| 00:10.51 | rektide | any one follow displayport? |
| 00:11.07 | rektide | supposedly the ability to daisy chain displayport peripherials is in the works |
| 00:54.03 | MonMotha | I have not been following displayport much |
| 00:54.07 | MonMotha | looks promising, though |
| 01:26.42 | *** join/#edev kergoth (n=kergoth@ip24-251-170-95.ph.ph.cox.net) |
| 01:36.36 | rektide | besides leading the rebel alliance, what else do you do MonMotha ? |
| 01:39.22 | MonMotha | heh, wrong origin for the name |
| 01:39.22 | MonMotha | it came from "Mothra" originally, but then got combined with some stuff to end up here |
| 01:39.40 | MonMotha | as for what else I do, I design computer/electronic hardware and I collect arcade games |
| 01:41.03 | rektide | well, i've greatly enjoyed having your presence to rely on in here, thought i'd ask |
| 01:41.17 | MonMotha | I've been hanging out in here for a while |
| 01:41.45 | MonMotha | basically since the channel "forked" from #elinux (of sorts) and I was in #elinux when it was #tuxscreen |
| 01:42.19 | rektide | never knew #elinux. i just got sick of ##electronics and eventuall found someone to recommend me come here for a little more interesting / high level systems |
| 01:42.49 | rektide | been here for a while, just mostly idling. and sans internet connection across the summer. |
| 01:44.10 | MonMotha | linux is actually what got me into embedded stuff, and many (though by no means all) of my embedded things run Linux |
| 01:44.29 | landley | Why did this channel fork from #elinux? |
| 01:44.39 | MonMotha | landley: it wasn't exactly a fork |
| 01:45.26 | MonMotha | I seem to recall that there were topics coming up not strictly linux related but still very embedded that had no forum of discussion, and many of the people in #elinux either migrated to here or joined both (I'm currently sitting in both) |
| 01:46.30 | MonMotha | #elinux was what become of #tuxscreen, and it was a little specific to that sort of thing at the time |
| 01:46.39 | SpeedEvil | ##electronics is interesting - but in many ways I'm not helping. |
| 01:46.47 | SpeedEvil | does the off-topic thing too readily. |
| 01:46.55 | landley | I used to hang out in #elinux but there was never really any traffic there. |
| 01:47.05 | rektide | ##electronics doesnt have interesting SoC discussions very often |
| 01:47.09 | rektide | its usually low level components |
| 01:47.27 | SpeedEvil | yeah |
| 01:48.18 | MonMotha | I pretty much deal with component selection, board level stuff, drivers, and base system |
| 01:48.28 | MonMotha | application is generally, though not always, out of my purview |
| 01:48.40 | MonMotha | software application, that is, not system application |
| 01:49.40 | SpeedEvil | I'm currently trying to get on with designing and selling stuff for ebay |
| 01:49.50 | MonMotha | I also have to do some occasional reverse engineering to make things work together |
| 01:49.55 | SpeedEvil | I've got half-baked designs for laser-ranger, and other stuff. |
| 01:50.37 | landley | I'm currently frowning at the aranym and qemu sources, trying to figure out how to add proper m68k support to qemu. |
| 01:50.47 | landley | I have no _idea_ what I'm doing, but since when has that stopped me? |
| 01:50.55 | SpeedEvil | :) |
| 01:51.11 | SpeedEvil | And I'm investigating my new shiny toy - n900. |
| 01:51.16 | rektide | i havent made hardware in a long time. :( |
| 01:51.38 | MonMotha | argh, I wish there was a CDMA2000 N900 |
| 01:51.39 | rektide | i want to re-do the home automation network i built, with some kind of wireless comm |
| 01:51.44 | MonMotha | I'd have one by now :) |
| 01:52.45 | rektide | n900 is just the start |
| 01:52.47 | rektide | of something great |
| 01:53.00 | SpeedEvil | MonMotha: plug into home network, youtube et al all play fine. Have 5 browsers open, and the 'tasks' screen shows live thumbnails. |
| 01:53.07 | rektide | gimme a cell phone with dual 2GHz Cortex A9 and a 16 core SGX543 |
| 01:53.10 | rektide | oohhhh man |
| 01:53.17 | SpeedEvil | rektide: Not the point. |
| 01:53.29 | rektide | yes, the linux part is integral to the whole notion |
| 01:53.30 | SpeedEvil | rektide: The CPU isn't really the constraint ImO. |
| 01:53.37 | SpeedEvil | It's the software. |
| 01:53.53 | rektide | 3d capabilities mean a lot to me |
| 01:53.55 | SpeedEvil | For example - MMS and video call are not constrained by hardware - but by them not being polished. |
| 01:54.46 | SpeedEvil | (to the extent that they are not included as they are not done) |
| 01:55.24 | SpeedEvil | I'm planning on openstreetmap based apps - for which I wish I could get 1cc of volume in the phone and an I2C bus - and a diet/kitchen inventory one. |
| 01:55.30 | solar | phones feel more mem bound vs cpu bound. |
| 01:55.36 | SpeedEvil | yeah. |
| 01:56.09 | solar | so you picked up the n900 ? |
| 01:57.02 | SpeedEvil | I got it at a multiply discounted price - with multiple voucher codes - ending up at about half price. |
| 01:57.04 | rektide | swap space usually is enough to alleviate my memory concerns |
| 01:57.12 | solar | plan to keep Mer/Mameo on it or maybe opting to snag the motorola andriod port? |
| 01:57.13 | SpeedEvil | swap to flash needs fixed though |
| 01:57.47 | rektide | cant wait to get Mer running on my OpenPandora |
| 01:57.50 | SpeedEvil | Swap doesn't know that it should simply write linearly, not in blocks. |
| 01:59.06 | rektide | cant wait to get my OpenPandora |
| 01:59.35 | rektide | please please please, may Nokia's work getting texture_to_pixmap working on the N900 make the transition to the OpenPandora |
| 01:59.53 | rektide | give me composited x or give me different hardware! |
| 02:00.08 | MonMotha | my phone choices are unfortunately constrained by my carriers' use of CDMA2000 |
| 02:00.16 | SpeedEvil | I decided that I'm quite willing to compromise openness for shiny. |
| 02:00.46 | SpeedEvil | Especially given that I expect mer-on-n900 to get a lot more polished with tens of thousands of them in the wild. |
| 02:01.22 | rektide | verizon doesnt get my business until i can use lte everywhere |
| 02:01.29 | SpeedEvil | My last phone had schematics - and it diddn't really help. |
| 02:01.32 | MonMotha | I'm not on Verizon |
| 02:01.52 | rektide | i figured the case would somehow not be that simple or fathomable |
| 02:02.04 | solar | I'm on AT&T and they really have no offerings outside of iphones or winmo |
| 02:02.14 | MonMotha | maybe not from AT&T |
| 02:02.22 | MonMotha | but AT&T is GSM/WCDMA. Lots of options. |
| 02:02.29 | MonMotha | An N900 would work, but you won't get 3G due to band limitations. |
| 02:02.37 | solar | exactly ;/ |
| 02:02.52 | SpeedEvil | MonMotha: you won't get 3G on other than t-mobile |
| 02:02.57 | MonMotha | right |
| 02:02.59 | MonMotha | AT&T uses a strange band |
| 02:03.08 | MonMotha | though you can buy some HTC stuff I think with support for it |
| 02:03.34 | rektide | i'm still on an old as hell Motorola |
| 02:03.46 | rektide | on AT&T |
| 02:04.13 | rektide | maybe Google Wireless Data will launch before i am compelled to upgrade |
| 02:05.24 | rektide | i had two conversations with non techies this summer about how long before we switch to packet networks, both prompted by the non-technical person. little did i expect google wireless data rumblings to start. i thought it would be 4-6 years. |
| 02:05.40 | rektide | and that LTE alliance declaring they wanted to do everything packet based. |
| 02:06.26 | MonMotha | Sprint's solution to "4G" has been very simple: WiMax |
| 02:07.05 | solar | SpeedEvil: so have you seen the motorola driod hands on? Being that it and the n900 are pretty much the same basic hardware(chip wise). which do you like better? |
| 02:07.32 | SpeedEvil | solar: no |
| 02:07.55 | solar | darn. well are you pleased with the phone so far? |
| 02:08.00 | rektide | droid is not an 800MHz Cortex A8 with PowerVR |
| 02:08.06 | SpeedEvil | Many people simply don't get that wireless networks are shared. |
| 02:08.23 | SpeedEvil | And try to generalise from DSL or cabled internet models onto them. |
| 02:08.27 | SpeedEvil | solar: very |
| 02:08.51 | rektide | my bad, it is another omap3. powervr, 600MHz |
| 02:09.02 | SpeedEvil | It feels solid and weighs -7Kg. |
| 02:09.11 | SpeedEvil | (I partially paid for it by going on a diet) |
| 02:09.20 | MonMotha | somebody needs to tell Nokia to put a darned CDMA2000 radio in the N900 |
