| 00:53.26 | *** join/#edev davidc__ (n=davidc__@S01060002b360aacd.vc.shawcable.net) |
| 01:19.44 | *** join/#edev bsdfox (n=h36sa@unaffiliated/bsdfox) |
| 01:48.58 | *** join/#edev pocek (i=pock@87-205-72-142.adsl.inetia.pl) |
| 02:39.38 | *** join/#edev roxfan (n=dunno@91.182.229.85) |
| 03:29.39 | *** join/#edev JoeLlama (n=snork@206.169.63.10) |
| 04:12.58 | *** join/#edev MonMotha (n=monmotha@71.188.171.50) |
| 05:41.30 | *** join/#edev harpal (n=Harpal@121.246.75.165) |
| 06:12.44 | *** join/#edev hw_ (n=hw@p578b3905.dip0.t-ipconnect.de) |
| 06:12.48 | *** join/#edev tsjsieb (n=tsjsieb@dejongbeheer.nl) |
| 06:17.44 | *** join/#edev TodWulff (n=TodWulff@c-76-122-186-40.hsd1.mi.comcast.net) |
| 06:18.47 | TodWulff | Good day. Does anyone have knowledge of an OEM that has produced a MCU/SOC where they married ARMv7+ IP with the F-RAM memory? |
| 06:23.19 | *** join/#edev lyakh (n=lyakh@p57BD1B96.dip0.t-ipconnect.de) |
| 06:51.35 | *** join/#edev tsjsieb (n=tsjsieb@dejongbeheer.nl) |
| 07:03.29 | *** part/#edev TodWulff (n=TodWulff@c-76-122-186-40.hsd1.mi.comcast.net) |
| 07:11.26 | *** join/#edev mbuf (n=user@61.16.248.242) |
| 07:11.55 | *** join/#edev damunix (n=damunix@rps919.ovh.net) |
| 07:14.47 | *** join/#edev pat123 (n=pat@99-41-81-18.lightspeed.mssnks.sbcglobal.net) |
| 07:23.07 | *** join/#edev devilsadvocate (n=quassel@202.3.77.11) |
| 07:25.27 | *** join/#edev davidc__ (n=davidc__@S01060002b360aacd.vc.shawcable.net) |
| 07:45.33 | *** join/#edev JoeLlama (n=snork@206.169.63.10) |
| 08:09.34 | chiques | I have a general idea on how to use an IDE and c with my mcu. I've noticed that my MCU came with examples in asm format. Is it common for MCU users to mix c and asm examples? |
| 08:16.25 | roxfan | yes |
| 08:21.00 | *** join/#edev mbuf (n=user@61.16.248.242) |
| 08:23.01 | chiques | Must my IDE support assembly or can I link it somehow? |
| 08:30.37 | *** join/#edev damunix_ (n=damunix@rps919.ovh.net) |
| 09:10.20 | *** join/#edev damunix (n=damunix@rps919.ovh.net) |
| 09:12.41 | *** join/#edev florian_kc (n=fuchs@port-217-146-132-69.static.qsc.de) |
| 10:09.47 | *** join/#edev pirho (i=pirho@gateway/gpg-tor/key-0x2CEEC9CB) |
| 11:13.12 | *** join/#edev harpal_ (n=Harpal@121.246.75.165) |
| 11:20.47 | *** join/#edev harpal_ (n=Harpal@122.169.108.196) |
| 11:22.11 | *** part/#edev harpal_ (n=Harpal@122.169.108.196) |
| 11:31.34 | *** join/#edev likewise (n=Leon_Woe@atwork-193.r-212.178.107.atwork.nl) |
| 11:36.10 | *** join/#edev harpal__ (n=Harpal@121.246.75.165) |
| 11:38.35 | *** join/#edev g1powermac (n=g1powerm@unaffiliated/g1powermac) |
| 11:46.27 | *** part/#edev harpal__ (n=Harpal@121.246.75.165) |
| 12:24.04 | *** join/#edev miknix (n=miknix@gentoo/developer/miknix) |
| 12:34.01 | *** join/#edev linac (n=lin@122.90.101.10) |
| 13:22.10 | *** join/#edev g1powermac_EPC (n=g1powerm@unaffiliated/g1powermac) |
| 13:40.58 | *** join/#edev methril|work (n=Methril@213.27.233.98) |
| 13:42.33 | *** join/#edev prpplague (n=dave@mail.americanmicrosystems.com) |
| 13:45.13 | *** mode/#edev [+o prpplague] by ChanServ |
| 13:51.04 | methril|work | yo |
| 13:52.50 | prpplague | methril|work: greethings |
| 13:53.21 | methril|work | prpplague: greetings!! nice to see you again here :) |
| 14:00.06 | chouimat|work | morning |
| 14:00.42 | *** join/#edev GPSFan (n=kenm@64.92.145.112) |
| 14:00.42 | *** mode/#edev [+o GPSFan] by ChanServ |
| 14:01.11 | prpplague | GPSFan: greetings |
| 14:01.23 | GPSFan | prpplague: g'day |
| 14:01.56 | *** join/#edev roxfan (n=dunno@91.182.229.85) |
| 14:29.22 | *** join/#edev damunix (n=damunix@rps919.ovh.net) |
| 14:42.47 | *** join/#edev roxfan (n=dunno@91.182.229.85) |
| 14:50.39 | *** join/#edev kergoth (n=kergoth@98.167.215.86) |
| 14:58.05 | *** join/#edev damunix_ (n=damunix@rps919.ovh.net) |
| 14:59.26 | *** join/#edev kergoth_ (n=kergoth@98.167.215.86) |
| 15:08.13 | *** join/#edev kergoth (n=kergoth@ip98-167-215-86.ph.ph.cox.net) |
| 15:17.16 | *** join/#edev JoeLlama (n=snork@206.169.63.10) |
| 15:17.35 | *** join/#edev kergoth_ (n=kergoth@ip98-167-215-86.ph.ph.cox.net) |
| 15:20.24 | prpplague | methril|work: thanks bud, glad to be back |
| 15:21.10 | methril|work | prpplague: you're wellcome :) You are the soul of the channel ;) |
| 15:21.29 | prpplague | methril|work: hehe, i don't know about that, but thanks anyway, hehe |
| 15:21.52 | prpplague | wonders if people just come to hear me whine and complain |
| 15:21.56 | methril|work | prpplague: i see a lot of silence when you wasn't here :) |
| 15:21.57 | solar | he might be right. When you are /away the traffic in here is cut down to 1/5th at best |
| 15:22.12 | methril|work | nods last solar sentenc |
| 15:24.23 | prpplague | solar: and the other 4/5 is me complaining about whatever i'm working on |
| 15:24.48 | solar | heh. Not really. It just gets quiet around here. |
| 15:25.18 | methril|work | anyway is good to se you back :) |
| 15:25.25 | prpplague | methril|work: hehe thanks |
| 15:26.01 | prpplague | well, probably get more busy around here in coming months |
| 15:27.01 | chouimat|work | ARRGGG |
| 15:28.26 | prpplague | chouimat|work: spill hot coffee while trying to download pr0n? |
| 15:29.24 | prpplague | i need to get back to posting stuff on elinux.org |
