#touchbook IRC Log


IRC Log for 2009-11-03

Timestamps are in GMT/BST.

[0:00] <dpb> gregoiregentil: does the install-sd.sh script only work with 8GB sd-cards?
[0:00] <gregoiregentil> yes and no. Read the script. there are different options and sections
[0:00] <gregoiregentil> it's working perfectly with the Silicon Power card (only this model)
[0:01] <gregoiregentil> then it can run for any other card but you need to read the comment
[0:01] <hyc> huh, how can the DSS2 driver think it's always suspended...
[0:01] <gregoiregentil> and you may have one action to do yourself
[0:01] <gregoiregentil> hyc: if only I knew ;-)
[0:01] <dpb> gregoiregentil: I have a script that formats sd-cards, and should work on any sd-card
[0:02] <dpb> gregoiregentil: I took the script from http://www.xora.org.uk/2009/09/07/omap3-card-formatter-license/, but added a third partition as swap: http://koti.kapsi.fi/~dpb/tmp/mkcard.sh.txt (if you're interested in it)
[0:04] <gregoiregentil> interesting. please read my comment in my script
[0:04] <gregoiregentil> I hit a bug with sfdisk and this script is a little bit similar
[0:04] <Corsac> gregoiregentil: hmhm, can I easily import a file from another git branch?
[0:04] <gregoiregentil> wonder if the guy found a solution to the problem I hit
[0:04] <Corsac> gregoiregentil: btw, mounting the vfat part with the sync option might be a good idea
[0:05] <Corsac> (to be sure that the mlo is the first file on the part without running sync after each copy)
[0:05] <gregoiregentil> yes. the umount for mlo is still the safest path to make sure that mlo is the first file
[0:05] * leinir (n=leinir@amarok/usability/leinir) Quit (Read error: 113 (No route to host))
[0:05] <dpb> gregoiregentil: well that's weird if you can't boot it. this script I use has worked quite alright booting ubuntu for me
[0:06] <gregoiregentil> That's really fun this script
[0:06] <gregoiregentil> it's very similar to what I do (I was not aware of this script BTW ;-) )
[0:07] <dpb> :)
[0:07] <gregoiregentil> there is one tiny difference which was very close to the source of the problem I identified but was not able
[0:07] <gregoiregentil> to fix!!!
[0:07] <dpb> well maybe you can fix it now ;)
[0:07] <gregoiregentil> If you compare, the guy does:
[0:07] <gregoiregentil> { echo ,9,0x0C,* echo ,,,- } | sfdisk -D -H 255 -S 63 -C $CYLINDERS $DRIVE
[0:07] <gregoiregentil> I do
[0:07] <gregoiregentil> sfdisk -C${CYL} -H255 -S63 -f ${DEST} << EOF
[0:07] <gregoiregentil> 0,5,c,*
[0:07] <gregoiregentil> ,62,
[0:07] <gregoiregentil> ,62,82
[0:07] <gregoiregentil> ,
[0:07] <gregoiregentil> EOF
[0:08] <gregoiregentil> and the problem with my script is that the card was not booting because the very first sector was not really written correctly
[0:08] <gregoiregentil> the only difference is that I have 0,5... and he has ,9
[0:08] <dpb> that doesn't matter
[0:08] <gregoiregentil> I'm wondering if my mistake was not to force 0 as the starting sector instead of letting sfdisk doing it
[0:08] <dpb> I changed mine to 5
[0:08] * leinir (n=leinir@amarok/usability/leinir) has joined #touchbook
[0:09] <gregoiregentil> not the 5 vs. 9
[0:09] <gregoiregentil> but the 0 vs. nothing as the very first argument
[0:09] * diroots (n=diroots@ Quit (Remote closed the connection)
[0:09] * diroots (n=diroots@ has joined #touchbook
[0:09] <dpb> true
[0:09] <gregoiregentil> I think that it matters because if you do fdisk after running my script, I realized that I was not writing at the right place for the first sector
[0:09] <gregoiregentil> I should try...
[0:09] <dpb> might also be because you have too few ,-marks on the other lines, maybe it needs those too
[0:11] <hyc> why doesn't just fdisk work?
[0:11] <hyc> that's what I used on all my cards
[0:11] <gregoiregentil> fdisk was working but fdisk, you need to do it manually
[0:11] <gregoiregentil> on sfdisk you can do it in full command line
[0:11] <hyc> hm, you can't make it read from stdin?
[0:12] <gregoiregentil> fdisk: no, I don't think so
[0:13] <dpb> gregoiregentil: will there be newer kernels coming anytime soon?
[0:14] <gregoiregentil> I would like to bring linux-omap-2.6.29-pm very soon
[0:14] <dpb> great
[0:15] <DJWillis> Corsac: bumped up xfce-panel to 4.6.2 and pushed to my OP tree. Seems ok, I wanted a bug fix from it.
[0:18] <Corsac> \o/
[0:20] <Corsac> gregoiregentil: new kernel means compiling the kernel, generating the uImage, put it on the vfat part, put the modules on the ext part, and voila ?
[0:21] <dpb> yes
[0:21] <gregoiregentil> correct
[0:21] <Corsac> might try to build that too
[0:21] <Corsac> though I'll first let it build webkit+midori during the day
[0:21] <Corsac> and go to work :)
[0:21] <Corsac> cheers,
[0:21] <gregoiregentil> OK. Try the webkit patch referenced by the Midori guy if you have time
[0:24] <koen> webkit patch?
[0:26] <gregoiregentil> there is a bug in Midori / webkit when we scroll with finger. it's been acknowledged by Midori's author
[0:26] <gregoiregentil> we are trying to figure out where the problem is.
[0:27] <gregoiregentil> http://www.alwaysinnovating.com/tmp/midori.jpg
[0:27] <koen> ah, I haven't tried ts mode yet
[0:28] <gregoiregentil> With a few fixes, it can be really cool and to my mind, it beats crazy fennec (which doesn't work at all)
[0:28] <gregoiregentil> it's just very simple (not too many fancy features), like an iphone
[0:28] <koen> gregoiregentil: I added a patch to gecko-media-player to have configurable backends
[0:29] <koen> gregoiregentil: brijesh and I are looking at using his gst-player as a proof of concept
[0:29] <gregoiregentil> OK. definitely very interesting
[0:32] * fooq (n=reini@cm56-152-97.liwest.at) has joined #touchbook
[0:36] <DJWillis> gregoiregentil: yep, seen the Midori issues in my testing with the TS and fingerscroll.
[0:36] <ryuo> gregoiregentil, any idea whats causing the seemingly random black outs of the screen?
[0:37] <DJWillis> gregoiregentil: also seen crashes that I am still trying to track down, what is stability like on the TB?
[0:37] <gregoiregentil> I have some ideas. working on it
[0:37] <ryuo> i hope its software.
[0:37] <gregoiregentil> DJWillis: what kind of crashes?
[0:37] <ryuo> so it can just be patched
[0:38] <DJWillis> gregoiregentil: the odd error where the app just disappears sometimes when using the finger scroll. Odd and I can't get a good trace out of it.
[0:39] <gregoiregentil> No. I haven't see that
[0:39] <gregoiregentil> which version of Midori? 0.2.0?
[0:40] <ryuo> greg, how would we enable the default Xfce menu? i can't find it.
[0:40] * azaghal (n=azaghal@ Quit (Read error: 113 (No route to host))
[0:40] <ryuo> the one that lets you change xfwm4 themes, etc.
[0:40] <ryuo> hm
[0:41] <gregoiregentil> try xfce4-settings-show in terminal
[0:41] <ryuo> ah, thanks.
[0:41] <DJWillis> gregoiregentil: yep, i'll keep looking into it, sure it's something local if it's good for you, are you still on 4.4.2 Xfce or trying 4.6.1 underneath out of interest?
[0:41] <ryuo> 4.6.x fixes many bugs..
[0:42] <gregoiregentil> DJWillis: I did the 4.6 port but I never completed all the testing so I didn't push it and stayed with 4.4
[0:42] <gregoiregentil> it's on the todo list obviously
[0:42] <ryuo> i'd say hardware bugs are a bigger priority though
[0:42] <ryuo> namely some of the wierd behavior
[0:42] <gregoiregentil> ouahh, I found the bug in my script for any other card. It was the -D option in sfdisk to be dos compatible!!!
[0:43] <gregoiregentil> They guy at http://www.xora.org.uk/2009/09/07/omap3-card-formatter-license/ is really good ;-)
[0:43] <ryuo> boy, this thing is so different...
[0:44] <ryuo> i've only known computers in the x86 department.
[0:44] <DJWillis> gregoiregentil: yep, we found out about your Xfce 4.6 work a day or so after I released the recipies I did for the OP ;). Support in mainline OE for stock Xfce 4.6 is not bad now really. Well pretty solid on the few devices I have tried (ARM and others).
[0:44] <ryuo> I'm used to BIOS, CD, USB booting...
[0:44] <ryuo> grub
[0:44] <ryuo> MBR
[0:44] <gregoiregentil> DJWillis: yup
[0:44] <ryuo> just need to get networking to work...
[0:45] <ryuo> i have no wireless AP right now
[0:46] <DJWillis> gregoiregentil: I notice in some of your branding patches that you could do some of them without resorting to source tweaks, I was pondeirng a clean way for a machine and/or distro in OE to provide a set of branding for Xfce Panel etc. - any thoughts on that?