| 02:09.42 | rektide | someone needs to tell CDMA2000 carriers to get with the program |
| 02:10.06 | MonMotha | well, at this point, I'd agree, though CDMA2000 offered a lot of advantages over GSM until WCDMA started to supplant it |
| 02:10.31 | MonMotha | however, there's too much infrastructure (including handsets!) to do a wholesale conversion |
| 02:13.05 | rektide | i have this warm fuzzy picture of lte coming along and making hte old distinctions irrelevant |
| 02:13.17 | rektide | that i wont have to buy a phone with a Verizon logo stamped on it to use a verizon network |
| 02:13.39 | rektide | of course, thats bunk, as LTE will only be in metropolises |
| 02:14.27 | SpeedEvil | Thankfully in the EU - that's generally not the case. |
| 02:15.54 | MonMotha | Sprint is pretty nice about activating anything compatible with their network, but the fact that they use CDMA2000 is limiting |
| 02:19.17 | calculus | SpeedEvil: 7Kg?, that seems like a lot for a phone |
| 02:22.12 | SpeedEvil | calculus: -7Kg - I'm counting the weight I've lost. |
| 02:22.15 | Mirell | It's the brick phone. |
| 02:42.35 | calculus | oh |
| 02:42.40 | calculus | that makes more sense |
| 03:00.57 | *** join/#edev monkey_ (n=monkey@216-243-6-35.users.condointernet.net) |
| 03:01.05 | monkey_ | hey guys |
| 03:33.12 | landley | Hi monkey_ |
| 03:34.23 | landley | SpeedEvil: swap to flash has inherent problems. Limited write cycles and _insane_ granularity. |
| 03:35.16 | MonMotha | swap to NOR flash could be an interesting idea, at least |
| 03:35.19 | landley | solar: does AT&T really _need_ an offering outside iphone? :) (WinCE phones don't matter and never have.) |
| 03:35.23 | MonMotha | since it's addressable as memory |
| 03:35.49 | MonMotha | I think the limited write cycles thing is still blown out of proportion, though swap can be pretty hard |
| 03:36.04 | MonMotha | I suspect it would just be so slow as to be something not worth considering, though |
| 03:37.06 | landley | MonMotha: eventually we'll just get more dram in the suckers. Or commercially viable mram or similar will show up. |
| 03:37.14 | landley | Flash always struck me as a hack. |
| 03:38.03 | MonMotha | I'd like to see a machine with just tons of MRAM |
| 03:38.06 | MonMotha | instant-on/off |
| 03:38.20 | MonMotha | basically zero-power STR |
| 03:38.49 | MonMotha | it's getting to where you could theoretically make a handheld using MRAM as it's primary form of RAM at least, but mass storage still is not an option due to cost |
| 03:41.59 | monkey_ | is there an equivilant of flash magic for linux? |
| 03:42.02 | monkey_ | my board got corrupted |
| 03:42.58 | monkey_ | or something, i need to reflash it with something other than openocd |
| 03:43.07 | landley | MonMotha: these guys are trying: http://www.crocus-technology.com/pr-10-01-09.html |
| 03:43.20 | landley | What's "flash magic"? |
| 03:43.42 | landley | Why do you need something other than openocd to flash it? |
| 03:43.55 | landley | (What's wrong with using openocd here? Other than that it's slow...) |
| 03:43.55 | monkey_ | well, last night openocd/gdb were working |
| 03:44.06 | monkey_ | and now it had a bad flash. so now openocd isn't working |
| 03:44.18 | landley | openocd should still be working fine, it's just that gdb isn't. |
| 03:44.28 | monkey_ | i get a ton of errors before i even start gdb |
| 03:44.29 | landley | openocd is essentially a driver that talks to your jtag. |
| 03:44.36 | landley | Hooking up gdb is _one_ thing you can do with this driver. |
| 03:44.40 | landley | Reflashing the board manually is another. |
| 03:44.52 | landley | hey prpplague_afk, you here? This is your area. |
| 03:44.57 | landley | (Or possibly solar could help, not sure.) |
| 03:45.02 | monkey_ | well, otherstudens had this problem, and teacher says use flash magic |
| 03:46.07 | landley | I've used openocd to reflash my nail board, but it requires a config file for the board. |
| 03:46.26 | landley | In theory you already have a config file to hook up gdb via openocd, but I dunno how to adapt it to reflash. |
| 03:46.29 | landley | prpplague would. |
| 03:46.47 | landley | I could probably figure it out in an hour or so if I had the existing setup and the board in front of me, but I don't. |
| 03:47.13 | monkey_ | yeah |
| 03:47.16 | monkey_ | i understand |
| 03:49.07 | landley | Ha! Somebody made an fpga out of mram. http://www.edn.com/professormemory/blog/1640000764/post/330050033.html |
| 03:49.16 | landley | That's somewhere between "funny" and "scary". |
| 03:52.48 | MonMotha | that actually makes sense |
| 03:52.56 | MonMotha | all the benefits of SRAM only FPGAs but non-volatile |
| 03:53.07 | landley | I'm aware it makes sense. |
| 03:53.18 | landley | It's also one of those "I can't predict where this is going" things. |
| 03:54.04 | rektide | so the voltage keeps going down ddr->ddr2->ddr3; how much of that power savings applies in STR? |
| 03:54.08 | landley | The new thermal phase of mram writing (which makes it faster and lower power by electrically heating the cells in the write stage) has unfortunate associations with magneto-optical drives for me. |
| 03:54.17 | landley | But I suppose it would have to, with the physics being the same and all... |
| 03:54.29 | landley | rektide: apparently about half what it used to take. |
| 03:54.53 | landley | Not that I know what it used to take. Less power-hungry than flash on writes, but I don't remember by how much. |
| 03:55.11 | landley | Apparently where mram really shines is in reads. They've done 1ns reads in the lab. |
| 03:55.21 | landley | It's faster than dram for reads. |
| 03:55.23 | rektide | so what would be interesting |
| 03:55.30 | rektide | shuffle all the data to one ram chip |
| 03:55.41 | rektide | i guess most systems dont really run with free ram now a days |
| 03:55.53 | landley | rektide depends on your definition of "free". |
| 03:56.07 | landley | Linux uses anything unallocated for clean disk caches it can drop at will. |
| 03:56.50 | rektide | overallocation is the rule in app land |
| 03:56.58 | landley | It's still possible something could come out of left field and render this moot, the way it did for MO drives. |
| 03:57.14 | landley | This particular manufacturer is looking to produce commercial quantities in 2011. |
| 03:57.25 | landley | Which is ages away in hardware terms. :) |
| 03:57.29 | solar | landley: do they need it? well sure. My work will pay any plan AT&T offers, as long as they don't have to switch plans. So everybody is ending up with iphones |
| 03:57.36 | rektide | i really really really wonder what kind of GHz/mW CortexA9 is going to have |
| 03:57.48 | rektide | i dont lose sleep over power savings from MRAM |
| 03:58.10 | rektide | particularly how linear A9 will scale |
| 03:58.51 | landley | solar: I meant does AT&T need any phone _other_ than the iPhone. |
| 03:59.01 | landley | If all they sold was iPhones, would they be doing noticeably worse than they are today? |
| 03:59.09 | rektide | has anyone ever physically seen an E4690 chip? |
| 03:59.12 | solar | landley: I'm still rocking an older HTC Wizard(omap-850) phone where we ported the basics over we needed for linux. |
| 03:59.26 | solar | landley: for profit. Probably nope |
| 03:59.44 | landley | My phone still hasn't got a noticeable brain. |
| 04:00.02 | landley | I got it to provide internet to my laptop via bluetooth, and just use my laptop. |
| 04:00.22 | landley | Fade has an iPhone though. She loves it, of course. |
| 04:00.25 | solar | when you do upgrade you will enjoy not always having to go to the coffee shop for the little stuff like irc |
| 04:00.43 | landley | Oh I'd love to have an iPhone, I just can't justify it. |
| 04:00.53 | landley | And it's a pain to crack 'em to get a shell prompt. |
| 04:01.02 | solar | I'm not a fan of softkeys |
| 04:01.11 | landley | None of the Linux-based phones I've seen really interest me. |
| 04:01.27 | solar | the motorola driod and nokika n900 are starting to impress me |
| 04:01.38 | landley | Don't think I've seen those yet. |
| 04:01.44 | landley | Does android let you run native code yet? |
| 04:01.55 | solar | both are cortex-a8 based. |
| 04:02.05 | solar | yes. Sure you can |
| 04:02.26 | solar | it's still a linux kernel. I think the command 'adb' breaks you out to a shell |
| 04:02.52 | solar | the UI is all the hype really. |
| 04:03.06 | solar | nice thing is they are coding drivers to make hardware work |