| 15:29.51 | prpplague | i noticed that the Didj isn't selling well, wal-mart has pulled them from the shelves and target has them marked down to $69.00 |
| 15:34.36 | *** join/#edev workrat (n=chatzill@adsl-76-234-174-169.dsl.sfldmi.sbcglobal.net) |
| 15:36.52 | chouimat|work | prpplague: no new board is pissing me off |
| 15:38.51 | prpplague | chouimat|work: ahh |
| 15:39.50 | chouimat|work | at least I don't have to do it for last week so I'm ok :) |
| 16:43.12 | *** join/#edev brolin (n=brolin@200.24.16.55) |
| 16:53.38 | prpplague | i think this is the attitude that alot of developers in the OSS community have: http://www.wordinfo.info/words/index/info/view_unit/1/?letter=B&spage=3 |
| 16:53.45 | *** join/#edev landley (n=landley@adsl-99-169-44-161.dsl.austtx.sbcglobal.net) |
| 16:53.45 | *** mode/#edev [+o landley] by ChanServ |
| 16:53.57 | prpplague | landley: hey bud |
| 16:54.00 | landley | waves at prpplague. |
| 16:54.02 | landley | How goes? |
| 16:54.16 | prpplague | landley: its going, still hanging in for the next round of layoffs |
| 16:54.19 | prpplague | <prpplague> i think this is the attitude that alot of developers in the OSS community have: http://www.wordinfo.info/words/index/info/view_unit/1/?letter=B&spage=3 |
| 16:54.27 | prpplague | landley: see if you agree |
| 16:54.38 | m4t | hey |
| 16:54.43 | m4t | any of you guys have a dht-walnut? |
| 16:55.10 | prpplague | m4t: there are a number of folks in the channel that do, however you may have to wait for someone to reply |
| 16:55.37 | m4t | cool, i need to find out what size barrel plug they use for their dc input |
| 16:55.50 | prpplague | landley: i'm still struggling with the wince api's |
| 16:56.00 | prpplague | m4t: should be on the wiki |
| 16:56.11 | *** join/#edev albech (n=albech@124.157.238.9) |
| 16:56.14 | prpplague | or atleast i think it is |
| 16:57.13 | landley | prpplague: the dns here is screwed up and not resolving www.wordinfo.info. Sorry. :P |
| 16:57.27 | landley | (I can't get it to resolve landley.net either, which is a bit awkward, but I know the IP of that one.) |
| 16:58.02 | prpplague | m4t: http://elinux.org/DHT-Walnut |
| 16:58.06 | landley | There it went (on the ffifth try). |
| 16:58.06 | prpplague | landley: doh |
| 16:58.07 | m4t | prpplague 12v is there, the picture shows a 2a fuse |
| 16:58.19 | m4t | but i read through a few dht-walnut pages |
| 16:58.25 | m4t | and none of them show the size in mm |
| 16:58.42 | landley | prpplague: yes, I'm familiar with the blind man and the elephant fable. (Since second grade.) |
| 16:59.29 | prpplague | landley: i guessed that most people would be familiar with it, just curious if you find that is how the OSS community is |
| 17:00.12 | landley | Insular to the point of xenophobia? Of course. More BSD than Linux, though. A lot of Linux guys came from Java and OS/2 and such. And these days, an awful lot of OSS developers are encountering the macintosh. |
| 17:00.28 | landley | Mirell's learning objective C, for example. |
| 17:00.50 | landley | Val Henson uses a macbook air as her primary development platform. (Running MacOS X.) |
| 17:01.14 | landley | Last Ottawa Linux Symposium I was at, I'd estimate at least 1/3 of the laptops were macs. Probably gone up since then. |
| 17:01.27 | prpplague | interesting |
| 17:01.52 | landley | If you're referring to "my part of the system is the important part, and everything's really just a variation on what I do", that's why I HATE GIT's UI. |
| 17:02.10 | prpplague | yea |
| 17:02.28 | landley | Mercurial was written by people who understood userspace and source control user interfaces. Git was written by people who did in-kernel filesystem development and thought UI was somebody else's problem. |
| 17:03.13 | prpplague | hehe indeed |
| 17:03.37 | prpplague | landley: whatcha working on today? |
| 17:11.57 | prpplague | ho ho hum |
| 17:12.39 | landley | prpplague: finally biting the bullet and upgrading FWL to use the last GPLv2 gcc release, 4.2.1. |
| 17:12.56 | prpplague | landley: lovely |
| 17:12.56 | landley | Then I need to swap in the new kernel release and update my perl removal patches and submit 'em upstream while the merge window's open. |
| 17:13.04 | landley | 4.2.1 breaks things. Sigh. |
| 17:13.12 | prpplague | landley: you ever use the nail board anymore? |
| 17:13.18 | landley | It didn't until the canadian cross stuff went in. The canadian stuff _shatters_ when you switch gcc versions. Grrr. |
| 17:13.28 | landley | prpplague: no, but I've got it with me in my backpack. |
| 17:13.39 | landley | I haven't had an excuse. |
| 17:13.54 | landley | I thought tincantools had semi-closed-down with the recession and switch of focus to wince. |
| 17:13.55 | prpplague | yea i figured as much |
| 17:14.06 | prpplague | landley: two different companies |
| 17:14.15 | landley | I'd be happy to help push it, and provide software for it, if I saw anybody else playing with it. |
| 17:14.23 | landley | (Part of it's I've been offline a lot the past month or two.) |
| 17:14.32 | prpplague | landley: AML is scaling back, but TinCanTools is still doing lots of biz |
| 17:14.52 | landley | Cool. |
| 17:15.10 | landley | Did the nail board actually sell? Last I heard there was no real marketing or anything behind it... |
