**** BEGIN LOGGING AT Wed Feb 22 02:59:59 2012
Feb 22 03:39:46 <multiplex> i'm learning oe angstrom, i typed bitbake systemd-image. now how can i clean/remove  systemd-image.
Feb 22 03:41:56 <multiplex> i suppose it's stupid but i dont yet understand all.
Feb 22 03:42:22 <multiplex> may be i'm stupid too
Feb 22 04:09:36 <multiplex> for now it will be rm -rf build/*
Feb 22 04:51:33 <robinswan>  I want to build angstrom-2010, but fetch ftp://invisible-island.net/ncurses/5.7/ncurses-5.7-20110108-patch.sh.bz2 failed, Can any one give me valid link of the source package?
Feb 22 07:56:41 <Russ> ot: how does one escape '[foo]' on github wiki
Feb 22 07:58:18 <aholler> by running away? I assume you have tried \[ and [[
Feb 22 08:05:26 <Russ> yup
Feb 22 08:18:04 <aholler> maybe html-letters will help
Feb 22 08:18:18 <aholler> &foo;
Feb 22 08:18:46 <aholler> &slash; or &#47;
Feb 22 08:19:08 <aholler> uups, no bracketleft -right is what you need
Feb 22 08:20:16 <av500> trying to program the bone in wiki markup? is .js not good enough now?
Feb 22 08:24:08 <aholler> the next generation scada system programed using .js will be funny ;)
Feb 22 08:29:07 <aholler> sorry boss, a web-spider just went through our system, which made some funny things with our machines
Feb 22 08:30:06 <kblin> aholler: just wait a few more years until all the kids know is Objective-J or the like
Feb 22 09:08:28 <paa> hello friends...I am using beagle board which i boot from sd/mmc card with kernel image and file system.MMC card is made bootable correctly.  I had already boot beagle board,mounted the file system and worked with it,It was working properly.  But suddenly due to some unknown reasons, Now as I setup the hardware and try to boot the beagle board it shows following statements and repeatedly shows it.Looks like u-boot is not booting properly
Feb 22 09:08:52 <paa> der 1.5.0  Beagle xm >>  This two statements are repeatedly shown by the beagle board,it doesn't boot...please help me out...
Feb 22 09:09:24 <av500> what did you change?
Feb 22 09:09:55 <paa> nothing....
Feb 22 09:10:44 <paa> i dont know...it was sudden behaviour
Feb 22 09:11:07 <paa> Looks like it gets reset repeatedly
Feb 22 09:13:17 <av500> have a 2nd card to try?
Feb 22 09:14:04 <paa> no...i tried with 2nd card too...but it doesnt work..
Feb 22 09:15:18 <av500> what is the power supply?
Feb 22 09:15:38 <paa> 5 v dc
Feb 22 09:16:42 <paa> i had already boot from same sd card..and was working properly..
Feb 22 09:16:54 <av500> pastebin the full boot log
Feb 22 09:17:55 <paa> sorry...i dont get  what u said..
Feb 22 10:41:51 <perter-jim> hi,
Feb 22 10:43:27 <perter-jim> omap3 DQS ,Why not high impedance ?
Feb 22 13:17:44 <amin_> hi
Feb 22 13:18:07 <amin_> Hello everyone
Feb 22 13:19:52 <amin_> I am a newcomer and I dont know how can I state my problem and how I being answered?!
Feb 22 13:20:23 <LetoThe2nd> amin_: in short: just ask as precisely as possible. if someone knows, he/she/it will answer.
Feb 22 13:22:29 <amin_> LettoThe2nd: thanks. I'll try..
Feb 22 13:24:56 <amin_> I ordered a beagleborad-xm and it will be in my hand in nex week. Shortly, my question is that how can I execute my codes on the board?
Feb 22 13:25:37 <av500> like on any linux computer
Feb 22 13:26:12 <LetoThe2nd> av500: exact steps provided already on beagleborad.org?
Feb 22 13:26:24 <av500> shh
Feb 22 13:27:08 <jkridner_> amin_: are you familiar with programming computers?
Feb 22 13:27:14 <amin_> av500: if I dont use linux, is there any solution? I mean If I dont use OS, how can I access to the different parts of the board in my code?
Feb 22 13:27:17 <jkridner_> gcc runs natively.
Feb 22 13:27:29 <av500> amin_: if you dont use an OS, buy an arduino
Feb 22 13:28:09 <amin_> what is arduino?
Feb 22 13:28:15 <jkridner_> av500: don't want more folks in the Linux club?
Feb 22 13:28:16 <amin_> is it a simple microcontroller?
Feb 22 13:28:22 <jkridner_> amin_: yeah...
Feb 22 13:28:23 <av500> yes
Feb 22 13:28:33 <jkridner_> geared towards people who don't know how to program yet.
Feb 22 13:28:33 <av500> the beagle is not a simple microcontroller
Feb 22 13:28:58 <jkridner_> yeah, a BeagleBoard/xM/Bone is a full computer.
Feb 22 13:29:10 <jkridner_> typically running Linux.
Feb 22 13:29:18 <jkridner_> Bone has "StarterWare" as well.
Feb 22 13:29:35 <LetoThe2nd> jkridner_: once upon a time, people also thought of z80/6502/68ks as full computers ;)
Feb 22 13:29:39 <amin_> jkridner:
Feb 22 13:29:50 <av500> LetoThe2nd: you confuse CPU with device
Feb 22 13:29:58 <jkridner_> I was on a call with a customer using StarterWare yesterday.  wow.
Feb 22 13:30:05 <LetoThe2nd> av500: right, i did. mea culpa.
Feb 22 13:30:06 <av500> jkridner_: wow
Feb 22 13:30:14 <amin_>  I could not find startware for DM3730
Feb 22 13:30:31 <jkridner_> I don't think there is starterware for the xM (DM3730).
Feb 22 13:30:38 <LetoThe2nd> shouldn't it be pronounced StartWhere?
Feb 22 13:30:40 <amin_> starterware exist for AM3359
Feb 22 13:30:49 <jkridner_> there is puppybits, but it is also a bit advanced in it out-of-box-experience.
Feb 22 13:30:55 <jkridner_> and u-boot, of course.
Feb 22 13:30:59 <av500> and mlo
Feb 22 13:31:16 <jkridner_> yeah, StarterWare exists for AM3359.
Feb 22 13:31:50 <av500> and for AM3358?
Feb 22 13:31:57 <jkridner_> LetoThe2nd: is that a comment about the people using it or the software itself?
Feb 22 13:32:35 <mdp> if I were to name something geared for people that don't know how to program it would be Scratch and then the related Mindstorms NXT-G language…not C-like Arduino sketches
Feb 22 13:33:28 <LetoThe2nd> jkridner_: neither. isn't it the thing that comes with current bones called "sdk", or is that again something different?
Feb 22 13:33:47 <jkridner_> while avoiding syntax errors is nice, textual languages add few barriers to starter learning.  you should really see how an 8 year old can take to an Arduino.
Feb 22 13:34:04 <jkridner_> it is something different.
Feb 22 13:34:14 <jkridner_> the "SDK" is Linux-based.   StarterWare is non-OS.
Feb 22 13:34:43 * av500 wonders is no-os is the same kind of fad as no-sql
Feb 22 13:34:46 <av500> if
Feb 22 13:34:51 <jkridner_> http://processors.wiki.ti.com/index.php/StarterWare
Feb 22 13:35:12 <jkridner_> it is not where i'd start.
Feb 22 13:35:18 <jkridner_> I'd rather teach people to use Linux...
Feb 22 13:35:35 <mdp> av500, I vote for no s/w
Feb 22 13:35:52 <jkridner_> so, I'd point amin_ to http://www.beagleboard.org/~arago/education/presentations/Learning_Linux_Tops_Down.pdf
Feb 22 13:36:03 <av500> mdp: you'd vote for no-bama as well :)
Feb 22 13:36:08 <mdp> lol
Feb 22 13:36:14 <koen> jkridner_: http://dominion.thruhere.net/koen/images/capeviewer.png
Feb 22 13:36:29 <av500> koen: get a job!
Feb 22 13:36:43 <jkridner_> sweet.  we can still use colors to reflect the live values of the pins too. :)
Feb 22 13:37:06 <av500> the DVI cape should have a 720x576 pixel matrix with live color values
Feb 22 13:37:29 <jkridner_> koen tried that one already.
Feb 22 13:38:19 <koen> jkridner_: https://github.com/koenkooi/bonescript/commits/capeviewer
Feb 22 13:42:45 <amin_> Dear freind, finally I dont know I should  forget no OS or there is a solution?
Feb 22 13:43:36 <av500> why not use an os?
Feb 22 13:43:46 <av500> you are free to program the omap3 on the beagle without OS
Feb 22 13:44:02 <jkridner_> amin_: on xM, I'm not aware of a simply no-OS option.  I've vote for learning Linux.
Feb 22 13:44:18 <jkridner_> There are some great embedded Linux books, including the Embedded Linux Primer.
Feb 22 13:44:18 <amin_> ok jkridner
Feb 22 13:44:47 <amin_> can I customize linux? I mean i need to my kernel is small
Feb 22 13:44:50 <jkridner_> the learning curve is steep, but the pay-off is large as well.
Feb 22 13:45:00 <LetoThe2nd> jkridner_: btw: are you coming to embedded world next week?
Feb 22 13:45:01 <jkridner_> depends on what you call small...
Feb 22 13:45:07 <av500> why small?
Feb 22 13:45:13 <jkridner_> ~2MB is certainly possible.
Feb 22 13:45:17 <av500> the smallest beagle has 128MB
Feb 22 13:45:23 <jkridner_> yeah, I'll be at Embedded World.
Feb 22 13:45:47 <av500> when?
Feb 22 13:45:59 <LetoThe2nd> jkridner_: cool, time to poke all you guys then ;)
Feb 22 13:46:30 <amin_> jkridner: I designed embedded system by powerpc, microblaze witout os
Feb 22 13:46:31 <jkridner_> I bought one of those pre-paid SIM cards....  I should have my german telephone number around here somewhere.
Feb 22 13:46:50 <amin_> but TI Processor is my first experience
Feb 22 13:47:18 <jkridner_> TI has a huge variety of processors.
Feb 22 13:47:35 <jkridner_> you might find MSP430s very suitable.
Feb 22 13:47:36 <jkridner_> just depends on what you need...
Feb 22 13:47:56 <amin_> DM3725 or DM3730 are my final points and I will not use thier graphical properties.
Feb 22 13:47:57 <jkridner_> this class of processor is great for running Linux/Android/QNX/etc.
Feb 22 13:49:21 <amin_> jkridner : do you know aboy sys/bios?
Feb 22 13:49:30 <av500> amin_: dsp stuff
Feb 22 13:49:52 <av500> amin_: if you designed no-os stuff before, then you will succeed on the omap3 too
Feb 22 13:50:06 <jkridner_> at one point, there was an attempt to host it on ARM as well, but I'm not aware of any nicely packaged solution to run it on the ARM.
Feb 22 13:51:56 <amin_> why starterware exist for AM3359 but no for DM3730?
Feb 22 13:54:23 <jkridner_> different target markets.
Feb 22 13:54:52 <mdp> everybody on DM3730 loves Linux
Feb 22 13:55:40 <amin_> I dont need Mulrimedia feature of it, I just need its speed.
Feb 22 13:56:03 <av500> then go on
Feb 22 13:59:06 <amin_> av500: what do you mean?
Feb 22 14:00:06 <jwinnebeck> koen: that capeviewer png, does this mean you have a DCAN cape?
Feb 22 14:00:56 <av500> amin_: well, what do you need answered still?
Feb 22 14:02:27 <amin_> thanks av500,
Feb 22 14:03:31 <amin_> I will back my freinds
Feb 22 14:05:40 <koen> jwinnebeck: yes
Feb 22 14:06:00 <jwinnebeck> hmm
Feb 22 14:06:18 <jwinnebeck> This is something built or designed or something that can be bought?
Feb 22 14:06:55 <koen> jwinnebeck: google for the text in the graphic :)
Feb 22 14:07:44 <jwinnebeck> hmm I didn't think there were any actual production capes yet
Feb 22 14:07:48 <jwinnebeck> just the DVI one which is coming
Feb 22 14:08:07 <ant_work> koen: may I send you another patch for the new layer? You remain the one who pulls, isn't?
Feb 22 14:08:44 <koen> yes, yes
Feb 22 14:08:57 <ant_work> great, thx
Feb 22 14:10:25 <jwinnebeck> OK so it's time for me to ask a question about an elephant in the room in regards to my beaglebone. I'm surprised to see that DHCP doesn't work out of the box (I've used several systemd images, including one I built on Monday)
Feb 22 14:10:45 <jwinnebeck> I'm still new to systemd, but there is a sysv init script for "busybox-udhcpc"
Feb 22 14:11:30 <jwinnebeck> however, this script is odd, it runs udhcpc only if /proc/cmdline contains "ip=dhcp". However, the system is booting into "runlevel 1" and in runlevel1, that script doesn't even run
Feb 22 14:12:32 <jwinnebeck> So in my uEnv.txt for uboot I put ip=dhcp and kernel does DHCP on boot, which is not that great. However as udhcpc isn't running, nothing is renewing the lease and eventually IP address is lost and bone goes offline
Feb 22 14:12:54 <jwinnebeck> I can run udhcpc manually, or manually do ifup -a and that will run it
Feb 22 14:13:09 <jwinnebeck> But is it expected that something will run it automatically?