[0:48] <DJWillis> ryuo: use the OTG USB for networking to a PC?
[0:48] <ryuo> DJWillis, huh?
[0:48] <ryuo> i only have ethernet right now
[0:49] <ryuo> but i can probably make a wireless software AP with my x86 laptop broadcom chipset
[0:49] <ryuo> i am a little disappointed lol. my favorite editor won't link on ARM
[0:50] <DJWillis> ryuo: if you don't have Wireless then use the on the go USB client port (the TB still has the OTG port right?) to connect to a PC and network via that. Saying that, if you can make a virtual access point, go for that ;-)
[0:50] <gregoiregentil> Good to have this kind of commit: http://git.alwaysinnovating.com/cgit.cgi/ai.openembedded.dev/commit/?id=83d1e5b3d9643f05a25eb886793d0221ca6e3897 how to fix a bug that has been opened for six months or so....
[0:50] <ryuo> the problem is some compatibility in the wide character types
[0:50] <ryuo> wchar_t
[0:50] <ryuo> i'll notify the maintainer so maybe he can fix it
[0:51] <DJWillis> gregoiregentil: got to like those sort of fixes, simple and rewarding.
[0:51] <gregoiregentil> DJWillis: not totally sure to see what you mean though i get the feeling. Yes many patches (that came from Zonbu) to have some branding stuff
[0:51] <gregoiregentil> one of my main complaints against xfce4-panel, there is no support for transparency
[0:51] <gregoiregentil> it's really too bad that this feature has been left on the side for such a long time
[0:52] <koen> gregoiregentil: tried doing 'inherit gconf' to get the proper postinst automatigically for onboard?
[0:53] <gregoiregentil> arghhh!!!!
[0:53] <gregoiregentil> it took me a while to fix this bug too. At least, it was interesting to understand the schema stuff
[0:54] <gregoiregentil> OE is pretty impressive for those kinds of stuff. to generate the postint just from one line in the makefile... Congrats! ;-)
[0:56] <ryuo> gregoiregentil, i believe fbpanel supports some kind of transparency? but i think it only uses the root window image
[0:56] <DJWillis> gregoiregentil: don't you find that transparency kills the performance with the current backend X drivers?
[0:57] <gregoiregentil> with xfwm4, it's not too bad. xfwm4 has always been very impressive (even when I was working on crappy VIA CPU)
[0:57] <ryuo> DJWillis, you mean 'composite'?
[0:57] <gregoiregentil> Macer: yes. it's accessible in the desktop settings
[0:58] <gregoiregentil> Olivier's work has always been very impressive
[0:58] <ryuo> fbpanel is interesting but there is no GUI config for it
[0:58] <ryuo> composite is slow.
[0:59] <ryuo> well it should be the video chipset
[0:59] <gregoiregentil> major performance hit: I wouldn't say that
[0:59] <gregoiregentil> performance degradation: yes
[0:59] <ryuo> hm
[0:59] <gregoiregentil> no, it doesn't use SGX
[0:59] <ryuo> well your the ARM expert, greg =p
[1:00] <jvs> gregoiregentil, when can we expect news about the next batch?
[1:00] <gregoiregentil> jvs: please read the log within a couple of hours. I have replied to a similar question a little bit earlier
[1:01] <ryuo> gregoiregentil, you need someone to moderate spam? I saw some on the forums earlier.. and still there
[1:01] <DJWillis> <ryuo> DJWillis, you mean 'composite'? << I was thinking more about fake transparancy stuff.
[1:01] <jvs> gregoiregentil, thanks (and sorry, couldn't find this message)
[1:01] <ryuo> oh.
[1:01] <ryuo> composite is "real" transparency but very performance intensive
[1:02] <ryuo> http://www.alwaysinnovating.com/forums/viewtopic.php?f=12&t=404 <-- if this isn't spam i don't know what is :P
[1:02] <DJWillis> ryuo: well aware of that ;) but you would want better unlaying support (be it NEON, ARM or SGX) for that to really work well IMHO ;-)
[1:03] <ryuo> gregoiregentil, the magnets you sent with my TB are so strong... might as well call them "samson magnets" :D
[1:03] <hyc> Macer: lay off the caffeine, man...
[1:03] <ryuo> i can't even pull them apart with my hands
[1:03] <DJWillis> Macer: extreme ;-), it would have be playing with the DSS2 /sys nodes to see if something is awol in there but 'you takes your choice'
[1:03] <gregoiregentil> ryuo: browse the wiki
[1:04] <ryuo> thanks
[1:04] <gregoiregentil> I give the solution how to "easily" separate them
[1:05] <ryuo> haha. i love these super magnets
[1:05] <DJWillis> Hmm, are they neodymium magnets?
[1:05] <ryuo> at least my TB won't fall down :D
[1:06] <ryuo> kudos to the guy who can turn the TB into a toaster :P
[1:07] <fooq> gregoiregentil: i can't find anything regarding the October batch in today's logs... i know that you talked about it on Sunday, do you mean these logs?
[1:08] <fooq> hehe
[1:09] <jvs> he is used to it
[1:09] <jvs> believe me
[1:09] <gregoiregentil> yes. one second
[1:09] <ryuo> gregoiregentil, still. great work I say. I've never seen a netbook/laptop device you could detach the screen with, mount on a magnetic surface, etc...
[1:10] <ryuo> my 16G SD card will finally see some use
[1:10] <ryuo> Macer, i think theres a reason its not.
[1:11] <ryuo> perhaps their afraid it'd come out in the case during shipment?
[1:11] <gregoiregentil> ** Blanking screen problem ** I think that I may have a fix for the blanking screen bug #39 on Bugzilla. You may try this kernel on your first partition of the SD card: http://www.alwaysinnovating.com/tmp/uImage
[1:11] <ryuo> gregoiregentil, I'll try it too. i'm having the same problem
[1:12] <ryuo> hm
[1:13] <koen> gregoiregentil: any patch we can try?
[1:14] <ryuo> gregoiregentil, i know of another company that sells linux hardware. perhaps your companies could have a partnership or something? http://www.zareason.com/shop/home.php
[1:14] <jvs> gregoiregentil, I was just asking because last time you mentioned to put some news online on monday
[1:14] <ryuo> i came across them when i heard about the cloudbook from the now defunct everex
[1:14] <gregoiregentil> You have them. this kernel is the the OE-touchbook kernel I asked you to push upstream
[1:14] <koen> gregoiregentil: ok, so I don't have to add an extra patch to the -pm kernel :)
[1:15] <gregoiregentil> jvs: yup. I was too busy today and not sure how to best formulate the news about what I said earlier
[1:15] <gregoiregentil> koen: yes. no new patch for you
[1:15] <gregoiregentil> ryuo: interesting company
[1:15] <koen> gregoiregentil: tried the -pm kernel yet?
[1:15] <ryuo> gregoiregentil, heh. =p
[1:16] <gregoiregentil> koen: but note that the state of dss2 doesn't seem to work
[1:16] <ryuo> its focus is x86 though
[1:17] <jvs> gregoiregentil, we are just impatient and want to develop
[1:17] <gregoiregentil> we have this patch and it was working but ti doesn't seem to work anymore. It always reports 0
[1:17] <gregoiregentil> http://git.alwaysinnovating.com/cgit.cgi/ai.openembedded.dev/tree/recipes/linux/linux-omap-2.6.29/omap3-touchbook/dss2-export-status.patch
[1:17] <gregoiregentil> it's interesting to know the state of DSS2. It would make sense to have this in DSS2
[1:17] <jvs> gregoiregentil, formulate a sentence with smthg like "you will be in the next batch" :)
[1:18] <ryuo> jvs, why not grab qemu ARM emulation in the mean time?
[1:18] <jvs> ryuo, its like playing need for speed while waiting for your bugatti
[1:18] <ryuo> heh.
[1:19] <fooq> wont help with touchscreen jitter problems, and the like ryuo
[1:19] <ryuo> oh.
[1:19] <fooq> we know the forums by heart, and well, emulated it wont be of much help
[1:19] <ryuo> i thought you guys were still waiting on your orders
[1:19] <fooq> correct
[1:19] * azaghal (n=azaghal@ has joined #touchbook
[1:20] <Corsac> qemu-omap3 doesn't really work anyway
[1:20] <ryuo> i got my order yesterday
[1:20] <fooq> gratz
[1:20] <ryuo> i meant qemu-versatile :x
[1:20] <Corsac> qemu-versatile won't boot the kernel
[1:20] <ryuo> Oh. hm
[1:20] <Corsac> see my wiki page on qemu
[1:20] <gregoiregentil> koen: don't you want to try something like that http://git.alwaysinnovating.com/cgit.cgi/ai.openembedded.dev/tree/recipes/linux/linux-omap-2.6.29/omap3-touchbook/dss2-export-status.patch ?
[1:20] <Corsac> you can start from there, but...
[1:21] <ryuo> i had used qemu with debian
[1:22] <DJWillis> gregoiregentil: considered mailing Tomi with DSS2 feature requests? Not sure what the 'standard' way to expose a display status is but I can see it's a good idea.
[1:22] <gregoiregentil> yup. Tomi is nice. Probably very busy too.
[1:22] <ryuo> i found the /sys nodes for power, but its not like ACPI battery info i'm used to
[1:22] <dpb> gregoiregentil: so what did you change in the kernel for bug #39?