| 17:15.11 | prpplague | landley: yea, i'm kinda of ambivalent about TinCanTools right now |
| 17:15.30 | prpplague | landley: yea its been selling pretty good from what i hear |
| 17:15.36 | landley | Did you notice that FWL grew hw- targets a few months back? |
| 17:16.00 | prpplague | landley: like you, i've been kinda offline for awhile trying to get things in order |
| 17:16.06 | landley | We can now specify different kernels for existing toolchain/root filesystem combinations. Hardware configurations with a "base configuration". :) |
| 17:16.14 | prpplague | landley: after i finish this wince stuff i hope to get back doing some personal projects |
| 17:16.20 | landley | Woot. |
| 17:16.24 | landley | Keep me posted on that, you have my email. |
| 17:16.29 | landley | You've heard about my pet project qcc, right? |
| 17:16.37 | prpplague | landley: my father left me some money, so i want to try to do something with it |
| 17:16.47 | prpplague | landley: yea we discussed it the other day |
| 17:17.33 | landley | prpplague: sorry to hear about your father. |
| 17:17.49 | prpplague | landley: thanks |
| 17:18.10 | landley | I need to get FWL caught up, and toybox properly dealt with, so I have _time_ to properly poke at qcc. |
| 17:18.22 | landley | I've been going over toybox to see what I need to submit upstream to busybox. |
| 17:18.44 | landley | I've been porting my patch command (although I haven't touched it in a week now), and I need to do work on netcat too... |
| 17:18.49 | prpplague | http://www.legacy.com/obituaries/news-journal/obituary.aspx?n=patrick-j-anders-pat&pid=127654401 |
| 17:19.11 | prpplague | landley: you going to pass stuff up to busybox? |
| 17:19.14 | landley | What brought up the elephant thing for you, by the way? |
| 17:19.21 | landley | prpplague: that's the general idea, yeah. |
| 17:19.38 | landley | Busybox has an active development community, I don't have _time_ to give toybox the attention it needs. |
| 17:19.57 | prpplague | landley: just thinking about all the mailing list battles i've seen recently |
| 17:20.10 | landley | Yes I could reproduce everything busybox does, if I wanted it to be my full time hobby to the exclusion of all else for the next 5 years. |
| 17:20.19 | landley | Busybox? I haven't been reading the list much. |
| 17:20.57 | landley | If you've read my blog you'll notice I've finally given up on desktop linux amounting to anything, which means Linux is going to be stuck as a niche product indefinitely ala OS/2 or the Amiga. |
| 17:21.08 | landley | Sure it's big in embedded, but so was the Z-80. |
| 17:23.05 | landley | What mailing list battles were you referring to? |
| 17:35.38 | prpplague | landley: nothing specific, just stuff i see between RMK and some of the other subsystem maintainers |
| 17:40.37 | kergoth_ | after using my macbook for most daily tasks, i cannot stand how annoying and unintuitive my Linux desktop at work is. I'm much happier using an Mac OS X laptop and windows 7 desktop and both with ssh connections to a Linux box for embedded coding |
| 17:40.38 | kergoth_ | heh |
| 17:42.14 | prpplague | kergoth_: funny, i'm doing almost all of my wince dev from a linux desktop |
| 17:42.23 | prpplague | kergoth_: rdesktop to a winblows box |
| 17:42.52 | prpplague | but then, i'm much less demanding on the desktop than most people |
| 17:44.11 | prpplague | <landley> Sure it's big in embedded, but so was the Z-80. |
| 17:44.33 | prpplague | yea well if the GUI stuff doesn't keep up with the core linux kernel and utils, it won't last there either |
| 17:45.43 | prpplague | landley: our share of terminals running linux has dropped of dramatically in the last 2 years due to the fact the gui and development environments aren't as easy for VARs to use as the wince stuff |
| 17:48.10 | ds2 | morning |
| 17:49.56 | prpplague | ds2: greetings |
| 17:50.02 | prpplague | ~lart ds2 |
| 17:50.02 | purl | squeezes ds2 till ds2 turns blue like papa smurf |
| 17:50.14 | ds2 | blue enough already :D |
| 17:50.19 | prpplague | ~emulate abcs |
| 17:50.19 | purl | if we can't sell it for 500% markup we aren't interested! |
| 17:51.00 | prpplague | ho ho hum |
| 17:51.01 | ds2 | :) |
| 17:51.11 | prpplague | wonders if he will ever be done with this wince stuff |
| 17:51.14 | prpplague | ~emulate abcs |
| 17:51.14 | purl | selling used products as new isn't unethical! |
| 17:51.20 | ds2 | getting irked at google more and more each day |
| 17:51.32 | ds2 | prpplague: did you ever get a chance to ping bill? |
| 17:52.02 | prpplague | ds2: no not yet |
| 17:52.33 | ds2 | searching for a DRG (a topo map format) viewer returned tons of hits... for a DWG viewer @$@#$@#%@#$@#$@#$ blasted peice of crap |
| 17:52.33 | prpplague | ds2: i'm up against a friday EOD deadline for this wince stuff |
| 17:52.44 | ds2 | prpplague: EOD for you means 1700? |
| 17:53.03 | prpplague | ds2: 18:30 |
| 17:53.34 | ds2 | out here, it can easily mean 2359 ;) |
| 17:53.40 | prpplague | hehe |
| 17:54.04 | prpplague | ds2: yea well this is the first place i ever worked that actually closes and locks the doors at 18:31 |
| 17:54.42 | prpplague | ds2: we have a major retailer that is evaluating a product but needs wince on it |
| 17:55.16 | ds2 | prpplague: that is pretty strange... every place I been to, I was given a key for 24/7... they even did it for the contract positions |
