**** BEGIN LOGGING AT Thu Sep 01 02:59:57 2011
Sep 01 03:05:16 <halfhalo> basically setting up a dyndns thing but purely for devices.  I already have the server dns portion setup, just need a proper client
Sep 01 03:07:19 <PuffTheMagic> client?
Sep 01 03:07:22 <PuffTheMagic> u dont need a client
Sep 01 03:07:25 <PuffTheMagic> u can do it with curl
Sep 01 03:09:21 <ka6sox> scoutcamper, where were we on xapps?
Sep 01 03:09:34 <ka6sox> oh, I needed to ask about git.
Sep 01 03:10:32 <scoutcamper> ka6sox, idk, and im at lan party now, so not in ubuntu to look
Sep 01 03:10:59 <ka6sox> okay nm
Sep 01 03:11:12 <halfhalo> PuffTheMagic: part of it can be done with curl only.  but also plan on having the client update the server right after its ip changes/connects to the internet, plus am tempted to tunnel termlib.js through it as well.
Sep 01 03:12:05 <PuffTheMagic> termlib.js ?
Sep 01 03:12:15 <PuffTheMagic> is this different than the tty stuff in node?
Sep 01 03:12:19 <halfhalo> yeah
Sep 01 03:12:26 <PuffTheMagic> whats it do?
Sep 01 03:12:36 <halfhalo> termlib.js is a client side terminal emulator
Sep 01 03:13:46 <PuffTheMagic> so we could make a unjailed JS term app for webos?
Sep 01 03:14:54 <halfhalo> termlib.js is only for frontend, but if you connect it to say, a node server running as root, yes.
Sep 01 03:15:26 <PuffTheMagic> node server?
Sep 01 03:15:42 <PuffTheMagic> so termlib plus node tty could do it then?
Sep 01 03:15:51 <PuffTheMagic> so termlib does the VT100 processing and shit?
Sep 01 03:16:05 <halfhalo> there are a few other parts inbetween, but basically yes.
Sep 01 03:16:20 <PuffTheMagic> key event processing?
Sep 01 03:16:35 <halfhalo> termlib to node server running a http server and websocket server to tty
Sep 01 03:16:44 <halfhalo> termlib should handle key event processing
Sep 01 03:16:51 <PuffTheMagic> why is a server necessaru
Sep 01 03:17:00 <PuffTheMagic> and why http
Sep 01 03:17:05 <PuffTheMagic> im missing something
Sep 01 03:17:45 <PuffTheMagic> seems like this could be a normal webos app
Sep 01 03:17:50 <PuffTheMagic> with an unjailed node service
Sep 01 03:20:40 <halfhalo> server means you can load it up from a different device.  like windows.  which has a pretty bad cmd line env.  but i guess a normal app is doable assuming you deal with the node memory limit
Sep 01 03:22:54 <PuffTheMagic> which is?
Sep 01 03:23:21 <PuffTheMagic> i guess a C service could work too
Sep 01 03:23:41 <PuffTheMagic> i really wish C plugins didnt crash outside of jails
Sep 01 03:23:53 <PuffTheMagic> rwhitby, ^^ has this been tested on 302?
Sep 01 03:23:58 <PuffTheMagic> i last tried them on 300
Sep 01 03:24:18 <halfhalo> node has something like a 23MB ram limit by default on webos to keep services from sucking up all the ram
Sep 01 03:26:14 <halfhalo> its trivial to increase it using a command line flag though
Sep 01 03:34:03 <halfhalo> found that out the hard way with novapack (node palm-package basically), which as of 20 minutes ago is about to be both open sourced and made public.
Sep 01 03:34:30 <PuffTheMagic> nice
Sep 01 03:34:40 * PuffTheMagic sees new Tide version coming out
Sep 01 03:37:02 <halfhalo> mebe.  it all depends on how public it will be.  but I have concrete confirmation that it will become open source sooner rather than later.