| 04:03.17 | solar | so you can really slap any UI on it you want |
| 04:03.46 | solar | but currently the UI's they offer a pretty well rounded in terms of functionality |
| 04:04.29 | landley | Ah, "you can slap any UI on it you want". the battle cry of Unix on the desktop for the past 30 years. :) |
| 04:04.29 | landley | Open source is structurally incapable of doing user interface development. |
| 04:04.38 | landley | Any time "shut up and show me the code" is not a valid response to the problem at hand, the process melts down. |
| 04:04.43 | landley | Which is sad, but I've come to accept it. |
| 04:04.47 | rektide | android is not a linux syste,m |
| 04:04.52 | solar | it's pretty nice imo to see winmo take a backseat to smartphones imo :) |
| 04:05.05 | rektide | just dont even pretend. its a stack, and the stack has very little to do with linux |
| 04:05.05 | solar | I'm happy |
| 04:05.25 | landley | stays out from between rektide and solar on this one. |
| 04:05.54 | rektide | http://thefowle.livejournal.com/399906.html chromeos |
| 04:05.58 | landley | I talked to a couple iPhone developers to ask why they didn't do android ports of their apps. |
| 04:06.04 | rektide | chromeos is at least a linux, and its a web linux |
| 04:06.09 | rektide | theres really nothing more necessary |
| 04:06.14 | rektide | the only |
| 04:06.16 | rektide | the ONLY |
| 04:06.16 | rektide | question |
| 04:06.17 | solar | landley: free vs pay? |
| 04:06.27 | landley | Apparently, the screen's aspect ratio isn't consistent, the amount of memory you have to work with isn't consistent, whether you're using a real keyboard or a touch screen isn't consistent... |
| 04:06.36 | landley | They told me there's no such thing as android. |
| 04:06.50 | landley | solar: didn't ask, it was one of those door64 events I think. |
| 04:07.22 | solar | if I was coding higher level. dunno maybe before I invoked a type of functionality I might check to see if it existed |
| 04:07.42 | solar | dunno. maybe thats just every coder I knoew :) |
| 04:08.00 | landley | I'm interested in the possibility of chromeos allowing the cheap $250 laptop-lets (successors to OLPC and eeepc) to go down under $100, possibly arm based. |
| 04:08.06 | landley | That would be cool, although it would need a flash plugin. |
| 04:08.16 | MonMotha | screw flash |
| 04:08.17 | landley | (I don't care if it's the fsf's flash clone or what, it just needs to _work_.) |
| 04:08.29 | landley | MonMotha: I'll believe in html5 when I see it. |
| 04:08.44 | MonMotha | did I say anything about HTML5? :) |
| 04:09.12 | landley | solar: the iPhone guys are doing game console development. The android guys are doing PC development. |
| 04:10.00 | landley | The thing is, people have fallen out of the habit of doing PC development. The "everything on the web" stuff is all for one or two browsers and you get to specify your one and only server. |
| 04:10.42 | landley | So the "test the result on 30 slightly different deployment platforms ala DOS" thing doesn't really seem to appeal to the current crop of developers. |
| 04:10.54 | landley | Perhaps I've been speaking to an unrepresentative subset. |
| 04:11.07 | solar | DOS = ? |
| 04:11.08 | landley | It wasn't serious research, just casual "so why not" mingling. |
| 04:11.46 | landley | solar: back under DOS your app had to deal with every sound card and every graphics card itself. |
| 04:11.50 | solar | like os DOS? Xerox/ibm? |
| 04:11.51 | landley | The high water mark of non-uniformity. |
| 04:11.59 | landley | Yup. |
| 04:12.02 | solar | ok. |
| 04:12.10 | landley | Then there was Windows "dll hell", and so on. |
| 04:12.32 | landley | Android has manufacturers removing/modifying part of the base set of libraries to try to cause vendor lock-in. |
| 04:12.38 | landley | What they've managed is developer lock-out. |
| 04:13.04 | landley | That's pretty much anecdotal and 6 months old, though, so take it with a grain of salt. |
| 04:13.05 | solar | are you seeing that? |
| 04:13.22 | landley | solar: talked to I think 3 different developers who had seen it. |
| 04:13.34 | landley | I don't think they knew each other, but they were all here in Austin. |
| 04:13.52 | solar | for me it's hard to tell. I've only seen a few guys going from something like an andriod enabled phone to something like gpe and custom kernel |
| 04:13.56 | landley | Not exactly a scientifically rigorous sample, I'll admit. |
| 04:14.27 | landley | I haven't seen people leaving android after committing resources to it, I've just seen people not picking it up beyond the experimental phase. |
| 04:14.34 | solar | all HTC msm-qualcom based phones |
| 04:15.03 | *** join/#edev kergoth (n=kergoth@ip24-251-170-95.ph.ph.cox.net) |
| 04:15.27 | landley | I've seen lots of people making iPhone apps. I'm still waiting for android to take off in that regard. |
| 04:15.46 | MonMotha | why does google allow android phones with the google branding to be shipped without basic libs/ |
| 04:15.46 | landley | Perhaps it's just iPhone having a couple year headstart... |
| 04:15.48 | MonMotha | ? |
| 04:16.05 | landley | Have they tried to stop it? Not my area... |
| 04:16.16 | landley | I haven't worked for a company doing android, and I don't personally own an android phone. |
| 04:16.21 | MonMotha | that just seems odd |
| 04:16.37 | landley | The only people I can think of who _do_ are Neil Gaiman (whose son works at Google) and Eric Raymond (who got one on principle, and loves it, but doesn't develop for it). |
| 04:16.43 | landley | It has mame with pac-man. |
| 04:16.45 | MonMotha | obviously, being mostly GPL, they can't stop people from shipping something based on android that isn't "complete", but they should be able to enforce their trademarks to cause some sort of stability to the platform |
| 04:17.09 | landley | (Actually I think it was a java port of the original pac-man, or some such. Wasn't mame. Dunno, not my device...) |
| 04:17.27 | solar | my day to day cell is still winmo and side project port going on for kix |
| 04:17.51 | landley | MonMotha: Steve Jobs went to the phone companies the same way he did to the RIAA for iTunes and the MPAA for Pixar. He out-schmoozed them on their own turf, and negotiated hollywood contracts with hollywood teeth. |
| 04:18.03 | landley | I dunno if Google is up to that kind of black belt in legal slime. |
| 04:19.32 | MonMotha | landley: but they should at least be able to pretty easily do the whole "we won't give you our build or let you use our trademarks if you don't at least do this" bit |
| 04:20.10 | MonMotha | apparently one of the big draws of android is that it's "cheaper than free" since google gives the carriers and phone makers a chunk of search ad revenue on android phones |
| 04:20.13 | MonMotha | they don't have to do that |
| 04:24.17 | landley | I'm not the person to ask. |
| 04:24.25 | landley | Perhaps the problems I saw were just early teething troubles. I dunno. |
| 04:24.53 | rektide | andriod is another bullshit stack and the last thing we need is more stacks |
| 04:28.35 | landley | I tend to stick with busybox, uClibc, and the Linux kernel. |
| 04:28.53 | landley | Other stacks are distro issues to me. |
| 04:30.27 | rektide | android is _not_ a distro issue, if absolutely nothing else |
| 04:30.46 | solar | landley: so what phone would you opt for if work offered you any phone+plan@free ? |
| 04:30.46 | rektide | thats why i think chromeos is important. its a distro. a web distro. write web, be ok. |
| 04:31.15 | rektide | the only question is how well and how qucikly google & html5 expose system resources |
| 04:31.27 | rektide | OMFG |
| 04:31.51 | rektide | new EVE client is a web browser. i'm sure i can implement an in game geolocation API system. |
| 04:31.54 | landley | solar: an iPhone. |
| 04:32.04 | solar | really? |
| 04:32.08 | landley | Yup. |
| 04:32.15 | landley | I'd crack it and get a shell prompt on it. |
| 04:32.31 | landley | Might pick up my old work on tinycc and get it to produce arm and mach-o executables. |
| 04:32.39 | solar | sure but it's drivers only exist under a BSD based kernel. |
| 04:32.52 | landley | Most likely I'd just use it and not develop for it much. |
| 04:32.56 | landley | solar: so what? |
| 04:32.57 | solar | you ready to port the extras over? |
| 04:33.04 | landley | Its kernel is bsd based, its userspace is vaguely posix-ish. |
| 04:34.46 | landley | It's a bit like asking if I was going to do web app development, would I do it in flash? Probably yes. |