| 17:55.32 | prpplague | ds2: yes indeed |
| 17:55.41 | ds2 | prpplague: thought Wince is just drop and go with the ready made BSPs? |
| 17:55.59 | prpplague | ds2: no more so than linux is |
| 17:56.38 | ds2 | prpplague: heh... I have been hearing all about how painful Linux is ;) |
| 17:56.48 | ds2 | and supposely the BSPs solve all that |
| 17:56.50 | prpplague | ds2: just like in linux for the s3c2410 you have the smdk2410 bsp, but that doesn't mean it has everything configured for the Hammer, for example |
| 17:57.12 | prpplague | ds2: it all depends on what level you are on |
| 17:57.13 | landley | ds2: I've hit their "near miss" stuff a lot myself. Every time I search for something I've done, I have to add "-langley" to the search terms. |
| 17:57.17 | ds2 | prpplague: I sort of expected that but I was told it was not the case |
| 17:57.28 | prpplague | ds2: one you have a "BSP" for a hardware platform, wince is very simple to configure |
| 17:58.10 | ds2 | landley: so I am not the only one... I seem to run into it 80% of the time and usually don't notice til I am deep wading through links and wonder why it doesn't seem to resembly anything like what I was expecting |
| 17:58.23 | ds2 | prpplague: I see. |
| 17:58.55 | prpplague | ds2: the feature that people like about the wince BSP's is that it is treat much like a project file, something you'd see in orcard |
| 17:59.04 | prpplague | s/treat/treated |
| 17:59.19 | prpplague | ds2: file->open->project |
| 17:59.53 | landley | ds2: the ability to -exclude search terms you don't want is useful. |
| 18:00.01 | prpplague | ds2: once the project is open, you can basically drag-n-drop the features you want |
| 18:00.02 | landley | It's a workaround for google being _broken_, of course... |
| 18:00.13 | ds2 | prpplague: so if someone were to do that for Linux, would people want it then? |
| 18:00.20 | landley | Software that tries to "help in the kitchen" is almost never a good thing. (Think "clippy".) |
| 18:00.22 | prpplague | ds2: i would think so |
| 18:00.56 | landley | Gotta wrap up soon, battery low. |
| 18:01.00 | ds2 | landley: indeed... |
| 18:01.01 | prpplague | landley: later |
| 18:01.04 | ds2 | later landley |
| 18:01.28 | chouimat|work | I knew it won't last ... the office was quiet for 2 minutes .... |
| 18:01.41 | ds2 | prpplague: whatever happen to Linux is great because you can do anything you want with it? the only way to resolve the difference is to build a genie |
| 18:01.48 | prpplague | ds2: the attitude with VARs is that they don't want to know about the guts of the OS, they just want to be able to customize the device with the features they need without learning a bunch of stuff |
| 18:02.20 | prpplague | ds2: it is still that way, but i think that there should besome layers developed |
| 18:02.33 | ds2 | layers == overhead |
| 18:02.40 | prpplague | ds2: the arduino stuff is a prime example, but on a smaller scale |
| 18:02.46 | ds2 | once you do that, you toss out the low end support |
| 18:03.49 | prpplague | ds2: not really, the support is there, its no more different that what OE does |
| 18:04.16 | prpplague | ds2: just more of a application feel to it, like you were working with a powerpoint presentation |
| 18:04.42 | ds2 | prpplague: and that is bad, IMO.... what if you needed to build a device that runs in 2M of RAM? |
| 18:05.01 | ds2 | (not that he current 2.6 folks are doing anything to kill that off </sarcasm>) |
| 18:05.24 | prpplague | ds2: that is completely doable, you can do that with wince, you have various modes and base builds to work with |
| 18:06.33 | ds2 | I donno.... I am not convinced OE makes anything easier |
| 18:06.52 | ds2 | if anything, it compounds issues and introduces its own problems |
| 18:07.36 | prpplague | ds2: hehe |
| 18:08.00 | prpplague | ds2: well now having some experience with the wince build process, i do see why alot of people are more comfortable with it |
| 18:08.22 | ds2 | prpplague: I did just as an experiment - run the OE/Angstrom build on the Overo and the Beagle platforms.... the Overo blows up... the Beagle finishes after a lot of tweaking and alot of time |
| 18:08.40 | ds2 | and the result is a lot of stuff I did not care for |
| 18:08.51 | prpplague | ds2: indeed, one of my major complaints of OE |
| 18:09.13 | prpplague | ds2: which i have discussed with kergoth_ many many times |
| 18:09.46 | ds2 | prpplague: then there is the abuse of git that OE does |
| 18:09.59 | prpplague | yep |
| 18:10.07 | ds2 | prpplague:how long does it take to do a build in CE? |
| 18:10.20 | kergoth_ | abuse of git? what are you talking about? |
| 18:10.26 | ds2 | if it is quick, I might have someone run through it |
| 18:11.00 | ds2 | kergoth_: I think someone described in a mailing list but - right now rather then using git branches etc... OE pulls multiple repos and manually patch things in |
| 18:11.04 | kergoth_ | also, small images and avoiding pulling in unncessary crap for images is improving, now that people are working on the micro distribution |
| 18:11.06 | prpplague | ds2: clean sysgen build is about 25-35 minutes, bsp build is about 2 minutes |
| 18:11.17 | kergoth_ | ds2: uh, no |
| 18:11.26 | ds2 | kergoth_: so you are saying the Angstrom distro is wrong? |