Feb 22 14:13:36 <koen> jwinnebeck: when I connect an ethernet cable, dhcp just works on all my beaglebones
Feb 22 14:13:49 <jwinnebeck> I have the cable in on boot, but that shouldn't matter?
Feb 22 14:13:54 <koen> no
Feb 22 14:14:02 <jwinnebeck> Actually I HAVE to, because bone doesn't boot without ehternet plugged in
Feb 22 14:14:05 <tasslehoffwrk> if I get an Oops when doing 'modprobe -r smsc95xx && modprobe -r musb-hdrc && modprobe -r omap2430', is that considered a bug? If I put some sleeps in between, it works much better.
Feb 22 14:14:15 <koen> jwinnebeck: RMA your bone
Feb 22 14:14:25 <jwinnebeck> Well it's because of ip=dhcp in uEnv
Feb 22 14:14:29 <jwinnebeck> which I have now
Feb 22 14:14:40 <jwinnebeck> kernel won't start init until DHCP lease is had
Feb 22 14:14:41 <tasslehoffwrk> but it kinda feels like a module should have cleaned up properly when the modprobe call returns
Feb 22 14:14:47 <jwinnebeck> it's not a hardware defect
Feb 22 14:15:08 <jwinnebeck> But I did that because the busybox script already on the bone looked for it in proc cmdline so I thought maybe that's what I was supposed to do
Feb 22 14:15:22 <jwinnebeck> koen: are you using the systemd image or the "cloud9" one?
Feb 22 14:15:27 <koen> both
Feb 22 14:15:31 <jwinnebeck> hmm
Feb 22 14:15:45 * jkridner_ is very confused by the 0002-f_rndis... patch and the fact that in my kernel tree it seems that this code has been commented out by David Brownell since 2008.
Feb 22 14:15:56 <jwinnebeck> Let me unplug and replug the cable and see if the ifup scripts run
Feb 22 14:16:26 <mru> see if it f's anything up
Feb 22 14:16:53 <jwinnebeck> koen: does your /proc/cmdline have ip=dhcp in it?
Feb 22 14:17:11 <koen> no
Feb 22 14:17:18 <koen> I don't use the in-kernel dhcp
Feb 22 14:17:23 <jwinnebeck> Well unplugging and plugging in the cable didn't make it do anything, but the kernel did see the link state change.
Feb 22 14:17:33 <jwinnebeck> I'll try reverting my dhcp kernel thing
Feb 22 14:17:51 <jwinnebeck> If the busybox script in angstrom checks for it in one place, maybe it checks for it elsewhere that I don't know
Feb 22 14:19:25 <koen> that script isn't used
Feb 22 14:19:38 <jwinnebeck> it's not because the runlevel is 1 when I boot
Feb 22 14:19:48 <jwinnebeck> there is an S## script in a higher RC
Feb 22 14:19:54 <jwinnebeck> I forget which because I'm trying to reboot
Feb 22 14:20:05 <jwinnebeck> But apparently halt doesn't work, and neither does reboot
Feb 22 14:20:11 <jwinnebeck> It's never worked for me...
Feb 22 14:20:15 <jwinnebeck> this is weird
Feb 22 14:20:47 <jwinnebeck> I see "Stopped /mnt/boot" then nothing after that... But I am on the serial console. I should get messages to taht all the way to the end
Feb 22 14:20:53 <jwinnebeck> oh wait it actually did reboot
Feb 22 14:21:00 <jwinnebeck> it just took a very long time
Feb 22 14:22:15 <snowrichard> hi
Feb 22 14:22:43 <jwinnebeck> Well then, I turn off kernel dhcp and it still got an address, that's good
Feb 22 14:22:47 <jwinnebeck> but udhcpc isn't running
Feb 22 14:23:05 <jwinnebeck> koen: is that expected? Does systemd somehow manage the leases?
Feb 22 14:23:22 <mru> systemd (mis)manages most everything
Feb 22 14:23:37 <jwinnebeck> Yeah I'm starting to learn these things
Feb 22 14:23:53 <koen> jwinnebeck: in angstrom you can if ifupdown, connman or networkmanager to do your networking, it depends on the image which one is used
Feb 22 14:23:57 <jwinnebeck> So I took out kernel DHCP and it boots and DHCPs which is good, it is what I want, and bootup is far faster now
Feb 22 14:24:34 <jwinnebeck> Well ifupdown only runs scripts when the interfaces change, in a typical system I thought it runs a dhcp client which stays resident
Feb 22 14:24:35 * dwery loves capeviewer
Feb 22 14:24:38 <jwinnebeck> which I don't see
Feb 22 14:24:43 <dwery> jwinnebeck: http://www.towertech.it/en/products/hardware/tt3201-can-cape/ is in production
Feb 22 14:24:46 <jwinnebeck> ifup really does to DHCP which is good
Feb 22 14:25:24 <jwinnebeck> But you can't just dhcp once
Feb 22 14:28:24 <jwinnebeck> hmm my image has a bunch of connman* installed, so that must be what I am using
Feb 22 14:32:36 <jwinnebeck> ok connmand is running, so hopefully it is doing DHCP renews
Feb 22 14:34:17 <tasslehoffwrk> since you are on the topic: anyone know to prevent a dhcp renewal on ifup if an interface is configured as static in /etc/network/interfaces?
Feb 22 14:35:10 <jwinnebeck> tasslehoffwrk: well ifup just runs a series of scripts
Feb 22 14:35:21 <jwinnebeck> the script normally reads that file and skips dhcp if it's listed static
Feb 22 14:35:32 <jwinnebeck> but don't know what scripts angstrom is using
Feb 22 14:35:33 <mru> if it's statically configured, dhcp should stay the hell away from it
Feb 22 14:35:49 <jwinnebeck> Well dhcpc normally doesn't look at /etc/network/interfaces
Feb 22 14:35:55 <mru> but I guess mr poettering doesn't believe in static configuration
Feb 22 14:35:57 <jwinnebeck> the ifup script does that, if I understand it properly
Feb 22 14:38:23 <tasslehoffwrk> I tried out pump and udhcpc, but both of them "steal" my static IP by failing when trying to renew the lease on a static ip
Feb 22 14:38:57 <mru> why would you run a dhcp client if you have a static configuration?
Feb 22 14:39:50 <tasslehoffwrk> mru: I'm trying to figure out how to avoid running it by tracking down why ifup is trying to start it
Feb 22 14:41:56 <tasslehoffwrk> when I write a static interfaces file I kill the dhcp client, but on next reboot it comes back, and I haven't figured out how to prevent it
Feb 22 14:42:10 <jwinnebeck> So there exists an opkg "connman-plugin-udhcp" but not installed and also an older version though. But DHCP is working so far
Feb 22 14:42:18 <jwinnebeck> I wish there was some kind of connman status thing
Feb 22 14:42:28 <jwinnebeck> I guess I just need to wait for my lease to expire and see if I go offline
Feb 22 14:43:02 <mru> why don't you two simply swap configs? you seem to be having opposite problems
Feb 22 14:43:12 <tasslehoffwrk> :)
Feb 22 14:43:42 <jwinnebeck> Hahaha yeah
Feb 22 14:43:58 * tasslehoffwrk must go pickup the kids in kindergarden, but will read any clever replies later
Feb 22 14:44:07 <jwinnebeck> What I find interesting is that I don't have any if-up.d scripts but ifup does stuff, so it must be connman "replaces" the scripts?
Feb 22 14:44:20 <parana> hi i have beagleboard b7 and angstrom 2.6.32, so i use usb otg port which is connected to externally powered usb hub via usb mini B (pin4 and pin5 short circuited) cable. But i cant see usb devices.. even if loading g_ether or g_zero modules after boot up, i cant see. But when i plug out and plug in cable, then modules automatically been loaded, and i can see usb devices. So in this case whats the problem?
Feb 22 14:46:21 <parana> btw, i tried to adding usb otg's adresses to uEnv.txt .. It also does not affects..
Feb 22 15:23:15 <jkridner_> GSoC site seems to be down: http://google-opensource.blogspot.com/2012/02/google-summer-of-code-2012-is-on.html
Feb 22 15:32:54 <dev_> How to access gpio of beagleboard (ubuntu 10.04 installed)
Feb 22 15:33:10 <dev_> using C programming
Feb 22 15:39:43 <jsabeaudry> Anyone managed to install librxtx on the bone?
Feb 22 15:40:42 * av500 has a groundhog moment
Feb 22 15:44:42 <jwinnebeck> RXTX again...
Feb 22 15:44:55 <jwinnebeck> I'm telling you all, free yourselves from RXTX
Feb 22 15:45:07 <av500> and use RTSCTS
Feb 22 15:45:19 <jsabeaudry> jwinnebeck, How do you do serial comm from java
Feb 22 15:45:42 <jwinnebeck> InputStream is = new FileInputStream( "/dev/ttyUSB0" );
Feb 22 15:46:01 <jwinnebeck> and use stty
Feb 22 15:46:21 <jwinnebeck> Sounds like a hack, but I've used RXTX a lot, and it's certifiably shitty
Feb 22 15:46:29 <jwinnebeck> and the stty hasn't failed me
Feb 22 15:47:05 <jwinnebeck> I agree that it doesn't settle well with me...
Feb 22 15:47:23 <jwinnebeck> I'm usually very by the book
Feb 22 15:47:25 <juhsis> av500 if you have a time for reading and may be replying this, I would appreciate..
Feb 22 15:47:33 <jsabeaudry> jwinnebeck, I'll take your advice, I'll build my own stuff, it just seems like reinventing the wheel to write a lib generating events from the serial port in 2012
Feb 22 15:47:33 <juhsis> i have beagleboard b7 and angstrom 2.6.32, so i use usb otg port which is connected to externally powered usb hub via usb mini B (pin4 and pin5 short circuited) cable. But i cant see usb devices.. even if loading g_ether or g_zero modules after boot up, i cant see. But when i plug out and plug in cable, then modules automatically been loaded, and i can see usb devices. So in this case whats the problem?
Feb 22 15:47:40 <koen> jsabeaudry: poke woglinde to add rxtx to meta-jave
Feb 22 15:47:46 <koen> java*
Feb 22 15:47:52 <jwinnebeck> jsabeaudry: there is nothing to build
Feb 22 15:48:02 <jwinnebeck> It's 2-3 lines of code
Feb 22 15:48:17 <jwinnebeck> Now I'm assuming you don't need to twiddle RTS/CTS and all of that
Feb 22 15:48:35 <jsabeaudry> jwinnebeck, I don't, I just need to read the port and act on the data received
Feb 22 15:48:35 <jwinnebeck> If you do, then mabye it's worth figuring out how to get RXTX to work and compiled
Feb 22 15:48:43 <jwinnebeck> OK then you aren't reinventing anything
Feb 22 15:48:47 <jwinnebeck> RXTX gives you an InputStream
Feb 22 15:48:55 <jwinnebeck> In reality what I do is support RXTX and Linux directly
Feb 22 15:49:03 <jwinnebeck> and use an interface to obtain an InputStream
Feb 22 15:49:21 <jwinnebeck> for Linux it is just one line to run stty and one line to open the dev file
Feb 22 15:49:29 <jwinnebeck> then it's just an InputStream from there, just like RXTX
Feb 22 15:49:30 <jsabeaudry> jwinnebeck, Ah interesting, then I guess I can hack a library built on RXTX and just give it the fileinputstream
Feb 22 15:49:31 <av500> juhsis: sorry, I have been banned from commenting on OMAP3 and MUSB
Feb 22 15:49:47 <jwinnebeck> Oh you aren't writing your own code? It's a lib already using RXTX?
Feb 22 15:49:58 <av500> juhsis: and .32 is quite old
Feb 22 15:50:03 <jwinnebeck> If that's true then of course you need to weigh how hard it is to get in there
Feb 22 15:50:41 <juhsis> i know but, it is more stable and easily works for me..
Feb 22 15:50:50 <jwinnebeck> On Windows or elsewhere I just put in the implementation that gets the stream from RXTX
Feb 22 15:50:54 <av500> for your definition of "works"
Feb 22 15:51:27 <aracanescu_> what is the correct defintion of "works" ?
Feb 22 15:51:49 <aracanescu_> i mean the universal one?
Feb 22 15:51:50 <av500> whatever works for you :)
Feb 22 15:52:11 <av500> but stating a problem and saying it works at the same time I find amusing
Feb 22 15:52:13 <juhsis> i tried 3.0.14+ but, my program's fps decreases from 30 to 5
Feb 22 15:52:28 <av500> increase CONFIG_FPS in the kernel
Feb 22 15:52:37 <aracanescu_> av500: thats usually the case, there are cases where one guy said I have a problem and then said nvm it "works"
Feb 22 15:52:39 * av500 has a groundhog moment
Feb 22 15:52:54 <koen> av500: did you see your own shadow?
Feb 22 15:53:01 <jwinnebeck> jsabeaudry: http://pastebin.com/v958TxFy
Feb 22 15:53:03 <aracanescu_> the point being that only stating the problem can lead to "works" solution :
Feb 22 15:53:11 <jwinnebeck> Taht's the code I use
Feb 22 15:53:15 <juhsis> av500 i mean process time of my program
Feb 22 15:53:47 <av500> juhsis: I know, we have been there before
Feb 22 15:54:02 <juhsis> it is image processing project, it process 30 frames /sec in 2.6.32 but 5 frame/sec in 3.0.14+
Feb 22 15:54:05 <av500> but since you are unwilling to find out why your SW is slower, yes you are stuck with .32
Feb 22 15:54:10 <av500> yes, I know
Feb 22 15:54:18 <juhsis> hmm ok.