| 04:34.53 | landley | That's what people actually use. |
| 04:34.56 | MonMotha | what attracts me to Android is that I can get a phone that works out of the box but that I can (usually) hack on a little if it's doing something I don't like |
| 04:35.11 | landley | And the things out there trying to clone it are 5 years behind at least. |
| 04:35.21 | landley | MonMotha: sounds cool. |
| 04:35.31 | MonMotha | and while people use flash, I don't think users in general like it |
| 04:35.35 | MonMotha | they put up with it because everything uses it |
| 04:35.35 | *** join/#edev mbuf (n=mbuf@118.102.130.6) |
| 04:35.35 | landley | I haven't got one, don't know anyone who's actually recommended one who wasn't a programmer. |
| 04:35.49 | landley | MonMotha: millions of people play facebook games, they're flash. |
| 04:35.59 | MonMotha | I'd probably have a HTC Hero right now if it weren't for the fact that it lacks a hard keyboard |
| 04:36.01 | landley | Youtube's player is implemented in flash. |
| 04:36.19 | landley | Hulu too. |
| 04:36.27 | MonMotha | yes, lots of people USE flash, but I'm not sure I know anyone who things flash is especially good |
| 04:36.32 | landley | Flash is pretty much required for the modern web experience. I may not _like_ that, but it's the truth. |
| 04:36.38 | MonMotha | and I know many who lament the poor "not a PC running Windows" support of flash |
| 04:36.46 | solar | yep |
| 04:36.46 | MonMotha | these are normal users |
| 04:36.49 | landley | I don't _like_ the qwerty keyboard. It's not going anywhere any time soon. |
| 04:36.58 | MonMotha | some are technically inclined, while others just want to know wtf flash is so bad on their phone or netbook |
| 04:37.17 | landley | The iPhone doesn't support flash, largely because the battery would go dead in an hour if it did. |
| 04:37.20 | solar | I have a legit gripe with flash |
| 04:37.27 | MonMotha | the qwerty keyboard at least gets the job done pretty well and is easily implemented anywhere allowing for good compatibility |
| 04:37.34 | solar | it requires memory segments to to be rwx |
| 04:37.36 | landley | I have a bunch of legit gripes with flash. I'm just not convinced they're relevant. |
| 04:38.13 | MonMotha | flash barely gets the job done in most applications it is used for, and it is vendor proprietary which causes use cases that the vendor ignores (some of which are fairly popular) to suck |
| 04:38.26 | MonMotha | flash games are generally OK, but it makes a LOUSY media player |
| 04:38.40 | MonMotha | it's barely accelerated on Windows and totally unaccelerated on other platforms |
| 04:38.50 | landley | If I want to write an app, letting myself get distracted into creating a new platform with zero deployment and convincing people to use it _distracts_ from writing and deploying my app. |
| 04:39.17 | landley | MonMotha: and they solve that on the PC by throwing hardware at the problem, as usual. |
| 04:39.36 | landley | Flash worked great for me on Unbuntu 8.10. In 9.04 they switched to the new intel 945 drivers that DON |
| 04:39.36 | MonMotha | right, but you can't do that on a phone |
| 04:39.54 | landley | 'T WORK and my 3d became unaccelerated, and now flash is a slideshow full screen and needs a small window to be smooth. |
| 04:40.01 | landley | But I blame Ubuntu for that, not flash. |
| 04:40.11 | landley | Even though VLC can do fullscreen video smoothly. |
| 04:40.20 | landley | Because it USED to work, and Ubuntu broke it by messing with drivers. |
| 04:40.37 | landley | (I suspect that's why my desktop locks solid and needs to be rebooted twice a week, it's X11 freezing when trying to do something with 3d.) |
| 04:40.39 | MonMotha | that's weird - the GMA945 drivers are generally pretty good |
| 04:40.57 | MonMotha | now, the GMA500 (aka Poulsbo) drivers are terrible |
| 04:41.04 | landley | MonMotha: they were, yes. But in Ubunt 9.04 they switched to a new DRM implementation in X11, and it doesn't work with the Intel drivers. |
| 04:41.11 | MonMotha | I see... |
| 04:41.11 | landley | I need to install 9.10 and see what they broke this time. |
| 04:41.19 | MonMotha | I have 9.10 on my laptop, and everything works |
| 04:41.20 | landley | (Each new version is full of regressions. It's exhausting.) |
| 04:41.27 | MonMotha | well, that's Ubuntu in a nutshell |
| 04:41.37 | MonMotha | they have a set in stone release schedule, and they just "make it work" |
| 04:41.48 | landley | For a certain definition of "work". |
| 04:41.48 | MonMotha | either it works great for you or you give up and install real Debian |
| 04:42.04 | landley | I will never install real Debian. |
| 04:42.10 | MonMotha | what's wrong with Debian? |
| 04:42.14 | MonMotha | I've run it for years. Works great. |
| 04:42.19 | landley | The debian mailing list made me a loyal Red Hat user in 1998. |
| 04:42.30 | MonMotha | oh, well, ignore all their policy/politics BS |
| 04:42.45 | MonMotha | it's a technically sound distro |
| 04:42.52 | landley | MonMotha: I refuse to get any of debian on me. I only use Ubuntu because it INSULATES me from debian's development community. |
| 04:43.10 | MonMotha | I stay away from the development community. I just use it. |
| 04:43.17 | landley | Many things are technically sound and completely unusable. |
| 04:43.46 | MonMotha | dunno, it's always worked well enough for me |
| 04:43.56 | landley | You're welcome to it. |
| 04:46.42 | rektide | thanks i'll take it |
| 04:46.52 | rektide | now you go back to your opinions |
| 05:08.58 | landley | rektide: never left 'em, thanks. |
| 05:09.08 | landley | Wake me when debian gets preinstalled on anything. |
| 05:10.52 | landley | Debian is the technical equivalent of a subsistence economy. |
| 05:10.52 | landley | Everybody farms. If you don't farm, you don't count. |
| 05:10.57 | landley | By programmers, for programmers, at the expense of anybody who doesn't program. |
| 05:11.40 | landley | The "look at how well we do our thing" argument is kind of pointless if nobody other than the in crowd can use it. |
| 05:24.09 | rektide | Debian is a distro that just works. Anything else is superflous. |
| 05:24.24 | rektide | its not better |
| 05:24.25 | solar | heh |
| 05:27.36 | MonMotha | yeah distro wars! |
| 05:27.41 | MonMotha | I <3 holy wars. |
| 05:28.50 | MonMotha | I also use vi >;) |
| 05:29.36 | rektide | the fridge is on our side |
| 05:30.02 | *** join/#edev archae0pteryx1 (n=snewman@207.47.42.130.static.nextweb.net) |
| 05:51.36 | Mirell | TeeFury is my new favorite addiction... |
| 05:52.21 | Mirell | Also, |
| 05:52.59 | *** mode/#edev [+o Mirell] by ChanServ |
| 05:53.01 | *** topic/#edev by Mirell -> A channel for the discussion of embedded systems (Hardware/Software/Linux/WinCE/Commercial Device Repurposing) : Firmware Linux Discussion : Busybox and Buildroot Discussion : Gold 1200.00 : Silver 18.81 : Your Mom Is The Best Linux Distro |
| 05:53.08 | Mirell | Hmm... |
| 05:53.12 | Mirell | prpplague might get annoyed. |
| 05:53.19 | *** topic/#edev by Mirell -> A channel for the discussion of embedded systems (Hardware/Software/Linux/WinCE/Commercial Device Repurposing) : Firmware Linux Discussion : Busybox and Buildroot Discussion : Gold 1200.00 : Silver 18.81 |
| 06:24.05 | landley | gold? silver? what? |
| 06:24.34 | landley | Mirell: what's TeeFury? |
| 06:25.15 | *** join/#edev SandGorgon (n=OmNomNom@122.160.41.129) |
| 06:29.52 | *** join/#edev PureSine (n=PureSine@87.247.172.211) |
| 06:37.23 | luke-jr | o hai thar! |
| 06:37.23 | *** join/#edev dsr (i=0k2yo1v7@76.115.94.244) |
| 06:37.30 | luke-jr | oops, wrong channel for that XD |
| 06:48.45 | *** join/#edev RemoteSine (n=chatzill@87.247.172.211) |
| 06:53.30 | *** join/#edev PureSine (n=PureSine@87.247.172.211) |
| 06:59.08 | Mirell | landley: https://www.teefury.com/ |
| 07:05.10 | landley | Eh, not my thing. I didn't really get into http://www.warrenellis.com/?p=7999 either. |
| 07:28.38 | Mirell | landley: http://www.teefury.com/archive/642/Hothmas |
| 07:29.03 | Mirell | landley: http://www.teefury.com/archive/609/TRAP |
| 07:30.18 | Mirell | landley: http://www.teefury.com/archive/554/Transminator |
| 07:30.28 | ncopa | landley: hi, have yo tried build a native toolchain with uclibc-nptl? |
| 07:30.31 | Mirell | landley: What you linked was just lame t-shirts printed via Cafe Press. |
| 07:34.12 | *** join/#edev cbx (n=cbx@unaffiliated/cbx) |
| 07:49.29 | Mirell | Hmm.. |