[1:22] <gregoiregentil> but it would be better to send to Tomi a working patch...
[1:23] <gregoiregentil> dpb: the DSS2 set of patches was a little bit old and probably not correct.
[1:23] <dpb> ok
[1:23] <gregoiregentil> so I can't say what was wrong but this set of DSS2 patches is clean and should work better
[1:23] <dpb> great
[1:23] <gregoiregentil> I have a device with xset off for more than 30 minutes and no blanking (though this unit was blanking before)
[1:24] <gregoiregentil> I will leave it for the night
[1:24] <gregoiregentil> 1:30am: sorry! it's late here. going to bed
[1:24] <gregoiregentil> bye now
[1:24] <ryuo> k
[1:24] <hyc> 'night
[1:25] <DJWillis> night gregoiregentil
[1:29] <jvs> nite
[1:29] <ryuo> the only real bug i've found is the screen blanking lol
[1:50] * alongst (n=alon@ has joined #touchbook
[1:51] <Corsac> hmhm, python fails
[1:51] <Corsac> http://paste.debian.net/50559/
[1:55] <DJWillis> Corsac: yep, not shipping the lib would = fail ;-). Is that from mainline?
[1:57] <Corsac> nop
[1:57] <Corsac> but I'm more concerned with:
[1:57] <Corsac> arm-angstrom-linux-gnueabi-objcopy:/home/corsac/projects/touchbook/aios/tmp/work/armv7a-angstrom-linux-gnueabi/python-2.6.1-ml5/image/usr/lib/python2.6/config/stpZo3FS/getbuildinfo.o: cannot create debug link section `/home/corsac/projects/touchbook/aios/tmp/work/armv7a-angstrom-linux-gnueabi/python-2.6.1-ml5/image/usr/lib/python2.6/config/.debug/libpython2.6.a': Invalid operation
[2:00] <DJWillis> Corsac: did you rebase with the 'strip-libs' stuff, I wonder if it is that, koen would know for sure.
[2:00] <Corsac> didn't rebase myself on that one
[2:00] <Corsac> I just pull from ai
[2:05] <DJWillis> Corsac: ahhh, not built from the AI tree.
[2:05] <Corsac> sure it is
[2:06] <DJWillis> Corsac: do your push your tree?
[2:06] <Corsac> nop
[2:06] <Corsac> (I don't really have somewhere to push to)
[2:08] <DJWillis> Corsac: GITHub ;) - Not that it matters but it's nice to comment and see others cool work.
[2:08] <Corsac> didn't really do any work atm
[2:08] <Corsac> DJWillis: is there -dev packages in OE?
[2:09] <Corsac> like, where should the libpython2.6.so and .a files go?
[2:09] <DJWillis> Corsac: yep.
[2:09] <Corsac> don't think so
[2:11] <Corsac> http://cgit.openembedded.org/cgit.cgi/openembedded/tree/recipes/python/python_2.6.1.bb there's no python-dev package
[2:14] <DJWillis> Corsac: no, the one BB will produce it with a ${PN}-dev (note, I am stuck at work so can't check the recipie it's self right now).
[2:14] <Corsac> ok, that's in the python.inc stuff I guess
[2:15] <Corsac> hmhm, not really
[2:15] * fooq (n=reini@cm56-152-97.liwest.at) Quit ()
[2:16] <Corsac> but python 2.6.2 recipe has the -dev stuff
[2:17] <Corsac> (now I don't know why bitbake midori tries to build 2.6.1 and not 2.6.2
[2:19] <DJWillis> Corsac: default-pref = -1 somewhere.
[2:19] * Vito89 (n=quassel@ has joined #touchbook
[2:20] <koen> OE .dev build 2.6.2
[2:21] <Corsac> grep -c default-pref python/python_2.6.2.bb
[2:21] <Corsac> 0
[2:24] <koen> grep python conf/distro/includes/angst*
[2:24] <Corsac> conf/distro/include/angstrom-2008-preferred-versions.inc:PREFERRED_VERSION_python = "2.6.1"
[2:25] <Corsac> :)
[2:25] <Corsac> I guess I should put a PREFERRED_VERSION_PYTHON = "2.6.2" in my local conf?
[2:26] * andrewgodwin (n=andrew@hydrae.aeracode.org) Quit (Excess Flood)
[2:26] * andrewgodwin (n=andrew@hydrae.aeracode.org) has joined #touchbook
[2:27] <koen> or remove that line
[2:30] * Vito89 (n=quassel@ Quit ("No Ping reply in 90 seconds.")
[2:31] * Vito89 (n=quassel@ has joined #touchbook
[2:39] <Corsac> ERROR: QA Issue with libpython2: non -dev package contains symlink .so: libpython2 path '/work/armv7a-angstrom-linux-gnueabi/python-2.6.2-ml6.0/install/libpython2/usr/lib/libpython2.6.so'
[2:39] <Corsac> hinhin
[2:53] * alongst (n=alon@ Quit (Read error: 110 (Connection timed out))
[3:06] <koen> Corsac: do you know what the webkit patch is gregoire was talking about?
[3:16] <koen> DJWillis: tracker 0.5.x has buildproblems, do you have a WIP tracker 0.6.x recipe?
[3:18] <DJWillis> koen: I dropped tracker from my builds for now (as part of cleaning out Gnome bits that I carried in Xfce), I have a WIP 0.6 somewhere however, not sure if it's on this box, i'll have a look.
[3:18] <dpb> anyone else tried the kernel from bug #39 yet?
[3:31] <Vito89> gregoiregentil: hi, ad. october batch, if there are still units available, otherwise consider it as dump, on sunday(i know, sunday... but it's tuesday now) i've send email to contac@alwa.., regarding my order (ZPAPP), yet you(they) didn't respond. Shall i wait till they respond? Or you can push it faster to complete order? I just want to be sure:) Thanks
[3:32] <dpb> Vito89: are you sure you didn't send it to contact@ ? you're missing a t..
[3:35] * Vito89 (n=quassel@ Quit (Read error: 104 (Connection reset by peer))
[3:35] * Vito89 (n=quassel@ has joined #touchbook
[3:36] <Vito89> stupid wifi, yeah i'm sure, the automat responded with an automatic confirmation
[3:53] * elysion (n=mikoki@81-175-179-250.bb.dnainternet.fi) has joined #touchbook
[3:53] <elysion> hello!
[3:54] <elysion> is there anyone here who could tell me what are the most critical issues in the touchbook at the moment?
[3:54] <leinir> Random black screen? :)
[3:54] <leinir> which comes back on after tapping the power button a couple of times
[3:55] <elysion> it looks like a really interesting piece of equipment, but I tend to get quite frustrated with a device that doesn't work properly
[3:55] <leinir> Very odd, but workaroundable for now :)
[3:55] <elysion> ok, so if you just press the button a few times, it comes back without needing to reboot or anything?
[3:55] <leinir> Yeah :)
[3:55] <elysion> ok, doesn't sound that bad
[3:55] <elysion> what kind of other issues does it have?
[3:56] <elysion> how is the software support for it?
[3:56] <leinir> Software is under heavy development, but for the most part is quite pleasant :)
[3:56] <elysion> which linux distro is it based on?
[3:56] <elysion> easy to port regular linux software to it?
[3:56] <leinir> i'm not so keen on the design they've decided to go with, and the technology underlying it, but since there's people working on getting Mer on there, i'm quite happy :)
[3:56] <leinir> It runs OpenEmbedded :)
[3:57] <elysion> oh, ok, thats nice :)
[3:57] <elysion> guess there should be a few guys here and there that could help with the software issues
[3:57] <leinir> yeah :)
[3:57] <koen> leinir: you mean "it runs angstrom, built with openembedded", right?
[3:58] <leinir> koen: sorry, yes :)
[3:58] <leinir> ??ngstr??m, right? ;)
[3:58] <elysion> leinir: what do you mean by "design they've decided to go with"?
[3:58] <elysion> and that is Mer?
[3:59] <dpb> elysion: it also can run other distros, like ubuntu and gentoo
[3:59] <leinir> Mer is the free version of Maemo - they're working on getting it running on the touchbook :)
[3:59] <dpb> It already runs
[3:59] <dpb> dunno if it's usable though..
[4:00] <leinir> depends on what the progress is on the 3d driver, really.
[4:00] <leinir> ..
[4:03] <leinir> personally i hope to get plasma-netbook running on it, but that again depends entirely on the 3D driver :)
[4:24] * diroots (n=diroots@ Quit (Remote closed the connection)
[4:24] * diroots (n=diroots@ has joined #touchbook
[4:26] * shiznebit (n=shiznebi@ool-18b99096.dyn.optonline.net) Quit ("Leaving")
[4:28] * lcuk (i=lcuk@Maemo/community/contributor/lcuk) has joined #touchbook
[4:34] <elysion> leinir: plasma-netbook sounds quite heavy for a device like this
[4:34] <elysion> thought that mer sounded familiar :)
[4:34] <elysion> but that would be very nice to be able to run the same applications
[4:35] <elysion> does anyone know about the order queue?
[4:35] <elysion> how long would it take to get the device?