Feb 22 15:54:31 <juhsis> actually i only have this problem in 2.6.32
Feb 22 15:54:38 <jwinnebeck> jsabeaudry: so obviously this only works for "simple" things where you just need to send and receive bytes, and you don't need any wonky timeout stuff beyond what Java already gives you, and also that you don't need to twiddle the serial lines up and down. If you had to change baud rate in the middle, you could do another stty call, but then it really does start to look like a hack
Feb 22 15:54:55 <juhsis> so i wonder if it is a problem which has no solution or a known problem..
Feb 22 15:55:08 <aracanescu_> the best way for " fast" code is assembly
Feb 22 15:55:24 <aracanescu_> :)
Feb 22 15:55:27 <av500> juhsis: does it work .14?
Feb 22 15:55:30 <av500> juhsis: does it work with .14?
Feb 22 15:55:33 <jsabeaudry> jwinnebeck, I guess my needs are pretty elementary, I'll see if I can KISS and not use any lib
Feb 22 15:55:53 <jsabeaudry> jwinnebeck, Thanks for the input
Feb 22 15:56:00 <juhsis> 3.014? yes it works after setting uEnv.txt with adresses of usb otg port
Feb 22 15:56:10 <jwinnebeck> Yeah like I said I've used RXTX a lot, and I spent more pain making sure its so and jar are in the right place and in classpath, etc etc and plus the random bugs in RXTX and their shitty (polling) input stream, etc etc
Feb 22 15:56:10 <juhsis> and of course after loading g_ether module
Feb 22 15:56:12 <av500> adresses?
Feb 22 15:56:28 <juhsis> g_ether.host_addr=16:0F:15:5A:E1:21
Feb 22 15:56:28 <juhsis> g_ether.dev_addr=16:0F:15:5A:E1:20
Feb 22 15:56:42 <av500> so, it has been fixed
Feb 22 15:56:47 <av500> in later kernels
Feb 22 15:56:49 <juhsis> in 3.0.14+ yes
Feb 22 15:57:14 <av500> so you can backport stuff to .32 or you can try to run with a newer kernel
Feb 22 15:57:16 <juhsis> is it means no solution for this in 2.6.32?
Feb 22 15:57:27 <av500> of course, backport the right fix
Feb 22 15:57:40 <jwinnebeck> And to use RXTX in a closed source project is a little scary because due to the limited enumeration capabilities in it, you have to revert to importing their class to construct it manually, although only a single line it goes against the "GPL exception" in the library so it might force your code to be open source
Feb 22 15:58:14 <av500> and how many lines does it have?
Feb 22 15:58:18 <av500> (just curious)
Feb 22 15:58:25 <juhsis> which one?
Feb 22 15:58:30 <juhsis> uEnv.txt?
Feb 22 15:58:34 <av500> what?
Feb 22 15:58:49 <juhsis> you asked how many lines does it have?
Feb 22 15:59:26 <av500> juhsis: yes, but not you :)
Feb 22 15:59:35 <juhsis> aah ok, sory :)
Feb 22 16:00:07 <jwinnebeck> who did you ask?
Feb 22 16:00:23 <av500> you
Feb 22 16:00:38 <jwinnebeck> oh... um how many lines in what
Feb 22 16:01:00 <jwinnebeck> I didn't get that you were talking to me
Feb 22 16:03:39 <koen> kkeller, jkridner_: (untested) http://dominion.thruhere.net/koen/angstrom/beaglebone/0001-gpio-check-exported-gpios-first-before-trying-to-re-.patch
Feb 22 16:04:09 <jwinnebeck> av500: ?
Feb 22 16:04:36 <kkeller> koen sweet… I'll try out this afternoon
Feb 22 16:05:58 <av500> jwinnebeck: I just wanted to know how large that rxtx thing is
Feb 22 16:06:01 <koen> it doesn't fix the problem you're having, but it will give more clues on the failure more :)
Feb 22 16:06:15 <jwinnebeck> av500: oh I don't know, it's pretty large in C and Java code
Feb 22 16:06:22 <jwinnebeck> I don't know how many lines it has
Feb 22 16:06:26 <jwinnebeck> but the C code is pretty bad
Feb 22 16:06:35 <av500> ic
Feb 22 16:06:36 <kkeller> exactly… a simple console message will go a long way
Feb 22 16:06:46 <koen> av500: 36kiB compressed arm binary
Feb 22 16:07:04 <av500> koen: well, a hello world is huge these days :)
Feb 22 16:08:53 <jwinnebeck> well there is 500k of C code wow
Feb 22 16:09:01 <jwinnebeck> one of the files is 164kb and 6010 lines
Feb 22 16:09:08 <jwinnebeck> I love 6010 line C files
Feb 22 16:10:18 <jwinnebeck> the thing is that it's riddled with ifdefs because it works on windows and wince and macos and linux and unix*1000...
Feb 22 16:10:51 <jwinnebeck> hmm there is a file called "i2cimp.c" in addition to serial, parallel and rs485
Feb 22 16:11:10 <jwinnebeck> rs485 you don't see that often... but how is rs485 that different other than voltages and speeds
Feb 22 16:12:05 <av500> rs485 defines the electrical layer
Feb 22 16:12:23 <jwinnebeck> Yeah
Feb 22 16:12:31 <jwinnebeck> Well this RXTX has an "RS485Port" class
Feb 22 16:12:38 <jwinnebeck> which is different from serial port
Feb 22 16:13:10 <av500> and it offers different methods?
Feb 22 16:13:26 <jwinnebeck> it's hard to tell from the C code right away
Feb 22 16:13:57 <jwinnebeck> it almost looks like a copy/paste from the RS232
Feb 22 16:14:08 <av500> who'd have thought :)
Feb 22 16:14:09 <jwinnebeck> But the classes are not related by inheritance
Feb 22 16:14:24 <jwinnebeck> there is a new method "setRS485PortParams"
Feb 22 16:14:43 <jkridner_> koen: is the check necessary?  I have no issue with adding correctness to code, but I want to know if there were unanticipated side effects.
Feb 22 16:14:53 <jwinnebeck> But that method takes speed, data bits, stop bits, and parity
Feb 22 16:15:04 <av500> thats so unusual :)
Feb 22 16:15:07 <jwinnebeck> I know :)
Feb 22 16:15:15 <jwinnebeck> except that RS485 I thought is restricted to 9600 bps
Feb 22 16:15:21 <av500> no
Feb 22 16:15:34 <jwinnebeck> oh I'm thinking of J1708
Feb 22 16:15:51 <jwinnebeck> RS485 is just the signaling physical layer
Feb 22 16:15:54 <koen> jkridner_: kkeller problem was that the pgio was already used by the kernel, with my patch you get better error logging in the console
Feb 22 16:16:06 <jwinnebeck> but the default speed for RS485 in RXTX is 9600
Feb 22 16:16:13 <av500> "...A message is composed of up to 21 characters, unless the engine is stopped and the vehicle is not moving in which case transmitters are allowed to exceed the 21 byte max message length..."
Feb 22 16:16:19 <koen> jkridner_: since it avoids a common cause of errors: re-exporting a gpio
Feb 22 16:16:33 <jwinnebeck> and every RS485 I've seen is 9600 but my experience is limited to J1708/J1587
Feb 22 16:16:40 <mru> gpios have export restrictions?
Feb 22 16:16:41 <jkridner_> k.  thanks.
Feb 22 16:16:47 <jwinnebeck> What I find interesting is that there is a "I2CPort" class in RXTX
Feb 22 16:16:50 <av500> mru: some are not allowed in Iran
Feb 22 16:16:51 <mru> bitbanged missile guidance?
Feb 22 16:17:02 <jkridner_> is there any way to assign an application to be owner, rather than just all of userspace?
Feb 22 16:17:04 <av500> prpplague: http://www.gif.tv/#/giftv-1212-baconsizzle
Feb 22 16:17:30 <prpplague> av500: hehe blocked
Feb 22 16:17:31 <av500> jwinnebeck: DMX512 is RS485 with 250kbaud
Feb 22 16:17:48 <av500> and my RS485 parts from TI arrived today :)
Feb 22 16:17:59 <jwinnebeck> ha, well I don't know DMX512
Feb 22 16:18:02 <koen> jkridner_: kkeller and I worked out the error and then set out to improve error reporting for future users
Feb 22 16:18:04 <jwinnebeck> my world is small
Feb 22 16:18:05 <av500> prpplague: imaging bacon, sizzling
Feb 22 16:18:20 <jwinnebeck> av500: and what is up with the quote from J1708 on message length?
Feb 22 16:18:26 <koen> jkridner_: although it does mean more people will find out that I am responsible for messing up their app :)
Feb 22 16:18:31 <jwinnebeck> That's .... true but relevant to what?
Feb 22 16:18:53 <prpplague> av500: hehe
Feb 22 16:19:39 <av500> jwinnebeck: I found it amusing
Feb 22 16:20:00 <av500> like the ECUs starting to exchange gossip while the vehicle is stopped
Feb 22 16:20:17 <jwinnebeck> hah yeah well the annoying part about J1708 is that when the engine is off ther eis NO limit to the message length
Feb 22 16:20:26 <jwinnebeck> And there's no message length in the frame
Feb 22 16:20:28 <kkeller> jkridner_:  with out this, a failed export just ends up as a undefined path during the writeFileSync() in the fs.js file
Feb 22 16:20:36 <jwinnebeck> the frame ends when there's an 11ms pause or whatever the length is
Feb 22 16:20:45 <av500> jwinnebeck: ok, so I find it now amusing and stupid
Feb 22 16:20:47 <jwinnebeck> So to be per the spec you need to have a buffer of infinite size
Feb 22 16:21:16 <av500> jwinnebeck: I know people talking on the phone that expect the same from me :)
Feb 22 16:21:17 <jwinnebeck> Although in practice it's not an issue, and it's no different from something like HTTP/SMTP and such that dont' really have a message length
Feb 22 16:21:51 <SilicaGel> av500: DMX is the protocol used for controlling lights and animitronic displays and such, yes?
Feb 22 16:22:18 <av500> yes
Feb 22 16:22:26 <jwinnebeck> you just need to set a reasonable "largest message" that you ever care to expect to read and stick to it
Feb 22 16:23:06 <SilicaGel> av500: do you work in this area?  or some other reason you mentioned dmx?
Feb 22 16:23:55 <av500> no
Feb 22 16:23:59 <SilicaGel> oh ok
Feb 22 16:24:01 <av500> I plan to play with it
Feb 22 16:24:01 <SilicaGel> i was just curious
Feb 22 16:24:12 <koen> SilicaGel: it's also a famous actor/musician
Feb 22 16:24:21 <SilicaGel> ever sincE I saw that youtube thing with the halloween house that sings michael jackson's "Thriller"... I have been interested in taht stuff, slightly
Feb 22 16:24:51 <jwinnebeck> thanks for posting the gif.tv link, I've never heard of this site, and now my whole friggin day is shot
Feb 22 16:24:55 <SilicaGel> I decided I don't want to spend that much $ on it though so instead I bought about 15' of EL wire and plan to recreate that singing house in an 8.5 x 11 picture frame
Feb 22 16:24:58 <jwinnebeck> I can't stop lauging
Feb 22 16:25:20 <koen> av500: https://plus.google.com/115547683951727699051/posts/C4ALVny7j5Q
Feb 22 16:25:20 <av500> jwinnebeck: same here, but I thought I ruin somebody else'S too
Feb 22 16:25:48 <av500> koen: yep
Feb 22 16:26:26 <jwinnebeck> wow I read that whole page yesterday and I didn't see the goatse!
Feb 22 16:26:37 <jwinnebeck> wow canonical!
Feb 22 16:27:03 <av500> https://tbe.taleo.net/NA3/ats/careers/requisition.jsp?org=CANONICAL&cws=1&rid=217
Feb 22 16:27:17 <av500> they need it more than ever
Feb 22 16:27:37 <jwinnebeck> "Candidates must have a proven track record in producing stunning concepts and converting these into final artworks"
Feb 22 16:27:42 <jkridner_> kkeller, koen: thanks.  I'll either apply the given patch and test it or accept a git merge request.
Feb 22 16:27:45 <jwinnebeck> well there's not much more stunning than goatse
Feb 22 16:27:50 <av500> and final
Feb 22 16:27:56 <jwinnebeck> that's true
Feb 22 16:28:11 <av500> maybe it was his final artwork before he left :)
Feb 22 16:28:49 <koen> I bet he works for the prague public transit org nowadays
Feb 22 16:28:57 <av500> and btw, since when do people "talk" by touching each others fingertips?
Feb 22 16:29:06 <kkeller> jkridner_:  I can test this afternoon if you wish… I have several test cases already set up (Just have to go out for several errands)
Feb 22 16:29:07 <jwinnebeck> Since ET you insensitive clod
Feb 22 16:29:08 <av500> ET: yes, *you* do
Feb 22 16:29:10 <jwinnebeck> you racist
Feb 22 16:29:17 <jwinnebeck> or speciest
Feb 22 16:29:20 <jwinnebeck> whatever it is
Feb 22 16:29:27 <av500> I prefer specialist
Feb 22 16:29:56 <jwinnebeck> speaking of old stuff, I wonder if...
Feb 22 16:30:01 <jwinnebeck> hmm nope
Feb 22 16:30:16 <jwinnebeck> I used to use mirc years ago you could slap people with trouts
Feb 22 16:30:21 <jkridner_> kkeller: thanks.   I'll look for your ack before applying to master.