| 07:49.55 | Mirell | If I have control-3452.blah, and control-4325.blah, how can I just get the highest numbered file easier. |
| 08:00.37 | MonMotha | I think the sort command has a "numeric" option |
| 08:00.50 | MonMotha | but you may have to strip off the prefix then add it back on |
| 08:09.20 | landley | ncopa: I've poked at it a little. |
| 08:09.56 | landley | Mostly trying to get uclibc-nptl to work at all has been a bit of a pain. |
| 08:10.11 | landley | Now that there's a new rebased nptl branch I've been poking at it, but the underlying master doesn't even build a busybox defconfig for i686 or x86_64, so there hasn't seemed to be much point just yet. |
| 08:10.27 | landley | The uClibc repository has been in pieces on the floor for ages. Diseased development model... |
| 08:11.06 | landley | I go back and poke at it and debug another thing or two every once in a while. It's about #5 on my todo list right now, which is near the top. |
| 08:11.39 | ncopa | ok |
| 08:11.44 | ncopa | wow |
| 08:11.55 | ncopa | system-image-i586/image-i586.sqf |
| 08:11.55 | ncopa | system-image-i586/run-emulator.sh |
| 08:11.55 | ncopa | === Packaging complete |
| 08:12.12 | ncopa | i just finished build a toolchain with uclibc-nptl |
| 08:12.18 | ncopa | using your fwl :) |
| 08:12.21 | landley | ncopa: cool. |
| 08:12.32 | ncopa | there are some broken stuff in there |
| 08:12.35 | landley | Grabbing the current origin/nptl? |
| 08:12.37 | ncopa | rpc in uclibc is broken |
| 08:12.40 | ncopa | yes |
| 08:12.51 | ncopa | so i had to disable mount in bb etc |
| 08:12.55 | landley | Huh, they must have fixed stuff since last I looked. |
| 08:12.59 | ncopa | yes |
| 08:13.03 | ncopa | well |
| 08:13.05 | landley | Ah, so they fixed _some_ stuff. :) |
| 08:13.08 | landley | Still, yay progress. |
| 08:13.15 | ncopa | i used a hand made config |
| 08:13.18 | landley | The result won't be too happy about booting without "mount", but it did build. :) |
| 08:13.23 | ncopa | also static linking is broken |
| 08:13.36 | landley | In my scripts, or in uClibc? |
| 08:13.38 | ncopa | in uclibc |
| 08:13.53 | landley | My current scripts in -hg need the static toolchain build fixed too... |
| 08:13.57 | landley | Working on it. :) |
| 08:14.10 | landley | Trying to get a release out monday, as I just posted to the FWL mailing list. |
| 08:14.19 | landley | (Are you on that list?) |
| 08:14.26 | ncopa | no |
| 08:14.33 | landley | shrugs. |
| 08:14.38 | ncopa | i'm no too many lists already |
| 08:14.48 | ncopa | what i found sligly annying was that it rebuilds too much on failure |
| 08:14.56 | landley | I post status things to it once or twice a week, but you can get that through the web thingy just as easily. |
| 08:15.16 | ncopa | i mean if toybox build fails, it will still rebuild busybox |
| 08:15.18 | landley | http://lists.impactlinux.com/pipermail/firmware-impactlinux.com/2009-December/000503.html was today's. |
| 08:15.38 | landley | ncopa: I actually just fixed that the day before yesterday. I believe it's in current -hg. |
| 08:15.46 | ncopa | oh cool |
| 08:15.57 | ncopa | otherwise, the scripts are cool |
| 08:16.03 | landley | My current quest is to break up all the sources/sections blobs into individual package builds. |
| 08:16.10 | ncopa | yeah |
| 08:16.13 | ncopa | that makes sense |
| 08:16.23 | landley | Which is how I broke the static toolchain: breaking up the binutils, gcc, ccwrap, uClibc, and uClibc++ builds is DARN FIDDLY. |
| 08:16.52 | ncopa | btw, the scripts fails when buildin on uclibc host |
| 08:17.01 | landley | The non-static toolchain works, the native toolchain works, but the static toolchain (which is the native toolchain with a different host compiler) doesn't. Sigh. |
| 08:17.08 | landley | It'll be some stupid one line fix, it's just a question of finding it... |
| 08:17.16 | landley | ncopa: that sucks. Fails how? |
| 08:17.49 | ncopa | host tools e2fsprogs |
| 08:17.49 | landley | You mean it's not rebuilding inside a system-image? |
| 08:17.49 | landley | Does it do that with the 0.9.30.1 uclibc release? |
| 08:17.49 | ncopa | yes |
| 08:17.58 | ncopa | also the cross compiler segfaults |
| 08:17.59 | landley | tries to remember the last time he tested that... spring, possibly? |
| 08:18.15 | landley | I need to track down and fix both of these problems. |
| 08:18.25 | landley | Actually the cross compiler segfaulting may be a problem I already fixed, what release are you using? |
| 08:18.27 | ncopa | could be my gcc |
| 08:18.41 | landley | It could be that ccwrap used to go nuts if you didn't have $PATH exported. |
| 08:18.50 | ncopa | aha |
| 08:18.51 | landley | (Tested for null _after_ using it. Doh!) |
| 08:18.57 | ncopa | lol |
| 08:18.57 | landley | I fixed that. |
| 08:19.05 | ncopa | cool! |
| 08:19.37 | landley | Commit 839, october 8th. |
| 08:19.46 | ncopa | i suppose i should use latest hg then |
| 08:20.02 | landley | http://impactlinux.com/hg/hgwebdir.cgi/firmware/rev/839 |
| 08:20.14 | landley | ncopa: latest hg is, as I mentioned, slightly horked. :) |
| 08:20.24 | ncopa | hum ok |
| 08:20.26 | landley | I broke it on... saturday, I think. |
| 08:20.35 | landley | It works fine as long as you don't try to make a static cross compiler. |
| 08:20.46 | ncopa | well, i think I just got my toolchain built |
| 08:20.51 | landley | (I.E. the STATIC_CROSS_COMPILER_HOST=i686 stuff.) |
| 08:20.55 | ncopa | no |
| 08:21.14 | ncopa | uclibc-nptl is broken when it comes to static |
| 08:21.18 | landley | Of course if you don't do that, your cross compiler won't have thread support, or uClibc++. |
| 08:21.36 | ncopa | so i had to build bb and tybox dynamic |
| 08:21.39 | landley | I am on a quest to track them down and fix them, but I have a few other quests I'm on first. :) |
| 08:21.46 | ncopa | ok |
| 08:21.48 | landley | (Lotsa yellow exclamation points...) |
| 08:22.03 | landley | ncopa there's a config switch for that. |
| 08:22.13 | ncopa | toybox does not respect it |
| 08:22.27 | landley | ncopa: that's a bug. |
| 08:22.35 | landley | goes to check that bug and fix it... |
| 08:22.53 | Mirell | stares at an endless sea of backticks |
| 08:23.27 | landley | Mirell: $() is your friend. |
| 08:23.38 | landley | It nests, and does so semi-coherently. |
| 08:23.40 | ncopa | landley: heres my workaround: http://pastebin.com/d56b5a150 |
| 08:23.43 | landley | Although the quoting semantics are crazy. :) |
| 08:24.01 | Mirell | landley: I know, I'm converting scripts with the backticks to $() among other things |
| 08:24.15 | landley | ncopa: ouch. |
| 08:24.45 | landley | I need to test whether breaking up sources/sections/toybox.sh fixed that already. I'm unsure about environment variable scope here... |
| 08:24.55 | landley | (Since they're sourced, but they're sourced from within a shell function...) |
| 08:26.02 | landley | I'd rather make the first set have the proper scope than explicitly remove debris afterwards. |
| 08:26.50 | landley | ncopa: the 0.9.8 release came out November 7th. It should have the ccwrap segfault fix. |
| 08:26.58 | ncopa | ok |
| 08:27.06 | landley | I'll test rebuilding the scripts under themselves though, you mentioned multiple problems and I want to find them and fix them. |
| 08:27.45 | ncopa | i coudl probably provide patches for bulding the host tools on uclibc host |
| 08:27.56 | landley | cross compiler segfaulting, and e2fsprogs not building... |
| 08:28.01 | landley | ncopa: sure. |
| 08:28.22 | landley | The main host tool is busybox. |
| 08:28.32 | ncopa | my xcomplie segfaulting could be my host toolchain |
| 08:28.45 | ncopa | using a hardened toolchain |
| 08:28.46 | landley | It only builds e2fsprogs so it can do ext2 packaging in system-image. |
| 08:28.57 | ncopa | i figured so i commented that out |
| 08:29.02 | landley | I still want to track it down. |
| 08:29.03 | ncopa | hum |
| 08:29.13 | ncopa | i think --disable-nls should to the trick |
| 08:29.16 | landley | Which release are you using? |
| 08:29.20 | ncopa | i need hg to make patches? |
| 08:29.24 | landley | I thought I was doing --disable-nls. Oh, you mean ext2. |
| 08:29.36 | ncopa | yes --disble-nls to e2fsprogs |
| 08:29.41 | landley | e2fsprogs has ./configure. It should be able to figure _out_ that it hasn't got nls support, shouldn't it? |
| 08:29.55 | *** join/#edev luke-jr (n=luke-jr@2002:62b3:1d4c:0:20e:a6ff:fec4:4e5d) |
| 08:30.11 | landley | But sure, hit ./configure with a bigger hammer, that fixes things half the time... |
