| 00:19.39 | *** join/#edev chouimat|busy (n=dieu@r2351064.cidc.net) |
| 00:32.05 | *** join/#edev chouimat|away (n=dieu@r2351064.cidc.net) |
| 00:45.33 | ds2 | another complain about DBUS and Bluetooth - http://gnumonks.org/~laforge/weblog/2007/ |
| 00:56.09 | *** join/#edev diJenerate (n=dijenera@72.51.97.189) |
| 01:17.13 | *** join/#edev landley (n=landley@cpe-72-177-114-214.austin.res.rr.com) |
| 01:20.08 | *** join/#edev ashikase (n=ashikase@softbank218131026001.bbtec.net) |
| 01:22.01 | ashikase | I'm using an ARM board with an unaccelerated framebuffer... |
| 01:22.49 | ashikase | what is the best (fastest) graphics backend/library for creating an animated GUI? (SDL, DirectFB, etc) |
| 01:30.45 | landley | That's like asking what's the fastest programming language in which to write a database. |
| 01:31.01 | landley | Mostly depends on how you write it and what exactly you're trying to do. |
| 01:39.42 | dbtid_ | landley: C ;) |
| 01:40.08 | landley | dbtid_: I asked a misleading question. |
| 01:40.22 | landley | It's 10 times faster to write your program in python. |
| 01:40.30 | landley | Or grab a prepackaged database and use sql. |
| 01:40.42 | landley | In terms of _speed_, hand crafted assembly still beats C. |
| 01:40.58 | ashikase | landley, true, but some are more suitable than other.... for example, GTK is not. |
| 01:41.05 | landley | Does "faster" mean "how long it'll take", or "how fast will it run"? |
| 01:41.11 | landley | In this case, it sounds like "how fast will it run". |
| 01:41.42 | landley | The fastest GUI is one written to the bare metal with an intelligent cacheing strategy, but by asking for a library he pretty much ruled out that approach. (As would sanity, these days.) |
| 01:42.05 | landley | QT embedded kinda sucks too if you take into account the fact it's written in C++, a language that's not well-suited to doing _anything_. |
| 01:42.47 | landley | I have no idea what ashikase wants this gui to _do_. |
| 01:42.52 | landley | What does "animated" mean here? |
| 01:42.56 | landley | Animated gifs? |
| 01:43.04 | ashikase | landley: any library that will allow smooth transitions (eye-candy) without accelerated hardware... |
| 01:43.05 | landley | Buttons blink when you press them? |
| 01:43.20 | ashikase | though doing it all in assembly is not an option |
| 01:43.20 | landley | Any library will. |
| 01:43.31 | landley | I was doing smooth scrolling in MCGA under DOS. |
| 01:43.46 | landley | SDL and DirectFB have both been used to do cool things. |
| 01:43.55 | landley | Asking which is _better_ involves knowing a lot more about what you plan to do. |
| 01:44.24 | ashikase | I guess I should add that I would prefer a library that provides GUI elements/objects |
| 01:44.25 | landley | I haven't used DirectFB much, but I believe it just gives you the frame buffer as a big chunk of memory you mmap. |
| 01:44.28 | landley | Drawing on it is your problem. |
| 01:44.56 | landley | QT embedded at least pays attention to embedded. I dont' like C++, but I've heard good things about that toolkit. |
| 01:45.27 | ashikase | QTE isn't bad, but it's still generally too slow (at least 2.x was) |
| 01:45.48 | landley | Does this help? It's kind of oldish, I think: http://www.linuxdevices.com/articles/AT9202043619.html |
| 01:46.21 | ashikase | thanks, I'll have a look at that |
| 01:47.55 | landley | Somebody put together a bigger list here: http://www.geocities.com/SiliconValley/Vista/7184/guitool.html |
| 01:48.24 | landley | SDL and DirectFB are both pretty much drawing libraries. You need a toolkit on top of that providing gui widgets. |
| 01:48.47 | *** join/#edev redblue (i=star@ppp190.109-253-207.mtl.mt.videotron.ca) |
| 01:49.25 | landley | I wrote my own widget set once long ago (in java, shortly before the horror that is swing came out). |
| 01:49.29 | landley | It's not that hard. |
| 01:49.33 | landley | It's also reinventing the wheel... |
| 02:13.03 | *** join/#edev chouimat|busy (n=dieu@r2351064.cidc.net) |
| 02:21.30 | Mirell | It's somewhat amazing this account on this machine is still active. |
| 02:28.13 | landley | Mirell: without knowing what machine that is or why, that's a content-free statement to the rest of us. |
| 02:28.28 | *** part/#edev ashikase (n=ashikase@softbank218131026001.bbtec.net) |
| 02:34.45 | *** join/#edev redblue (i=star@ppp066.109-253-207.mtl.mt.videotron.ca) |
| 02:35.26 | *** join/#edev Retro1974 (i=Retro197@bas13-toronto63-1177851706.dsl.bell.ca) |
| 02:37.38 | Mirell | landley: quiche.ling.utexas.edu |
| 02:37.42 | Mirell | CompLing lab machine |
| 02:37.46 | *** join/#edev roxfan2 (n=dunno@212.196-201-80.adsl-dyn.isp.belgacom.be) |
| 02:38.46 | landley | Ah. No idea there. |
| 02:40.14 | Mirell | It's because accounts are manually managed. |
