot build/pre2-p103ueu-wr-2.0.0/rootfs Nov 04 06:55:57 make -f Makefile.new VERSION=2.0.0 DEVICE=pre2 CUSTOM_DEVICETYPE=castle CUSTOM_BOOTLOADER=~/dr/145/resources/webOS/nova-cust-image-castle.rootfs/boot/boot.bin CUSTOM_INSTALLER=1 CUSTOM_INSTALLER_UIMAGE=~/dr/145/resources/webOS/nova-installer-image-castle.uImage REMOVE_CARRIER_CHECK=1 REMOVE_MODEL_CHECK=1 DISABLE_MODEM_UPDATE=1 DISABLE_UPLOAD_DAEMON=1 BYPASS_FIRST_USE_APP=1 ENABLE_DEVELOPER_MODE=1 patch pack Nov 04 06:56:26 destinal: you'll want /lib/modules too Nov 04 06:57:01 destinal: BTW, VERSION=2.0.0 is the default for DEVICE=pre2 Nov 04 06:57:38 destinal: CUSTOM_INSTALLER points to the uImage, there is no CUSTOM_INSTALLER_UIMAGE Nov 04 07:00:21 right, somehow I'd left something in from an earlier attempt Nov 04 07:04:43 rwhitby: ok, someone with nothing set up could do this: (provided they had extracted 145 doctor directory laid out like mine Nov 04 07:04:44 http://webos.pastebin.com/HfVzVA05 Nov 04 07:04:52 seems to be running Nov 04 07:06:03 and we have a doctor. let's see what it does Nov 04 07:06:45 destinal: I'll be able to test the doctor in about 30 mins too Nov 04 07:08:23 rwhitby: I'm doctoring PuffTheMagic's development pre from on the other side of the US over x11 tunnel :) Nov 04 07:08:36 if it works he should have an interesting surprise when he wakes up Nov 04 07:09:00 if not, you'll loose contact with it Nov 04 07:09:20 yup Nov 04 07:10:07 don't disconnect your phone Nov 04 07:10:12 don't worry, I can't :) Nov 04 07:12:19 failed Nov 04 07:12:20 http://webos.pastebin.com/4sRQwsDg Nov 04 07:12:35 Nov 4, 2010 3:05:00 AM com.palm.nova.installer.recoverytool.CardController handleFailCase INFO: exiting handleFailCase() java.io.FileNotFoundException: nova-installer-image-castle.uImage (No such file or directory) Nov 04 07:12:44 must have failed to copy it Nov 04 07:13:02 bbiab Nov 04 07:16:30 ah, typo, easy enough to fix Nov 04 07:21:02 rwhitby: nope, we never renamed nova-installer-image-roadrunner.uImage Nov 04 07:21:58 err and never copied the castle one either as they have different md5sum. odd. Nov 04 07:24:11 hiya Mercury Nov 04 07:27:05 destinal: diagnosis? Nov 04 07:28:05 rwhitby: I haven't yet understood how your CUSTOM_INSTALLER directive is working Nov 04 07:28:38 doesn't seem to be doing anything at this point Nov 04 07:29:33 destinal: mine got way past that point Nov 04 07:30:30 make CUSTOM_DEVICETYPE=castle CUSTOM_BOOTLOADER=/srv/doctors/unpacked/webosdoctorp100ueu-wr-1.4.5/rootfs/boot/boot.bin CUSTOM_INSTALLER=/srv/doctors/unpacked/webosdoctorp100ueu-wr-1.4.5/webOS/nova-installer-image-castle.uImage clobber-wr all-wr Nov 04 07:30:53 failed after ramdisk had loaded and it did the carrier check - I'll add those now Nov 04 07:32:21 I run it with: Nov 04 07:32:23 java -jar ./build/pre2-p103ueu-wr-2.0.0/webosdoctorp103ueu-wr-2.0.0.jar --auto Nov 04 07:33:03 right, cli mode would be sensible too. :) my problem was that the uImage wasn't getting integrated though it's there in the command line I used Nov 04 07:33:23 I'll try a simpler list of commands Nov 04 07:35:54 booted into installer Nov 04 07:36:27 INFO: DataPreCheckStage: no status updates within the 30sec window, closing stream and exiting stage!! Nov 04 07:36:27 are you using -f Makefile.new also? Nov 04 07:36:40 yep Nov 04 07:37:24 hmm, I wonder if they're using a much different installer image in 2.0, wonder what it's looking for on the stream Nov 04 07:37:40 can you novacom in manually? Nov 04 07:37:44 yep Nov 04 07:38:21 so I need to disable this data migration step Nov 04 07:38:23 k, good. let me see where that error