[4:37] <leinir> What makes you think plasma-netbook would be too heavy for a device like this? :)
[4:37] <leinir> And before you answer that, lemme grab you a video... :)
[4:38] <leinir> http://www.youtube.com/watch?v=040BbTRLKdM
[4:43] * robclark (n=robclark@nat/ti/x-e77f263f02b55035) has joined #touchbook
[4:59] <dpb> Yeah
[5:16] <elysion> leinir: wow, that's quite impressive :)
[5:16] <leinir> Nah, that's just gnome 3 ;)
[5:16] <leinir> elysion: *nods* That's what i thought too - has a lot of people going "Holy crap, it's NOT a resource hog?!" ;)
[5:16] * Macer puts the old kernel on it
[5:21] <dpb> Macer: so you get the same as me
[5:21] <dpb> (I already commented about that in the bug)
[5:22] <dpb> You really *don't* want to build the kernel on your TB.. that's insane..
[5:23] <dpb> Macer: maybe gregoires upload failed or something, it's 0.4MB smaller than the 09.c kernel too..
[5:23] * Vito89 (n=quassel@ Quit ("No Ping reply in 90 seconds.")
[5:23] <dpb> Macer: building a kernel is so much faster on a desktop machine
[5:23] * Vito89 (n=quassel@ has joined #touchbook
[5:26] <Corsac> Macer: didn't you forgot to copy the modules?
[5:29] <dpb> Corsac: yeah, it doesn't boot so far
[5:29] <dpb> Macer: the kernel can do a uImage straight if you have the right tools
[5:52] * fooq (n=reini@cm56-152-97.liwest.at) has joined #touchbook
[6:04] * Martix_ (n=Martix_@ has joined #touchbook
[6:08] <DJWillis> Macer: make uImage will do it as long as you have a u-boot-image package installed on your device (assuming your trying to make inside of Ubuntu).
[6:08] * Vito89_ (n=quassel@ has joined #touchbook
[6:09] * Vito89 (n=quassel@ Quit (Read error: 104 (Connection reset by peer))
[6:14] * Martix_ (n=Martix_@ has left #touchbook
[6:15] * Martix_ (n=Martix_@ has joined #touchbook
[6:16] <ryuo> DJWillis, i wonder... if greg is successful with the TB, will he come out with ARM hardware for desktops? lol.
[6:16] <ryuo> as a low power computer
[6:16] * Martix_ (n=Martix_@ has left #touchbook
[6:17] <DJWillis> ryuo: I can use my Beagle as a desktop, you can even get really nice cases for them ;). I guess you lack some miniIDE or mSATA connetion to get lots of storage on there without wrecking USB but it's a viable solution.
[6:17] <ryuo> heh. :P
[6:18] <ryuo> it could work..
[6:18] <ryuo> could fit it all in a small box
[6:18] <ryuo> ARM is so tiny...
[6:18] <DJWillis> ryuo: we had some debate about doing something on those lines with a Pandora vairent but it's not something any of us want to drop time as getting the Pandora done would be nice 1st ;-)
[6:18] <ryuo> theres already such contraptions for x86
[6:19] <ryuo> would be pretty cool if it had DVI/VGA out too
[6:19] <ryuo> so you could choose what external hardware you use it with :P
[6:19] <ryuo> perhaps even eSATA
[6:20] <ryuo> eSATA is a standard for external SATA if i recall
[6:21] <ryuo> DJWillis, in fact... the TB is rather cheap for a ARM mobile device.
[6:21] <ryuo> some can run like $1000
[6:21] <ryuo> i'd say its pretty good for the money you pay
[6:22] <ryuo> how many netbooks have a working touchscreen at $400? ($300 for cheaper model)
[6:22] <ryuo> lol
[6:25] <DJWillis> ryuo: A Beagleboard and one of these make for a very good under TV box ;) https://specialcomp.com/beagleboard/DSCN0403.JPG - Yep, the TB is fair value for what it is. I would love to get my hands on one someday.
[6:27] <ryuo> ah.
[6:28] <mturquette> gregoiregentil: is there a public git tree where I can find the work to push tb board support to upstream LO? I'd like to contribute to the effort.
[6:28] <ryuo> i think greg could design an innovative tiny ARM device for embedding in various computing environments that are always plugged in
[6:28] <ryuo> etc, router.
[6:29] <DJWillis> mturquette: http://git.alwaysinnovating.com/cgit.cgi/ai.openembedded.dev/ but I think gregoiregentil is working with newer patchsets off mainline or linux-omap-pm
[6:30] <mturquette> DJWillis: that is a git repo for OE.
[6:30] <mturquette> i'm looking for a Linux source tree where development is happening to rebase the tb support against upstream LO
[6:30] <mturquette> thanks anyways.
[6:32] <ryuo> mturquette, this ain't x86 :/
[6:32] <ryuo> it would be nice if you could use a vanilla kernel but..
[6:32] <DJWillis> mturquette: I was trying to say that, he is just working off a patchset in that OE tree, I don't think gregoiregentil maintains a kernel tree like we do for the Pandora.
[6:32] <ryuo> and have fully functional hardware
[6:32] <DJWillis> ryuo: nothing wrong with help to push upstream :)
[6:33] <ryuo> DJWillis, would be nice but ARM is so varied that some things are very HW specific.
[6:33] <DJWillis> ryuo: on a good day Linux-OMAP can get most things working on most OMAP3 boards, it's getting a LOT better and Tony's feed tree gets a lot of love.
[6:33] <mturquette> think i'll start a branch for it on my github repo.
[6:33] <ryuo> I'm no kernel hacker.
[6:34] <ryuo> I just write user space apps
[6:34] <ryuo> then again, I don't know hardware as intimately as Greg does.
[6:34] <DJWillis> ryuo: in a lot of ways things are in better shape then you suspect but TI really need to feed more into the mainline trees but they are getting better.
[6:35] <mturquette> DJWillis: indeed we are getting better.
[6:35] * sconemad (n=wedge@82-69-254-38.dsl.in-addr.zen.co.uk) has joined #touchbook
[6:35] <mturquette> it takes a lot of rudder to move a ship this big.
[6:35] <mturquette> but its happening.
[6:35] <ryuo> i think i'm going to end up...
[6:35] <ryuo> porting FW's userspace to ARM and just let people compile their own kernels since
[6:35] <ryuo> a stock kernel is not very effective it seems
[6:36] <mturquette> ryuo: what is FW?
[6:36] <ryuo> FW = frugalware
[6:36] <DJWillis> mturquette: :-), I did not realise your Ti, yep, a definate change of direction over the last 2 years or so (and the last 6 months or so that mainline has played a role have been good).
[6:37] <ryuo> thats why i wanted to get debian running because
[6:37] <ryuo> it has all the tools i need to port my distro to ARM
[6:38] <ryuo> the base system
[6:38] <ryuo> i'm working on a native branch because i dont feel comfortable with doing cross compiling
[6:39] <ryuo> cant get any serious work done until we fix our main machine.
[6:39] <ryuo> so anyone here willing to help test my port once i get it functional?
[6:40] <ryuo> our main machine has been having HW issues that cause it to go down...
[6:40] <ryuo> only the Bug tracker and forums are online right now
[6:41] <ryuo> i bought a TB mainly for learning purposes...
[6:41] <ryuo> so I could learn how another architecture works compared to what i'm used to wee
[6:59] * leinir (n=leinir@amarok/usability/leinir) Quit (Read error: 104 (Connection reset by peer))
[7:22] * leinir (n=leinir@amarok/usability/leinir) has joined #touchbook
[7:30] * shiznebit_TB (n=ai@touchbook.poly.edu) has joined #touchbook
[7:42] * post_j (n=poster@75-1-214-255.lightspeed.snantx.sbcglobal.net) has joined #touchbook
[7:46] * fooq (n=reini@cm56-152-97.liwest.at) Quit ()
[7:48] * setanta (n=setanta@ has joined #touchbook
[7:53] * shiznebit_TB (n=ai@touchbook.poly.edu) Quit (Read error: 110 (Connection timed out))
[8:01] * srobertson (n=steven@2001:470:8:3ec:0:0:0:1) Quit (Read error: 110 (Connection timed out))
[8:04] * lcuk2 (i=lcuk@cpc3-oldh7-0-0-cust590.manc.cable.ntl.com) has joined #touchbook
[8:10] * lcuk (i=lcuk@Maemo/community/contributor/lcuk) Quit (Read error: 60 (Operation timed out))
[8:11] * lcuk2 is now known as lcuk
[8:14] [Vito89_ PING]
[8:20] * MMlosh (n=MMlosh@2001:470:1f0b:b78:915f:1d10:15ee:4c70) has joined #touchbook
[8:27] * dirk2 (n=dirk@p5B040C44.dip0.t-ipconnect.de) has joined #touchbook
[8:35] * alexandre (n=alexandr@c-76-21-41-103.hsd1.ca.comcast.net) has joined #touchbook
[8:44] <jvs> hey alexandre
[8:49] <alexandre> hello
[8:49] * Meizirkki (n=Meizirkk@bbwirelessgw2-feeedc00-64.dhcp.inet.fi) has joined #touchbook
[9:09] <Corsac> hey there
[9:11] * pwie42 (n=paul@cm56-156-107.liwest.at) has joined #touchbook
[9:27] <alexandre> hello
[9:42] * Geep (n=JoeBelow@bas21-toronto12-1242364376.dsl.bell.ca) has joined #touchbook
[9:58] * dirk2 (n=dirk@p5B040C44.dip0.t-ipconnect.de) Quit (Remote closed the connection)
[10:06] * Vito89_ (n=quassel@ Quit (Read error: 110 (Connection timed out))
[10:09] * dirk2 (n=dirk@p5B040C44.dip0.t-ipconnect.de) has joined #touchbook
[10:16] * snlemons (n=snlemons@s9-4348.unh.edu) has joined #touchbook
[10:16] * gregoiregentil (n=gregoire@c-98-210-206-210.hsd1.ca.comcast.net) Quit (Read error: 110 (Connection timed out))
[10:35] * snlemons (n=snlemons@s9-4348.unh.edu) Quit (Read error: 60 (Operation timed out))
[10:42] * snlemons (n=snlemons@s9-4348.unh.edu) has joined #touchbook
[11:17] <DJWillis> koen: found that half finished Tracker 0.6.95 recipe, it's a bit of a mess as Tracker hates being cross compiled ;-). I'll try and munge it up after lunch ad chuck i up for you to look at.