| 02:40.34 | Mirell | Theoritically, it should have been removed over 13 months ago. |
| 02:46.30 | cajaka | Any one here seen other thermal sensors like this: http://www.robotshop.ca/home/products/robot-parts/sensors/thermal-array-sensors/index.html |
| 02:50.24 | Retro1974 | never seen anything like that before |
| 02:58.29 | cajaka | Ah found the search phrase for them nothing as easy to interface as the one I linked so far |
| 03:38.12 | *** part/#edev Retro1974 (i=Retro197@bas13-toronto63-1177851706.dsl.bell.ca) |
| 03:56.53 | *** join/#edev chouimat|away (n=dieu@r2351064.cidc.net) |
| 04:46.20 | *** join/#edev chouimat|busy (n=dieu@r2351064.cidc.net) |
| 05:15.46 | *** join/#edev DrZimmerman (n=theo@2001:5c0:8fff:fffe:0:0:0:6af1) |
| 05:16.40 | *** join/#edev DrZimmerman (n=theo@2001:5c0:8fff:fffe:0:0:0:6af1) |
| 05:39.45 | Mirell | I suppose I should get some sleep |
| 07:03.24 | *** join/#edev _lifeless (n=patrick@89.20.120.165) |
| 07:28.26 | *** join/#edev harpal (n=Harpal@124.125.79.206) |
| 08:07.36 | *** join/#edev emil|k (n=emil@e179233137.adsl.alicedsl.de) |
| 08:17.18 | *** join/#edev lyakh (n=lyakh@p57BD2BD1.dip0.t-ipconnect.de) |
| 08:35.15 | *** join/#edev harpal (n=Harpal@124.125.79.206) |
| 09:54.48 | *** join/#edev bbradley (n=bbradley@82-32-35-42.cable.ubr05.azte.blueyonder.co.uk) |
| 10:09.12 | Shingoshi | Where and how can I change the version of any source file needed by "make source" with buildroot? |
| 10:16.53 | Shingoshi | Never mind. I found it. |
| 11:31.06 | *** join/#edev harpal (n=Harpal@124.125.79.127) |
| 12:15.34 | *** join/#edev howtoo (n=howtoo@195.159.62.66) |
| 12:39.29 | *** join/#edev dijenerate (n=dijenera@66.54.118.16) |
| 13:15.27 | *** join/#edev __lifeless (n=patrick@90.151.44.181) |
| 14:08.45 | *** join/#edev hapchap (n=hapchap@mail.americanmicrosystems.com) |
| 14:14.07 | sunrunner20|prom | YAYYAYAAYAYA |
| 14:14.20 | sunrunner20 | GO MICROSOFT!!!!!!!!!!!!!!!!!!!!!! |
| 14:14.21 | sunrunner20 | http://weblogs.asp.net/scottgu/archive/2008/01/16/net-framework-library-source-code-now-available.aspx |
| 14:14.41 | *** join/#edev howtoo (n=howtoo@195.159.62.66) |
| 14:26.16 | *** join/#edev chouimat|work (n=mchouina@209.217.106.98) |
| 14:26.45 | *** join/#edev GPSFan (n=kenm@12.10.255.246) |
| 14:26.45 | *** mode/#edev [+o GPSFan] by ChanServ |
| 14:54.02 | Mirell | Heh |
| 14:56.28 | *** join/#edev g1powermac (n=g1powerm@unaffiliated/g1powermac) |
| 14:59.34 | *** join/#edev redblue (i=star@ppp132.109-253-207.mtl.mt.videotron.ca) |
| 15:31.22 | Mirell | Hmm... |
| 15:38.28 | Mirell | I'm tempted when I deposit my first paycheck (Stupid Direct Deposit prenotification crap) to get back $100 in $1s, and carry it around as a money roll |
| 15:45.14 | *** join/#edev emil|k (n=emil@e179239080.adsl.alicedsl.de) |
| 15:46.46 | *** join/#edev CosmicPenguin (n=nobody@163.181.251.103) |
| 15:46.46 | *** mode/#edev [+o CosmicPenguin] by ChanServ |
| 16:04.54 | hapchap | Mirell: as long as you get a $20 for the outside |
| 16:05.16 | Mirell | hapchap: Exactly. |
| 16:06.50 | *** join/#edev hw_ (n=hw@p578b3905.dip0.t-ipconnect.de) |
| 16:15.31 | dbtid | cbrake: did you get an answer to your lockup question? |
| 16:33.30 | Mirell | Woot, just got my first paycheck |
| 16:47.11 | SpeedEvil | :) |
| 17:03.49 | *** join/#edev ico2 (n=ico2@modem-1369.cougar.dialup.pol.co.uk) |
| 17:09.33 | Mirell | Speaking of laptops, need to order that Dell XPS M1330 now. |
| 17:18.58 | Mirell | Mmm, stupid taxes. |
| 17:25.22 | hapchap | hehe... my two byte checksum on a micro's firmware is BABE |
| 17:25.37 | hapchap | its the small things in life that keep me happy |
| 17:28.45 | *** join/#edev prpplague (n=dave@70.249.150.251) |
| 17:28.57 | prpplague | ho ho ho |
| 17:29.22 | prpplague | hapchap: ping |
| 17:32.33 | prpplague | ds2: ping |
| 17:48.45 | *** join/#edev dijenerate (n=dijenera@72.22.133.141) |
| 18:04.02 | *** join/#edev landley (n=landley@cpe-72-177-114-214.austin.res.rr.com) |
| 18:10.03 | *** join/#edev Shingoshi (n=shingosh@c-67-189-36-133.hsd1.or.comcast.net) |
| 18:16.42 | *** join/#edev DrZimmerman (n=theo@2001:5c0:8fff:fffe:0:0:0:6af1) |
| 18:17.49 | hapchap | prpplague: pong |
| 18:43.06 | *** join/#edev dijenerate (n=dijenera@66.54.118.16) |
| 18:58.34 | sunrunner20|afk | lol hapchapp |
| 18:58.41 | sunrunner20|afk | don't forget DEADBEEF |
| 19:01.44 | *** join/#edev prpplague (n=dave@mail.americanmicrosystems.com) |
| 19:29.20 | *** join/#edev ico2 (n=ico2@modem-439.chameleon.dialup.pol.co.uk) |
| 19:30.33 | *** join/#edev dijenerate_ (n=dijenera@66.54.118.16) |
| 19:36.56 | ds2 | prpplague: pong (sorry, late) |
| 19:40.14 | prpplague | ds2: np, i was just curious if you'd look any more at the lcd fb stuff on the hammer |