comes from Nov 04 07:39:39 that phrase does not exist in 1.4.5 doctor Nov 04 07:39:44 but it does in 2.0 of course Nov 04 07:39:57 Binary file ./nova/installer/core/stages/DataPreCheckStage.class matches Nov 04 07:39:57 Binary file ./nova/installer/core/stages/DataRestoreStage.class matches Nov 04 07:39:57 Binary file ./nova/installer/core/stages/DataSaveStage.class matches Nov 04 07:41:12 ah, of course. the 2.0 installer uImage has some scripts in it which the 1.4.5 installer uImage does not Nov 04 07:41:13 it's looking for "pid is" Nov 04 07:41:31 like /usr/local/sbin/recovery/pre_check.sh Nov 04 07:42:04 that's going to be a general problem - mismatch between 2.0 doctor and 1.4.5 installer uImage Nov 04 07:42:07 yup it's usr/local/sbin/recovery/pre_check.sh" Nov 04 07:42:32 we may as well build a new installer image by taking the 2.0 one and swapping the 1.4.5 kernel into it Nov 04 07:44:51 /usr/local/sbin/recovery/pre_check.sh and /usr/local/sbin/recovery/cleanup_script.sh and /usr/local/sbin/recovery/data_save.sh and /usr/local/sbin/recovery/data_restore.sh -- looks like doctor can do a lot more than blow things away now Nov 04 07:46:07 but I have to work in the morning and before that, sleep. I have a feeling you'll have this pretty quickly though Nov 04 07:47:51 rwhitby: good night Nov 04 08:14:44 destinal-sleep: it's creating filesystems and untarring Nov 04 08:34:13 reflashing now with 1.4.5 kernel and 2.0 rootfs Nov 04 08:36:15 <[AD]Turbo> yo Nov 04 08:39:46 rebooting Nov 04 08:42:37 <0>Kernel panic - not syncing: No init found. Try passing init= option to kernel. Nov 04 08:44:13 are you using the 1.4.5 init or the 2.0 init? Nov 04 08:44:30 1.4.5 kernel, 2.0 rootfs Nov 04 08:44:46 which initramfs? Nov 04 08:45:16 this is real boot, but installer Nov 04 08:45:24 s/but/not/ Nov 04 08:45:24 rwhitby meant: this is real boot, not installer Nov 04 08:45:52 the initramfs has to match the kernel build. Nov 04 08:46:06 what are the bootie boot options? Nov 04 08:46:25 1.4.5 bootie Nov 04 08:46:38 which initramfs are you referring to? Nov 04 08:47:20 interesting...its saying it can't find /sbin/init I think...misread it. Nov 04 08:49:46 that's correct Nov 04 08:55:59 do the pre's use an initrd or initramfs? I left my USB cord at the girlies place otherwise I would check Nov 04 09:00:01 did they move the partitions around in the LVM? Nov 04 09:01:35 shouldn't matter. /boot is not in LVM, and /boot/sbin/init mounts by name Nov 04 09:03:57 rwhitby, what is keeping us from grabbing the source for the newest kernel available, applying the palm patches (modifying them where they need to be to get them to apply), and compiling it? Nov 04 09:07:19 I haven't seen anything in opensource.palm.com with 2.0 in it. Nov 04 09:10:14 correct Nov 04 09:21:08 http://forums.precentral.net/showthread.php?p=2744809 Nov 04 09:21:47 who is able to verify the procedure in that post for me ? Nov 04 14:40:31 rwhitby: nice work -- now I wonder what's going on with the kernel and init, hmm Nov 04 14:50:14 rwhitby: I definitely won't try it on PuffTheMagic's dev pre since I'd lose the ability to talk to it Nov 04 14:51:41 now if only I can find a victim device :) Nov 04 15:08:36 * halfhalo will try it on his pre later today Nov 04 15:08:56 Since this particular pre has not been non in months Nov 04 15:10:02 Hmmm, gotta find the tokens for it first though... Never flashed it to vzw after the Coms swap Nov 04 15:59:01 halfhalo: you could just use dd to back up a large chunk of the flash disk Nov 04 15:59:11 tokens are in nvram area at the beginning Nov 04 15:59:30 Well, it has the wrong tokens