| 18:11.28 | kergoth_ | the OpenEmbedded metadata is stored in a single git repository |
| 18:11.45 | kergoth_ | I'm sayign that angstrom has a different target audience than the folks that hang out in this channel likely expect |
| 18:11.48 | ds2 | no... not OE but the way it builds things... like U-boot |
| 18:12.00 | kergoth_ | everything built by OpenEmbedded is fetched down and patched and built |
| 18:12.06 | kergoth_ | how its fetched varies from recipe to recipe |
| 18:12.17 | kergoth_ | it isn't abuse, it's the way every single recipe works, not unlike buildroot, actually |
| 18:13.10 | ds2 | they way they have done the U-boot build makes it very difficult to find the issues (most likely it is an upstream thing) |
| 18:13.29 | ds2 | s/^they/the/ |
| 18:14.04 | kergoth_ | so you're attributing "oe abuse of git" due to how one recipe among like 6,000 is done? |
| 18:14.08 | kergoth_ | brilliant, truly |
| 18:14.22 | chouimat|work | :) |
| 18:14.32 | ds2 | fair counter argument... what is a distro that does it better then? |
| 18:14.44 | kergoth_ | distro defines global policy |
| 18:14.49 | kergoth_ | it isn't going to magically fix a single recipe |
| 18:15.05 | ds2 | and it isn't just one.... it just happens to be this one is blowing up :D |
| 18:15.08 | kergoth_ | the micro distro will result in images that pull in less extra stuff, because the priority is that of small images |
| 18:15.33 | kergoth_ | once again, it just pulls down from a source and patches it. there's nothing special about fetching from git as opposed to any other fetch method |
| 18:16.44 | ds2 | sure there it is, it clones, wipes out all the stuff except for metadata... bring it back, and does all sorts of other weirdness to it |
| 18:16.52 | ds2 | anyways |
| 18:17.23 | kergoth_ | OpenEmbedded has its issues, git isn't one of them. |
| 18:19.23 | kergoth_ | i think what you'd prefer is probably a general improvement for the active developer use case. instead of fetch+patch and doing so buried deep in tmp/work, put the recipe in the source tree, pull that down and build it *there*, inside the source tree, in a working copy of whatever scm they use upstream... or something along those lines.. it's something I'm hoping MV will let me fix :\ |
| 18:20.28 | kergoth_ | shrugs |
| 18:21.49 | ds2 | I'd prefer it to be more troubleshootable |
| 18:22.03 | kergoth_ | in what way? |
| 18:22.15 | ds2 | and to do that, I rather it not do the git juggling but rather keep the tree intact and apply things layer by layer |
| 18:22.35 | kergoth_ | ah |
| 18:22.42 | ds2 | like I said, they way it is setup to use GIT is it goes in, clones the repo... then wipes out all but the metadata and then goes to do various integrations |
| 18:23.08 | ds2 | instead, if it had created alocal branch and started growing from there, any failures can be tracked and a figure pointed to |
| 18:23.09 | kergoth_ | the reason it works the way it does is that a tarball is produced after the fetch from upstream, that tarball can then be fetched from mirrors |
| 18:23.42 | ds2 | a prestine tarball or a patched one? |
| 18:23.46 | kergoth_ | pristine. |
| 18:23.56 | kergoth_ | then that tarball is what's unpacked, then patched in those tasks |
| 18:24.01 | ds2 | why is that necessary? |
| 18:24.22 | kergoth_ | many people want all sources fetched from their own mirrors, that's why the tarball is necessary |
| 18:24.47 | ds2 | I see... why not just keep the git metadata around? |
| 18:24.49 | kergoth_ | the use of that tarball in the build really isn't, its just that there are assumptions that it's fetch, then unpack, then patch in the system. it's a good idea though, i'll add it to my list and see what can be done about it |
| 18:24.58 | kergoth_ | that's a good question |
| 18:24.59 | kergoth_ | :P |
| 18:25.13 | kergoth_ | thanks for taking the time to explain it, thats helpful |
| 18:25.18 | ds2 | 'k |
| 18:26.05 | ds2 | from my experiment - It appears to me that OE either works or you are AFU because it is a lot of data to go through (yes, I know about the verbose flags to bitbake) |
| 18:26.40 | ds2 | it gives a lot of data for people who are curious but doesnt seem to give me the real data w/o me digging through it (i.e. ./configure failures) |
| 18:27.35 | kergoth_ | I've been trying to explain how the pieces fit together to people lately. it helps if you understand that the original intent was for the configuration metadata to be layered. most generic to most specific, with later overriding earlier. bitbake.conf includes separate files for each layer, in that order. target arch, target os, distro, machine, site, local, ... then with OVERRIDES, you can do conditional variable overriding at each of t |
| 18:29.18 | *** join/#edev sav1o (n=savio@jagat.xored.org) |
| 18:31.18 | kergoth_ | if you want to see a ./configure failure |
| 18:31.23 | kergoth_ | just set BBINCLUDELOGS = "yes" |
| 18:31.29 | kergoth_ | if a task fails, it automatically dumps the log file from that task |
| 18:32.03 | kergoth_ | the corresponding run.* script that goes with it is the exact task that was run that produced that log, you can do a bitbake -c devshell to drop into the source tree of the package within the bitbake environment, with our vars set |