| 19:40.43 | ds2 | prpplague: eh? I think you have me confused with someone else. I never looked at the FB stuff on hammer |
| 19:41.15 | ds2 | well, maybe not never but not much |
| 20:07.28 | *** join/#edev RobotGuy (i=linuxguy@dsl093-038-070.pdx1.dsl.speakeasy.net) |
| 20:08.29 | *** join/#edev RobotGuy (i=linuxguy@dsl093-038-070.pdx1.dsl.speakeasy.net) |
| 20:09.37 | prpplague | ds2: i could have sworn we were discussing using a tft display on the hammer the other day |
| 20:09.47 | prpplague | maybe it was SpeedEvil |
| 20:10.25 | RobotGuy | prpplague: I think I started a discussion like that one day. |
| 20:10.56 | prpplague | RobotGuy: yea, but one of the folks was going to do a proof on concept on it |
| 20:12.03 | RobotGuy | prpplague: Ahhh. It would be nice to be able to have one of those 320x240 LCDs (or TFT) even in greyscale mode. |
| 20:12.38 | prpplague | RobotGuy: yea, it would take some tweaking but it looks like we can run a 3-3-2 color |
| 20:13.04 | RobotGuy | prpplague: What kind of tweaking? |
| 20:13.40 | prpplague | RobotGuy: the tft controller and frame buffer for the s3c2410 is geared towards 5-6-5 or 6-6-6 interface |
| 20:14.37 | RobotGuy | prpplague: Ahhh, so that driver would need to be forked to create a new monochrome driver? |
| 20:15.26 | prpplague | RobotGuy: not monochrome, color |
| 20:15.26 | RobotGuy | prpplague: I mean greyscale driver. |
| 20:15.32 | prpplague | RobotGuy: 3-red -3blue -2 green |
| 20:16.11 | prpplague | RobotGuy: 256 colors |
| 20:16.42 | RobotGuy | prpplague: Ahhhhh, OK. That should take less RAM, right? |
| 20:16.53 | prpplague | RobotGuy: ram is not a real issue |
| 20:18.57 | ds2 | prpplague: nope. the last time I mention it was last year |
| 20:19.14 | ds2 | prpplague: there is a fundamental problem of me not having a hammer unit |
| 20:22.12 | prpplague | ds2: naw, this was a proof of concept idea |
| 20:22.52 | ds2 | I do remmeber seeing you and someone else talk about it... think he was saying something about just truncating bits |
| 20:23.05 | RobotGuy | prpplague: Is the issue one of how much the hammer can process for a display? |
| 20:23.21 | prpplague | ahh yea it was SpeedEvil |
| 20:23.45 | prpplague | RobotGuy: no, its a limitation of which pins are brought out |
| 20:23.54 | ds2 | the sledge is suppose to fix it, IIRC |
| 20:24.04 | prpplague | RobotGuy: if all the pins for the tft lcd were brought out then you could |
| 20:24.15 | RobotGuy | prpplague: Ah, OK. |
| 20:24.19 | prpplague | ds2: not really a "fix" just different i/o |
| 20:24.45 | prpplague | ds2: the hammer was mainly designed to operate with small greyscale lcds |
| 20:24.53 | RobotGuy | prpplague: I would be happy with 256 colors or 256 grey levels. |
| 20:25.08 | prpplague | RobotGuy: grey scale is easy |
| 20:25.26 | prpplague | RobotGuy: you can get a 160x160 that interfaces to the hammer easy |
| 20:25.49 | prpplague | RobotGuy: http://www.elinux.org/M5900 |
| 20:25.59 | RobotGuy | prpplague: I am thinking a bit larger though, like 320x240. |
| 20:26.47 | RobotGuy | I can't do URLs right now - no network from my main PC. :( |
| 20:27.22 | ds2 | while we are talking about LCDs - what are the differences between those gray scale LCDs and a full blown color LCD? from what I can tell, the signals are basically the same except for color you have extra lines |
| 20:27.45 | prpplague | ds2: thats about it |
| 20:27.57 | ds2 | oh |
| 20:28.11 | prpplague | ds2: for the difference between greyscale stn and color stn |
| 20:28.24 | prpplague | ds2: but tft is different in how the data is formated |
| 20:28.34 | ds2 | prpplague: in what way? |
| 20:28.52 | ds2 | <-- didn't noticed any significant difference from the data sheet |
| 20:31.32 | prpplague | ds2: usually the color data for a 8-bit cstn is formated like this on 8 data lines, RRRBBBGG |
| 20:31.46 | prpplague | ds2: all the color is on data lines 0 thru 7 |
| 20:32.18 | prpplague | ds2: now for most tft controlers in soc's are configured as 24 bit |
| 20:33.15 | prpplague | ds2: where if you are still using the 3-3-2 configuration you would need to connect line 0-2 for red, 8-10 for blue, and 16-17 for green |
| 20:33.39 | prpplague | ds2: because inside the memory it is formated to go across all of the data lines, even if you aren't using them |
| 20:34.07 | ds2 | oh that's a hookup stuff... it could be described as 8-8-8 and you need to connect up the LSBs used for each color |
| 20:34.17 | prpplague | ds2: correct |
| 20:34.28 | ds2 | that doesn't seem that different to me |
| 20:34.49 | prpplague | ds2: now here is the problem with the hammer and tft, only the data line 0-7 are brought out, which is fine for stn |
| 20:35.04 | prpplague | ds2: but in tft mode that means only the 0-7 for red is available |