as it is Nov 04 15:59:52 that must be an interesting story.. Nov 04 16:00:09 Its a vzw pre - Nov 04 16:02:11 ah, does that require a com board swap or just putting a different MEID on it? Nov 04 16:02:26 I did a coms swap Nov 04 16:02:46 Since I wanted my sprint phone to be a pre plus Nov 04 16:03:02 ah Nov 04 16:04:07 I want to do the same thing, either pre+ or better yet pre2 -> sprint Nov 04 16:04:40 which will leave me with a dev pre with a UMTS board in it I imagine Nov 04 18:56:10 Hi everyone Nov 04 18:57:00 just a question, building the FrankenDoctor, right now it won't boot, it won't install and it won't work at all, leaving my Pre in a state in which I'll would have to use another Doctor to restore it, am I right? :) Nov 04 19:02:51 lol Nov 04 19:02:59 that's what it looks like to me Nov 04 19:05:28 Yeah, thought so also. I'll have to wait just a bit longer, I don't have a spare Pre to test :( Nov 04 20:01:09 rwhitby: ping Nov 04 20:02:57 Loudergood: fyi, I know you didn't ask but Nutty is gone. The frankendoctor may render your hardware an actual brick, so anyone not possessing a spare pre and willing to do arcane things to try to fix it if it's even fixable should not attempt Nov 04 20:03:09 ok, not an actual brick, but actually brick it :) Nov 04 20:03:21 * halfhalo wants a brick Nov 04 20:03:48 what's a frankendoctor Nov 04 20:03:55 http://forums.precentral.net/webos-internals/267880-frankendoctor-1-4-5-2-0-0-a.html Nov 04 20:04:03 haha yes, I figured as much Nov 04 20:04:15 egaudet_work: a 2.0 doctor for roadrunner hacked to flash castle Nov 04 20:04:31 I happened to stumble on the twitter post about 30 seconds after it was up last night so I looked at it then Nov 04 20:05:30 destinal: is this an attempt to get a webOS 2.0 metadoctor ? Nov 04 20:06:09 egaudet_work: yup. I think we're pretty close with all the work rod put into metadoctor last night Nov 04 20:06:19 you know they couldn't resist trying that Nov 04 20:07:11 why not just rip the userland and castle.xml out and put in regular doctor if there are brickable parts this way? Nov 04 20:08:32 egaudet_work: that's not a bad idea but there were some changes to how 2.0 is set up and changes to both client (doctor) and server (pre installer image) sides to make it work. Nov 04 20:09:18 more than just the xml descriptors, though those are significantly different too (different partitions etc) Nov 04 20:10:20 What would the doctor be doing new that was absolutely needed aside from the .xml partitioning and userland Nov 04 20:10:43 I mean 'needed' for old hardware Nov 04 20:11:14 needed for laying out the system for 2.0 not necessarily for new hardware. Nov 04 20:11:23 from what I gathered it was more of a kernel issue? Nov 04 20:11:25 are you talking what is needed to get the image running on older hardware? thats what I also asked myself while extracting the image.... Nov 04 20:11:25 and tbh I'm not sure, it may work Nov 04 20:11:56 Can the new image even run on old hardware? Nov 04 20:12:12 Hopefully I'll be able to get a stable internet connection going tonight Nov 04 20:12:14 egaudet_work: userland should Nov 04 20:12:26 yeah userland not kernel Nov 04 20:12:38 egaudet_work: with the pre's kernel I don't see why it shouldn't Nov 04 20:13:36 is the userland still packed up as nova-cust-castle- etc.. Nov 04 20:16:29 destinal: is the kernel panic in that post still the current issue? Nov 04 20:16:31 stupid question maybe but they didn't build any of the userspace binaries targeting the new processor? Nov 04 20:18:40 anyone get more debugging information on that init failure? does it not *find* the file or does it fail when executing? Nov 04 20:27:48 egaudet_work: I never got as far as a kernel panic. rwhitby did but nobody else can even get the installer to boot since it doesn't appear to be storing with the right filename. with work today I haven't had a chance to do any more than reproduce the same issue they're having and post it to the forum thread Nov 04 20:28:21 bsm: yup that's what we're working on, think we're fairly close but it's slightly tricky Nov 04 20:28:32 So what are the "brick" worries ? Nov 04 20:29:04 egaudet_work: I think he's mainly trying to scare people off. it could theoretically flash something without a check with the wrong flash or something Nov 04 20:29:25 ok just checking if there was already a brick case Nov 04 20:29:32 since we don't know much about 2.0 or the new hardware to know if differences. no cases that I've heard of Nov 04 20:31:50 destinal, which thread are you referring to? its not the "Pre 2 webOS Doctor available - what can we learn from it?", is it? Nov 04 20:32:08 bsm: http://forums.precentral.net/webos-internals/267880-frankendoctor-1-4-5-2-0-0-a.html Nov 04 20:32:41 destinal thx Nov 04 21:14:27 egaudet_work: i won't brick anything, just make it unbootable. Nov 04 21:24:06 rwhitby: morning Nov 04 21:25:21 rwhitby: I haven't had time to investigate it but do you know why everyone else but yet is not having CUSTOM_INSTALLER work properly? Nov 04 21:25:29 s/yet/you/ Nov 04 21:25:31 destinal meant: rwhitby: I haven't had time to investigate it but do you know why everyone else but you is not having CUSTOM_INSTALLER work properly? Nov 04 21:26:27 we're all left with a doctor with a roadrunner named uimage and none with acastle name Nov 04 21:29:19 destinal: yeah, dunno. Maybe it's a sed thing. I'll test on Ubuntu today (been doing it so far on MacOSX) Nov 04 21:29:44 can you check the resulting installer.xml and see if it correct? Nov 04 21:32:47 rwhitby: Nov 04 21:33:01 so that looks correct Nov 04 21:33:25 I don't have a nova-installer-image-castle.uImage in there though Nov 04 21:33:29 in the directory Nov 04 21:49:43 rwhitby: you said "I've confirmed that /dev/mmcblk0p2 is being mounted and that /sbin/init exists on that partition." what about an init symlink in / ? Nov 04 21:52:30 destinal: do you see the lines where the Makefile moves that in? Nov 04 21:53:06 linuxjacques: I'll check later Nov 04 21:53:30 linuxjacques: I'm also wondering whether the toolchain change has any impact Nov 04 21:53:55 what's the toolchain change? Nov 04 21:54:23 I noticed in the OE control files on the rootfs that the toolchain has changed. Nov 04 21:54:25 tho, kernel build toolchain and userspace build toolchain are orthogonal Nov 04 21:54:31 looks like they've updated to cs09q1 Nov 04 21:54:47 yeah, they should be orthogonal Nov 04 21:54:52 as long as you build kernel and modules with same toolchain, you;re OK :-) Nov 04 21:55:10 if you don't, you can get all sorts of hard to debug problems Nov 04 21:55:21 yeah, we're copying in kernel and /lib/modules from the 1.4.5 doctor Nov 04 21:56:11 the init thing is curious because on my buildroot-built embedded images I need that symlink, but my desktops don't have it Nov 04 21:56:47 linuxjacques: I'm going to compare the 1.4.5 boot partition and the 2.0 boot partition later today. Nov 04 21:57:02 cool, that should be quite interesting Nov 04 21:57:02 i believe it's still in the boot partition at that point and has not done the pivot_root yet Nov 04 21:57:25 * rwhitby bbiab Nov 04 21:57:25 yeah I think so Nov 04 21:58:24 Tman, ping? Nov 04 22:01:04 rwhitby: back Nov 04 