| 18:32.29 | ds2 | that sounds useful |
| 18:33.05 | kergoth_ | it is, very, i often drop in there and use quilt to generate new patches, and copy them out, then add them to the recipe to get them used |
| 18:33.31 | kergoth_ | relative to the source tree, the logs and run scripts are in ../temp, you can run them, or even source them, in the dev shell, to reproduce the task execution |
| 18:33.38 | ds2 | w/o that it is a huge blackbox |
| 18:34.04 | ds2 | I found the logs by rummaging through the things it generates |
| 18:34.14 | kergoth_ | when a task fails, it gives you the path to the log |
| 18:34.29 | kergoth_ | i agree, i hate where they're located though, buried deep in tmp/, its a maze in there |
| 18:34.34 | kergoth_ | like 42 tab completes later, I'm at the log.. |
| 18:34.40 | kergoth_ | heh |
| 18:35.09 | ds2 | unless I record exactly what is printed, I have to guess at the right one since rerunning things generates yet another log |
| 18:35.16 | prpplague | imho build systems are a rather complex beast to begin with, the more you try to be everything to everyone, the more of a mess it will become |
| 18:36.19 | ds2 | what really needs to happen first is libtool/configure/pkgtool/etc needs to go away |
| 18:36.24 | kergoth_ | our recipes are really quite simple, rarely messy at all, unless you look at something like gcc. the classes are fairly simple too, and the entirety isn't that much of a mess, it just seems that way before you wrap your head around the thing |
| 18:36.28 | ds2 | then the build toolss can be simplier |
| 18:36.33 | prpplague | doesn't make much sense to add screwdriver tips to a hammer, does make sense to have difference claws to a hammer to remove different types of nails |
| 18:36.50 | kergoth_ | without configure, you'd just need a different tool to do the same thing, run tests and let the user manipulate the build |
| 18:37.14 | ds2 | kergoth_: while you are educating us... is there a way to ask it for a given package, what is the recipe file used (full path relative to the top of the recipes directory)? |
| 18:37.33 | prpplague | kergoth_: in a round about way, it sounds like ds2 is having same issues with OE i have when it comes to doing custom builds |
| 18:37.53 | ds2 | not really, things worked fine before autoconf |
| 18:38.18 | kergoth_ | the problem with ditchinig autoconf is then tools like OpenEmbedded have to figure out how to exert contorl over the build, differently, for every package |
| 18:38.24 | kergoth_ | you have to go wading through the makfeils and figure out whats what |
| 18:38.40 | kergoth_ | with autoconf, things suck, but they suck in a way that can be consistently worked aorund, and its a known quantity |
| 18:38.45 | ds2 | it isn't like the current ./configure stuff is that consistant |
| 18:38.53 | kergoth_ | it almost always is, though |
| 18:38.59 | kergoth_ | working around cross issues is easy |
| 18:39.04 | ds2 | I much rather have the old style config.h |
| 18:39.35 | kergoth_ | well, again, thatd generate a lot of extra work for distro folk, every package you'd have to figure out how i make *this* buildsystem obey my cflags, or turn off this piece of functionality |
| 18:39.49 | kergoth_ | arguably, they should know that anyway, though |
| 18:40.06 | kergoth_ | :) |
| 18:40.11 | ds2 | that sounds more like you have relented to ./configure and given up that battle |
| 18:40.19 | prpplague | kergoth_: you really should take a little time and have a short look at the wince platform builder |
| 18:40.23 | ds2 | things used to be SO much simplier |
| 18:40.55 | ds2 | things actually worked and you spend your time figuring out real problems not fighting configure/libtool/pkgtool |
| 18:41.08 | prpplague | ds2: yea, men were men, women were women, and we used buildroot :P |
| 18:41.31 | ds2 | prpplague: I predate buildroot so even that wasn't that useful ;) |
| 18:43.05 | prpplague | ds2: i too was doing linux embedded builds before buildroot |
| 18:43.26 | prpplague | heck i used to have some monster shell scripts |
| 18:43.34 | ds2 | heh |
| 18:44.06 | prpplague | checks his cvs web browser |
| 18:45.52 | prpplague | ds2: hehe, the first kernel i have in repo for a linux based product only has 148 machine device ids in the mach-types file |
| 18:46.12 | ds2 | prpplague: ARM kernel? |
| 18:46.14 | kergoth_ | there's a lot to be said for consistency. the ugliest, most painful builds in OE don't even use autoconf |
| 18:46.17 | kergoth_ | they're things like perl |
| 18:46.24 | kergoth_ | whose own ./Configure is .. ugh |
| 18:46.28 | prpplague | ds2: yes - Last update: Thu Dec 20 11:08:44 2001 |
| 18:46.31 | ds2 | well, perl does not like being x-compiled |
| 18:46.53 | prpplague | ds2: oh wait here is another one, november 1998, just 56 entries |
| 18:46.59 | ds2 | prplague: heh... I cut my teeth on x86 embedded stuff |
| 18:47.20 | *** join/#edev damunix (n=damunix@72.184.204-77.rev.gaoland.net) |
| 18:47.26 | prpplague | ds2: as did i, but ARM is where i found my passion |
| 18:47.49 | ds2 | prpplague: We do indeed have similar roots and interests |
| 18:48.28 | ds2 | prpplague: (different topic) - does it snow in you area? |
| 18:48.43 | prpplague | ds2: once every 2 or 3 years |