| 20:35.22 | ds2 | ultimately, from what I understand, you still need to give it N number of bits for each pixel and each pixel is clocked in by oen clock, and another clock for each line, and another clock per frame |
| 20:35.32 | ds2 | prpplague: that sounds like it can be worked around in software |
| 20:35.35 | prpplague | ds2: that is correct |
| 20:36.16 | prpplague | ds2: yea it seems possible |
| 20:37.08 | ds2 | thought I am surprise this chip doesn't let you reconfigure itself to define the lines differently. IIRC, OMAPs will let you do that |
| 20:37.13 | RobotGuy | prpplague: Seems like you could just "remap" the pins from the hammer to the display in the driver. |
| 20:37.26 | prpplague | RobotGuy: remap in what way? |
| 20:37.48 | prpplague | ds2: for lcd data lines? |
| 20:37.56 | prpplague | ds2: might want to double check that |
| 20:38.20 | RobotGuy | prpplague: Take the data from the hammer, and send it to different pins on the display,. but then you would probably need a translater between the two. |
| 20:38.21 | prpplague | ds2: i've yet to see a soc intergrated lcd controller that could define the data line structure |
| 20:38.31 | ds2 | prpplague: yeah... it is something like LCD0, LCD1, LCD2... and there is a register that puts the controller in "8bit" mode where LCD0-7 are used and 16bit mode with 2 different RGB varianets, etc |
| 20:38.36 | landley | prpplague: do you know how to set up a fully virtual x.org for use with just vnc? |
| 20:38.40 | prpplague | RobotGuy: sorry not understanding you |
| 20:38.45 | ds2 | prpplague: not completely free to define |
| 20:39.04 | ds2 | I know this is hte case cuz I screwed up some code and got funny colors |
| 20:39.34 | prpplague | ds2: sounds like that is the stn mode |
| 20:39.38 | ds2 | turns out hw was wired as if was a 24bit device (LCD was 15bit) and I had programmed the controller as 15bit |
| 20:39.39 | RobotGuy | prpplague: Take the display data pins on hammer and route them to the required pins on the display? |
| 20:40.18 | prpplague | RobotGuy: yea, thats what we are talking about, the problem is that the fb driver expects those pins to perform specific data functions, |
| 20:40.38 | prpplague | landley: i used to, but its been a while since i did any x hacking |
| 20:40.53 | prpplague | landley: i've done it in the past with my ipaq, and zaurus |
| 20:41.04 | ds2 | prpplague: what fb driver is it using? and is it in the stock K.O tree? |
| 20:41.11 | RobotGuy | prpplague: So the fb driver needs to be fooled then, eh? |
| 20:41.27 | prpplague | ds2: yea is the stock K.O and its the s3c2410fb |
| 20:41.30 | ds2 | RobotGuy: you really don't want to do it there in software |
| 20:41.33 | prpplague | RobotGuy: correct |
| 20:41.45 | landley | same here. |
| 20:41.58 | ds2 | prpplague: what version was it merged in? |
| 20:42.08 | ds2 | n/m |
| 20:42.11 | landley | I did extensive XFree86 micro-tweaking under OS/2, and under the 2.2 kerrnel. |
| 20:42.12 | prpplague | ds2: like 2.4.20 |
| 20:42.23 | landley | Stayed away since. It was _evil_... |
| 20:42.26 | ds2 | was expecting video/samsung/... =) |
| 20:43.12 | prpplague | landley: yea what i had in mind was when you plug in the nail board you'd be able to start a xnest session with the nail board |
| 20:43.45 | RobotGuy | prpplague: Nail can run a better display? |
| 20:43.51 | landley | I'm vaguely pondering something like this myself... |
| 20:44.01 | prpplague | RobotGuy: no, different concept |
| 20:44.17 | landley | Running xnest over ssh's X forwarding sounds like the kind of evil I'd enjoy... |
| 20:44.36 | ds2 | isn't Xvnc lighter weight? |
| 20:44.39 | prpplague | ds2: http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=blob;f=drivers/video/s3c2410fb.c;h=ad35033f1a48947832f8896ed3db0c9936344d06;hb=d8c89eb3a12f0da96d049bd515c7fa3702e511c5 |
| 20:44.54 | prpplague | landley: hehe |
| 20:44.58 | ds2 | prpplague: I found it already... looking for the S3C2410 docs |
| 20:45.09 | prpplague | ds2: i have them on the wiki page |
| 20:45.19 | ds2 | found it |
| 20:45.28 | prpplague | http://www.elinux.org/S3C2410 |
| 20:46.11 | prpplague | ds2: just a note, i have a bounty up for getting the tft working |
| 20:46.21 | prpplague | ds2: hammer kit with lcd stuff |
| 20:46.45 | RobotGuy | prpplague: For some reason that reminds me that mpfr won't build as a software selection for hammer. |
| 20:46.54 | landley | RobotGuy: what do you mean _go_to_? |
| 20:47.01 | landley | I have a summer home there. |
| 20:47.02 | prpplague | RobotGuy: we fixed that remember |
| 20:47.03 | landley | Built it myself. |
| 20:47.25 | RobotGuy | prpplague: Apparently there is still a problem. |
| 20:47.48 | prpplague | RobotGuy: pastebin.ca the log |
| 20:48.01 | RobotGuy | Hmmmmmm, I replaced the mpfr.mk file. |