Sep 01 03:51:20 <halfhalo> annnddd its not to be announced yet.  damn.  but I can use it on an app and distribute the resulting ipk for testing right now.  was hoping for more, but this is hp after all.
Sep 01 06:35:43 <ka6sox> cryptk, I'm chasing down the gpio bit that turns on and off wifi...BT is easy.
Sep 01 06:35:54 <cryptk> nice
Sep 01 06:36:12 <cryptk> once we have that GPIO bit for WiFi I think we have everything else that is needed to make it work
Sep 01 06:36:26 <cryptk> well, we have the rest of the technique.. then we just need a tool that does it
Sep 01 06:36:47 <cryptk> I have a stripped down version of bootie-config already that just pulls out and spits out the contents of the WIFICAL token
Sep 01 06:37:01 <ka6sox> iwconfig is an option for Angstrom from Narc.
Sep 01 06:55:21 <cryptk> well, for now the issue isn't with iwconfig or ifconfig... it's getting the kernel module properly loaded to get the interface available to be interacted with
Sep 01 06:55:47 <cryptk> what I may try and do is compile a version of the ar6000.ko module without the palm patches and see if we can just use that (which IMHO would be preferable)
Sep 01 06:56:27 <ka6sox> ya, but we still need to get power to it
Sep 01 06:56:44 <ka6sox> so I will continue to poke around and see what changes.
Sep 01 06:56:57 <ka6sox> ie turning wifi on and off.
Sep 01 06:57:23 <cryptk> yeah, but using the non-patched ar6000 module may make that easier or different
Sep 01 06:57:31 <cryptk> since the sysfs interfaces may be different
Sep 01 06:57:45 <cryptk> is there a tool that just monitors for things changing in sysfs?
Sep 01 07:01:26 <ka6sox> not really
Sep 01 07:01:37 <ka6sox> you kind of have to know where to look
Sep 01 07:03:00 <ka6sox> the incantations in the kernel that create sysfs entries are full of Dragons and Pit Vipers.
Sep 01 07:03:28 <ka6sox> and its kind of tough to figure out exactly where it will show up
Sep 01 07:03:28 <cryptk> snakes... why'd it have to be snakes...
Sep 01 07:03:34 <ka6sox> yup
Sep 01 07:03:46 <ka6sox> I did mine 1 way but most do it another.
Sep 01 07:03:51 <ka6sox> (even MORE convoluted)
Sep 01 07:04:27 <cryptk> but if we are using an un-patched ar6000.ko then many things become easier
Sep 01 07:04:40 <cryptk> such as not needing the WIFICAL token data and such
Sep 01 07:04:49 <ka6sox> sstate(foo->argv(blah))
Sep 01 07:05:18 <ka6sox> we would still want that I would think
Sep 01 07:05:28 <ka6sox> we may just need to process it into a file
Sep 01 07:05:37 <ka6sox> using your base64 decoding
Sep 01 07:05:45 <cryptk> well, one thing that I know is in there is a MAC that palm is assigning to override the one built into the wifi chip
Sep 01 07:05:58 <cryptk> other than that... I am not sure what is in there
Sep 01 07:05:58 <ka6sox> right
Sep 01 07:06:02 <ka6sox> but we *know* that
Sep 01 07:06:09 <cryptk> likely what channels it is allowed to use etc
Sep 01 07:06:44 <cryptk> what we need is to modify bootie-config to not just be able to list the tokens, but also be able to do the equivalent of tokens --mount
Sep 01 07:06:51 <ka6sox> and we don't even need to turn on the device to find out what the mac addr is.
Sep 01 07:06:54 <cryptk> then we can access them through /dev/tokens etc
Sep 01 07:16:32 <cryptk> so what it looks like is the tokens --mount just created /tmp/tokens with a file for each token... then a symlink is made at /dev/tokens that points to /tmp/tokens
Sep 01 07:16:42 <cryptk> that should be pretty easy to work up
Sep 01 07:16:52 <cryptk> I will work on getting bootie-config to be able to do that
Sep 01 07:17:11 <cryptk> then we can do bootie-config --mount-tokens and have it do the same thing
Sep 01 07:17:29 <cryptk> then we will have an FOSS version of the tokens binary
Sep 01 07:18:02 <ka6sox> yup
Sep 01 07:18:07 <ka6sox> sounds good.