| 18:48.52 | kergoth_ | i think more people who try to get into OpenEmbedded should just create their own collection that only has the recipes they want, and stop using upstream OpenEmbedded directly completely. they branch off at that point, and it'll take much less time to parse, and be much less annoying, and less likely to break from underneith you. better to cherry pick bits into that collection instead |
| 18:49.18 | kergoth_ | i kind of miss arm |
| 18:49.25 | ds2 | prpplague: hmmm so if I come through there in the winter, is there anything bad? (mud, heavy rain, etc) |
| 18:49.33 | prpplague | kergoth_: agreed, the problem is that most people don't know how to properly cherry pick |
| 18:49.54 | solar | you need a tool for it |
| 18:50.06 | prpplague | ds2: just cold and dreary |
| 18:50.43 | kergoth_ | i created a new collection with a git repository and generated a fake merge commit at the point that i branched off, so git knows where i started from, then i can just use git log and git cherry for the most part to review what's new upstream in the recipes i copied |
| 18:50.52 | kergoth_ | works fairly well, but git's cherry picking tracking is weak |
| 18:50.55 | ds2 | prpplague: what's cold? thinking of doing a long road trip in the winter out to florida and back and camping along the way |
| 18:51.03 | kergoth_ | if a cherry pick results in you having to fix something first, its not seen as the same commit as the other anymore |
| 18:51.15 | prpplague | ds2: 30F to 50F |
| 18:52.46 | prpplague | kergoth_: yea the one thing that i do really like about the wince side is that i never have to worry about the userland apps building |
| 18:54.35 | kergoth_ | oftentimes things build inconsistently due to build environment, even when crosscompiling, because of autoconf enabling features based on something existing, rather than being more explicit |
| 18:54.39 | kergoth_ | so the builds aren't deterministic |
| 18:54.45 | kergoth_ | they build based on whats available.. bad |
| 18:55.07 | kergoth_ | *some* autoconf setups you can explicitly override everything, but not all |
| 18:55.51 | ds2 | and some just ignore it even though it acts like you can override |
| 18:57.17 | kergoth_ | true, that's a good point |
| 18:57.41 | ds2 | the most insidious thing I have found is libtool subverting stuff |
| 18:57.44 | kergoth_ | they try to make everything automatic and transparent, and lose determinism as a result, it's bad behavior on the part of upstream buildsystem maintainers.. course, building things in chroots helps |
| 18:58.01 | kergoth_ | it helps if you patch libtool to stop leaking vars from .la files into the library search path |
| 18:58.28 | ds2 | it used to be simple.... -L's and -l's in the LDFLAGS and you are done |
| 18:58.47 | ds2 | now it wants stuff embedded inside files which may reside in different places then where it will eventually run |
| 18:59.04 | kergoth_ | you know, dolt is cool, but i bet we could make something like it that works for the link, as opposed to just the compile bits that dolt does |
| 18:59.18 | kergoth_ | especially if you're on a linux-elf system, knowing you have a sane linker |
| 18:59.46 | kergoth_ | ds2: for pkg-config, it supports sysroot, which resolves that for the most part. sadly, libtool doesn't support sysroot |
| 19:00.41 | ds2 | kergoth_: I am not a fan of having to setup a full staging area for installs just to keep everything happy. My rpeference is to point dependant packages back into the build tree but all that crap makes it impossible to do |
| 19:01.14 | ds2 | disks may be cheap but I find 12-14G of stuff per build to be a bit excessive |
| 19:01.23 | kergoth_ | i can see why libtool does the .la files and search path stuff, but it was due to old linkers not supporting libs linking to other libs, and such.. with a modern linker, it's completely unnecessary |
| 19:01.38 | kergoth_ | i wonder what it'd take to just kill its lib search path stuff, only using the -L's from the commandline |
| 19:01.54 | ds2 | the GNU linker seems to be doing it on most platforms |
| 19:02.31 | kergoth_ | yeah |
| 19:03.58 | kergoth_ | i saw a neat project a while back that basically gives you configure like functionality, but was done entirely in make |
| 19:04.03 | kergoth_ | tries to remember what it was |
| 19:08.29 | *** join/#edev kaza_lite (n=mohsin@94-193-39-116.zone7.bethere.co.uk) |
| 19:09.00 | kaza_lite | any suggestions to reduce linux kernel foot print and bootup time? |
| 19:09.34 | ds2 | what board? |
| 19:09.42 | kaza_lite | MIPS |
| 19:09.46 | kaza_lite | MIPS4KC |
| 19:10.09 | kaza_lite | distor: Monta Vista |
| 19:10.31 | ds2 | kaza_lite: I did a squeeze everything + kitchen sink talk at ESC... |
| 19:10.55 | kergoth_ | ds2: http://code.google.com/p/quagmire/source/browse/trunk/README - kind of neat, scroll down to the Example: |
| 19:11.06 | kergoth_ | weird, but neat |
| 19:11.17 | *** join/#edev likewise_ (n=chatzill@82-171-51-231.ip.telfort.nl) |
| 19:11.21 | kaza_lite | what do you mean by it? |
| 19:11.58 | ds2 | http://www.hy-research.com/ |
| 19:12.13 | roxfan | kaza_lite: some ideas http://elinux.org/Boot_Time |
| 19:12.20 | ds2 | follow the link for ESC... ESC-246 |
| 19:12.32 | ds2 | that has my slides |
| 19:12.43 | kaza_lite | ahh...right |
| 19:13.48 | ds2 | kaza_lite: the talk is on the topic you are asking about |