Feb 22 16:30:23 <jwinnebeck> apparently pidgin can't /slap
Feb 22 16:30:30 * av500 slaps jwinnebeck with a huge trout
Feb 22 16:30:35 <jwinnebeck> oh no!!! :)
Feb 22 16:30:46 <jwinnebeck> or did you cheat with a /me
Feb 22 16:31:08 <koen> it's hard to be a racist since all living humans belong to the same hominid subspecies
Feb 22 16:31:16 <mru> jwinnebeck: maybe he created an irssi alias, but does it really matter?
Feb 22 16:31:25 <av500> slapped is slapped
Feb 22 16:31:33 <jwinnebeck> Yes, mru, it does obviously. There's no slap as authentic as the original one
Feb 22 16:41:20 <SilicaGel> the original woudl ahve said "about a bit"
Feb 22 16:41:24 <SilicaGel> it slaps you about a bit with a large trout
Feb 22 16:43:03 <pkim> hello
Feb 22 16:46:18 <pkim> I have a question
Feb 22 16:46:25 <pkim> about memory
Feb 22 16:46:42 <pkim> data group DQ and DQS
Feb 22 16:48:07 <pkim> I use Agilent 9000 Test DDR, it was found that the DQ DQS no Hi-z state
Feb 22 16:48:33 <pkim> thanks
Feb 22 16:49:26 <av500> ask Gerald on the mailing list
Feb 22 16:49:28 <pkim> I want to ask what reason is this cause
Feb 22 16:52:17 <pkim> I config SDRC_DATA0 1 ......register, enable input , disable pull up and pull down
Feb 22 16:53:04 <av500> ask Gerald on the mailing list
Feb 22 16:54:05 <pkim> ok,
Feb 22 16:54:08 <mru> av500: ever considered a career as a parrot?
Feb 22 16:54:26 <av500> been there, done that
Feb 22 16:54:31 <av500> been there, done that
Feb 22 16:54:49 <pkim> :)
Feb 22 17:32:25 <Marc_> Hello, I'm looking to purchase a BeagleBoard-xM, and I was wondering if anyone here could provide some insight into the required hardware to setup the proper enviroment. Is a separate monitor/mouse/keyboard really required for testing/development?
Feb 22 17:33:23 * mranostay yawns
Feb 22 17:47:11 <mranostay> ah my mug is in a few ELC photos I see
Feb 22 17:47:26 <mranostay> at least i have proof that i went :)
Feb 22 17:48:17 * koen checks http://www.flickr.com/photos/13825348@N03/sets/72157629428790651/
Feb 22 17:49:18 <koen> I spot Tartarus in the first one
Feb 22 17:49:36 <Tartarus> heh
Feb 22 17:49:45 * prpplague looks for a good picture to add spock into
Feb 22 17:49:54 <Tartarus> only my back!
Feb 22 17:49:56 <prpplague> hehe zach
Feb 22 17:51:04 <Tartarus> heh, there's one of Russ
Feb 22 17:51:15 <koen> http://www.flickr.com/photos/13825348@N03/6773574904/in/set-72157629428790651/lightbox/
Feb 22 17:51:30 <Tartarus> was about to say, koen
Feb 22 17:51:32 <bradfa> Marc_, You don't NEED a monitor, keyboard or mouse so long as you don't want to use any of those.  I made a list (it's a bit old now) of what I got: http://www.bradfordembedded.com/2011/01/it-s-alive-the-beagleboard-xm-that-is/
Feb 22 17:51:40 <Tartarus> and mdp and mranostay
Feb 22 17:51:46 <koen> Tartarus, mdp, mranostay, me
Feb 22 17:52:14 <koen> slightly to the left, galak?
Feb 22 17:52:22 <Tartarus> yes
Feb 22 17:52:32 <Tartarus> and reaperofsouls
Feb 22 17:53:06 <mranostay> reaperofsouls?
Feb 22 17:53:11 <koen> and sakoman to the far left
Feb 22 17:53:12 <Tartarus> jeremy
Feb 22 17:53:15 <koen> mranostay: jeremy
Feb 22 17:53:55 <Tartarus> So, of the 6000 shots of beckyb, none online? conspiracy!
Feb 22 17:54:09 <mranostay> private collection :)
Feb 22 17:54:51 <koen> isn;t becky in  http://www.flickr.com/photos/13825348@N03/6773574904/in/set-72157629428790651/lightbox/ ?
Feb 22 17:54:56 <mdp> Tartarus, proof to mgmt that we were there…*whew*
Feb 22 17:55:20 <mranostay> mdp: that is what i was thinking as well :)
Feb 22 17:55:23 <Tartarus> Ah, true
Feb 22 17:55:27 <Tartarus> Just all of the closeups
Feb 22 17:55:46 <koen> I think even dsaxena is in there
Feb 22 17:55:48 <Marc_> Thanks bradfa for the great link. I'm looking to run the board off of minicom on a Linux desktop (to cut costs) and I wasn't sure if it was possible.
Feb 22 17:55:53 <mdp> mranostay, we just have to explain the amount of time spent in the "Champagne Room"
Feb 22 17:56:02 <bradfa> Marc_, Yes, definitely possible, that's how I run my xM
Feb 22 17:56:23 <Marc_> Awesome, I really appreciate bradfa, thank you for the help, I think I'll definitely purchase an xM now.
Feb 22 17:56:47 <mdp> koen, I think I see Waldo in that one now!
Feb 22 17:57:03 <bradfa> Marc_, Just be aware, I think the power supply I have listed in my blog post is no longer available.
Feb 22 17:57:16 <mranostay> mdp: charity event for wayward women? :)
Feb 22 17:57:30 <mdp> HR has no humor
Feb 22 17:57:35 <mdp> like Legal
Feb 22 17:57:43 * mranostay should dress up as Waldo for the next conference
Feb 22 17:58:19 <Marc_> Thanks bradfa, I believe we have a suitable power-source available.
Feb 22 17:58:24 <Marc_> Take care!
Feb 22 17:58:33 <mdp> we went to a AA baseball game some time back and there was a guy dressed like Waldo wandering about….kept switching seats at the field
Feb 22 17:58:56 <mdp> had the proper glasses and everything, really, his intention was to be real Waldo
Feb 22 18:00:15 <koen> mdp: luckily "champagne room" is only a loaded term in the us
Feb 22 18:00:42 <mdp> heh
Feb 22 18:01:06 <mdp> I'm easily amused too
Feb 22 18:01:10 <mranostay> koen: "luckily"? you guys are missing out :P
Feb 22 18:02:21 <koen> as long as you take your vitamines
Feb 22 19:02:37 <cogman> What is the best way to ensure that my webcam is working? I have a Logitech C110 and a dmesg shows that the driver is present and the device is being registered.
Feb 22 19:41:20 <the_student> How easy is it to write an OS for Beagle*?
Feb 22 19:41:52 <djlewis> ask, how difficult is it to write linux..  :)
Feb 22 19:41:56 <jwinnebeck> probably not harder than anything else ?
Feb 22 19:42:21 <jwinnebeck> I get the impression (I'm not super knowledgable) that ARM is a bit easier than x86
Feb 22 19:42:35 <jwinnebeck> but I've no real experience with that
Feb 22 19:42:38 <jwinnebeck> I just use Linux :)
Feb 22 19:42:43 <the_student> djlewis: The thing is that I want to write an ARM OS, and I am not aiming for a Linux competitor
Feb 22 19:42:55 <jwinnebeck> of course once you get past the architecture, a lot of the OS is common
Feb 22 19:42:56 <djlewis> with my present knowledge base, given the choice I'd pick x86 to bring up to an os
Feb 22 19:42:58 <jwinnebeck> like process management
Feb 22 19:43:27 <_av500_> what is an ARM OS?
Feb 22 19:43:46 <_av500_> opposite of LEG OS?
Feb 22 19:44:02 <the_student> _av500_: An OS running on an ARM processor, in this case the ARM Cortex-A8/A9
Feb 22 19:44:52 <the_student> djlewis: But osdev requires asm, and no sane person would consider writing x86 ASM
Feb 22 19:45:04 <_av500_> the_student: you need a few lines of asm
Feb 22 19:45:05 <SilicaGel> LegOS ... that's Legolas's brother, right?
Feb 22 19:45:36 <the_student> _av500_: But ARM asm is atleast half-sane
Feb 22 19:45:59 <_av500_> that all nice
Feb 22 19:46:02 <_av500_> thats
Feb 22 19:46:28 <_av500_> but it wont help you much to write an os
Feb 22 19:46:38 <mdp> can we ask Gerald on the list?
Feb 22 19:47:36 <the_student> _av500_: I also don't have to bother with being compatible with all those weird and legacy combinations
Feb 22 19:48:18 <_av500_> you are writing a toy os for yourself
Feb 22 19:48:25 <_av500_> you dont have any legacy
Feb 22 19:48:33 <_av500_> to care for
Feb 22 19:48:48 <the_student> I was talking about legacy hardware
Feb 22 19:48:56 <mdp> you don't want ISA support?
Feb 22 19:49:20 <the_student> Not really
Feb 22 19:49:51 <the_student> Neither do I want to have to mess around with the nasty x86 tables either
Feb 22 19:50:24 <mdp> there's nastiness everywhere
Feb 22 19:50:34 <mdp> ARM is wrong endian…that's ugly
Feb 22 19:50:44 <_av500_> lol
Feb 22 19:50:48 <the_student> ARM is Bi-endian
Feb 22 19:51:00 <_av500_> no, ISO9660 is
Feb 22 19:51:14 <mdp> the_student: put that in quotes :)
Feb 22 19:51:32 <mru> the_student: arm hasn't been bi-endian for a long time
Feb 22 19:51:42 <mdp> and I've worked on BE IXP too :)
Feb 22 19:51:49 <the_student> So it's 6502 endian?
Feb 22 19:51:57 <mru> it's little endian
Feb 22 19:52:15 <mru> there is a mode where _data_ accesses are big endian
Feb 22 19:52:26 <mru> instructions are always LE
Feb 22 19:52:39 <mru> everything is simpler if you just stick with the full LE mode
Feb 22 19:52:44 <mdp> there's the "quotes" I wanted
Feb 22 19:53:33 <mru> armv5 had full bi-endian support
Feb 22 19:53:43 <mru> this was deprecated and made optional in armv6
Feb 22 19:53:48 <mru> and it's gone in v7
Feb 22 19:55:12 <the_student> Although I am used to big-endian, little-endian seems practical for my uses
Feb 22 19:55:52 <mru> for most uses it doesn't matter at all
Feb 22 19:56:42 <mru> if data needs to be read or written in the "wrong" byte order, there are fast byte-reverse instructions for that
Feb 22 19:56:51 <mru> it's highly unlikely that this will be a bottleneck
Feb 22 19:57:23 <_av500_> some people have principles
Feb 22 19:57:35 <_av500_> and wont write on OS for the wrong endiannes
Feb 22 19:58:16 <jwinnebeck> So if ARM is LE, then that means the vast majority of processors out there are LE
Feb 22 19:58:22 <jwinnebeck> since x86 is LE
Feb 22 19:58:27 <_av500_> yep
Feb 22 19:58:28 <jwinnebeck> and sparc is sort of dead
Feb 22 19:58:38 <jwinnebeck> that's the only BE arch I know about
Feb 22 19:58:45 <jwinnebeck> except PPC... I can't remmeber
Feb 22 19:59:03 <mdp> right, and there's only a couple ppcs around
Feb 22 19:59:09 <the_student> That's great, some of the stuff I'll interface is meant for the PDP-11 series
Feb 22 19:59:13 <mdp> they aren't used anywhere
Feb 22 19:59:28 <mdp> like freebsd
Feb 22 19:59:44 <thurbad> isn't xbox 360 ppc?
Feb 22 19:59:46 <jwinnebeck> does ARM require aligned reads/writes?
Feb 22 20:00:03 <_av500_> not any more
Feb 22 20:00:16 <the_student> thurbad: Kinda
Feb 22 20:00:16 <jwinnebeck> hmm OK so is it more like x86 where misalignment is OK but at a penalty?
Feb 22 20:00:24 <jkridner_> _av500_: no?
Feb 22 20:00:30 * _av500_ would rather discuss important stuff like java uart access than endianness in 2012
Feb 22 20:00:41 <jwinnebeck> Java UART access?
Feb 22 20:00:46 <jwinnebeck> not RXTX again! :)
Feb 22 20:00:51 <mdp> _av500_, microkernel?
Feb 22 20:01:19 <mdp> risc versus cisc?
Feb 22 20:01:21 <jwinnebeck> I'm looking at Java PRU access right now
Feb 22 20:01:35 <jwinnebeck> trying to figure out the best way
Feb 22 20:01:43 <the_student> thurbad: 3.2 GHz PowerPC Tri-Core Xenon
Feb 22 20:02:19 <mdp> *ahem*…that's Power(™)
Feb 22 20:02:24 <mdp> the PC is gone
Feb 22 20:02:44 <_av500_> jwinnebeck: wire the PRU to be a uart and use RXTX :)
Feb 22 20:02:54 <jwinnebeck> oh come on
Feb 22 20:02:59 <jwinnebeck> don't make me throw up
Feb 22 20:03:00 <mru> jkridner_: arm can do unaligned accesses since v6
Feb 22 20:03:21 <_av500_> mru: arm took out all the fun out of arm
Feb 22 20:03:25 <jkridner_> but do they result in what is expected?
Feb 22 20:03:28 <jkridner_> k.
Feb 22 20:03:48 <the_student> _av500_: What do you mean?