Sep 01 07:18:21 <ka6sox> we should be able to publish that.
Sep 01 07:18:33 <cryptk> yep
Sep 01 07:18:44 <cryptk> that will be useful for us and the CM7 guys
Sep 01 07:19:06 <cryptk> I am on board with helping out the CM7 guys personally... I like their mentality much better than some of the other groups
Sep 01 07:27:01 <ka6sox> yup
Sep 01 07:28:22 <cryptk> there are several things that are closed source and we are just going to need to figure out what they do and come up with FOSS ways of doing them
Sep 01 07:31:46 <ka6sox> indeed
Sep 01 07:32:05 <ka6sox> I am looking at a variation of this to allow us to switch this for now
Sep 01 07:32:08 <ka6sox> https://github.com/slyon/bootr
Sep 01 07:32:24 <ka6sox> switch between angstrom and webos
Sep 01 07:32:36 <ka6sox> till we get a way to boot using buttons.
Sep 01 07:34:50 <ka6sox> this seems to swap kernels too.
Sep 01 07:51:15 <cryptk> nice
Sep 01 07:52:08 <ka6sox> what we need is a 3rd stage that allow us to swap kernels if the original one fails.
Sep 01 07:52:13 <ka6sox> (or replace bootie)
Sep 01 07:52:50 <cryptk> or figure out how to have bootie chainboot to a different bootloader
Sep 01 07:53:14 <ka6sox> thats what I mean.
Sep 01 07:53:54 <ka6sox> okay I now have a DEAD TP to take apart starting tommorrow.
Sep 01 07:59:04 <cryptk> how did it die?
Sep 01 07:59:46 <ka6sox> they replaced P6 with something from the "Test Code" TP image.
Sep 01 08:00:43 <ka6sox> P6 is modem code
Sep 01 08:00:52 <ka6sox> and it works in concert with P3
Sep 01 08:35:44 <cryptk> I find it funny that the option for the tokens command is --mount... it isn't actually mounting anything as far as I can tell
Sep 01 08:35:54 <cryptk> nothing shows up in mtab or anything
Sep 01 08:37:52 <cryptk> wow... just realized that tokens and trenchcoat are the same file...
Sep 01 08:38:26 <cryptk> looks like it just acts differently based on it's name... like how busybox works
Sep 01 08:45:56 <rwhitby> yep
Sep 01 08:46:40 <cryptk> well, I'm not looking to replicate all of trenchcoat... but adding functionality similar to tokens --mount to bootie-config should be pretty easy
Sep 01 08:46:56 <cryptk> just a directory creation... some file writes and a symink
Sep 01 08:50:43 <ka6sox> rwhitby, adding shim for xapps to chroot app or separate app?
Sep 01 08:54:53 <rwhitby> ka6sox: I have no idea what you are talking about
Sep 01 08:56:14 <ka6sox> I want to add this: http://pastebin.com/uTUxBRAu
Sep 01 08:56:27 <ka6sox> to /usr/sbin on the chroot during install of the chroot.
Sep 01 08:58:29 <cryptk> so... why is WIDK telling me that it can't find cc
Sep 01 09:04:17 <cryptk> ok WTF scratchbox...
Sep 01 09:04:41 <ka6sox> there is your answer.
Sep 01 09:04:59 <ka6sox> rwhitby what I just said up above.
Sep 01 09:11:07 <rwhitby> ka6sox: then you send it over in the package, and install it idempotently in the postinst
Sep 01 09:11:28 <ka6sox> correct
Sep 01 09:11:55 <rwhitby> since it's distro-dependent, it goes in the chroot package
Sep 01 09:12:16 <ka6sox> thats what I thought.