| 19:14.08 | ds2 | kergoth: nice.... would have to try it to really get a feel one way or another |
| 19:15.07 | *** join/#edev sjhill (n=sjhill@home.bethel-hill.org) |
| 19:15.07 | *** mode/#edev [+o sjhill] by ChanServ |
| 19:16.05 | kaza_lite | thnks guys....am reading the slide first |
| 19:16.30 | ds2 | hope it helps... I squish possibly too much stuff in there |
| 19:27.14 | prpplague | throws wince platform builder cds at sjhill |
| 19:29.29 | *** join/#edev theDarkAura (i=Administ@c-24-34-200-228.hsd1.ma.comcast.net) |
| 19:30.06 | kergoth_ | prpplague: where can i find that, by the way? |
| 19:30.32 | prpplague | kergoth_: platform builder? |
| 19:31.15 | roxfan | you still use cds? |
| 19:31.33 | prpplague | kergoth_: http://www.microsoft.com/downloads/details.aspx?displaylang=en&FamilyID=7e286847-6e06-4a0c-8cac-ca7d4c09cb56 |
| 19:31.36 | prpplague | kergoth_: eval edition |
| 19:31.52 | prpplague | kergoth_: fully functional, just doesn't produce licensable editions |
| 19:33.25 | *** join/#edev samrobb (n=samrobb@75-149-8-243-Pennsylvania.hfc.comcastbusiness.net) |
| 19:34.24 | prpplague | kergoth_: full edition of platform builder will run you $1000 |
| 19:34.56 | kergoth_ | k, i'll check that out, thanks |
| 19:35.31 | prpplague | kergoth_: there is also an eval for platformbuilder for wince5 if you are interested |
| 19:35.31 | kergoth_ | I'm looking to do major revamping of oe stuff, so any examples of good ways to do it are helpful |
| 19:35.31 | kergoth_ | cool |
| 19:36.04 | prpplague | kergoth_: the platform builder ships with a qemu like emulator |
| 19:36.17 | kergoth_ | neat |
| 19:36.17 | prpplague | kergoth_: lets you test different base builds |
| 19:41.34 | *** join/#edev pat123 (n=pat@99-41-81-18.lightspeed.mssnks.sbcglobal.net) |
| 19:41.47 | pat123 | hi |
| 19:54.03 | *** join/#edev damunix (n=damunix@72.184.204-77.rev.gaoland.net) |
| 19:56.43 | *** part/#edev kaza_lite (n=mohsin@94-193-39-116.zone7.bethere.co.uk) |
| 20:08.50 | *** join/#edev miknix (n=miknix@81.193.83.213) |
| 20:09.10 | prpplague | pat123: greetings |
| 20:09.36 | pat123 | hi prpplague |
| 20:10.22 | pat123 | hows it going |
| 20:12.09 | prpplague | its going, fighting with irq handling in wince :( |
| 20:12.35 | pat123 | ic |
| 20:38.52 | *** join/#edev damunix_ (n=damunix@128.173.204-77.rev.gaoland.net) |
| 20:57.09 | *** join/#edev flo_lap (n=fuchs@f049081084.adsl.alicedsl.de) |
| 21:13.20 | prpplague | argh i have an irq problem driving me nuts! |
| 21:23.50 | *** join/#edev kaza_lite (n=mohsin@94-193-39-116.zone7.bethere.co.uk) |
| 21:23.56 | kaza_lite | hi all |
| 21:24.17 | kaza_lite | how can i measure boot up time? |
| 21:24.30 | kaza_lite | of linux kernel |
| 21:24.53 | ds2 | clocks? watches? time machines? hour glass? :D |
| 21:25.43 | roxfan | kaza_lite: http://elinux.org/Boot_Time has some ideas :) |
| 21:27.10 | ds2 | roxfan: did the CEforum site move to elinux? |
| 21:27.18 | roxfan | no idea |
| 21:27.44 | roxfan | i don't really follow the wiki, just remembered this from my browsing some time ago |
| 21:39.28 | theDarkAura | what are the chances i can find CE 5 drivers for a device to connect to a a BT HID KB |
| 21:39.39 | theDarkAura | ? :) |
| 21:50.40 | *** join/#edev JoeLlama (n=snork@206.169.63.10) |
| 21:51.03 | prpplague | theDarkAura: slim to none |
| 21:51.13 | prpplague | theDarkAura: its a bluetooth HID keyboard? |
| 21:52.05 | prpplague | theDarkAura: its not good to use IRC when logged in as administrator |
| 21:57.20 | g1powermac_EPC | as I'm kinda curios to this myself and since he provided more info on ##microcontrollers, theDarkAura wants to use this keyboard http://www2.dealextreme.com:8080/details.dx/sku.16685 with this gps device http://www2.dealextreme.com:8080/details.dx/sku.21038 |
| 21:58.10 | g1powermac_EPC | of course, the first thing that went through my mind with that device is how hard it would be to boot linux on it :-) |
| 21:59.33 | g1powermac_EPC | its a quite interesting device for only just over $100 |
| 22:03.33 | rootaccess | ALL OF YOUR IRC CLIENTS ARE BELONG TO US |
| 22:03.35 | roxfan | http://msdn.microsoft.com/en-us/library/ms881906.aspx looks like it is supported by CE5 |
| 22:03.39 | rootaccess | ;) |
| 22:04.25 | rootaccess | nice ! |
| 22:04.34 | roxfan | should be possible to build a driver with platform builder |
| 22:04.44 | prpplague | looks |
| 22:05.14 | prpplague | rootaccess: hid support is there, not sure about the HID over bt |
| 22:05.25 | roxfan | To add the Bluetooth-enabled keyboard or mouse support to your OS design, set the SYSGEN_BTH_HID_KEYBOARD or SYSGEN_BTH_HID_MOUSE Sysgen variables. |
| 22:05.28 | prpplague | rootaccess: interesting |
| 22:05.47 | prpplague | tests |
| 22:06.02 | rootaccess | sweet :) |
| 22:49.03 | rootaccess | how goes the testing prpplague ? |
| 23:01.07 | Madsy | How can I make autoconf find my custom toolchain? It's installed in /opt and all the binaries are prefixed with arm-linux- |
| 23:01.14 | Madsy | It seems to ignore --host |
| 23:01.59 | Madsy | And prefixes can only be set for the install script `make install`, not for the toolchain used in the compilation process. |
| 23:39.34 | solar | add /opt/ to PATH |
| 23:39.45 | solar | and check what sysroot that toolchain was compiled with. |