| 20:48.17 | landley | You're talking to someone who actually wonders what happens if there's an embedded NUL byte in the output of a command line in called from the shell via $() |
| 20:48.28 | RobotGuy | prpplague: I can't do that right now. I'm on my server with no mouse for X. |
| 20:48.29 | landley | Answer: it gets filtered out. |
| 20:48.46 | prpplague | hehe |
| 20:48.49 | landley | Does not prematurely end the stream, but I don't remember if it counts as a token separator in the absence of other whitespace... |
| 20:48.52 | landley | Yes, I tested this. |
| 20:49.00 | landley | (toysh! When I say "bash replacement", I _mean_ it...) |
| 20:49.08 | prpplague | landley: hehe |
| 20:49.26 | landley | When i say "I broke echo", I'm not joking... |
| 20:49.42 | landley | I'm the guy who needs to _understand_ why the shell "input i" command can't be suspended with ctrl-z. |
| 20:50.14 | prpplague | RobotGuy: hard to since you are using a ton of his code on both the hammer board and your linux desktop |
| 20:50.45 | landley | Most people only test code under conditions it's expected to behave well in. |
| 20:50.57 | landley | I try to figure out what the nastiest evil corner cases are. |
| 20:51.11 | RobotGuy | landley: I am not one of those. I break things frequently. |
| 20:51.47 | landley | This is why busybox sed won't add an \n when editing a file where the last line doesn't end with \n. |
| 20:52.03 | landley | Unless you're editing multiple files, in which case it adds an \n to distinguish it from the first line of output in the next file. |
| 20:52.25 | landley | Unless no lines are output in the second file, in which case it still doesn't output an \n if the last line of the first file was output and didn't end with one. |
| 20:52.43 | landley | Now: ask yourself, when editing multiple files, what does the "match last line" search mean? |
| 20:52.49 | landley | The last line of each file? The last line of the last file? |
| 20:53.02 | RobotGuy | last line of current file? |
| 20:53.36 | ds2 | prpplague: how's the bounty work? wouldn't I have to buy one first then get it to work and if I get it, i'll have 2 kits? |
| 20:53.49 | landley | sed -e "blahblahblah" one.txt two.txt three.txt |
| 20:53.53 | landley | Define "current file". |
| 20:54.07 | landley | Yes, I made this work correctly. No, the spec doesn't give a _clue_ about this. |
| 20:54.20 | ds2 | prpplague: and what performance does it have to have? ;) |
| 20:54.26 | landley | er, sed -e "1,$blahblahblah" perhaps... |
| 20:54.36 | prpplague | ds2: performance isn't a big issue |
| 20:54.36 | landley | <PROTECTED> |
| 20:54.45 | landley | Once upon a time, there was /bin. |
| 20:54.54 | RobotGuy | landley: Why? |
| 20:54.57 | prpplague | ds2: basically if you buy the kit, get it working, we'll refund your money |
| 20:55.00 | landley | And then Ken and Dennis filled up their RK05 disk pack, and the OS leaked into /usr/bin. |
| 20:55.14 | landley | And then distros happened, and /usr/local/bin was introduced for locally installed software. |
| 20:55.17 | ds2 | ah |
| 20:55.25 | landley | And user home directories also commonly had a ~/bin once upon a time. |
| 20:55.42 | landley | And then some people decided that /usr/local/bin wasn't good enough, and added /opt. |
| 20:55.46 | ds2 | prpplague: any TFT display of my choice? |
| 20:55.55 | landley | It's only a matter of time. |
| 20:56.09 | ds2 | landley: isn't that a solaris-ism? |
| 20:56.23 | prpplague | ds2: yea, since it really doesn't matter, as long as the functionality is there |
| 20:56.27 | RobotGuy | landley: Well, I would have to have my thumb drive mounted as /usr to do a regular build. |
| 20:56.39 | landley | ds2: every directory people install software into becomes cluttered. |
| 20:56.42 | ds2 | prpplague: and you want it as a fb device I am guessing? |
| 20:56.49 | prpplague | ds2: correct |
| 20:56.57 | landley | And people decide to invent a _new_ directory to install software into, which won't be cluttered because nobody's using it yet. |
| 20:57.04 | ds2 | hmmm hack virtfb to do nasty stuff... }:-) |
| 20:57.04 | landley | As soon as people start using it, it gets cluttered. |
| 20:57.08 | landley | And they come up with a new one... |
| 20:57.14 | prpplague | ds2: http://www.sparkfun.com/commerce/product_info.php?products_id=8335 |
| 20:57.43 | ds2 | prpplague: the connectors is what I am don't like about it |
| 20:57.43 | prpplague | ds2: http://www.sparkfun.com/commerce/product_info.php?products_id=8490 |
| 20:57.51 | prpplague | ds2: about what? |
| 20:57.57 | ds2 | the PSP LCD |
| 20:58.29 | ds2 | I have a few documented LCDs at home that I want to get working but they all use funny connectors and I been too lazy to tack little wires on them |