[11:17] * snlemons (n=snlemons@s9-4348.unh.edu) Quit ("Leaving")
[11:30] * Vito89 (n=quassel@gw.loccal.net) has joined #touchbook
[11:36] <koen> DJWillis: yeah, it checks for /usr/bin/tracker :(
[11:36] <koen> scratchbox using idiots
[11:38] <DJWillis> koen: that and the stupid unpatched 'try and complie this test' that just errors if it finds a cross compile, mental.
[11:39] <DJWillis> koen: think I have most sorted now, just adding in patches from Debian as most look approperate.
[11:42] * koen is neck deep in sgx driver testing
[11:42] <DJWillis> koen: make it a good release with nice working X and framebuffer ;-)
[11:43] * Meizirkki_ (n=Meizirkk@bbwirelessgw2-feeedc00-64.dhcp.inet.fi) has joined #touchbook
[11:43] <koen> my us trip and 2 weeks of prevented a thorough test, so I'm going to try getting a decent makefile and initscript in
[11:43] <koen> and stop it from overwriting /lib/libstc++
[11:44] <DJWillis> koen: cool :), so back at home for a bit?
[11:44] <DJWillis> koen: are you over for OEDEM?
[11:44] * Meizirkki (n=Meizirkk@bbwirelessgw2-feeedc00-64.dhcp.inet.fi) Quit (Read error: 60 (Operation timed out))
[11:44] <koen> I could just compile the DDK myself and not use the binary SDK, but I wouldn't be allowed to distribute the result
[11:44] <koen> DJWillis: nope, I arrive in the UK the 9th, after OEDEM
[11:45] <DJWillis> ahhh
[11:45] <koen> allows me to go to my mums birthday on the 6th :)
[11:46] <DJWillis> koen: well that seems a very sensible thing to do.
[11:47] * diroots (n=diroots@ Quit (Remote closed the connection)
[11:47] * diroots (n=diroots@ has joined #touchbook
[11:48] <koen> DJWillis: stock beta sdk requires 2.6.31, btw
[11:48] <koen> DJWillis: it has patches to fixup .29
[11:49] <koen> DJWillis: dunno if they work on .27, so pandora might need to update its kernel
[11:50] <DJWillis> koen: the .27 is because that is what we validated against, we have .29 running well and .31 but .27 got most of the testing, I am not sure shipping the newer kernel would be that clever, hmmm, something to ponder ;-)
[11:50] <DJWillis> koen: tracker is building and packing up, I can't test right now but I can push up the recipe if you want to try.
[11:50] <koen> no hurry, I disabled it in nautilus for now )
[11:51] <DJWillis> Hmmm, I tell a lie, do_package failed on the QA, damm, this is one crappy build setup.
[11:53] * Meizirkki_ (n=Meizirkk@bbwirelessgw2-feeedc00-64.dhcp.inet.fi) Quit (Remote closed the connection)
[11:58] <koen> root@beagleboard:/usr/demos/OGLES/ChameleonMan# cat /proc/pvr/version
[11:58] <koen> Version (release) /OE/angstrom-dev/work/beagleboard-angstrom-linux-gnueabi/omap3-sgx-modules-
[11:58] <koen> System Version String: SGX revision = 1.2.1
[11:58] <koen> there we go :)
[12:00] <koen> poor chameleon man, he just keeps running
[12:01] * Vito89 (n=quassel@gw.loccal.net) Quit (Read error: 60 (Operation timed out))
[12:03] <DJWillis> Very nice, yep, not the mention the poor guy in the hot air balloon that always goes up;)
[12:03] <koen> :)
[12:18] * Vito89 (n=quassel@gw.loccal.net) has joined #touchbook
[12:39] * MMlosh (n=MMlosh@2001:470:1f0b:b78:915f:1d10:15ee:4c70) Quit ("Bye...")
[12:45] * srobertson (n=steven@2001:470:8:3ec:0:0:0:1) has joined #touchbook
[12:46] * dirk2 (n=dirk@p5B040C44.dip0.t-ipconnect.de) Quit (Remote closed the connection)
[12:48] * srobertson (n=steven@2001:470:8:3ec:0:0:0:1) Quit (Remote closed the connection)
[12:48] * srobertson (n=steven@2001:470:8:3ec:0:0:0:1) has joined #touchbook
[12:58] * pwie42 (n=paul@cm56-156-107.liwest.at) Quit (Read error: 60 (Operation timed out))
[13:09] <DJWillis> koen: http://git.openpandora.org/cgi-bin/gitweb.cgi?p=openembedded.git;a=commit;h=ac4aa2b84639c215fd295e48d6d723be4f7517f1
[13:11] * koen takes a look
[13:13] <DJWillis> koen: don't really think the xsession file is needed as there is an autostart .desktop file in the upstream package now. But it's doing no harm ;)
[13:13] <koen> right
[13:37] * pwie42 (n=paul@cm56-156-107.liwest.at) has joined #touchbook
[13:57] * Vito89 (n=quassel@gw.loccal.net) Quit (Remote closed the connection)
[14:07] * azaghal (n=azaghal@ Quit (Read error: 113 (No route to host))
[14:15] * Vito89 (n=quassel@gw.loccal.net) has joined #touchbook
[14:30] * srobertson (n=steven@2001:470:8:3ec:0:0:0:1) Quit ("Leaving")
[15:00] * tommd (n=Thomas_D@dhcp-223-19.seas.pdx.edu) has joined #touchbook
[15:06] * srobertson (n=steven@2001:470:8:3ec:0:0:0:1) has joined #touchbook
[15:06] * srobertson (n=steven@2001:470:8:3ec:0:0:0:1) Quit (Remote closed the connection)
[15:07] * srobertson (n=steven@2001:470:8:3ec:0:0:0:1) has joined #touchbook
[15:08] * srobertson (n=steven@2001:470:8:3ec:0:0:0:1) Quit (Remote closed the connection)
[15:09] * srobertson (n=steven@2001:470:8:3ec:0:0:0:1) has joined #touchbook
[15:09] * srobertson (n=steven@2001:470:8:3ec:0:0:0:1) Quit (Client Quit)
[15:10] * srobertson (n=steven@2001:470:8:3ec:0:0:0:1) has joined #touchbook
[15:10] * srobertson (n=steven@2001:470:8:3ec:0:0:0:1) Quit (Client Quit)
[15:11] * shiznebit (n=shiznebi@ool-18b99096.dyn.optonline.net) has joined #touchbook
[15:16] * DJWillis (i=djwillis@82-46-19-72.cable.ubr02.bath.blueyonder.co.uk) Quit (Client Quit)
[15:17] * alongst (n=alon@60-242-15-156.static.tpgi.com.au) has joined #touchbook
[15:28] * snlemons (n=snlemons@c-24-60-77-165.hsd1.nh.comcast.net) has joined #touchbook
[15:34] * ryuo (n=ryuo@frugalware/developer/ryuo) Quit (Remote closed the connection)
[15:35] * ryuo (n=ryuo@frugalware/developer/ryuo) has joined #touchbook
[15:42] * gregoiregentil (n=gregoire@adsl-99-138-86-218.dsl.pltn13.sbcglobal.net) has joined #touchbook
[15:45] <ryuo> gregoiregentil, evening!
[15:45] <gregoiregentil> Hello
[15:45] * gregoiregentil (n=gregoire@adsl-99-138-86-218.dsl.pltn13.sbcglobal.net) has left #touchbook
[15:45] <ryuo> .... LOL
[15:45] * gregoiregentil (n=gregoire@adsl-99-138-86-218.dsl.pltn13.sbcglobal.net) has joined #touchbook
[15:45] <shiznebit> evening gregoiregentil
[15:45] <ryuo> O_o
[15:51] <hyc> ryuo: it was obviously your tone of voice
[15:51] <ryuo> hyc, lol.
[15:53] * setanta (n=setanta@ Quit (Read error: 60 (Operation timed out))
[15:54] * srobertson (n=steven@2001:470:8:3ec:0:0:0:1) has joined #touchbook
[15:57] <hyc> someone really needs to permaban klmn361 on the forum and delete their posts...
[15:58] * gregoiregentil1 (n=gregoire@ has joined #touchbook
[16:04] <ryuo> hyc, that spammer?