Feb 22 20:03:55 <jkridner_> does it even do it in v4/v5 code?
Feb 22 20:04:07 <_av500_> next thing you know, arm will run windows...
Feb 22 20:04:26 <jkridner_> the_student: _av500_ means that it has become easy to program on ARM and therefore not as fun a challenge.
Feb 22 20:04:36 <the_student> mru: Isn't that required for Thumb?
Feb 22 20:04:44 <the_student> jkridner: k
Feb 22 20:04:53 <mru> the_student: no
Feb 22 20:05:34 <the_student> mru: I thought it was required since Thumb is 16-bit, not 32-bit
Feb 22 20:05:43 <mru> jkridner_: on v6 and up, unaligned accesses behave exactly as expected
Feb 22 20:05:52 <_av500_> the_student: registers are still 32bit
Feb 22 20:06:00 <mru> the_student: thumb _instructions_ are 16-bit aligned
Feb 22 20:06:16 <mru> so yes, a 32-bit thumb2 instruction can be misaligned
Feb 22 20:06:34 <mru> and that's a cause of a host of nasty bugs, both in sw and hw
Feb 22 20:06:37 <the_student> _av500_: Sure, since thumb isn't really a separate instruction set
Feb 22 20:07:02 <mru> but then again thumb2 is only in (some) v6 and later
Feb 22 20:07:07 <the_student> And why do you seem so negative to the Win-ARM idea?
Feb 22 20:08:10 <_av500_> oh, i am not
Feb 22 20:08:26 <_av500_> win on arm has the same use as linux on netbooks
Feb 22 20:08:31 <_av500_> make intel reduce chip prices
Feb 22 20:08:44 <_av500_> like it forced m$ to reduce XP prices
Feb 22 20:08:56 <the_student> And make Vista less crappy
Feb 22 20:09:06 <thurbad> win 7 prices are up again
Feb 22 20:09:41 <_av500_> thurbad: yes, linux on netbooks is dead :)
Feb 22 20:10:02 <_av500_> joe the plumber did not grok it
Feb 22 20:10:26 <thurbad> yeah I have to install linux myself after paying for the windows license on a new netbook :/
Feb 22 20:11:55 <thurbad> joe the plumber doesn't really understand windows or mac either but it comes with preconfigured shortcuts ~.~
Feb 22 20:12:48 <the_student> AFK
Feb 22 20:12:52 <the_student> Bye
Feb 22 20:12:53 <jwinnebeck> What really gets me is the people who can use a computer and NOT realize it's mac or windows or Linux
Feb 22 20:12:56 <jwinnebeck> in some cases
Feb 22 20:13:23 <_av500_> why?
Feb 22 20:13:30 <thurbad> I had a friend that grduated with a CS degree that didn't know if she had a mac or a pc
Feb 22 20:13:33 <jwinnebeck> then there are some that can see the difference but then assume they can't do anything at all, like it's mac now so they forget how to move windows
Feb 22 20:14:11 <_av500_> my wife does not realize when i leave the pc running linux
Feb 22 20:14:15 <mranostay> thurbad: *facepalm*
Feb 22 20:14:18 <_av500_> she clicks the firefox icon as before
Feb 22 20:14:23 <jwinnebeck> _av500_: exactly... so
Feb 22 20:14:26 <thurbad> desktop OS's are fairly similar as long as you're a superficial user
Feb 22 20:14:34 <jwinnebeck> well if there ar epeople that can't tell linux from Windows, then why is Linux a non-starter on desktop?
Feb 22 20:14:43 <_av500_> jwinnebeck: office
Feb 22 20:14:59 <thurbad> this was back in the days before firefox.. when netscape was still its own company
Feb 22 20:15:07 <jwinnebeck> OK you can get office on wine
Feb 22 20:15:21 <jwinnebeck> or open office
Feb 22 20:15:31 <thurbad> I tend to use open office most of the time
Feb 22 20:15:32 <jwinnebeck> I Switched to OOo on Windows
Feb 22 20:15:38 <jwinnebeck> MS Office is totally gone
Feb 22 20:15:40 <mranostay> thurbad: what do they teach in CS i wonder
Feb 22 20:15:46 <thurbad> it's not as good, but it's not $500 either
Feb 22 20:15:57 <jwinnebeck> For what I do as a home user it's perfectly fine
Feb 22 20:16:02 <jwinnebeck> I really only use calc
Feb 22 20:16:17 <_av500_> oh, and printer drivers
Feb 22 20:16:22 <jwinnebeck> I can probably count on one hand the number of times I used Word in the last decade at home
Feb 22 20:16:33 <_av500_> and scanner drivers
Feb 22 20:16:37 <_av500_> stuff my dad needs
Feb 22 20:16:53 <jwinnebeck> And now I use google docs instead
Feb 22 20:16:56 <thurbad> I only have Office because I occasionally do consulting that requires Access... ick
Feb 22 20:16:57 <_av500_> heck, even i boot windows to scan
Feb 22 20:17:05 <jwinnebeck> gdocs fixes a lot of the problems I've had with office for a long time
Feb 22 20:17:34 <jwinnebeck> it's collaborative and versioned
Feb 22 20:17:39 <thurbad> gdocs has been flowcharting than standard office
Feb 22 20:18:05 <jwinnebeck> I've been very dissapointed in MS Office's lack of innovation for a long time
Feb 22 20:18:22 <jwinnebeck> Office 2003 isn't THAT different from Office now
Feb 22 20:18:34 <jwinnebeck> they add a little useful thing now and then but still
Feb 22 20:18:35 <thurbad> not on the level of Visio.. but it's good enough for most of the things I need use it for
Feb 22 20:18:51 <jwinnebeck> Visio is one tool that's been hard to replace
Feb 22 20:19:02 <jwinnebeck> But I also "hate" it
Feb 22 20:19:05 <jwinnebeck> it seems to be a time suck
Feb 22 20:19:07 <thurbad> except for the changed UI, which was unnecessary
Feb 22 20:19:11 <jwinnebeck> although there's nothing better
Feb 22 20:19:35 <jwinnebeck> I mean there hav ebeen improvements, like in Excel they added the highlighing of cells used in a formula
Feb 22 20:19:41 <jwinnebeck> and easier conditional formatting
Feb 22 20:19:51 <jwinnebeck> but that's all they did in like 3 years between a release?
Feb 22 20:20:04 <jwinnebeck> Office is a horrible suite they need a whole paradigm shift
Feb 22 20:20:23 <jwinnebeck> collaboration... also more introduction of WYMIWYG concepts
Feb 22 20:20:39 <jwinnebeck> which HTML and LaTeX have figured out years before Office even existed
Feb 22 20:20:55 <thurbad> I think visual paradigm is a nice tool that can do most everything Visio can ,and then some but not cheap
Feb 22 20:20:59 <jwinnebeck> and even to this day you have people that spend 60 minutes going through a document and bolding and coloring things one line at a time
Feb 22 20:21:12 <jwinnebeck> then Word gets all screwed up
Feb 22 20:21:19 <jwinnebeck> and they fight against it to fix it
Feb 22 20:22:40 <thurbad> I really hate when word starts trying to 'fix' my formatting
Feb 22 20:22:52 <koen> latex ftw
Feb 22 20:23:01 <koen> or lately: plain text ftw
Feb 22 20:23:05 <jwinnebeck> ha
Feb 22 20:23:13 <thurbad> never learned LaTex
Feb 22 20:23:29 <jwinnebeck> I think "everyone" should learn LaTeX
Feb 22 20:23:41 <jwinnebeck> for the same reason "everyone" should learn LISP
Feb 22 20:23:47 <jwinnebeck> although I don't really know LISP that well :)
Feb 22 20:23:58 <jwinnebeck> latex and winword are opposites
Feb 22 20:24:11 <jwinnebeck> winword is the epitome of wysiwyg
Feb 22 20:24:18 <jwinnebeck> latex is the epitome of wymiwyg
Feb 22 20:24:58 <jwinnebeck> Although I could say learning proper HTML/CSS is about the same
Feb 22 20:25:10 <thurbad> I've seen word open documents differently than they appeared when they were saved.. that's not fun
Feb 22 20:25:26 <jwinnebeck> it's the same concept. Define the semantics of the text and style based on that rather than the style of the text without context
Feb 22 20:29:53 <aholler> latex is imho a must for every document with more than 10 lines, if it can't be html or plain ascii. ;)
Feb 22 20:30:04 <jwinnebeck> haha
Feb 22 20:30:14 <jwinnebeck> well years ago I always thought HTML was a better word processor than Word
Feb 22 20:30:27 <jwinnebeck> and I would use Dreamweaver as a word replacement in the early 200s
Feb 22 20:30:28 <jwinnebeck> 2000s
Feb 22 20:30:34 <jwinnebeck> everyone can open HTML
Feb 22 20:30:41 <jwinnebeck> and HTML can do everything Word can do and more
Feb 22 20:31:06 <jwinnebeck> and so that's another reason I "like" gdocs in that it appears that the docs are just HTML
Feb 22 20:31:09 <aholler> I've used html as a replacement for something like a wiki.
Feb 22 20:31:25 <jwinnebeck> And since HTML is text-based that enables proper version control and collaboration (3-way merge)
Feb 22 20:31:29 <aholler> not as something someone should write texts with
Feb 22 20:32:07 <jwinnebeck> Well I wouldn't try a book in HTML
Feb 22 20:32:12 <jwinnebeck> but I also wouldn't write a book in Word either
Feb 22 20:32:16 <jwinnebeck> I would use LaTeX
Feb 22 20:32:34 <aholler> txt2tags is nice too
Feb 22 20:32:44 <jwinnebeck> So I don't really find Word useful for anything outside of my professional life where Word is expect
Feb 22 20:32:51 <jwinnebeck> and actually the majority of my work documents are really wiki
Feb 22 20:33:05 <jwinnebeck> But I have to use Outlook/Word with outsiders
Feb 22 20:33:17 <mranostay> I usually use Google Docs with outsiders
Feb 22 20:33:30 <aholler> latex -> pdf ;)
Feb 22 20:33:30 <mranostay> they at least they can handle those
Feb 22 20:33:31 * woglinde wonders sometimes if jwinnebeck is a woman
Feb 22 20:33:56 <jwinnebeck> Yeah my big problem with gdocs is the fact I don't like to store private data in cloud, including word stuff
Feb 22 20:34:10 <mranostay> aholler: so emacs or vim? :)
Feb 22 20:34:16 <jwinnebeck> woglinde: why would you think I am one gender or another?
Feb 22 20:34:35 <woglinde> talking so much
Feb 22 20:34:43 <mranostay> heh
Feb 22 20:34:56 <mranostay> that is sexist and totally unacceptable
Feb 22 20:34:58 <jwinnebeck> woglinde: since you are here, you don't happen to have JNA in meta-java, do you? I didn't find it
Feb 22 20:35:07 <jwinnebeck> And I do talk a lot
Feb 22 20:35:07 <SilicaGel> koen: are you alive?
Feb 22 20:35:30 <djlewis> some blame it on too much caffiene
Feb 22 20:35:30 <jwinnebeck> I have a lot of qualities that are those typically associated with women
Feb 22 20:35:35 <Russ> latex is the way to go. Its just too bad about the syntax
Feb 22 20:35:43 <jwinnebeck> I also have a lot of qualities that men typically have
Feb 22 20:35:50 <SilicaGel> I'm actually the one who drinks too much caffeine, not him.
Feb 22 20:35:56 <jwinnebeck> Well I really use Lyx when I do latex
Feb 22 20:36:14 <jwinnebeck> I used to do latex in vim...
Feb 22 20:36:30 <jwinnebeck> Maybe one thinks lyx is cheating ,but it's still "what-you-mean" instead of "what you see"
Feb 22 20:36:48 <jwinnebeck> in that lyx doesn't have formatting commands
Feb 22 20:36:51 * mranostay is on a health kick. limiting caffeine and alcohol
Feb 22 20:37:03 <mranostay> of course ELC screwed that up last week :)
Feb 22 20:37:16 <aholler> what I need from latex are usually not more than 10 tags or such. everything else can be found in l2kurz
Feb 22 20:37:56 <woglinde> jwinnebeck send patch
Feb 22 20:38:17 <mranostay> mru: now every time you say something troll-like i hear it in a accent :)
Feb 22 20:39:07 * mdp imagines a troll's accent
Feb 22 20:39:12 <jwinnebeck> woglinde: well I don't think I need it yet
Feb 22 20:39:23 <woglinde> than why you asked me
Feb 22 20:39:43 <mranostay> well it is suprisely british
Feb 22 20:39:54 <jwinnebeck> was thinking about using it for PRU but there is already JNA for ARM, I might just use that
Feb 22 20:39:59 <jwinnebeck> if I do end up using JNA
Feb 22 20:40:22 <jwinnebeck> SilicaGel and I are working on it
Feb 22 20:40:26 <_av500_> JNA is JNI for ARM?