| 08:30.13 | ncopa | how many autotools packages have that part actually working do you think? |
| 08:30.21 | landley | ncopa: few or none? |
| 08:30.27 | landley | ./configure is an abomination. |
| 08:30.27 | ncopa | some |
| 08:30.48 | ncopa | i need hg to make pathes? |
| 08:30.54 | landley | If you're ever testing for more than a couple dozen things, something is deeply wrong. |
| 08:30.56 | MonMotha | oh dear...autoconf |
| 08:31.03 | landley | ncopa: nah, just email me a diff -u. |
| 08:31.05 | MonMotha | runs away screaming in horror |
| 08:31.34 | Mirell | whispers softly in MonMotha's ear...nested recursive makefiles |
| 08:31.47 | MonMotha | you know, they seem to work better than autoconf... |
| 08:32.02 | landley | See "damning with faint praise". |
| 08:32.17 | ncopa | landley: where di my toolchain go? |
| 08:32.38 | ncopa | an build/root-filesys |
| 08:33.00 | ncopa | ah* |
| 08:33.15 | landley | Under "build"? |
| 08:33.22 | ncopa | yeah i found it |
| 08:33.41 | landley | I need to write up a tips and tricks file. |
| 08:34.28 | landley | Does the result boot under qemu? |
| 08:34.33 | landley | what architecture are you targeting, anyway? |
| 08:34.44 | ncopa | x86 |
| 08:34.57 | landley | uclibc-nptl doesn't support x86 yet. |
| 08:35.02 | ncopa | well |
| 08:35.05 | landley | Well, except using old threading. |
| 08:35.14 | ncopa | i just build hello.c |
| 08:35.23 | ncopa | with a native x86 nptl toolchain |
| 08:35.45 | landley | Look at sources/native/src/thread-hello2.c |
| 08:35.59 | landley | That's an egregiously threaded hello world. :) |
| 08:36.41 | landley | I believe it exercises everything except thread cancellation. :) |
| 08:36.51 | MonMotha | hello, world requires threads/ |
| 08:36.52 | MonMotha | ? |
| 08:37.02 | ncopa | landleys version does :) |
| 08:37.08 | luke-jr | lol |
| 08:37.40 | ncopa | bah... the root-fs uses bash |
| 08:37.57 | ncopa | thats why my auto-complete didnt work |
| 08:38.11 | ncopa | bb ash is fairly competent nowdays |
| 08:38.28 | landley | http://impactlinux.com/hg/hgwebdir.cgi/firmware/file/tip/sources/native/src/thread-hello2.c |
| 08:38.38 | landley | An explicit test of the C library's threading infrastructure. |
| 08:38.48 | landley | If it prints out "Hello world" it means threading is working. |
| 08:38.51 | MonMotha | gah, I have a PDF with a table in it, and I want to actually get it into a usable tabular format. It's proving to be difficult. |
| 08:39.03 | landley | ncopa: oneit should be running ash. |
| 08:39.18 | ncopa | i run it now |
| 08:39.26 | ncopa | it was my chroot command |
| 08:39.44 | landley | Ah. |
| 08:40.07 | landley | You can still have chroot run sbin/init.sh, see lines 69-83 of http://impactlinux.com/hg/hgwebdir.cgi/firmware/file/tip/sources/native/sbin/init.sh |
| 08:40.23 | landley | But running ash directly works too. :) |
| 08:40.32 | ncopa | i will probably run it in a vserver guest |
| 08:41.04 | landley | is targeting qemu with the system-images. |
| 08:41.09 | *** join/#edev MonMotha (n=monmotha@cpe-65-29-71-209.indy.res.rr.com) |
| 08:41.25 | landley | Which presumably means kvm, but my laptop hasn't got the darn vt extensions kvm needs. :( |
| 08:41.33 | ncopa | :( |
| 08:41.36 | ncopa | kvm helps |
| 08:41.38 | landley | (Yes, there is an x86_64 processor that hasn't got the vt extensions. They made it for a few months, and I have one.) |
| 08:41.48 | landley | It's a core 2 duo even. |
| 08:41.57 | ncopa | kqemu? |
| 08:42.03 | ncopa | helps too |
| 08:42.08 | landley | ncopa: in theory, yes. In practice, I've never gotten kqemu to work. |
| 08:42.09 | landley | Ever. |
| 08:42.14 | ncopa | lol |
| 08:42.29 | landley | I've tried, on and off, for over 3 years. |
| 08:42.34 | landley | It just doesn't _like_ me. |
| 08:42.38 | ncopa | :) |
| 08:42.48 | landley | Occasionally it's claimed to be working... and provided no acceleration whatsoever. |
| 08:42.55 | landley | (Or hung. That's always fun too.) |
| 08:43.14 | ncopa | i have not got it working on my alpine linux either |
| 08:43.23 | ncopa | but qemu works |
| 08:43.27 | MonMotha | landley: IIRC, Intel is still making C2Ds without VT |
| 08:43.34 | ncopa | yup.. qemu on uclibc host |
| 08:43.46 | MonMotha | in fact, they sell them under the same marketing model numbers as other procs that DO have VT. You have to check the stepping to figure out if it supports it or not. |
| 08:45.55 | ncopa | qemu running on my uclibc laptop: http://dev.alpinelinux.org/~ncopa/alpine/Screenshot.png |
| 08:48.11 | ncopa | hum |
| 08:48.21 | ncopa | [root@ncopa-desktop /src]# ./thread-hello2 |
| 08:48.21 | ncopa | Can't modify /lib/libpthread.so.0's text section. Use GCC option -fPIC for shared objects, please. |
| 08:48.44 | ncopa | seems like libpthread is broken |
| 08:52.45 | landley | ncopa: not surprised. |
| 08:52.58 | landley | Did it work on your uClibc host? |
| 08:53.08 | ncopa | um well yes |
| 08:53.26 | ncopa | oh you mean the pthread hello? |
| 08:53.30 | ncopa | didnt test that yet |
| 08:53.44 | landley | It worked for me, but I'd make sure... |
| 08:53.59 | ncopa | but i did compile uclibc-nptl on my uclibc host |
| 08:54.18 | landley | Running gcc doesn't use threads. |
| 08:54.22 | landley | Neither does make -j. |
| 08:55.04 | ncopa | and built a phtread hello with CFLAGS=-I/path/to/my/nptl LDFLAGS=-L/path/to/mynptl |
| 08:55.16 | ncopa | and chrooted |
| 08:55.23 | ncopa | that worked |
| 08:55.31 | ncopa | after some patching |
| 08:55.47 | ncopa | i thought the patches was applied upstream... |
| 08:56.59 | *** join/#edev hw_ (n=hw@p578b3905.dip0.t-ipconnect.de) |
| 08:57.18 | landley | The dynamic loader is horked then? |
| 08:57.29 | landley | Run ldd on it and see where ld-uClibc lives. |
| 08:57.49 | landley | (It could quite easily be the dynamic loader being horked. That requires some changes for nptl.) |
| 08:58.54 | ncopa | [root@ncopa-desktop /src]# ldd thread-hello2 |
| 08:58.54 | ncopa | Can't modify application's text section; use the GCC option -fPIE for position-independent executables. |
| 08:59.18 | ncopa | I'll trywith -fPIE |
| 08:59.55 | MonMotha | mmm...pie |
| 09:00.19 | ncopa | i think ldd gets confused by the non PIC libpthread |
| 09:00.23 | landley | What the... I thought your host system was uClibc? |
| 09:00.35 | ncopa | not this time |
| 09:00.37 | landley | The uClibc ldd is a mutant "readelf". It shouldn't care about the application's sections... |
| 09:01.04 | ncopa | Dynamic section at offset 0x87c contains 15 entries: |
| 09:01.04 | ncopa | <PROTECTED> |
| 09:01.04 | ncopa | <PROTECTED> |
| 09:01.05 | ncopa | <PROTECTED> |
| 09:01.07 | landley | The uClibc ldd is in fact superior to the brain-damaged glibc one which tries to launch the program and then dump data about it instead of running main(). |
| 09:01.41 | landley | readelf -a a.out | grep ld- |
| 09:02.11 | landley | It should say "[Requesting program interpreter: thingy]" |
| 09:02.17 | landley | The "thingy" is the interesting bit. |
| 09:02.44 | ncopa | # readelf -d /lib/libc.so |
| 09:02.44 | ncopa | readelf: Error: Unable to read in 0x2074 bytes of section headers |
| 09:02.44 | ncopa | readelf: Error: Not an ELF file - it has the wrong magic bytes at the start |
| 09:03.00 | landley | file a.out |
| 09:03.19 | ncopa | im in the root-filesystem |
| 09:03.21 | ncopa | no file in there |
| 09:03.27 | landley | Run the "file" command on a.out |
| 09:03.37 | landley | There is a command named "file". |
| 09:03.45 | landley | It identifies the type of a file, often in some detail. |
| 09:04.01 | landley | $ file /bin/ls |
| 09:04.02 | landley | /bin/ls: ELF 64-bit LSB executable, x86-64, version 1 (SYSV), dynamically linked (uses shared libs), for GNU/Linux 2.6.8, stripped |
| 09:04.12 | ncopa | yes but the root-system does not have file i need to exit.. just a sec |
| 09:04.18 | landley | Ah. |
| 09:04.33 | ncopa | $ file thread-hello2 |
| 09:04.33 | ncopa | thread-hello2: ELF 32-bit LSB executable, Intel 80386, version 1 (SYSV), dynamically linked (uses shared libs), not stripped |
| 09:04.39 | landley | Yeah, it's not in my root-filesystem. It's a separate package. |
| 09:04.44 | *** join/#edev mbuf (n=mbuf@118.102.130.6) |
| 09:05.00 | landley | (Busybox hasn't got it either because its brains are a big data file that tell it all the various file magic to look for.) |
| 09:05.17 | ncopa | i have my own package manager and package building scripts i intend to use |