Sep 01 09:12:22 <rwhitby> ERROR_LOG looks like it has an extra space
Sep 01 09:12:31 <ka6sox> looking
Sep 01 09:12:46 <rwhitby> and why does line 13 have a \
Sep 01 09:13:22 <rwhitby> and you probably want to put ${APP_NAME} in the messages you send to the logs
Sep 01 09:13:33 <ka6sox> the paste put it back on 1 line
Sep 01 09:13:43 <ka6sox> but its 2 lines for ease of reading
Sep 01 09:13:47 <ka6sox> otherwise it wraps
Sep 01 09:13:58 <cryptk> ok, got my issues with CC fixed... working on modding bootie-config now
Sep 01 09:14:45 * rwhitby bbl
Sep 01 09:14:59 <ka6sox> ERROR_LOG does have an extra space, adding ${APP_NAME} to log messages is good idea.
Sep 01 09:20:07 <ka6sox> cryptk, http://www.gnu.org/s/hello/manual/libc/Creating-Directories.html
Sep 01 09:20:22 <ka6sox> its in glibc
Sep 01 09:21:33 <booto> apt-get install ${APP_NAME} >> ${APP_INSTALL_LOG_DIR}${INSTALLER_LOG} 2>&1
Sep 01 09:22:04 <ka6sox> after eh?
Sep 01 09:22:21 <ka6sox> oh I did the redirect wrong.
Sep 01 09:22:34 <ka6sox> too many things...I fix.
Sep 01 09:24:10 <cryptk> ka6sox, dirent.h was the include I was looking for
Sep 01 09:24:22 <ka6sox> ah....
Sep 01 09:24:32 <ka6sox> I wonder why gnu didn't say taht.
Sep 01 09:24:54 <cryptk> #include <dirent.h>
Sep 01 09:25:04 <cryptk> then you can just mkdir("/some/path/somedir");
Sep 01 09:26:08 <ka6sox> ya,
Sep 01 09:26:12 <ka6sox> nice
Sep 01 09:26:17 <ka6sox> I need to add time too.
Sep 01 09:26:30 <ka6sox> but that may not work
Sep 01 09:26:33 <ka6sox> I should check.
Sep 01 09:29:06 <booto> which buildenv is using dirent.h for mkdir?
Sep 01 09:29:31 <cryptk> booto, this is for a modification I am making to bootie-config for the angstrom port
Sep 01 09:29:31 <ka6sox> okay looks like chroot keeps good time...wrong TZ but good time.
Sep 01 09:29:55 <cryptk> ka6sox, that's expected... it's time will come from the host OS... modified by the TZ config in the chroot
Sep 01 09:30:31 <ka6sox> I just found another post install tweak I want to do.
Sep 01 09:30:41 <ka6sox> match TZs
Sep 01 09:37:59 <ka6sox> man date
Sep 01 09:38:08 <booto> cryptk: I think you might want <sys/stat.h> not <dirent.h> for mkdir (?)
Sep 01 09:38:44 <cryptk> I tested it with dirent.h, works fine
Sep 01 09:38:52 <cryptk> I am beyond that step now
Sep 01 09:39:23 <booto> weird :S
Sep 01 09:40:10 <ka6sox> I can never remember the right incantation for getting a nice compact date/time group.
Sep 01 09:44:46 <cryptk> date --rfc-3339=seconds?
Sep 01 09:45:56 <cryptk> what format are you looking for?
Sep 01 09:46:00 <cryptk> I can tell you what command to sue
Sep 01 09:46:08 <cryptk> s/sue/use/
Sep 01 09:47:32 <ka6sox> yyyymmddhhmm
Sep 01 09:50:35 <cryptk> date +%Y%m%d%H%M%S
Sep 01 09:51:00 <cryptk> that results in YYYYMMDDHHMMSS
Sep 01 09:51:15 <cryptk> 24 hour time
Sep 01 09:52:29 <ka6sox> okay
Sep 01 09:52:55 <ka6sox> let me add that.