[16:04] <hyc> yep
[16:09] * gregoiregentil (n=gregoire@adsl-99-138-86-218.dsl.pltn13.sbcglobal.net) Quit (Read error: 110 (Connection timed out))
[16:22] * robclark (n=robclark@nat/ti/x-e77f263f02b55035) Quit ()
[16:22] * alongst (n=alon@60-242-15-156.static.tpgi.com.au) Quit (Read error: 113 (No route to host))
[16:26] <tommd> Macer: Yes
[16:26] * Macer installs git
[16:26] <tommd> They have a git repo - see the support page most recent announcement
[16:27] <tommd> macer: http://www.alwaysinnovating.com/wiki/index.php/AI_OS_compilation
[16:27] <tommd> Be warned: THIS WILL NOT BUILD!
[16:27] <tommd> You need special (closed source) TI packages for the whole image to build
[16:27] <gregoiregentil1> it should work for the kernel
[16:27] <tommd> I have yet to be able to find those (though I am registered on TIs site)
[16:27] <gregoiregentil1> but for the kernel, it should work
[16:28] <tommd> Yes, and as I was also about to say - the kernel part should be fine.
[16:28] <tommd> Macer: My Core2 Duo 2.5GHz (Fedora) ran for hours on the 8800+ packages needed for the entire Image. Hopefully you are talking about building just the Kernel.
[16:29] <tommd> I couldn't imagine doing too much more on the TB directly.
[16:30] <tommd> So do I! I also try to use tools for what they are best at.
[16:30] <tommd> TB: Once hardened and provided with the right software would make a great tactical companion for the soldier on the go!
[16:30] <tommd> ;-)
[16:31] <tommd> Macer: That is exactly the type of crap I'd like to stop from being needed.
[16:31] <tommd> Macer: Nothing like being in a warzone and cursing some program that keeps crashing.
[16:31] * Vito89_ (n=quassel@gw.loccal.net) has joined #touchbook
[16:31] <tommd> ah
[16:32] * Vito89 (n=quassel@gw.loccal.net) Quit (Read error: 113 (No route to host))
[16:32] <tommd> I feel for you.
[16:32] <tommd> Nope, I've always been civ.
[16:35] <hyc> It took several hours for me to compile the kernel on the TB
[16:35] <hyc> I left it overnight
[16:35] <hyc> didn't time it exactly
[16:36] <tommd> hyc: And that is less than 1/100th of the code for ai-image... what do you think for the entire image 3 hours * 100 = 300 hours ~= 2 weeks?
[16:36] <tommd> Being nice.
[16:36] <hyc> yes, it worked
[16:37] <hyc> but right, I wouldn't consider building the entire distro that way
[16:37] <hyc> it already has ntpclient
[16:37] <hyc> oh, that has ntpdate
[16:38] <hyc> my ntpd startup script already does exactly that
[16:38] <hyc> you building 2.6.29 still, or newer?
[16:40] * tommd (n=Thomas_D@dhcp-223-19.seas.pdx.edu) has left #touchbook
[16:40] <hyc> AI is still 2.6.29
[16:40] <hyc> the PM stuff supposedly works, but I didn't get any improvement from it
[16:40] <hyc> well, newer versions are important for all kinds of other bug fixes
[16:41] <hyc> very likely, yes
[16:41] <shiznebit> Macer, yeap
[16:41] <shiznebit> yeap
[16:43] <hyc> my main install is a 16G
[16:43] <hyc> I have a pair of 4Gs for test images
[16:44] <hyc> not sure what the point is with ubuntu
[16:44] <hyc> they don't offer any more or better packages
[16:45] <hyc> and an actual ARMv7 binary is still better
[16:45] * pwie42 (n=paul@cm56-156-107.liwest.at) Quit (Read error: 60 (Operation timed out))
[16:45] <hyc> hm? that's just a matter of installing the driver
[16:45] <hyc> it's in the kernel build tree...
[16:45] <hyc> and you can create as many users as you want
[16:46] <hyc> I don't understand why people think particular distros are so important
[16:46] <hyc> when all the underlying source and packages are all the same
[16:46] * snlemons (n=snlemons@c-24-60-77-165.hsd1.nh.comcast.net) Quit ("Leaving")
[16:46] <hyc> and you can just alt-F1 to the TTY and login with any user you want
[16:47] <hyc> android is pretty far from useful, IMO
[16:47] <hyc> good for what?
[16:47] <hyc> but it's not lightweight at all
[16:48] <hyc> it's all heavyweight interpreted java crap
[16:48] <hyc> I seriously doubt that.
[16:49] <hyc> yeah, google has deep pockets
[16:49] <hyc> still thinking about whether it makes sense philosophically
[16:50] <hyc> I guess in a world where you can't trust any binaries you download, confining them to a sandbox makes sense
[16:50] <hyc> but I thought one of the reasons they were running the Market was to give you a safe place to get trustworthy binaries
[16:51] <hyc> and so many of the packages I see in the market wind up doing low level things anyway, so how effective is the sandboxing?
[16:51] <hyc> and right, the whole "universal" notion goes out the door because dalvik doesn't provide any of the normal java runtime
[16:52] <hyc> eh. not on a G1
[16:52] <hyc> when you have 2-3 background services active it slows to a crawl
[16:53] <hyc> and you have no control over the background services.
[16:53] <hyc> I kill them and they get respawned right away
[16:53] * tommd (n=Thomas_D@host-250-169.pubnet.pdx.edu) has joined #touchbook
[16:54] <hyc> I think that's the most unforgivable problem I have with it at the moment
[16:54] <shiznebit> Macer, are you using cyanogens rom's
[16:54] <hyc> their service manager isn't documented
[16:54] <hyc> and there are no tools to config/deactivate individual ones
[16:54] <shiznebit> Macer, have you been able to get USB tethering to work on the TouchBook
[16:55] <ryuo> Macer, flashlight app? haha.
[16:55] <hyc> I still have JF 1.5; I frequently find my battery drained because it keeps trying to download an update from google
[16:55] <ryuo> "New flashlight 2.0! Now with brighter illumination!"
[16:55] <ryuo> LOL
[16:55] <shiznebit> yeah but usb tethering = awesome
[16:55] <hyc> I have usb tethering working with my G1
[16:56] <shiznebit> hyc how did you do it
[16:56] <hyc> had to patch adb and recompile it
[16:56] <ryuo> seriously, how can you improve a -flashlight-?
[16:56] <ryuo> its only real use is to provide light
[16:56] <hyc> shiznebit: http://forum.xda-developers.com/showpost.php?p=4537323&postcount=14
[16:57] <ryuo> Macer, Oh yea? My flashlight can use any RGB 24 bit color! HA!
[16:58] <hyc> shiznebit: seems like the USB driver is buggy, it tends to disconnect for no obvious reason
[16:58] <ryuo> it still lacks that searing edge :P
[16:58] <shiznebit> hyc, yeah i noticed that too
[16:58] <hyc> also happens more often with 3G than with Edge
[16:59] <hyc> I was beginning to suspect there's something on the phone detecting network usage and killing the pppd when a quota is hit
[16:59] <ryuo> what does windows xp have to do with ARM?
[16:59] <hyc> but I haven't found anything like that
[16:59] <ryuo> Oh.
[16:59] <ryuo> Sync.
[17:00] <hyc> if you guys want my adb compiled for TB I'll post it somewhere
[17:00] <shiznebit> hyc, it be nice if the next aios update included support for the G1's usb tethering
[17:00] <ryuo> Macer, G1 reassigned to paperweight duty!
[17:01] <hyc> heh. I havent had to restore my backup yet, luckily
[17:01] <hyc> you and your damn highrises. :P
[17:02] <ryuo> how big is the whole ipkg/opkg?
[17:02] <ryuo> including everything it installs
[17:03] <hyc> we probably would have had this in an AI release already, but as luck would have it, it didn't work when I was trying to demo it for gregoire
[17:03] <ryuo> hm
[17:03] <hyc> I later discovered that adb will fail to start if the loopback interface isn't up
[17:03] <ryuo> our main package management library is 200k compiled on x86 and the front end is about 65k
[17:03] <hyc> and on the TB, I frequently see it boot up with lo missing
[17:03] <hyc> still dunno wtf that's all about
[17:04] <hyc> no idea. seems like whatever part of the boot process skips turning it on sometimes
[17:04] <shiznebit> yeah it just doesn't start for somereason
[17:04] <ryuo> hyc, i've noticed people use linked records/lists a lot in C. is it because they can allocate as many blocks as necessary "on the fly"?
[17:05] <ryuo> unlike a resized array
[17:05] <hyc> ryuo: in general, yes, linked lists are used because they can grow dynamically
[17:05] <hyc> I take it you weren't a computer science student, or never took a data structures class...
[17:05] <ryuo> i haven't had the chance yet
[17:06] <ryuo> it would be very useful, so i can understand why they do things this way
[17:06] <ryuo> or that way
[17:06] <ryuo> i can understand how it works but i dont understand the rationale
[17:06] <hyc> indeed. good of you to even recognize and ask the question.
[17:06] * Macer would love to play marvel vs capcom
[17:06] <hyc> most programmers these days don't even ask
[17:06] <ryuo> yea. i noticed it was faster than using a pointer array
[17:07] <ryuo> and reallocating when you used up the memory
[17:07] <hyc> they just use whatever crap they're familiar with in C++ or java and assume it's good enough
[17:07] <hyc> realloc is evil
[17:07] <ryuo> i do like OOP though..
[17:07] <hyc> it's one of the slower functions...