| 09:05.20 | landley | Hmmm... so readelf didn't think it was an elf file, but the file command did? |
| 09:05.24 | ncopa | once this toolchain is working... |
| 09:05.27 | landley | That's odd. |
| 09:05.35 | ncopa | well |
| 09:05.35 | landley | You ran "readelf" out of the host filesystem, right? |
| 09:05.41 | landley | Does your host have strace? |
| 09:06.08 | landley | Anyway, run "strings" on the a.out file and grep for "ld-" |
| 09:06.58 | ncopa | readelf on binary hello does work in chroot |
| 09:07.02 | ncopa | http://pastebin.com/d8020f0e |
| 09:07.18 | ncopa | readelf on /lib/libc.so does not |
| 09:07.57 | landley | The uClibc readelf isn't mentioning the dynamic loader? Odd. |
| 09:08.10 | landley | I'll have to fix that. |
| 09:08.11 | ncopa | heh |
| 09:08.21 | landley | Anyway, could you run "strings" on the binary and grep for "ld-" |
| 09:09.08 | ncopa | heh.. /lib/libc.so is a GNU ld script |
| 09:09.49 | landley | ncopa: yup. |
| 09:09.51 | landley | Has been for a while. |
| 09:09.57 | ncopa | [root@ncopa-desktop /src]# strings thread-hello2 | grep 'ld-' |
| 09:09.59 | ncopa | /lib/ld-uClibc.so.0 |
| 09:10.07 | landley | But in _my_ root filesystem, it doesn't have absolute paths hardwired into it. :) |
| 09:10.15 | landley | (One of the many, many tricks to make my cross compilers relocatable.) |
| 09:10.29 | landley | Ok, so it's using the dynamic loader out of your host system. |
| 09:10.37 | landley | When you run it out of the chroot, it should use the dynamic loader out of the chroot. |
| 09:10.58 | landley | So it looks like the dynamic loader is what's broken. |
| 09:11.35 | ncopa | its not libpthread that is compiled without -fPIC? |
| 09:11.49 | landley | You're welcome to try that and see if it fixes it. |
| 09:12.09 | ncopa | um.., i need to rebuild everything then? |
| 09:12.18 | landley | But I thought you'd build it against the chroot's libraries and it ran ok on the host? |
| 09:12.34 | ncopa | on other box |
| 09:12.37 | ncopa | other test |
| 09:12.49 | ncopa | i never got the gcc work that way |
| 09:12.50 | landley | The bit with CFLAGS=-I/path/to/my/nptl LDFLAGS=-L/path/to/mynptl |
| 09:13.01 | landley | Ah. |
| 09:13.02 | ncopa | yes, that was other chroot |
| 09:13.10 | ncopa | i coudl never get gcc work that way |
| 09:13.18 | ncopa | so i droppped that and went for fwl |
| 09:13.33 | landley | CFLAGS and LDFLAGS aren't used by the compiler, they're used by makefiles. |
| 09:13.41 | landley | This didn't have a makefile, so they wouldn't make a difference in a command line gcc invocation. |
| 09:13.53 | landley | You'd have to add the -I and -L on the gcc command line. |
| 09:14.13 | ncopa | i havent tested your ptherad-hello on that box |
| 09:14.29 | landley | Although why the uClibc build wouldn't be adding -fPIC to its build flags if it actually needed that, I have no idea. |
| 09:14.39 | landley | shrugs. |
| 09:14.51 | landley | Anyway, I find that thread-hello2.c to be a good smoke test for the threading subsystem. |
| 09:15.08 | ncopa | nice |
| 09:15.10 | ncopa | thanks |
| 09:15.22 | landley | You're welcome. I stagger vaguely bedwards now... |
| 09:15.23 | ncopa | i found fwl useful too |
| 09:15.30 | ncopa | yeah get some sleep |
| 09:15.32 | landley | Thanks. I plan to continue banging on it. |
| 09:15.39 | ncopa | I'll get some coffe |
| 09:15.41 | landley | Working on another release monday. |
| 09:15.47 | ncopa | ok |
| 09:15.56 | landley | I have an energy drink, but Fade would be a touch upset if I got _completely_ off a day schedule. Again. |
| 09:16.26 | ncopa | yeah. get soem sleep. it helps :) |
| 09:17.09 | landley | g'night. |
| 09:44.11 | *** join/#edev PureSine (n=PureSine@87.247.172.211) |
| 09:47.07 | mnemoc | i'm always late to bother landley :-\ |
| 09:50.04 | ncopa | what url should i use for checking out the firwmare linux from hg? |
| 09:51.48 | *** join/#edev PureSine (n=PureSine@87.247.172.211) |
| 10:07.24 | SpeedEvil | land |
| 10:07.24 | SpeedEvil | oh |
| 10:07.43 | SpeedEvil | insane granularity of erase which can be fixed largely by linear write |
| 10:42.20 | *** join/#edev keesj (n=keesj@ip49-193-210-87.adsl2.static.versatel.nl) |
| 10:42.42 | *** join/#edev florian_kc (n=fuchs@Maemo/community/contributor/florian) |
| 11:05.16 | *** join/#edev JoeLlama (n=snork@206.169.63.10) |
| 11:07.39 | *** join/#edev keesj (n=keesj@ip49-193-210-87.adsl2.static.versatel.nl) |
| 11:25.05 | *** join/#edev pirho (i=pirho@gateway/gpg-tor/key-0x2CEEC9CB) |
| 11:39.42 | *** join/#edev pirho (i=pirho@gateway/gpg-tor/key-0x2CEEC9CB) |
| 11:50.45 | *** join/#edev SpeedEvil (n=user@87.113.130.190) |
| 12:15.41 | *** join/#edev gustavoz (n=gustavoz@host248.201-252-31.telecom.net.ar) |
| 12:45.11 | *** join/#edev PureSine2 (n=PureSine@87.247.172.211) |
| 13:03.56 | *** join/#edev mbuf (n=mbuf@122.175.94.89) |
| 13:07.27 | ncopa | aha.... |
| 13:07.37 | ncopa | its ccwrap that does the segfault |
| 13:08.10 | ncopa | bah |
| 13:52.14 | *** join/#edev cbx (n=cbx@unaffiliated/cbx) |
| 13:54.44 | *** part/#edev dsr (i=0k2yo1v7@76.115.94.244) |
| 14:01.17 | *** part/#edev cbx (n=cbx@unaffiliated/cbx) |
| 14:04.40 | *** join/#edev roxfan (i=dunno@189.218-200-80.adsl-dyn.isp.belgacom.be) |
| 14:13.50 | *** join/#edev linac (n=lin@122.90.123.152) |
| 14:34.36 | *** join/#edev GPSFan (n=kenm@64.92.145.112) |
| 14:34.36 | *** mode/#edev [+o GPSFan] by ChanServ |
| 14:38.53 | *** join/#edev SpeedEvil (n=user@tor/regular/SpeedEvil) |
| 14:46.25 | *** join/#edev Crofton (n=balister@12.238.63.177) |
| 14:51.00 | *** join/#edev brolin_ (n=brolin@iner.udea.edu.co) |
| 15:02.49 | prpplague | http://www.nbcdfw.com/enjoy-this/North-Texas-Wakes-to-Big-Snowflakes-78310222.html |
| 15:07.49 | *** join/#edev kergoth (n=kergoth@nat/montavista/x-zogbxynntkfefjof) |
| 15:07.56 | *** part/#edev Reggie (n=reggie@89.243.158.195) |
| 15:12.08 | florian | prpplague: heh |
| 15:12.15 | prpplague | florian: greetings |
| 15:12.23 | prpplague | tries to get motivated today |
| 15:12.38 | florian | prpplague: we ususally have weather like this at this time of the year... but not this year |
| 15:13.22 | kergoth | morning |
| 15:13.28 | florian | hi kergoth |
| 15:13.34 | kergoth | hey |
| 15:13.57 | prpplague | hey kergoth |
| 15:14.05 | prpplague | kergoth: how's trix? |
| 15:14.45 | kergoth | sick, so it seems, have what appears to be a swollen lymph node on one side of my jaw, quite annoying |
| 15:14.47 | kergoth | otherwise not bad, you? |
| 15:14.49 | kergoth | heh |
| 15:14.58 | prpplague | kergoth: not too bad myself |
| 15:15.33 | prpplague | kergoth: cold? or do you have a "too-mer" |
| 15:15.47 | kergoth | heh, no idea what it is, guess i'll find out at the doc today |
| 15:17.07 | prpplague | kergoth: just put some windex on it, it will go away |
| 15:17.28 | kergoth | hehe |
| 15:18.11 | *** topic/#edev by prpplague -> A channel for the discussion of embedded systems (Hardware/Software/Linux/WinCE/Commercial Device Repurposing) : Firmware Linux Discussion : Busybox and Buildroot Discussion : Gold 1217.00 : Silver 19.01 |
| 15:18.22 | *** topic/#edev by prpplague -> A channel for the discussion of embedded systems (Hardware/Software/Linux/WinCE/Commercial Device Repurposing) : Firmware Linux Discussion : Busybox and Buildroot Discussion : Gold 1217.00 : Silver 19.07 |
| 15:20.27 | florian | watches prpplague becoming rich ;) |
| 15:20.56 | prpplague | doubt i'll be rich, just making sure that when the USD collapses i'm not broke |
| 15:21.14 | florian | :-) |
| 15:30.39 | SpeedEvil | Unless the government implements restrictions on the private holding of gold, and your banking provider caves. |
| 15:32.46 | prpplague | SpeedEvil: i doubt they'd go into my safety deposit boxes, plus i have some at home, and overseas in barbados |
| 15:32.59 | prpplague | SpeedEvil: so no matter what i'll have some somewhere |
| 15:33.05 | SpeedEvil | Ah. |
| 15:33.26 | SpeedEvil | Or unless someone works out a way to cheaply extract it from seawater. |
| 15:35.36 | chouimat | mornign |
| 15:36.46 | prpplague | chouimat: greetings |
| 15:36.50 | prpplague | SpeedEvil: indeed |
| 15:37.20 | chouimat | prpplague: 121x usd now? |
| 15:38.00 | prpplague | chouimat: yep |
| 15:38.10 | chouimat | nice |
| 15:38.33 | prpplague | chouimat: people are expecting it to hit $1250 by xmas and $1800 by the end of feb |
| 15:39.25 | chouimat | interesting |