| 20:58.56 | prpplague | ds2: ahh, np, you order the kit, we'll include a lcd board with an lcd |
| 20:59.20 | ds2 | prpplague: oh that means I have order the JTAG + serial break + everything else? |
| 20:59.32 | prpplague | ds2: yea |
| 20:59.41 | ds2 | that makes it less attractive... I was hoping to just invest in the hammer itself |
| 20:59.41 | *** mode/#edev [+o prpplague] by ChanServ |
| 21:00.21 | ds2 | that's a $300+ kit, isn't it? |
| 21:00.43 | prpplague | ds2: $249 |
| 21:01.34 | ds2 | prpplague: is there a wiki page describing the required functionality at completion? |
| 21:02.07 | prpplague | ds2: no, not at the momment, but i can generate one |
| 21:03.32 | ds2 | prpplague: list what is needed. it may be a specs-menship game in addition to actual work |
| 21:21.42 | CosmicPenguin | I am in such a crappy mood right now |
| 21:21.45 | CosmicPenguin | I just thought I would share |
| 21:21.58 | prpplague | CosmicPenguin: whats up bossman? |
| 21:30.59 | prpplague | landley: btw, i've been totally swampped at work and not feeling 100% well, just in case you are wondering about those patches |
| 21:31.01 | *** join/#edev ico2 (n=ico2@modem-3539.chimpanzee.dialup.pol.co.uk) |
| 21:32.30 | landley | No rush. I lost most of this week to something called "cedar fever", which is apparently some kind of pollen having nothing to do with cedar. :P |
| 21:33.10 | prpplague | landley: hehe, d00d, your jan15 blog entry reads exactly as if i'd written it |
| 21:33.31 | RobotGuy | Seems to be going around. I spent several hours at the ER last night - and they didn't even help me. |
| 21:34.18 | landley | RobotGuy: mine wasn't ER level, just really tired all week, as was Fade... |
| 21:34.23 | landley | And now, grocery store. Back in a bit... |
| 21:34.27 | prpplague | landley: later |
| 21:34.35 | ds2 | prpplague: what would it take to have a touch screen working on the hammer? |
| 21:34.56 | prpplague | ds2: just an adc , t0mw already has it working |
| 21:35.45 | ds2 | Hmmm... know if he has a board with the right connector for http://www.sparkfun.com/commerce/product_info.php?products_id=8448 |
| 21:36.17 | prpplague | ds2: he's using a cstn display with touchscreen |
| 21:36.31 | ds2 | with that, it would make a tempting project for a home automation control panel; wall powered so PM isn't a problem (that was my biggest thing with not using the hammer |
| 21:36.34 | ds2 | oh Hmmm |
| 21:37.14 | prpplague | ds2: http://store.earthlcd.com/EDMGRB8KJF?category=148 |
| 21:37.20 | prpplague | ds2: but the TS interface is the same |
| 21:37.40 | ds2 | prpplague: it is not the "Interface"; it is the connectors that has been concerned |
| 21:37.51 | ds2 | it is always the connectors that turn out to the hidden surprise in my projects :( |
| 21:38.08 | prpplague | ds2: yea, if you are refering to the TS panel from sparkfun, i have breakout boards underway |
| 21:38.28 | ds2 | prpplague: You have just made it a lot more interesting |
| 21:39.53 | ds2 | hammer + PSP LCD + touch screen + ethernet + 5V wall wart (or 5V from the unused ethernet pairs in a case the size of a wall thermostat |
| 21:40.06 | ds2 | running VNC client |
| 21:40.39 | ds2 | prpplague: does ALSA work on it yet? |
| 21:41.08 | prpplague | ds2: alsa as in audio? |
| 21:41.12 | ds2 | yeah |
| 21:41.20 | prpplague | ds2: the hammer doesn't have any audio support other than via pwm |
| 21:41.44 | ds2 | prpplague: is there an alsa driver for PWM output to a speaker via a single NPN (2N2222) driver? |
| 21:41.47 | prpplague | ds2: you can do a hiband filter to play audio, but the sample rate will be limited to around 22khz |
| 21:42.04 | prpplague | ds2: i don't know, but its something i've been meaning to look at |
| 21:42.07 | ds2 | don't care... it'd be for something like flite to play Home Automation related info |
| 21:42.15 | prpplague | ds2: right |
| 21:42.34 | ds2 | 8KHz telco quality would be good enough |
| 21:42.46 | ds2 | in fact, OSS driver would be even better... /dev/audio :) |
| 21:42.48 | prpplague | ds2: brb in 5 minutes |
| 21:48.05 | chouimat|work | prpplague: #3? |
| 21:48.07 | Mirell | Hmm.. |
| 21:48.21 | prpplague | chouimat|work: naw, but i have some at home |
| 21:48.31 | prpplague | chouimat|work: left over from my "tooth incident" |
| 21:48.33 | chouimat|work | prpplague: I'm currently on it :) |
| 21:49.03 | prpplague | chouimat|work: oh? cutting back on your crack usage to something a little lighter? |
| 21:50.35 | ds2 | an HA controller is a very tempting project |
| 21:53.29 | chouimat|work | prpplague: got a tooth removed yesterday ... about 1 hour with a jackhammer |
| 21:53.40 | prpplague | hehe |
| 21:54.52 | RobotGuy | I am still at a loss as to how to get software for hammer onto the thumb drive and be able to use it. /usr seems like a real pain. |