22:01:14 oh heh Nov 04 22:02:11 ain't that always the way? :-) Nov 04 22:03:57 400 Carleton Avenue Nov 04 22:04:03 whoops, mis-paste Nov 04 22:05:51 destinal: back for 10 Nov 04 22:06:01 infobot: launch ICMBs Nov 04 22:06:02 * infobot straps ICMBs to an ACME rocket, lights the fuse and ducks behind a rock Nov 04 22:11:35 make CUSTOM_DEVICETYPE=castle CUSTOM_BOOTLOADER=/srv/doctors/unpacked/webosdoctorp100ueu-wr-1.4.5/rootfs/boot/boot.bin CUSTOM_INSTALLER=/srv/doctors/unpacked/webosdoctorp100ueu-wr-1.4.5/webOS/nova-installer-image-castle.uImage CUSTOM_KERNEL_DIR=/srv/doctors/unpacked/webosdoctorp100ueu-wr-1.4.5/rootfs clobber-wr all-wr Nov 04 22:11:40 destinal: that's what I'm using now Nov 04 22:13:04 rwhitby: I'll try it Nov 04 22:15:03 rwhitby: doing now, only difference is I had to make -f Makefile.new DEVICE=pre2 at start of that line Nov 04 22:15:25 right. I have them in my local copy Nov 04 22:16:23 heh, actually my "ifeq (${LOGNAME},rwhitby)" bit does that :) Nov 04 22:16:48 destinal: pushed a better kernel copy (uses tar now) Nov 04 22:16:57 rwhitby: it's cp /srv/doctors/unpacked/webosdoctorp100ueu-wr-1.4.5/webOS/nova-installer-image-castle.uImage build/pre2-p103ueu-wr-2.0.0/webOS/./nova-installer-image-roadrunner.uImage Nov 04 22:16:57 why does it want to do that? Nov 04 22:17:19 the doctor is lookihng for a nova-installer-image-castle.uImage Nov 04 22:17:36 cp /srv/doctors/unpacked/webosdoctorp100ueu-wr-1.4.5/webOS/nova-installer-image-castle.uImage build/pre2-p103ueu-wr-2.0.0/webOS/nova-installer-image-castle.uImage Nov 04 22:17:50 try putting DEVICE=pre2 inside the Makefile.new Nov 04 22:18:03 maybe there's a var override difference Nov 04 22:18:24 * rwhitby bbiab 20 min Nov 04 22:19:12 hi all Nov 04 22:20:44 can anyone direct me to the link to get preware on my pc? Nov 04 22:21:06 Preware is not a PC app, it is for your phone. install.preware.org Nov 04 22:21:09 sorry noob question,I had it before I reinstalled my pc os Nov 04 22:21:40 preware is a webos app though, you definitely didn't have it installed on your PC, are you thinking of Quick Install? Nov 04 22:21:52 cryptk you do have to install it from your pc tho and the themes are done from your pc i think Nov 04 22:22:05 yeah quick install Nov 04 22:22:06 the themes can be isntalled from within preware Nov 04 22:22:20 really I havent found themes on mine Nov 04 22:22:37 it has themes in it but that appears to be keyboard themes Nov 04 22:22:41 they are disabled by default, go into the preware prefrences and enable the theme feeds Nov 04 22:22:51 awesome thanks Nov 04 22:23:07 preware is that thing with that feature that does that other thing, right? Nov 04 22:23:13 haha Nov 04 22:23:28 wait, thinking of my calculator Nov 04 22:23:32 I got my phone with all the preware on it gonna do my mothers in a little bit Nov 04 22:23:58 half are you maing fun of me ? Nov 04 22:24:09 no, making fun of preware Nov 04 22:24:10 lol its ok I do it often myself Nov 04 22:24:38 I have the cheapest web os phone the pixi Nov 04 22:24:48 I am ashamed to admit it here lol Nov 04 22:24:54 why? lol Nov 04 22:24:54 a sprint one at that Nov 04 22:24:57 I love my Pixi Nov 04 22:25:12 I love mine too but its low man on the totem pole Nov 04 22:25:15 the best thing I can say is... ~"haters gonna hate"~ Nov 04 22:25:40 Slap Preware on there, install Govnah, Warthog Pixi, and sail along at 787mhz. :p Nov 04 22:26:17 sinot I havent gotten that far Nov 04 22:26:38 I seen somewhere the pixi couldnt do over 600mhz tho Nov 04 22:26:56 DJNomad, mine runs @ 907mhz. Nov 04 22:26:59 it couldn't... but then Android handsets started using