Sep 01 10:06:08 <ka6sox> http://pastebin.com/ZpC02Dia
Sep 01 10:06:34 <ka6sox> and I don't know what it is with pastebin.com...they are messing up the formatting
Sep 01 10:11:09 <ka6sox> okay I am done with that for the night.
Sep 01 10:54:29 <cryptk> ok, I think I have bootie-config working with the --mount option now
Sep 01 11:10:36 <cryptk> ok, bootie-config officially knows how to do bootie-config --mount-tokens now
Sep 01 11:10:43 <cryptk> root@beagleboard:~# md5sum token_bdata.SD32.bin token_bdata.SD32.bin.2
Sep 01 11:10:43 <cryptk> e172c82b314162da1918ba846cbaaeaa  token_bdata.SD32.bin
Sep 01 11:10:43 <cryptk> e172c82b314162da1918ba846cbaaeaa  token_bdata.SD32.bin.2
Sep 01 11:11:14 <cryptk> the first is the file straight from webOS, the second is from bootie-config mounting the tokens, then be base64 decoding the WIFICAL token
Sep 01 11:11:25 <cryptk> we have a match... which means we have an FOSS version of tokens now
Sep 01 11:12:34 <cryptk> I'm not putting it in the bootie-config repo until I get rwhitby to look at it though... since my code is likely trash and he can clean it up now that the logic is there
Sep 01 11:35:36 <cryptk> rwhitby, when you get a chance, can you review this file http://git.cryptkcoding.com/bootie-config/tree/misc.c
Sep 01 11:35:49 <cryptk> new functions are mount_tokens and write_token_file
Sep 01 11:36:08 <cryptk> I don't want to push them to the main repo because my C code is likely horrid and could be cleaned up a lot
Sep 01 22:31:08 <ka6sox> scoutcamper, bring back up the Xapps shim
Sep 01 22:31:16 <ka6sox> need to add a removal tool.
Sep 01 22:32:21 <scoutcamper> ka6sox, done
Sep 01 22:34:43 <ka6sox> this would best be done with case...but....
Sep 01 22:34:54 <dtzWill> you guys pushing on the xapps front?! oh thank god :D
Sep 01 22:35:05 <ka6sox> dtzWill, yup
Sep 01 22:35:11 <dtzWill> if webos's 'window manager' worked with the keyboard
Sep 01 22:35:20 <dtzWill> that'd be the most amazing thing ever.
Sep 01 22:35:20 <dtzWill> xD
Sep 01 22:35:29 <dtzWill> wonder if we could fake enough mouse movement... :P
Sep 01 22:35:37 <scoutcamper> lol
Sep 01 22:35:56 <ka6sox> might be....
Sep 01 22:36:26 <ka6sox> I don't know if case is supported in our version of sh.
Sep 01 22:36:55 <dtzWill> ka6sox: dep on bash from optware, think it was preware-ified
Sep 01 22:36:57 <dtzWill> pretty sure :)
Sep 01 22:37:05 <dtzWill> no one will begrudge a proper shell
Sep 01 22:37:05 <dtzWill> :)
Sep 01 22:37:17 <scoutcamper> ka6sox, i would check but my TP is dead
Sep 01 22:37:32 <ka6sox> dtzWill, trying not to make the chroot depend on installing optware and bash
Sep 01 22:37:57 <ka6sox> scoutcamper, why dead?
Sep 01 22:38:14 <scoutcamper> as in battery is not charged
Sep 01 22:38:28 <ka6sox> go charge it mon
Sep 01 22:38:49 * scoutcamper just put it on the charger
Sep 01 22:39:17 <rwhitby> ka6sox: we don't want chroot depending on optware
Sep 01 22:46:34 <linuxjacques> which shells are in busybox?
Sep 01 22:46:44 <linuxjacques> (on the tp)
Sep 01 22:46:59 <linuxjacques> is it using ash?
Sep 01 22:50:19 <linuxjacques> ka6sox: FYI I am using case in ash
Sep 01 22:51:10 <ka6sox> ash
Sep 01 22:51:26 <ka6sox> okay so the way to do this is a case.