[17:07] <ryuo> i'm used to working with objects
[17:07] <ryuo> but linked objects do work well too
[17:08] <ryuo> but if you have a static buffer, its probably a better choice
[17:08] <ryuo> than a dynamic one
[17:08] <hyc> linked arrays also. there are many variations on the theme, to trade off search efficiency vs growth flexibility
[17:08] <ryuo> hm
[17:09] <ryuo> Macer, new nickname for your computer: Sir blinksalot!
[17:09] <hyc> well at least we know there's a bug in the DSS2 driver
[17:09] <hyc> gives us a starting point to look for a fix
[17:09] <shiznebit> DSS2 ?
[17:09] <ryuo> shiznebit, some OMAP terminology.
[17:10] <hyc> DSS=display subsystem
[17:10] <hyc> DSS2 is a rewrite of the driver
[17:10] <shiznebit> got it
[17:10] <ryuo> hyc, about the only place i use realloc in my code is for buffers i have to reallocate to use effectively. in my case, C char arrays.
[17:11] <ryuo> still, how does realloc know how large the original buffer is?
[17:11] <hyc> ryuo: sure. we used to have a lot of those calls in my code base. (OpenLDAP)
[17:11] <hyc> we profiled and found that 50% of our runtime was soaked up in libc alloc and str calls
[17:11] <hyc> a lot of that waste was in realloc
[17:11] <hyc> ryuo: the malloc library always records a bunch of bookeeping info for every block it allocates
[17:11] <hyc> including the start and size, at least
[17:11] <ryuo> Oh.
[17:12] <shiznebit> hyc, what can i do to make the watch hulu script work again
[17:12] <hyc> it has to, otherwise it can't free cleanly...
[17:12] <ryuo> I wondered how it knew because
[17:12] <hyc> shiznebit: I posted a working script in the forum
[17:12] <ryuo> it only asks you for the new requested size
[17:12] <ryuo> to help prevent bugs, i wrote my own macro expansion for my custom malloc
[17:13] <ryuo> #define xnew(T,N)
[17:13] <ryuo> (T*) xmalloc(sizeof(T) * (N))
[17:13] <ryuo> helps me to ensure i allocate the right amount
[17:13] <ryuo> the cast triggers a warning if its the wrong pointer type
[17:13] <ryuo> so that helps me too
[17:13] * robclark (n=robclark@nat/ti/x-a0c4c1e547dd8f34) has joined #touchbook
[17:14] <ryuo> i feel you ignore compiler warnings at your own peril
[17:14] <hyc> ryuo: sounds a bit paranoid, but whatever helps you...
[17:14] <hyc> We tend to dislike magic macros, and write everything out explicitly
[17:15] <ryuo> hyc, I see.
[17:15] <ryuo> I liked using macros for consistency.
[17:15] <shiznebit> hahah
[17:15] <hyc> and in general, I agree about warnings. most of our code is -Wall clean
[17:15] <shiznebit> the touchbook is an eye catcher though
[17:15] * diroots (n=diroots@ Quit (Remote closed the connection)
[17:15] * diroots (n=diroots@ has joined #touchbook
[17:16] <shiznebit> people can't help but look at what your typing
[17:16] <hyc> some stuff we still ignore, but they're special cases...
[17:16] <ryuo> yea..
[17:16] <shiznebit> Macer, i enjoy when the macbooks die out and i keep typing
[17:16] <ryuo> warnings usually mean somethings wrong, but it can be wrong
[17:16] <hyc> sometimes the compiler is trying to be too smart for its own good
[17:17] <ryuo> I'm just learning C, so..
[17:17] <ryuo> I don't have a lot of experience with programming this low level
[17:17] <ryuo> i'm used to my language handling a lot of low level details
[17:17] <hyc> I recall one warning that really pissed me off, gcc complained of mismatched types to a printf-style function
[17:17] <shiznebit> Macer, i get six hours
[17:17] <shiznebit> i think thats very good
[17:17] <hyc> but of course, it was our function, not libc's, so gcc was assuming the types meant something that they didn't
[17:18] <ryuo> hyc, you guys use any C99 features?
[17:18] <hyc> nope
[17:18] <ryuo> Oh.
[17:18] <hyc> there's some debate about whether we should by now
[17:18] <ryuo> even GCC only has partial C99
[17:18] <hyc> but afaik we can still compile cleanly on K&R C
[17:19] <hyc> and I see no reason to break that
[17:19] <ryuo> I'm considering trying objective-C...
[17:19] <ryuo> it doesn't add as much complexity as C++ does.
[17:20] <hyc> haven't looked at it. pretty much anything supported by Steve Jobs is automatically a bad idea.
[17:20] <shiznebit> Macer, it doesn't idle
[17:20] <shiznebit> thats all
[17:20] <ryuo> Oh?
[17:20] <shiznebit> if it would idle the thing would work really well
[17:20] <ryuo> people say the same thing about C# i bet :P
[17:21] <hyc> probably. which is why I stick to C.
[17:22] <shiznebit> blame x
[17:22] <shiznebit> its always using 3%
[17:22] <ryuo> X: fine, enjoy your TTY.
[17:22] <ryuo> "
[17:22] <ryuo> :P
[17:23] <hyc> sigh. still haven't had time to get MGR working
[17:23] <hyc> you can have a graphical environment without all the bloat of X...
[17:23] <shiznebit> Merry-Go Round
[17:23] <hyc> Bellcore Window ManaGeR
[17:23] <ryuo> i've noticed that C programs tend to be slow to develop but once they are mature their very good
[17:24] <hyc> Would be interesting to port Atari GEM over too
[17:24] <hyc> or maybe AmigaOS
[17:25] <shiznebit> Macer, im sure if a big company bought out AI and got more developers working on the code, we'd see impressive results
[17:25] <hyc> either one of those would be way more efficient than X
[17:25] <ryuo> Macer, put it in sleep mode. then it'll last indefinitely! :P
[17:26] <shiznebit> ryuo, your right, until you want to wake it up
[17:26] <hyc> MAcer: but it *has* power management
[17:26] <hyc> right now we haven't got any
[17:26] <shiznebit> then your in a world of pain
[17:26] <ryuo> shiznebit, haha.
[17:26] <ryuo> Hey, TB. Wake up.
[17:26] <shiznebit> well the usb's anyway
[17:26] <ryuo> TB: -snores- Go away.
[17:27] <ryuo> I didn't know what that was at first.
[17:27] <shiznebit> ryuo, try it out, echo -n mem /sys/power/state
[17:27] <ryuo> :P
[17:27] <ryuo> Macer, probably several gigs.
[17:28] <ryuo> if you get all of it
[17:28] <ryuo> i mainly use C to write X11 programs.
[17:31] <ryuo> hyc, i've noticed its generally safer to dereference a pointer like this (*ptr), would you agree?
[17:32] <hyc> ryuo: I can't think of a general case, no
[17:32] <ryuo> i learned that when i tried to access a pointer to a pointer that was an array
[17:32] <ryuo> *ptr[x] was causing segfaults
[17:32] <hyc> well, you just have to remember the precedence rules
[17:33] <ryuo> hm
[17:33] * shiznebit is not a programmer
[17:33] <hyc> in your example, you pretty much have no choice
[17:33] <ryuo> ah.
[17:33] <hyc> in general, I avoid parens unless they're required
[17:34] <ryuo> i never really figured out why i would want to use di and tri graphs.
[17:34] <ryuo> with digraphs, this is valid...
[17:34] <ryuo> %:include <stdlib.h>
[17:34] <hyc> a lot of folks argued that they should be deleted from the spec
[17:35] <ryuo> maybe, but might as well leave them
[17:35] <ryuo> they dont harm anything
[17:35] <hyc> they're a holdover from days when keyboards only produced a subset of ASCII
[17:35] <ryuo> Oh, i see.
[17:36] <ryuo> if i recall, void is like an incomplete struct definition?
[17:36] <ryuo> can only define pointers of incomplete types
[17:36] <hyc> void is a non-type
[17:37] <ryuo> typedef struct _Image Image;
[17:37] <ryuo> C compiler would only let me make a pointer variable out of this
[17:37] <ryuo> similar to void in that regard..
[17:37] <ryuo> but in this case its typed..
[17:38] <ryuo> excuse my misunderstanding then
[17:38] <ryuo> I read an ANSI C book before, but I'm still learning.
[17:39] <hyc> any type of pointer can be assigned to a void * (i.e., typechecking is pretty much turned off)
[17:39] <hyc> and a void * can be assigned to any type of pointer
[17:39] <ryuo> yea.
[17:39] <ryuo> its kind of like a union, if you use that to pass various types of data.
[17:39] <hyc> and, since it has no type, it also has no size.
[17:40] <ryuo> and partially defined structs, unions, etc are in the same boat.
[17:40] <hyc> you can't add or subtract offsets to a void* pointer
[17:40] <hyc> pretty much
[17:40] <ryuo> I use those for type safety when i want to obscure my internal components
[17:40] <hyc> right
[17:41] <hyc> standard technique
[17:41] <ryuo> its like private/protected/public in C++ classes
[17:41] <ryuo> and other OOP languages
[17:41] <hyc> personally I find it to be a dumb practice, from a performance standpoint
[17:42] <ryuo> it does hide implementation details that the user need not know
[17:42] <hyc> yes
[17:42] <hyc> but layering like that tends to hide details from programmers too, making it harder to track down wtf is really happening when you're reading source or inside a debugger
[17:43] <ryuo> Hm... I see your point.