Feb 22 20:40:34 <jwinnebeck> JNA is a "replacement" for JNI
Feb 22 20:40:37 <woglinde> av500 no
Feb 22 20:40:45 <woglinde> jan is jni with lower overhead
Feb 22 20:40:46 <jwinnebeck> it eliminates the need for you to write a JNI wrapper around C code and call C code directly
Feb 22 20:40:48 <woglinde> ups
Feb 22 20:40:55 <woglinde> like python-c foo
Feb 22 20:40:57 <woglinde> or so
Feb 22 20:41:04 <jwinnebeck> that's definitely not true woglinde. JNA has more overhead (3x direct or 30x standard)
Feb 22 20:41:12 <jwinnebeck> JNA is implemented on top of JNI
Feb 22 20:41:23 <woglinde> I meant the overhead of writing the c stubs foo
Feb 22 20:41:27 <jwinnebeck> JNA is more like P/Invoke from C#
Feb 22 20:41:29 * _av500_ proposes JNO in top of JNA
Feb 22 20:41:48 <aholler> just use c++ ;)
Feb 22 20:42:03 <jwinnebeck> JNA is basically a "generic" wrapper of C code written in JNI to remove the need for custom JNI code
Feb 22 20:42:17 <woglinde> jwinnebeck no I meant using lib functions directly in python
Feb 22 20:42:40 <jwinnebeck> oh yeah I don't know about python-c
Feb 22 20:42:58 <jwinnebeck> JNA is definitely less overhead for development but more overhead for CPU
Feb 22 20:43:04 <woglinde> learn more python
Feb 22 20:43:24 <woglinde> cpu's getting faster and faster
Feb 22 20:43:28 <aholler> na, python is broken by tabs
Feb 22 20:43:39 <woglinde> dont use tabs
Feb 22 20:43:42 <jwinnebeck> Yes CPUs are getting faster, maybe we are on 2 different conversations :)
Feb 22 20:43:53 <woglinde> .o(woman)
Feb 22 20:43:54 <aholler> and you will have to search tons of bugs through testing which is a total fail
Feb 22 20:44:06 <woglinde> aholler o.O
Feb 22 20:44:14 <jwinnebeck> groovy > python
Feb 22 20:44:15 <woglinde> maybee wrong debugging technic?
Feb 22 20:44:29 <mranostay> is it flame war day? :)
Feb 22 20:44:35 <woglinde> there is even springpython
Feb 22 20:44:37 <aholler> searching errors through debugging is wrong too
Feb 22 20:44:38 * woglinde runs
Feb 22 20:44:41 <jwinnebeck> Well someone started it earlier with emacs vs vi
Feb 22 20:44:46 <jwinnebeck> but emacs is an operating system
Feb 22 20:44:49 <jwinnebeck> vi is a text editor
Feb 22 20:45:30 <mranostay> init=/usr/bin/emacs   ? :)
Feb 22 20:45:36 <jwinnebeck> this became clear when I tried to bitbake angstrom and automake failed to compile because it was trying to use emacs to COMPILE something
Feb 22 20:45:50 <jwinnebeck> emacs is a compiler now apparently
Feb 22 20:47:31 <woglinde> hm lisp introduced many nice programming features
Feb 22 20:47:55 <jwinnebeck> also emacs was started by RMS
Feb 22 20:47:56 <aholler> besides brackets
Feb 22 20:48:06 <TonyB> for those that helped me out yesterday, I did manage to get a USB Bluetooth dongle working on beagleboard xm running android. Thanks a bunch!
Feb 22 20:48:11 <jwinnebeck> and I can't use a software from someone who doesn't know how to shower or get a haircut
Feb 22 20:48:23 <woglinde> TonyB what was the solution?
Feb 22 20:48:54 <woglinde> jwinnebeck jesus didnt know about haircut too
Feb 22 20:49:12 <jwinnebeck> seriously?
Feb 22 20:49:16 <djlewis> i suspect showers were hard to come by to.
Feb 22 20:49:30 <aholler> at least when you walk overwater
Feb 22 20:49:37 <jwinnebeck> https://www.google.com/search?q=richard+stallman&hl=en&prmd=imvnso&tbm=isch&tbo=u&source=univ&sa=X&ei=0FRFT_DWG4mIgwfFwNCaBA&ved=0CFMQsAQ&biw=1164&bih=845
Feb 22 20:49:37 <djlewis> lol
Feb 22 20:49:41 <jwinnebeck> that's the face of emacs!
Feb 22 20:49:50 <TonyB> The .config file was getting changed by omap3_beagleboard_android_defconfig even after I was using make menuconfig. So everything that I was changing with make menuconfig was being over written.
Feb 22 20:50:13 <woglinde> TonyB yes
Feb 22 20:50:16 <woglinde> hehe
Feb 22 20:50:25 <jwinnebeck> I can't say too much though he also worked on GCC
Feb 22 20:50:33 <woglinde> damn android build foo
Feb 22 20:50:55 <TonyB> I had to edit the omap3_beagleboard_android_defconfig to eventually get btusb to load.
Feb 22 20:51:28 <TonyB> no docs to support what I am saying, but it works
Feb 22 20:53:56 <aholler> btw. python is like perl, a language to friggle something fast together but not for real usage. for real stuff you want a static typed language where the compiler already finds errors.
Feb 22 20:54:19 <woglinde> aholler your design errors?
Feb 22 20:54:21 <jwinnebeck> I... disagree with that
Feb 22 20:54:28 <woglinde> or multithreading errors?
Feb 22 20:54:33 <aholler> get real
Feb 22 20:54:53 <jwinnebeck> After getting into code for things like Trac, I see that Python is perfectly fine for real projects
Feb 22 20:54:55 <woglinde> for errors you writes test anyway
Feb 22 20:55:00 <jwinnebeck> it's not my personal choice, but there is nothing wrong with it
Feb 22 20:55:21 <woglinde> hm ahollers tries to troll nice
Feb 22 20:55:22 <jwinnebeck> the dynamic vs static is like mac vs pc or emacs vs vi
Feb 22 20:55:32 <aholler> woglinde: next fail. it's almost impossible to write tests for every possible program flow
Feb 22 20:55:51 <SilicaGel> aholler: that's how I feel about python too
Feb 22 20:56:42 <woglinde> hm I wrote stuff in python faster than in java and works
Feb 22 20:57:03 <aholler> with how many lines? 100? 200?
Feb 22 20:57:04 <woglinde> and python has the better serial line bindings
Feb 22 20:57:05 <woglinde> haha
Feb 22 20:57:14 <jwinnebeck> oh no
Feb 22 20:57:18 <jwinnebeck> not RXTX again
Feb 22 20:57:22 <woglinde> ;)
Feb 22 20:57:26 <jwinnebeck> every conversation keeps coming back to RXTX
Feb 22 20:57:28 <jwinnebeck> I swear man
Feb 22 20:58:33 <SilicaGel> I wonder if some of those neato java text to speech libraries I was playing with on linux/x86 would work on bone with McASP
Feb 22 20:58:42 <SilicaGel> I don't know if there's native stuff involved.  Probably is. :(
Feb 22 20:58:50 <woglinde> SilicaGel it is
Feb 22 20:59:04 <SilicaGel> It is what?  Neat?  or native?
Feb 22 20:59:25 <woglinde> native
Feb 22 20:59:30 <SilicaGel> oh
Feb 22 20:59:32 <SilicaGel> I don't know waht I used now
Feb 22 20:59:35 <SilicaGel> let me look
Feb 22 21:00:03 <woglinde> btw you could use dbus for transmitting it to your speech service
Feb 22 21:00:18 <SilicaGel> and have to suck in libmatthew?  :(
Feb 22 21:00:27 <woglinde> ;)
Feb 22 21:00:44 <jwinnebeck> at least it's not RXTX...
Feb 22 21:00:58 <woglinde> in our project we runs the dbus service on side and java client on other side for communication
Feb 22 21:01:07 <woglinde> dbus service written in python
Feb 22 21:02:15 * SilicaGel curses at the am335x trm ... again
Feb 22 21:02:17 <aholler> my usual anti-python-example: http://fpaste.org/cirt/
Feb 22 21:02:19 <SilicaGel> I :heart: dbus
Feb 22 21:02:36 <SilicaGel> I used to have a palm pre, and the way they used the dbus on that was just fabulous
Feb 22 21:02:36 <woglinde> aholler anti-php example?
Feb 22 21:03:16 <SilicaGel> where in bloody hell is IMEM0 mapped into the global address space
Feb 22 21:03:20 <woglinde> aholler no option parser used you are out
Feb 22 21:03:28 <woglinde> ups argparser
Feb 22 21:03:30 <woglinde> since 2.7
Feb 22 21:04:07 <mranostay> does python still have that GIL?
Feb 22 21:04:07 <aholler> go and teach trolls
Feb 22 21:04:50 <woglinde> aholler sure make your c++ apps nobody can maintain
Feb 22 21:05:06 <SilicaGel>         pruramarea = (unsigned int *) prussdrv.pru0iram_base;
Feb 22 21:05:08 <SilicaGel> great
Feb 22 21:05:10 <SilicaGel> where the hell is THAT
Feb 22 21:05:29 <woglinde> yeah back to pruss
Feb 22 21:05:34 <SilicaGel> I'm sorry
Feb 22 21:05:34 <jwinnebeck> SilicaGel: the top of the prussdrv.c sets everything up
Feb 22 21:05:46 <jwinnebeck> the header has just a bunch of offsets
Feb 22 21:05:50 <jwinnebeck> all magical
Feb 22 21:07:14 <SilicaGel> PRU0IRAM_PHYS_BASE            0x01C38000
Feb 22 21:07:32 <SilicaGel> okular sucks ass, I need a better pdf reader for linux.
Feb 22 21:08:54 <jwinnebeck> I think I asked this yesterday but didn't get an answer... what does ./oebb.sh update do, is it the same as for each dir in sources do git pull --rebase?
Feb 22 21:08:55 <SilicaGel> Nice
Feb 22 21:08:56 <SilicaGel> that's wrong.
Feb 22 21:09:09 <SilicaGel> it's in the wrong place.
Feb 22 21:09:26 <SilicaGel> It's at 0x43a03400
Feb 22 21:09:36 <jwinnebeck> SilicaGel: it doesn't matter
Feb 22 21:09:51 <jwinnebeck> if you notice everything is subtracted from the PRU0BASE or whatever it is
Feb 22 21:09:55 <wmat> koen: do you have a PDF of your ELC talk?
Feb 22 21:10:08 <jwinnebeck> so the absolute addresses are there, but the prussdrv.c code only ever looks at the offsets
Feb 22 21:10:11 <wmat> jkridner: do you have a PDF of your ELC talk?
Feb 22 21:10:16 <jwinnebeck> why it's dont that way I don't know
Feb 22 21:10:27 <jkridner> I sent in the .pptx, but I'll get you a PDF.
Feb 22 21:10:56 <woglinde> hi wmat
Feb 22 21:12:26 <SilicaGel> yeah but it has to be wrong.  my program didn't get to IRAM0.
Feb 22 21:12:49 <wmat> jkridner: thx
Feb 22 21:12:56 <wmat> woglinde: hi
Feb 22 21:13:15 <wmat> most of the ELC2012 presentations are up on elinux.org now, btw
Feb 22 21:13:17 <kkeller> jkridner:  I tested the code from koen - it did not work. Failed one test case
Feb 22 21:14:41 <Crofton|work> video?
Feb 22 21:14:42 <prpplague> and the video is out now
Feb 22 21:15:04 <prpplague> http://video.linux.com/categories/2012-embedded-linux-conference
Feb 22 21:16:06 <wmat> wow, that was quick for the video turn around
Feb 22 21:16:29 <prpplague> wmat: greetings btw
Feb 22 21:16:37 <wmat> prpplague: greetings
Feb 22 21:16:58 <prpplague> wmat: do i need to add a link for my slides on elinux?
Feb 22 21:17:19 <wmat> prpplague: nope, I took care of it
Feb 22 21:17:34 <wmat> http://elinux.org/ELC_2012_Presentations
Feb 22 21:17:36 <prpplague> can we link to my resource page instead?
Feb 22 21:17:41 <wmat> sure
Feb 22 21:17:48 * prpplague edits
Feb 22 21:19:24 <prpplague> wmat: i'd like people to go to the resource page just so they are aware there is additional support material
Feb 22 21:20:48 <wmat> prpplague: sounds good
Feb 22 21:24:34 <wmat> did Free Electrons film the talks again?
Feb 22 21:25:05 <prpplague> wmat: yea they did the filming, but did it for LF iirc
Feb 22 21:28:48 <wmat> cool, thanks
Feb 22 21:33:08 <jwinnebeck> For some reason I thought that https://github.com/Angstrom-distribution/meta-ti/commit/d04a83f7ee7d4a5f51aebc97f444b1f0c19fb0a0 would get into my checkout, but I notice the setup-scripts repo at git://git.angstrom-distribution.org/setup-scripts has a layers file actually pointing to git://git.angstrom-distribution.org/meta-texasinstruments
Feb 22 21:33:20 <jwinnebeck> So does koen have his own branch or something?