Sep 01 23:09:46 <ka6sox> scoutcamper, makes sense?
Sep 01 23:10:43 <scoutcamper> ka6sox, tbh, i wasnt watching lol
Sep 01 23:10:51 <ka6sox> go look
Sep 01 23:14:32 <scoutcamper> ka6sox, yes, makes sense
Sep 01 23:23:45 <ka6sox> scoutcamper, the Makefile is in the dir we are working on?
Sep 01 23:24:17 <scoutcamper> yes
Sep 01 23:34:49 <ka6sox> I haven't found anything that "installs" or in any way packages up the "extra" files that are there.
Sep 01 23:35:26 <ka6sox> and puts them in the ipkg
Sep 01 23:36:19 <scoutcamper> what about the upstart files? they are packaged
Sep 01 23:37:25 <ka6sox> yes, thru the install here:
Sep 01 23:38:02 <scoutcamper> yes
Sep 01 23:38:13 <ka6sox> install -m 644 upstart/${APP_ID} build/armv7/etc/event.d/${APP_ID}
Sep 01 23:38:25 <scoutcamper> so write another line for the script
Sep 01 23:40:35 <ka6sox> install -m 744 files/usr/sbin/Xapps build/foo/usr/sbin/Xapps
Sep 01 23:41:04 <ka6sox> that other line was from the makefile for optware/openssh
Sep 01 23:41:43 <scoutcamper> install -m 744 files/usr/sbin/Xapps build/foo/media/ext3fs/ubuntu-natty-chroot/usr/sbin/Xapps
Sep 01 23:41:57 <ka6sox> no
Sep 01 23:42:22 <ka6sox> because before postinst and verifying the ext3fs is mounted and good etc....
Sep 01 23:42:25 <ka6sox> that doesnt' exist
Sep 01 23:42:55 <scoutcamper> oh, right
Sep 01 23:43:38 <scoutcamper> so in postinst you add cp /usr/sbin/Xapps /media/ext3fs/usr/sbin/Xapps
Sep 01 23:44:02 <ka6sox> yes after a:
Sep 01 23:44:08 <ka6sox> mkdir -p build/$*/usr/palm/applications/${APP_ID}/usr/bin
Sep 01 23:44:24 <scoutcamper> yes
Sep 01 23:44:26 <ka6sox> mkdir -p build/$*/usr/palm/applications/${APP_ID}/usr/sbin
Sep 01 23:47:17 <ka6sox> nstall -m 744 usr/sbin/*.sh build/$*/usr/palm/applications/${APP_ID}/usr/sbin/
Sep 01 23:47:22 <ka6sox> i+
Sep 01 23:48:48 <ka6sox> okay that should be all.
Sep 01 23:49:51 <ka6sox> need to create usr/sbin/ in the git repo add Xapps.sh to the dir and use this buildscript + the modded postinst to build then test.
Sep 01 23:51:44 <ka6sox> but before that...verify my substitutions (especially for any missing /'s)
Sep 01 23:52:00 <scoutcamper> allready did in a non screen
Sep 02 00:29:20 <ka6sox> okay, now that the shim works and installs....
Sep 02 00:29:36 <ka6sox> on to headless apps.
Sep 02 02:21:00 <PuffTheMagic> halfhalo: your js shit pushed to public yet?
Sep 02 02:48:04 <chpwn> when messing with Preware, is there a way to not have to sit at the changelog waiting screen every install?
Sep 02 02:48:41 <rwhitby> there's some comments in cookie.js which you can uncomment
Sep 02 02:52:05 <crimsonredmk> lol chpwn
Sep 02 02:52:28 <crimsonredmk> i dunno, take 3 hours or so to patch that up while i wait 3 seconds every 2 weeks or so :D
Sep 02 02:53:01 <chpwn> crimsonredmk: when building it and testing it, i mean, not installing it.
Sep 02 02:53:17 <crimsonredmk> ah, well then that makes sense
**** ENDING LOGGING AT Fri Sep 02 02:59:57 2011