| 21:54.57 | ds2 | remotely done buy another incarnation of the hammer board? ;) |
| 21:55.14 | ds2 | nfsroot it |
| 21:55.20 | ds2 | test it then worry about flash |
| 21:55.28 | prpplague | RobotGuy: plug it into your desktop, mount it, copy of the files you need, umount it |
| 21:55.43 | prpplague | ds2: its going into a bot |
| 21:55.47 | prpplague | ds2: needs to be mobile |
| 21:56.06 | ds2 | prpplague: for testing, NFS root is fine. once you got the right software, just CP to it and switch back to flash root |
| 21:56.15 | prpplague | ds2: true as well |
| 21:56.17 | RobotGuy | prpplague: Did that once, but mounting the thumb as /usr would conflict with the initrd image, wouldn't it? |
| 21:57.19 | prpplague | RobotGuy: i'm sorry i don't understand your statement |
| 21:57.34 | ds2 | UnionFS ;) |
| 21:57.37 | RobotGuy | prpplague: The initrd has stuff on /usr |
| 21:57.52 | prpplague | RobotGuy: don't put stuff on /usr in the initr |
| 21:58.17 | prpplague | RobotGuy: or mount the thumb drive in a different location |
| 21:58.42 | RobotGuy | prpplague: I don't know how to make an image with just the /usr mount point (nothing in it). |
| 21:59.16 | ds2 | don't matter ,just make sure you don't need to access to initrd /usr |
| 21:59.30 | RobotGuy | I tried what landley said a few days ago, but did not get it to work |
| 21:59.37 | ds2 | a mount can go on a populated directory but you will loose access to it |
| 21:59.49 | ds2 | or if you must, get unionfs working |
| 22:00.34 | RobotGuy | The thumb image will have everything the initrd has on /usr. |
| 22:01.13 | ds2 | then just mount it |
| 22:01.18 | ds2 | it won't conflict |
| 22:01.22 | prpplague | RobotGuy: mount the initrd image, copy the contents of the /usr from the initrd to the thumb drive |
| 22:01.35 | RobotGuy | OK, I will give it a go then! I was affraid of mount conflics. |
| 22:01.51 | RobotGuy | prpplague: That's what I will do now. |
| 22:01.53 | ds2 | nah... mount will just make the stuff underneath it in accessible; it is otherwise harmless |
| 22:02.07 | prpplague | RobotGuy: give me a few minutes i'll give you a script |
| 22:03.06 | ds2 | I wish I was back in school... this would have made a nice special projects class that is good for credit :( |
| 22:06.04 | RobotGuy | BTW, my build scripts are working pretty good now - they save all the built stuff and configs to a save dir. |
| 22:13.24 | landley | RobotGuy: hmmm? |
| 22:26.41 | prpplague | RobotGuy: http://pastebin.ca/859091 |
| 22:27.03 | prpplague | RobotGuy: build one initrd with just the basics |
| 22:27.20 | prpplague | RobotGuy: build a second image with everything turned of but what goes in /usr |
| 22:27.32 | prpplague | RobotGuy: execute that script in the dir where the image is |
| 22:27.50 | prpplague | RobotGuy: i.e. /usr/bin/thumbpop.sh /dev/sdc1 |
| 22:31.57 | ds2 | sdc? |
| 22:32.02 | ds2 | Hmmm |
| 22:34.26 | RobotGuy | YAY! Thumb drive is now mounted at /usr with software! Didn't need the script, prpplague. |
| 22:36.59 | prpplague | RobotGuy: don't forget to use a initrd with those options disable, that way your initrd is as small as it can be |
| 22:39.37 | RobotGuy | prpplague: Which options disabled? |
| 22:40.03 | prpplague | RobotGuy: build one image with everything except what needs to go into /usr |
| 22:40.06 | RobotGuy | prpplague: OK, wait, I think I know - busybox and such, right? |
| 22:40.16 | prpplague | RobotGuy: build a second image with ONLY what needs to be in /usr |
| 22:40.35 | ds2 | that is so 'wasteful'... |
| 22:40.57 | prpplague | ds2: what is? |
| 22:41.27 | ds2 | 2 builds |
| 22:41.30 | RobotGuy | I just copied the stuff from /usr to my thumb drive root |
| 22:41.37 | prpplague | ds2: why is that wasteful? |
| 22:41.51 | prpplague | ds2: takes just 20 seconds to do |
| 22:42.08 | RobotGuy | I have scripts that do it all for me. |
| 22:42.11 | prpplague | RobotGuy: that works, but you still have initrd thats too large |
| 22:42.32 | ds2 | 1 build and just manually pick and pull =) |
| 22:42.41 | RobotGuy | I turn off the stuff in my initrd build like busybox, right? |
| 22:42.54 | prpplague | ds2: i think you are confusing builds with compiles |
| 22:43.01 | *** join/#edev Heet (n=lucas@a82-92-5-118.adsl.xs4all.nl) |
| 22:43.27 | RobotGuy | I have always used the 2 image build method so it's no big deal for me. |
| 22:43.32 | ds2 | or it is just me using dumpster special 300MHz PII's as build machines ;) |
| 22:43.47 | prpplague | ds2: hehe, no you are confusing builds with compiles |
| 22:43.55 | prpplague | ds2: you compile once, build twice |
| 22:43.56 | RobotGuy | I build on an Athlon XP 2800+ |
| 22:44.09 | prpplague | ds2: the build is just the pick and populate portion |