[17:43] <hyc> it only works well when you're already a skilled and disciplined programmer
[17:43] <hyc> and if you are, then it's a crutch you don't actually need
[17:43] <ryuo> if you author a library, it might still be useful
[17:43] <hyc> if you're in the middle of writing some code that's only half working, it's a PITA
[17:44] <ryuo> hyc, expressions are evaluated the same as in C++?
[17:44] <hyc> mmmm. should be
[17:44] <ryuo> a "true" or "false" value?
[17:45] <hyc> 0=false, everything else=true
[17:45] <ryuo> need operators if you want it more specific
[17:45] <ryuo> i < 0
[17:45] <ryuo> more specific than just i
[17:46] <ryuo> do you use any of POSIX's standard functions?
[17:46] <hyc> probably
[17:46] <hyc> hard to get by without many of them
[17:46] <ryuo> I'm kind of divided on whether to use FILE *'s or file descriptors
[17:47] <hyc> common question...
[17:47] <ryuo> FILE *s seem to be higher level.
[17:47] <hyc> yes
[17:47] <hyc> and they're more standardized
[17:48] <hyc> main difference is that FILE *s are usually buffered
[17:48] <ryuo> any real reason to use wchar_t if you use utf8/ascii C char strings?
[17:48] <hyc> if you want to do a lot of incremental calls to generate output, FILE * is better
[17:48] <hyc> if you can do all your I/Os as a single record in a single call, fd's are more efficient
[17:49] <hyc> I pretty much never use wchar_t
[17:49] <ryuo> Hm.
[17:49] <ryuo> hyc, i did that once. i allocated the whole buffer in memory, made one call to fstat and read.
[17:49] <ryuo> fstat told me how many bytes i needed.
[17:49] <hyc> yes
[17:51] <ryuo> i found a use for strndup... i used it to copy an entire string from a buffer after i determined where it began and ended
[17:57] <hyc> psyco is the python compiler
[17:57] <hyc> it will all still work without it
[17:57] <hyc> just possibly slower
[18:03] <hyc> you're really going to build across the wifi? talk about making a slow job even slower... :P
[18:03] * gregoiregentil1 (n=gregoire@ Quit (Read error: 110 (Connection timed out))
[18:04] <hyc> tar ; scp ...
[18:12] <hyc> yeah, annoying
[18:12] * Macer sets up a symlink
[18:12] <hyc> I seem to recall that when I built these, it used a relative path
[18:12] <hyc> must be something special in the AI recipes
[18:13] <hyc> right, me neither
[18:14] <hyc> I build on ubuntu
[18:14] <hyc> no, on my laptop
[18:14] <hyc> on the TB I built on aios
[18:15] <hyc> mmm, have fun then
[18:16] <hyc> if you only want the kernel, then just use "bitbake virtual/kernel"
[18:16] <hyc> the kernel config file is preset, part of the recipe
[18:18] * gregoiregentil (n=gregoire@adsl-99-138-86-218.dsl.pltn13.sbcglobal.net) has joined #touchbook
[18:20] <hyc> you have swap enabled?
[18:20] <hyc> bitbake is a huge memory hog
[18:21] <hyc> oh mebbe it's not a memory problem then
[18:21] <hyc> good idea
[18:21] <hyc> ff is also a pig
[18:21] <hyc> and you really don't have a lot of resources to work with
[18:22] <ryuo> firefox : web browser or memory hog? lol
[18:22] <hyc> I would do it all from commandline, with X shut off
[18:22] <ryuo> my laptop has 2G ram
[18:22] <ryuo> still run x86, its not too bad
[18:28] * tommd (n=Thomas_D@host-250-169.pubnet.pdx.edu) Quit (Remote closed the connection)
[18:45] * Vito89_ (n=quassel@gw.loccal.net) Quit (Read error: 60 (Operation timed out))
[18:48] * snlemons (n=snlemons@c-24-60-77-165.hsd1.nh.comcast.net) has joined #touchbook
[18:50] <snlemons> is the wiki up to date about charging? does the maching have to be on to charge the tablet? if so, is that something that's likely to change in the future?
[18:51] <snlemons> *machine
[18:53] * lcuk (i=lcuk@Maemo/community/contributor/lcuk) Quit ("Leaving")
[18:58] <gregoiregentil> snlemons, theoretically, and according to TI datasheet, and according to our design, the machine could charge without being on for the top battery. Unfortunately, nobody has ever managed to make it work. So, for the moment, it's still an open question. I don't know.
[18:59] <snlemons> gregoiregentil: ah, I see. well, I have faith in you. :)
[19:00] <gregoiregentil> We are trying too but the corresponding TI part is really weird. And we are not the only company to have problem with it
[19:01] <snlemons> that's rough. I'd say I'll try to help when I get my TB, but I'm not so great with hardware. more of a user-level software guy.
[19:01] <shiznebit> corresponding TI part is really weird. <- how weird
[19:02] * tommd (n=Thomas_D@65-102-53-92.ptld.qwest.net) has joined #touchbook
[19:02] <shiznebit> rather,what makes it weird, gregoiregentil
[19:03] <gregoiregentil> that would be too long to explain and there is an nda with TI but I'm not convinced that it's the best part ever that TI has produced!
[19:05] <tommd> What part is being discussed?
[19:05] <gregoiregentil> TPS65950
[19:16] * gregoiregentil1 (n=gregoire@ has joined #touchbook
[19:32] * gregoiregentil (n=gregoire@adsl-99-138-86-218.dsl.pltn13.sbcglobal.net) Quit (Read error: 110 (Connection timed out))
[19:40] * mmurdock_ut (n=chatzill@c-98-202-154-226.hsd1.ut.comcast.net) has joined #touchbook
[19:45] * jknick (n=jknick@122x216x86x54.ap122.ftth.ucom.ne.jp) has joined #touchbook
[19:52] <drantin> non-disclosure agreement
[19:52] <drantin> specs for that part
[19:53] <drantin> obviously the best device produced by TI is their TI-85 graphing calculator
[19:54] * snlemons (n=snlemons@c-24-60-77-165.hsd1.nh.comcast.net) Quit ("Leaving")
[19:54] <drantin> it was the first graphing calculator hacked to run third-party asm
[20:00] * gregoiregentil1 (n=gregoire@ Quit (Read error: 110 (Connection timed out))
[20:07] * shiznebit (n=shiznebi@ool-18b99096.dyn.optonline.net) Quit ("Leaving")
[20:14] * mmurdock_ut (n=chatzill@c-98-202-154-226.hsd1.ut.comcast.net) Quit ("ChatZilla 0.9.85 [Firefox 3.5.4/20091016092926]")
[20:37] * robclark (n=robclark@nat/ti/x-a0c4c1e547dd8f34) Quit ()
[21:43] * swez (i=swez@sinux-eth0.sinus.cz) Quit (farmer.freenode.net irc.freenode.net)
[21:43] * Corsac (n=yap@ Quit (farmer.freenode.net irc.freenode.net)
[21:43] * mjr (i=mjr@ Quit (farmer.freenode.net irc.freenode.net)
[21:43] * honk (n=honk@ Quit (farmer.freenode.net irc.freenode.net)
[21:43] * honk (n=honk@ has joined #touchbook
[21:43] * Corsac (n=yap@ has joined #touchbook
[21:43] * mjr (i=mjr@ has joined #touchbook
[21:43] * swez (i=swez@sinux-eth0.sinus.cz) has joined #touchbook
[21:52] * Vito89 (n=quassel@gw.loccal.net) has joined #touchbook
[22:09] <hyc> you started bitbake in the wrong directory
[22:09] <hyc> so TMPDIR changed
[22:09] <hyc> and that whole dash thing is kind of a pain yah
[22:10] <hyc> unless you have it pointed at an absolute location in your config file,
[22:10] <hyc> bitbake always looks for tmp under your current working directory
[22:10] <hyc> so if you cd'd to somewhere else and then ran bitbake, it would look there
[22:11] <hyc> all in all the thing is pretty annoying
[22:12] <hyc> everything it does could be donw with gmake, a lot simpler and faster
[22:13] <hyc> that's not what it's looking for, pretty sure
[22:14] <hyc> but there should already be a setting in your local.conf
[22:30] * alexandre (n=alexandr@c-76-21-41-103.hsd1.ca.comcast.net) Quit ()
[22:36] * alexandre (n=alexandr@c-76-21-41-103.hsd1.ca.comcast.net) has joined #touchbook
[22:49] * Vito89 (n=quassel@gw.loccal.net) Quit (Read error: 113 (No route to host))
[22:59] <Corsac> Macer: just touch build/conf/sanity.conf
[23:00] <Corsac> well if you prefer touching your system config globally, sure, go ahead :)
[23:16] <Corsac> and on the touchbook?
[23:25] * diroots (n=diroots@ Quit (Remote closed the connection)
[23:25] * diroots (n=diroots@ has joined #touchbook
[23:30] * DJWillis (n=djwillis@82-46-19-72.cable.ubr02.bath.blueyonder.co.uk) has joined #touchbook
[23:51] * leinir (n=leinir@amarok/usability/leinir) Quit (Read error: 113 (No route to host))
[23:57] * leinir (n=leinir@amarok/usability/leinir) has joined #touchbook

These logs were automatically created by TouchBook-LogBot on irc.freenode.net using the Java IRC LogBot.