the same chipset Nov 04 22:27:06 so overclocking has become a reality Nov 04 22:27:08 I got govnah ..havwe no clue what warthog pixi is Nov 04 22:27:15 mine wasn't stable at 907mhz, ka6sox-work :( Nov 04 22:27:53 SineOt, bummers Nov 04 22:28:00 can anyone help me with preware prefs? Nov 04 22:28:03 mine runs at NaN Nov 04 22:28:24 I am going through it and "main scene" is set to "no" Nov 04 22:28:27 yeah, would trip into a full restart when it tried to hit 907... but 787 runs fine so far Nov 04 22:29:55 this "warthog pixi" is avail from preware on phone? Nov 04 22:30:06 DJNomad: that's what I have set too. You shouldn't need to touch the preferences much Nov 04 22:30:07 yes it is Nov 04 22:30:49 what category? Nov 04 22:30:56 SineOt, one pixi I have won't go above 800mhz but the other does 907mhz Nov 04 22:31:09 I should back it down for stability. Nov 04 22:31:55 DJNomad: you need to add the feed, first. Nov 04 22:32:21 I dont want to bother anyone but if someone has the time I would appreciate it Nov 04 22:32:26 name: webos-kernels-testing, url http://ipkg.preware.org/feeds/webos-kernels/testing/1.4.5/ Nov 04 22:32:45 only thing I have for trade would be secrets to the facebook game kingdoms of camelot lol Nov 04 22:33:19 sineot ,where do I go to add a feed ? Nov 04 22:33:30 under the Preware menu, "Manage Feeds" Nov 04 22:34:24 kk thx,I will have to manually add it ? Nov 04 22:35:00 and the "is compressed" would that be a yes or a no ? Nov 04 22:36:00 Hi! Im writting in search of anybody that please can help me with Quick Install installation. The problem, is that after getting done all the requirements (updating java, enabling devMode, etc), when i open quick install .jar file nothing happens. What should i do? does anybody knows? Nov 04 22:37:17 leave that yes, DJNomad Nov 04 22:37:23 the .jar file should open a program and run it Nov 04 22:37:36 Pablo_: WebOS Quick Install isn't supported by webos-internals... :p Nov 04 22:38:19 and do you know where to find support? Nov 04 22:38:28 Pablo_: but often we'll help where the answer is easy Nov 04 22:38:35 Pablo_: what host OS? Nov 04 22:38:47 win7 Nov 04 22:38:50 SineOt: I know I shouldnt have to ask this but what should I name the feed? Nov 04 22:39:15 Pablo_: ok, if you were following the instructions in the Preware Homebrew Documentation app in the app catalog, there is a link for win7 problems Nov 04 22:39:17 the problem is that the quick install does not run any program Nov 04 22:39:17 webos-kernels-testing is what I named mine Nov 04 22:39:24 I could never get it to find my phone using windows,I had to use linux to do it Nov 04 22:39:43 Pablo_: it points to a specific win7 thread on precentral Nov 04 22:40:12 rwhitby: shame you don't primarily use ubuntu any more. I tried changing your "rwhitby" section to my login name and for some reason it doesn't set the defaults right. could OSX's make be weird? Nov 04 22:40:18 Pablo_: it also has the support link for WOSQI Nov 04 22:41:22 destinal: it's a possibility. Most OSX/Xcode stuff uses LLVM now, so I don't know if they're even keeping GCC up to date any more, but who knows~ Nov 04 22:41:28 destinal: I just tried on Ubuntu and it worked, after changing the pre2 to pre in my login section Nov 04 22:42:06 +rwhitby: and what is really suppsed to do quickinstall? Nov 04 22:42:07 I'll try it again on Ubuntu with that section removed from the Makefile Nov 04 22:42:08 rwhitby: to pre? but version to 2.0.0 then? Nov 04 22:42:17 yes Nov 04 22:42:31 that's done by DEVICE=pre2 Nov 04 22:42:58 Pablo_: WOSQI is a java program Nov 04 22:43:44 I get: Nov 04 22:43:44