| 22:44.38 | RobotGuy | Yeah, I only compile once. |
| 22:44.41 | ds2 | it might be a gumstix-ism... but Is eem to recall it does both all the time |
| 22:44.44 | prpplague | RobotGuy: do one build with everything you need for the basic initrd |
| 22:44.49 | prpplague | RobotGuy: then turn everything off |
| 22:44.59 | prpplague | RobotGuy: then turn only the items you need to have in /usr on |
| 22:45.03 | RobotGuy | And add the software I want |
| 22:45.05 | prpplague | RobotGuy: then do the build |
| 22:45.30 | RobotGuy | I think I have it |
| 22:45.39 | prpplague | ds2: probably lame instructions |
| 22:45.59 | prpplague | ds2: if you simple remove the root dir thats populate it does the pick and place only |
| 22:46.24 | ds2 | could be |
| 22:46.33 | ds2 | my prefer build system is a bit more manual then what buildroot offers |
| 22:46.49 | ds2 | one day, I will have git setup and all these boards will be mere branches |
| 22:47.00 | ds2 | and userland will just dissolve in place |
| 22:49.32 | landley | ds2: actually, you don't lose access to the contents of a directory you overmount if you have a process which is using that directory as its cwd. :) |
| 22:49.39 | landley | In that case "." points to the overmounted directory... |
| 22:50.04 | ds2 | sure that still works? thought I tried that the other day and they closed that hole |
| 22:50.07 | landley | Old trick on shared unix systems for hiding things from the sysadmin... |
| 22:50.16 | landley | ds2: I last tried it... |
| 22:50.24 | landley | 2.6.19, I think? |
| 22:50.33 | landley | It's not _exactly_ a hole. |
| 22:50.46 | landley | Shared subtree might have changed the semantics. |
| 22:50.54 | prpplague | later folks |
| 22:50.57 | landley | bye prpplague. |
| 22:51.16 | ds2 | nope that won't work |
| 22:51.18 | ds2 | 2.6.21 |
| 22:51.33 | ds2 | lrwxrwxrwx 1 root root 0 Feb 4 23:23 /proc/364/cwd -> / |
| 22:51.45 | ds2 | symlinks are dereferenced at the time of usage not prior |
| 22:51.55 | ds2 | I was trying to use that to escape a chroot last week |
| 22:51.58 | landley | symlinks are, yes. |
| 22:52.35 | landley | I suspect you might now have to do all the 'openat" stuff with a filehandle opened before the mount. :) |
| 22:52.45 | landley | Close one door, another falls off the hinges... |
| 22:52.52 | ds2 | hahahah |
| 22:53.03 | landley | I_BREAK_STUFF. :) |
| 22:53.21 | landley | People keep telling me I should go into security... |
| 22:53.35 | ds2 | it was very handy when one is knee deep in chroots |
| 22:53.40 | landley | But I'm never comfortable thinking I've found _all_ the ways to break something. There's always more... |
| 22:53.46 | landley | I told you about my friend who found a bug in a zero byte file? |
| 22:53.54 | landley | I still can't top that. |
| 22:53.57 | ds2 | no |
| 22:54.16 | landley | Back on some old unix (irix, I think) they implemented "true" as an empty file with the executable bit set. |
| 22:54.18 | landley | Being clever. |
| 22:54.28 | landley | It was interpreted as a shell script, did nothing, and returned true. |
| 22:54.31 | ds2 | oh my |
| 22:54.36 | landley | Now try using that from .profile. |
| 22:54.40 | landley | Infinite loop. |
| 22:54.54 | landley | One bug, zero bytes: infinite bug density! |
| 22:55.24 | landley | That was bumped into by Lamont "Piggy" Yarroll, one of my old co-workers back when Timesys had an engineering department. |
| 22:55.30 | ds2 | all these bug closing is really wack-a-mole game |
| 22:55.32 | landley | I miss those days... |
| 22:55.52 | landley | This is why separate "stable" and "development" forks are nice. As well as a big regression test suite. |
| 22:56.31 | RobotGuy | What is tslib and what requires it? |
| 22:56.42 | ds2 | touch screen library |
| 22:57.01 | ds2 | touch screens requires it but X servers may be loading it just cuz it was config that way |
| 22:57.28 | RobotGuy | Something is requiring it for a build |
| 22:57.37 | ds2 | X11? |
| 22:57.42 | landley | Google said: http://tslib.berlios.de/ |
| 22:57.44 | RobotGuy | This is for a hammer build. Not X |
| 22:57.44 | landley | First hit. |
| 22:57.50 | ds2 | directfb? |
| 22:57.52 | ds2 | tscal? |
| 22:57.56 | *** join/#edev cajaka (n=none@cpe-24-58-193-200.twcny.res.rr.com) |
| 22:58.18 | landley | touch screen library |
| 22:58.50 | RobotGuy | I know it is touch screen library now. It was directfb requiring it. |
| 23:23.59 | *** join/#edev CosmicPe1guin (n=nobody@163.181.251.103) |
| 23:24.24 | *** mode/#edev [+o CosmicPenguin] by ChanServ |
| 23:32.00 | *** join/#edev ico2 (n=ico2@modem-3668.cheetah.dialup.pol.co.uk) |
| 23:55.25 | *** join/#edev jbevren (n=jbevren@dumt-04-044.dsl.netins.net) |
| 23:55.25 | *** mode/#edev [+o jbevren] by ChanServ |