Feb 22 21:33:30 <woglinde> its the same
Feb 22 21:33:47 <woglinde> name only differs
Feb 22 21:33:48 <jwinnebeck> oh man I'm really trying here with git, I better go back to my RTFMing
Feb 22 21:34:11 <jwinnebeck> I did a git fetch and git pull and didn't get the latest stuff
Feb 22 21:34:14 <jwinnebeck> at least based on git log
Feb 22 21:34:22 <jwinnebeck> I thought I was starting to get it
Feb 22 21:34:26 <aholler> git rebase
Feb 22 21:34:29 <woglinde> hm let me see
Feb 22 21:34:47 <jwinnebeck> OK I will read up on rebase soon
Feb 22 21:34:52 <woglinde> https://github.com/Angstrom-distribution/meta-ti
Feb 22 21:35:04 <woglinde> hm
Feb 22 21:35:33 <woglinde> http://git.angstrom-distribution.org/cgi-bin/cgit.cgi/meta-texasinstruments/
Feb 22 21:36:41 <jwinnebeck> Well here is what I did. I followed the directions at http://www.angstrom-distribution.org/demo/beaglebone/ on how to use git and bitbake to do systemd-image
Feb 22 21:36:47 <jwinnebeck> this was awhile ago, a month or 2 now
Feb 22 21:36:51 <jwinnebeck> I made simple changes
Feb 22 21:36:57 <jwinnebeck> I learned that with git is safe to commit
Feb 22 21:37:11 <woglinde> yes nearly
Feb 22 21:37:16 <jwinnebeck> so I watned to check out the updates, so I did a git branch xyz, git checkout xyz, git commit -a xyz
Feb 22 21:37:19 <jwinnebeck> git checkout master
Feb 22 21:37:25 <jwinnebeck> then I did git pull
Feb 22 21:37:35 <woglinde> do pull --rebase
Feb 22 21:37:37 <jwinnebeck> I figured I would get the latest code now
Feb 22 21:37:37 <woglinde> anytime
Feb 22 21:37:46 <jwinnebeck> Yeah I thought --rebase only matters if you have commits
Feb 22 21:37:51 <woglinde> depends what you pulled
Feb 22 21:37:52 <jwinnebeck> you todl me rebase so I did rebase almost always
Feb 22 21:38:08 <jwinnebeck> but when I made a branch, effectively I tagged what I had on my disk at that time, right?
Feb 22 21:38:08 <woglinde> you can do git fetch
Feb 22 21:38:19 <woglinde> and manuanlly select the branch to rebase
Feb 22 21:38:33 <woglinde> and yesterday I mastered rebase --onto
Feb 22 21:38:58 <woglinde> which lets you merge only a specific range patches
Feb 22 21:38:58 <jwinnebeck> Well what I wanted to do is take my version on disk, with my mods, and freeze them. I think I did that properly
Feb 22 21:39:09 <jwinnebeck> but git log HEAD..FETCH_HEAD still shows a bunch
Feb 22 21:39:14 <jwinnebeck> and get log shows older stuff
Feb 22 21:39:31 <jwinnebeck> but I'm also on meta-texasinstruments and not meta-ti
Feb 22 21:39:38 <jwinnebeck> I don't think they are the same although you said they were?
Feb 22 21:39:51 <woglinde> whats git branch -r saying?
Feb 22 21:40:15 <aholler> rebase does fast-forward if you don't have selfmade commits
Feb 22 21:40:23 <jwinnebeck> a lot of stuff, but the first line is origin/HEAD -> origin/master
Feb 22 21:40:33 <jwinnebeck> Yeah aholler: it did say it was fast forwarding when I did it the first time
Feb 22 21:40:36 <jwinnebeck> let me check something
Feb 22 21:40:43 <aholler> git rebase remotes/origin/HEAD
Feb 22 21:41:01 <jwinnebeck> ok my meta-ti is up to date
Feb 22 21:41:13 <jwinnebeck> it's just not inclusive of this patch: https://github.com/Angstrom-distribution/meta-ti/commit/d04a83f7ee7d4a5f51aebc97f444b1f0c19fb0a0
Feb 22 21:41:17 <jwinnebeck> because I'm on a different repo
Feb 22 21:42:05 <aholler> you might clone it on your pc and clone it on the beagle from the pc. so you can use gitk or something similiar
Feb 22 21:42:10 <jwinnebeck> so it's the setup-scripts git that has me stumped but I didn't make a branch there
Feb 22 21:43:24 <jwinnebeck> OK I have some sort of weird branch that's why
Feb 22 21:43:32 <jwinnebeck> I wonder if I made that branch
Feb 22 21:43:42 <woglinde> git branch
Feb 22 21:43:43 <jwinnebeck> there is master and oe-core local and I'm on oe-core
Feb 22 21:43:46 <woglinde> shows your local
Feb 22 21:43:48 <woglinde> branches
Feb 22 21:43:49 <jwinnebeck> Yeah
Feb 22 21:43:51 <jwinnebeck> I got that
Feb 22 21:43:55 <woglinde> git branch -r remotes one
Feb 22 21:43:59 <woglinde> git branch -a all
Feb 22 21:44:21 <jwinnebeck> Did oebb.sh create this branch?
Feb 22 21:44:24 <jwinnebeck> I don't think I did it
Feb 22 21:44:26 <woglinde> maybee
Feb 22 21:44:35 <jwinnebeck> in the setup-scripts, I decided a different approach, I did git commit myself
Feb 22 21:44:36 <woglinde> I dont use it
Feb 22 21:44:43 <woglinde> I update all with git pull --rebase
Feb 22 21:44:48 <jwinnebeck> I still haven't figured out what oebb.sh update does and if that's what I really want
Feb 22 21:45:00 <jwinnebeck> I thought it would git pull --rebase everything in layers.txt
Feb 22 21:45:02 <woglinde> and I hate this google repo tool
Feb 22 21:45:13 <jwinnebeck> layers.txt this is a bitbake thing it's NOT a git thing, right?
Feb 22 21:45:23 <woglinde> yes
Feb 22 21:45:26 <jwinnebeck> even though it's a reference to a bunch of gits
Feb 22 21:46:59 <jwinnebeck> ok I switch to master branch and git pull --rebase and it fast forwarded me
Feb 22 21:47:05 <jwinnebeck> so I or the oebb script created this branch
Feb 22 21:47:12 <jwinnebeck> and since it wasn't attached to the remote, there was nothing to pull into it
Feb 22 21:47:23 <jwinnebeck> my interpretation so far
Feb 22 21:47:33 <jwinnebeck> but now my commit is on oe-core branch :(
Feb 22 21:47:54 <jwinnebeck> these are the changes woglinde had me make for meta-java.
Feb 22 21:48:12 <jwinnebeck> OK I'll just port that back in and screw this oe-core branch whatever it is
Feb 22 21:48:49 <woglinde> sure
Feb 22 21:49:00 <woglinde> local branches should be short living
Feb 22 21:49:22 <jsabeaudry> jwinnebeck, btw, have you seen the post on oracle comparing the java performance on arm and subsequent comments by Xerces?
Feb 22 21:49:26 <jwinnebeck> well local branches, when I do "git branch" this is equivalent to svn cp right? It's effectively making a tag for me?
Feb 22 21:49:47 <woglinde> jsabeaudry btw xerces is xrandby
Feb 22 21:49:51 <jwinnebeck> jsabeaudry: I saw an oracle blog post that was more or less an advertisement for their JVM saying it's way faster than openjdk
Feb 22 21:49:56 <woglinde> here on freenode
Feb 22 21:49:58 <jwinnebeck> no way
Feb 22 21:50:01 <jwinnebeck> what the heck
Feb 22 21:50:06 <jwinnebeck> is everyone who is anyone on freenode?
Feb 22 21:50:11 <woglinde> yes
Feb 22 21:50:14 <jwinnebeck> I've been using xerces for years
Feb 22 21:50:16 <jsabeaudry> jwinnebeck, mostly yes
Feb 22 21:50:16 <woglinde> besides me
Feb 22 21:50:24 <jwinnebeck> This is why I love IRC
Feb 22 21:50:26 <woglinde> uhm not this xerces
Feb 22 21:50:35 <djlewis> one fiber cable == 60ns delay . ..
Feb 22 21:50:35 <jwinnebeck> oh?
Feb 22 21:50:44 <jwinnebeck> xerces, the Apache xml parser? the one merged into JDK?
Feb 22 21:50:46 <woglinde> xerces is a common name in sweden
Feb 22 21:50:51 <jwinnebeck> oh
Feb 22 21:51:03 <jwinnebeck> I didn't know that
Feb 22 21:52:46 <jwinnebeck> so then not http://xerces.apache.org/
Feb 22 21:53:23 <woglinde> hm didnt look up if he wrote it
Feb 22 21:53:28 <woglinde> but I bed not
Feb 22 21:54:37 <woglinde> hm xerces has same ammount of commits like llvm
Feb 22 22:02:21 <Russ> http://www.disinfo.com/2012/02/mythbusters-banned-from-discussing-rfid-by-visa-and-mastercard/ (TI element)
Feb 22 22:03:58 <woglinde> haha
Feb 22 22:16:02 <djlewis> just keeps the innocent folk from knowing about it
Feb 22 22:20:18 <jwinnebeck> what's the story behind that? They can ping the rfids without you knowing and then do tracking of some sort based on that?
Feb 22 22:20:26 <jwinnebeck> what is the distance?
Feb 22 22:21:00 <djlewis> several hundred feet
Feb 22 22:21:47 <jwinnebeck> wow
Feb 22 22:21:49 <woglinde> nothing new btw.
Feb 22 22:21:56 <woglinde> since last ccc congress
Feb 22 22:22:00 <jwinnebeck> I've heard the concept but...
Feb 22 22:22:06 <jwinnebeck> so this is very practical for tracking who enters and leaves a store
Feb 22 22:22:10 <jwinnebeck> for example
Feb 22 22:22:17 <woglinde> yes for example
Feb 22 22:22:23 <jwinnebeck> Maybe it is time to buy a tinfoil wallet
Feb 22 22:22:28 <Russ> sleep well knowing that the credit card companies are hard at work battling evil hackers
Feb 22 22:23:11 <jwinnebeck> so really what it means is that someone can figure out who is around them within 100m in an automated way
Feb 22 22:23:18 <jwinnebeck> But what information do they get?
Feb 22 22:23:30 <jwinnebeck> can they get actual names or clear info, or do they just get some opaque hash
Feb 22 22:23:58 <jwinnebeck> if it's a hash then they can tell that person A is the same as person A later but not immediately who person A is (unless they figure it out)
Feb 22 22:24:13 <mru> the only wireless card I ever carry is a london underground card
Feb 22 22:24:19 <mru> and it doesn't have my name associated with it
Feb 22 22:24:38 <jwinnebeck> my mastercard debit hard has "payexpress" or whatever it's called
Feb 22 22:24:43 <jwinnebeck> but you can tap the terminal and pay with it
Feb 22 22:24:43 <djlewis> there are howto's on how to remove the chip
Feb 22 22:24:51 <mru> must be an american thing
Feb 22 22:25:02 <jwinnebeck> But you are supposed to get within just a few millimeter
Feb 22 22:25:06 <jwinnebeck> and I've never used it once
Feb 22 22:25:20 <jwinnebeck> however, that distance could just be because of the power
Feb 22 22:25:27 <djlewis> i think it is getting mandatory in the US to put a chip in your pet.
Feb 22 22:25:33 <jwinnebeck> they are saying with higher power transmission they can go farther..?
Feb 22 22:25:37 <Russ> mru, it probably wouldn't be hard to associate a name with your underground card if you keep it for long periods of time
Feb 22 22:25:45 <jwinnebeck> djlewis: I've never heard of that
Feb 22 22:25:53 <mru> Russ: I only use it a few times a year
Feb 22 22:26:04 <mru> since I don't live in london
Feb 22 22:26:11 <Russ> ah, there is that
Feb 22 22:26:27 <mru> I'm not particularly concerned
Feb 22 22:26:34 <Russ> I was going to mention, if you scanned a credit card somewhere that had a reader for london underground cards too, they could associate the two
Feb 22 22:26:41 <Russ> just look at what target is doing
Feb 22 22:26:47 <mru> if someone wants to when I'm in london, they're better off monitoring irc traffic
Feb 22 22:26:52 <mru> there's a good chance I've told someone about it
Feb 22 22:28:56 <ds2> \\
Feb 22 22:31:05 <aholler> it starts getting evil, if your local shops are starting tracking based on your rfid's you carry around
Feb 22 22:32:19 <jwinnebeck> is that really that evil?
Feb 22 22:32:19 <mru> if you're a regular customer somewhere, they'll recognise you regardless
Feb 22 22:32:29 <jwinnebeck> I mean they also can use the security camera to do the same
Feb 22 22:32:32 <jwinnebeck> probably better than rfid
Feb 22 22:32:43 <jwinnebeck> facial recognition software is almost a commodity now
Feb 22 22:33:06 <jwinnebeck> and every major store around here in the US at least has a camera pointing at the door
Feb 22 22:33:10 <mru> besides, it's nice when you can simply order "the usual" at the pub
Feb 22 22:33:15 <jay6981> easily defeated by a hat and glasses
Feb 22 22:33:20 <Russ> jwinnebeck, most are such low quality, good luck
Feb 22 22:33:23 <jwinnebeck> and they can determine a LOT more information about you such as gender and race and a guess at socioeconomic status
Feb 22 22:33:37 <jwinnebeck> It's irrelevant if it is perfect
Feb 22 22:33:43 <jwinnebeck> it's irrelevant if it is defeated
Feb 22 22:33:51 <jwinnebeck> you don't need a 100% participation rate for market research
Feb 22 22:33:53 <aholler> facial recognition doesn't really work
Feb 22 22:34:14 <ds2> m:i
Feb 22 22:34:19 <mranostay> mru: forget that how when they start pouring when you walk in the door :)
Feb 22 22:34:31 * mranostay used to drink more than he does now
Feb 22 22:34:37 <jwinnebeck> Well even if you think it doesn't work, then it will soon and be better and easier than rfid
Feb 22 22:34:50 <jwinnebeck> So getting paranoid about rfid tracking is not that useful
Feb 22 22:34:56 <kkeller> koen and jkridner - FYI sent you an email at your beagleboard.org addresses...
Feb 22 22:35:05 <jay6981> that must be why those hipster stick-on mustaches are so popular now
Feb 22 22:35:05 <jwinnebeck> I say that assuming the RFID isn't so stupid that you can get names and credit card numbers from it
Feb 22 22:35:28 <mru> jay6981: stick-on, bah! real men _grow_ one
Feb 22 22:37:02 <mranostay> jay6981: they are?
Feb 22 23:18:51 <SilicaGel> wow, the PRU has a mode for the GPIO pins where it can clock bits in/out ... which is neat ... but it only works with 28 bit words .... which makes it useless.
Feb 22 23:18:56 <SilicaGel> I wonder what the hell it is intended for
Feb 22 23:19:39 <SilicaGel> i guess maybe it can do 16 bits, so ou can shift in/out 16 at a time...maybe
Feb 22 23:19:53 <ds2> kkeller?
Feb 22 23:20:01 <ds2> hmmm
Feb 22 23:20:49 <kkeller> ds2 huh?
Feb 22 23:21:25 <ds2> kkeller, trying to figure out ifyou are the ken
Feb 22 23:21:29 <jay6981> SilicaGel:  did you ever get the PRU working?
Feb 22 23:21:36 <SilicaGel> yeah
Feb 22 23:21:38 <SilicaGel> I have it running code!
Feb 22 23:21:43 <jay6981> sweet
Feb 22 23:21:53 <SilicaGel> loaded through a uio_pruss that I had to modify beacuse the addresses are all f***ed up
Feb 22 23:21:55 <kkeller> ds2 I am Ken Keller
Feb 22 23:21:56 <jay6981> i guess you found the assembler?
Feb 22 23:22:01 <SilicaGel> actually
Feb 22 23:22:05 <SilicaGel> the kernel module is fine, i said that wrong
Feb 22 23:22:08 <SilicaGel> uio_pruss.ko is fine
Feb 22 23:22:15 <kkeller> ds2 but there are several of us around in the world :)
Feb 22 23:22:17 <SilicaGel> what isn't fine is prussdrv which is the userspace library.
Feb 22 23:22:32 <mru> who needs userspace?
Feb 22 23:23:01 <SilicaGel> I agree
Feb 22 23:23:14 <SilicaGel> with a brain as powerful as mine, I prefer to execute everything in the kernel
Feb 22 23:23:19 <jay6981> wasn't there some problem with pru clocks not being setup right?
Feb 22 23:23:39 <SilicaGel> the latest angstrom kernel fixed up with bitbake is pretty darned close
Feb 22 23:23:45 <SilicaGel> it's correct for setting up the power domain and the clocks
Feb 22 23:23:52 <SilicaGel> it has one problem still
Feb 22 23:23:54 <jwinnebeck> SilicaGel: I thought you said the kernel was wrong?
Feb 22 23:23:58 <jwinnebeck> with the reset state?
Feb 22 23:24:03 <SilicaGel> yes
Feb 22 23:24:06 <SilicaGel> that's the one problem
Feb 22 23:24:27 <jwinnebeck> but you just said "what isn't fine is prussdrv"
Feb 22 23:24:33 <jwinnebeck> so there are 2 things wrong :)
Feb 22 23:24:38 <SilicaGel> oh
Feb 22 23:24:39 <SilicaGel> well yeah
Feb 22 23:24:41 <SilicaGel> haha
Feb 22 23:24:50 <SilicaGel> my brain assumed somebody else would fix the reset problem :D
Feb 22 23:25:01 <SilicaGel> jay6981: see  http://groups.google.com/group/beagleboard/browse_thread/thread/874bfed6fd714b26 starting at HOWEVER
Feb 22 23:25:10 <jwinnebeck> is the only problem with the prussdrv the addresses not being right for sitart?
Feb 22 23:25:12 <jwinnebeck> sitara?
Feb 22 23:25:26 <jwinnebeck> is there a way to get those values without just magically knowing you are on sitara and how the PRUs are mapped into memory?
Feb 22 23:25:49 <SilicaGel> yeah, the addresses are not right.  Actually, it's the offsets that aren't right.
Feb 22 23:26:08 <jwinnebeck> but that is because that code is for omap?
Feb 22 23:26:09 <SilicaGel> So what I did was told it the correct offsets.  So the pruss_base is correct, i tlearns that
Feb 22 23:26:13 <SilicaGel> yeah
Feb 22 23:26:14 <jwinnebeck> meaning, it was right for someone at some time?
Feb 22 23:26:25 <SilicaGel> right.
Feb 22 23:26:27 <SilicaGel> for a different pru
Feb 22 23:26:30 <jwinnebeck> yeah so is there a way to make that code right for both people?
Feb 22 23:26:35 <SilicaGel> you know the old PRUs only had 512 BYTES of DMEM
Feb 22 23:26:41 <SilicaGel> where we now have 8K of DMEM per PRU
Feb 22 23:27:02 <SilicaGel> apparently the address map (all the offsets) were significantly different on the davinci that all this code was written for
Feb 22 23:27:20 <SilicaGel> well hm
Feb 22 23:27:50 <SilicaGel> the only way I can think of to make it right for everybody is to have a couple of tables that you use instead of #defines to define the memory map, then have the software somewhere read REVID (assuming it can find it!) to figure out which pru it's on.
Feb 22 23:27:54 <SilicaGel> We'r eon PRU v2
Feb 22 23:28:36 <SilicaGel> Either that, or we suggest listen guys, modify uio_pruss to do all this work for us, and give us  separate mmappable pages for DMEM0, DMEM1, DMEM2, IMEM0, IMEM1
Feb 22 23:28:43 <SilicaGel> them mach-* figures out where all those things are
Feb 22 23:28:49 <SilicaGel> To be honest, I think taht'd be the cleanest solution
Feb 22 23:28:55 <jwinnebeck> I would think so
Feb 22 23:29:02 <SilicaGel> then userspace doesn't care
Feb 22 23:29:04 <SilicaGel> **AND**
Feb 22 23:29:18 <SilicaGel> if you did it taht way, it could return the correct length through /sys/uio/maps/map* for those things
Feb 22 23:29:24 <jwinnebeck> so then what do you think is the appropriate solution for US (since we are not maintainers/committers of the TI code)
Feb 22 23:29:32 <SilicaGel> so your user space code could easily figure out not only where, for example, IMEM0 is, but exactly how long it is !!!
Feb 22 23:29:45 <jwinnebeck> Yeah I sort of expected taht when reading the docs
Feb 22 23:29:49 <jwinnebeck> it's almost like the uio driver goes half way
Feb 22 23:29:59 <jwinnebeck> it maps out some areas but not all
Feb 22 23:30:13 <SilicaGel> I don't know.  I kind of think we should take uio_pruss.c and build it externally, and not use theirs.   But you know how that goes.  We fork it, it becomes our problem forever.
Feb 22 23:30:38 <SilicaGel> Maybe we should just live with it, and have the prussdrv userspace library handle all of this, including figuring out what mach we're on
Feb 22 23:31:09 <SilicaGel> that'd suck though; a lot of the point of these userspace drivers and UIO architecture is you don't have to know all of that.  You just mmap hunks of memory, all you have to know is what to expect to be there.
Feb 22 23:32:31 <jay6981> i think your original idea was better
Feb 22 23:32:43 <SilicaGel> yeah.
Feb 22 23:32:54 <SilicaGel> I can't really propose something to the community though that drastically changes the interface of uio_pruss
Feb 22 23:32:56 <jay6981> there are enough TI people around you should be able to work with
Feb 22 23:32:56 <SilicaGel> THOUGH
Feb 22 23:32:57 <SilicaGel> you know
Feb 22 23:32:59 <SilicaGel> what I could propose
Feb 22 23:33:14 <SilicaGel> is to take uio_pruss which adds maps 0, 1, 2... and just add ADDITIONAL mappings to them.
Feb 22 23:33:20 <SilicaGel> That wouldn't break backward compatibility with any existing code at all
Feb 22 23:33:51 <SilicaGel> There's no harm in having say 8 mmap entries per evt instead of 3 is there?
Feb 22 23:34:02 <jwinnebeck> no except if the upstream adds more maps
Feb 22 23:34:25 <SilicaGel> It somewhat annoys me that the 3 maps that are there now are duplicated once per event.
Feb 22 23:34:58 <SilicaGel> Sacha Baron Cohen Banned From Oscars... Developing...
Feb 22 23:35:03 <SilicaGel> oh man
Feb 22 23:35:11 <SilicaGel> that means the most beautiful actress in the world, isla fisher, won't be there either
Feb 22 23:40:14 <jwinnebeck> it's getting real late I better go
Feb 22 23:40:16 <jwinnebeck> tomorrow...
Feb 22 23:48:44 <SilicaGel> yeah, home &
Feb 23 00:27:27 <CanyonMan> hi
Feb 23 01:03:01 <SilicaGel_H> so how do I get a patch into the official angstrom distribution?
Feb 23 01:07:02 <mru> you get koen drunk, then slip a usb stick with the patch into his pocket
Feb 23 01:07:32 <SilicaGel2> is he the guy?
Feb 23 01:07:50 <SilicaGel2> Is he easy to get drunk?
Feb 23 01:08:00 <SilicaGel2> If this happens often, it could be quite expensive
Feb 23 01:08:19 <mru> depends on who pays for the beer
Feb 23 01:08:58 <SilicaGel2> wait, wait, now I think you're screwing with me.  If somebody else buys the beer, I have no argument that the beer TASTES better .... but it doesn't get you drunk faster!!!
Feb 23 01:10:03 <mru> well, the goal is to get koen drunk without paying too much, right?
Feb 23 01:10:08 <aholler> netherlands, use a cake ;)
Feb 23 01:10:18 <mru> so you get jkridner to buy the beer (or shots)
Feb 23 01:11:35 <SilicaGel2> right, that's the goal.  At least based on the premise that he won't let you put something into his pants unless you get him drunk first.
Feb 23 01:11:36 <SilicaGel2> Will he?
Feb 23 01:12:16 <mru> that might depend on your gender
Feb 23 01:12:28 <SilicaGel2> ok, so I need a female agent.
Feb 23 01:12:34 <SilicaGel2> This is getting very complicated.
Feb 23 01:12:45 <aholler> adding some hw to the green cake might have good results too ;)
Feb 23 01:12:48 <mru> then again, he has a gf so even that will require him being drunk
Feb 23 01:13:02 <SilicaGel2> unless I get her drunk instead
Feb 23 01:13:09 <SilicaGel2> that might be cheaper
Feb 23 01:13:33 <mru> I don't know, haven't met her
Feb 23 01:14:48 <aholler> a nice layout to keep a makerbot busy might help too.
Feb 23 01:14:57 <mranostay> mru: are we talking about taking advantage of koen ? :)
Feb 23 01:15:14 <aholler> bribing ;)
Feb 23 01:15:42 <ds2> Hmmmmmmm
Feb 23 01:17:50 <SilicaGel2> no
Feb 23 01:18:12 <SilicaGel2> I think the "get him drunk" plan is dfinitely "taking advantage," not "bribing"
Feb 23 01:18:28 <SilicaGel2> Bribing is however another route.  That may be cheaper than getting him drunk.
Feb 23 01:19:13 <SilicaGel2> .oO( unless "green cake" means bribery? I didn't understand that bit )
Feb 23 01:19:35 * mranostay pictures koen waking up the next morning "I SIGNED WHO'S GPG KEY?"
Feb 23 01:20:16 <aholler> green cake is a cake with some stuff which is legal in the netherlands but almost nowhere else
Feb 23 01:21:25 <aholler> don't know if he likes that, but I assume the effect is better than with beer
Feb 23 01:22:18 <jay6981> legal in california with dr. prescription
Feb 23 01:22:43 <mru> and widely available in most places, legal or not
Feb 23 01:23:07 <mranostay> -/* no comment */
Feb 23 01:24:41 <SilicaGel2> oooohhhhh no wonder I didn't get the reference
Feb 23 01:34:19 <aholler> but I'm just making fun about the netherlands.
Feb 23 01:34:58 <aholler> actually I like how easy they handle many things other countries have problems with.
Feb 23 01:35:24 <mranostay> bummer they are banning tourist from the coffee houses
Feb 23 01:35:38 <mranostay> you know it was one too many idiot tourist :)
Feb 23 01:45:27 <SilicaGel2> it's weird that they'd ban touists but not locals
Feb 23 01:45:37 <SilicaGel2> that kind of thing wouldn't fly in the US, I don't think
Feb 23 01:56:15 <timholum1> hello
Feb 23 01:57:14 <timholum1> Hello, I am wondering if anyone knows of a place to find a 4" lcd that works with the beaglebone?
Feb 23 02:00:26 <wmat> jkridner: sorry I didn't respond earlier, but you can send me or post yourself your presentation in whatever format it's currently in
Feb 23 02:00:33 <wmat> jkridner: http://elinux.org/ELC_2012_Presentations
Feb 23 02:03:18 <ds2> wmat: too tired to lug a tv this year? :D
Feb 23 02:04:05 <wmat> heh
Feb 23 02:04:16 <wmat> family obligations kept me home
Feb 23 02:05:10 <wmat> sad i missed it
Feb 23 02:05:34 <wmat> i'm really hoping to make elce this year
Feb 23 02:14:24 <ds2> :)
Feb 23 02:18:13 <mranostay> tv's are quite light these days
Feb 23 02:28:11 <mranostay> sweet accident center lane
Feb 23 02:28:58 <mranostay> can't ever be in the carpool lane can it :)
Feb 23 02:49:54 <ds2> not the sizes he usually lugs
Feb 23 02:59:12 <wmat> what did they give away at the closing ceremony at ELC2012?
**** ENDING LOGGING AT Thu Feb 23 02:59:58 2012