| 15:40.41 | chouimat | prpplague: saw this? http://www.reuters.com/article/wtUSInvestingNews/idUSTRE5B01Z920091201?sp=true |
| 15:40.58 | prpplague | looks |
| 15:47.04 | prpplague | hehe |
| 15:47.46 | kergoth | wishes he had the money to invest in much of anything, gold or otherwise.. maybe if his dad actually starts paying him his rent.. |
| 15:48.26 | chouimat | kergoth: ouch |
| 15:48.43 | kergoth | he owes me $13k now :\ |
| 15:48.43 | *** join/#edev cbx (n=cbx@unaffiliated/cbx) |
| 15:48.47 | kergoth | never rent to relatives. |
| 15:48.59 | prpplague | kergoth: indeed |
| 15:49.06 | prpplague | kergoth: relatives are evil |
| 15:49.25 | chouimat | prpplague: yup and they drink all your coffee and beer |
| 15:49.29 | kergoth | hehe |
| 15:57.35 | *** join/#edev Crofton (n=balister@12.238.63.177) |
| 15:58.07 | prpplague | Crofton: ! |
| 15:59.57 | SpeedEvil | New twist. The FBI mailing me to inform me that their investigation to prove that I've been cleared of being a terrorist, so the funds transfer can go through. |
| 16:00.16 | SpeedEvil | I just have to supply a small processing fee |
| 16:00.21 | prpplague | SpeedEvil: wow |
| 16:00.46 | SpeedEvil | their investigation into me being a terrorist has proved that I'm not, rather. |
| 16:02.27 | prpplague | SpeedEvil: is this some sort of phish scam or are you serious? |
| 16:02.40 | SpeedEvil | yes, it's a scam |
| 16:02.58 | SpeedEvil | it's a new twist on the normal 'you have been left x million dollars' scam |
| 16:05.31 | *** join/#edev davidc__ (n=davidc__@user-387g3od.cable.mindspring.com) |
| 16:09.09 | *** join/#edev olivdt (n=x@unaffiliated/olivdt) |
| 16:15.23 | *** join/#edev PureSine2 (n=PureSine@87.247.173.124) |
| 16:16.06 | *** join/#edev eox425 (n=root@213.133.177.11) |
| 16:29.55 | *** join/#edev flavioribeiro (n=flavio@187.64.55.50) |
| 16:35.53 | *** join/#edev dirk2 (n=dirk@p5B042325.dip0.t-ipconnect.de) |
| 16:36.26 | prpplague | dirk2: greetings |
| 16:45.19 | *** join/#edev rektide (i=rektide@voodoowarez.com) |
| 16:48.45 | prpplague | lordy i'm having trouble focusing today |
| 17:08.19 | *** join/#edev CosmicPenguin (n=nobody@66-162-99-237.static.twtelecom.net) |
| 17:10.02 | *** join/#edev PureSine2 (n=PureSine@87.247.173.124) |
| 17:11.33 | *** join/#edev PureSine (n=PureSine@87.247.173.124) |
| 17:16.49 | Mirell | Ah, nothing like starting out the morning sitting in a server room with a laptop and serial cable |
| 17:20.17 | rektide | i presume thats a usb-serial cable |
| 17:20.31 | rektide | i havent seen an honest to god serial port in oer half a decade |
| 17:29.22 | Mirell | Yah, USB serial |
| 17:29.52 | Mirell | SpeedEvil: ...wah? Terrorist? |
| 17:30.01 | SpeedEvil | Mirell: It was a scam email. |
| 17:30.05 | Mirell | Ah |
| 17:30.36 | SpeedEvil | Mirell: Claiming to be from the FBI, and that they had conducted an investigation into this money transfer, as it was suspected I was one. But it's been determined that I'm not, and for a small fee... |
| 17:30.52 | Mirell | Nice... |
| 17:43.53 | prpplague | hmm, need to order some xmas tree pcb's from bgmicro |
| 17:47.28 | SpeedEvil | ? |
| 17:47.40 | *** join/#edev pirho (i=pirho@gateway/gpg-tor/key-0x2CEEC9CB) |
| 17:48.38 | *** join/#edev RemoteSine (n=chatzill@87.247.173.124) |
| 17:49.55 | prpplague | SpeedEvil: http://www.bgmicro.com/index.asp?PageAction=VIEWPROD&ProdID=11924 |
| 17:50.31 | prpplague | SpeedEvil: this one is pretty good too - http://www.bgmicro.com/index.asp?PageAction=VIEWPROD&ProdID=12616 |
| 17:51.41 | prpplague | SpeedEvil: my slaves er i mean my kids love to put them together |
| 17:57.24 | mnemoc | :) |
| 19:05.27 | kergoth | prpplague: i think the correct term is minions :) |
| 19:05.40 | prpplague | kergoth: hehe |
| 19:21.23 | *** join/#edev florian (n=fuchs@Maemo/community/contributor/florian) |
| 19:52.14 | *** join/#edev SandGorgon (n=OmNomNom@122.173.249.152) |
| 20:01.03 | prpplague | wmat: tim is driving me nuts with his pages |
| 20:01.22 | Mirell | Tim? |
| 20:05.37 | davidc__ | prpplague: Page: "prpplague - just letting you know I'm going to send you a page" |
| 20:05.37 | davidc__ | Page: "Just to confirm that you got the pre-page page" |
| 20:06.02 | prpplague | Mirell: tim bird with CELF |
| 20:06.07 | prpplague | davidc__: hehe |
| 20:06.23 | prpplague | is OCD when it comes to the elinux wiki |
| 20:07.57 | prpplague | i end up spending about 10-12 hours a month working on orphaned pages and correcting stuff |
| 20:15.23 | prpplague | gah, 143 orphaned pages |
| 20:23.15 | wmat | prpplague: really, what's he doing? |
| 20:23.56 | wmat | looks |
| 20:23.58 | prpplague | wmat: just adding pages like: |
| 20:24.00 | prpplague | # (diff) (hist) . . Decrease X.org XFree86 server footprintâ; 19:07 . . (+32) . . Tim Bird (Talk | contribs | block) (âComments: add to category) [rollback] |
| 20:24.14 | prpplague | in the root wiki |
| 20:24.27 | prpplague | i've already sent an email to discuss changing things up |
| 20:24.39 | prpplague | he has CELF crap everywhere |
| 20:25.15 | wmat | CELF sponsors elinux.org |
| 20:25.32 | wmat | not that the site is CELF's site though |
| 20:26.16 | wmat | ugh, he's adding mailing list posts to the pages |
| 20:26.27 | wmat | i'd rather he pushed people toward the discussion pages |
| 20:27.12 | Mirell | I keep thinking in my mind Tim Riker when you say Tim Bird. |
| 20:28.12 | wmat | Tim R. is the original elinux.org guy, and still plays a big role |
| 20:29.29 | prpplague | wmat: don't forget i was one of the original elinux.org guys too |
| 20:29.43 | prpplague | wmat: i don't mind the CELF stuff, just as long as it is organized |
| 20:29.48 | wmat | of course |
| 20:29.56 | prpplague | wmat: problem is that it isn't organized |
| 20:30.29 | wmat | just tell Tim B. how you'd prefer it organized |
| 20:30.45 | prpplague | wmat: with the exception of flameman's stuff, all the remaining orphaned pages are CELF related |
| 20:31.17 | prpplague | wmat: basically for conferences and such they need to be organized in subdirs |
| 20:31.37 | prpplague | wmat: i.e. http://www.elinux.org/CELF/2005/presentations |
| 20:34.14 | wmat | that's not how anything else is organized though and not really how mediawiki handles organization; unless I'm missing something. |
| 20:34.36 | wmat | there could be a CELF category though |
| 20:34.55 | wmat | with CELF specific activities using it |
| 20:35.22 | prpplague | wmat: just that most of the upper pages are ok being flat |
| 20:35.32 | prpplague | wmat: an example is the hawkboard stuff |
| 20:35.43 | wmat | I've had a TODO to fix categories on the wiki for like a year now |
| 20:35.46 | prpplague | wmat: they have a sub page for their early adopters program |
| 20:36.37 | prpplague | wmat: like this - http://www.elinux.org/Hawkboard/earlyadopter |
| 20:37.16 | wmat | I see, but that's poor usage in that a page named "Hawkboard/earlyadopter" is not very intuitive |
| 20:38.46 | prpplague | wmat: but that forces you to make the page accessable from the Hawkboard pages making it easy to find |
| 20:39.11 | wmat | it's not really a subpage, it's a page name with a / in it |
| 20:39.33 | wmat | that could have been "Hawkboard---earlyadopter" |
| 20:39.35 | wmat | or whatever |
| 20:40.13 | wmat | [[Hawkboard/earlyadopter|Early Adopter Contest]] |
| 20:41.38 | wmat | that would work if subpaging was turned on though |
| 20:43.19 | prpplague | agreed |
| 20:43.40 | wmat | you can use it in talk and user pages now |
| 20:43.44 | prpplague | i know it isn't a true page, but it really helps with keeping the pages straight |
| 20:44.00 | wmat | perhaps we should turn it on |
| 20:44.26 | wmat | i'll look into it, as I'm curious why it's off by default |
| 20:44.29 | Mirell | I just hate wikis. |
| 20:53.51 | *** join/#edev JoeLlama (n=snork@206.169.63.10) |
| 20:57.12 | prpplague | Mirell: you probably feel about wiki's as i do about mailing lists |
| 21:04.14 | Mirell | prpplague: That they grow stale really fast? |
| 21:04.41 | prpplague | Mirell: yep |
| 21:04.55 | prpplague | prefers to use wiki's for documentation |
| 21:05.12 | prpplague | Mirell: i like the free form method that wikis allow |
| 21:11.30 | SpeedEvil | Wikis can be very broken very fast. |
| 21:11.40 | SpeedEvil | When you let users at them. |
| 21:20.56 | *** join/#edev likewise (n=likewise@82-171-51-231.ip.telfort.nl) |
| 21:38.36 | Mirell | prpplague: ...documentation on Wikis is the serious reason I hate them among others. STALE DOCUMENTATION |
| 21:40.44 | prpplague | Mirell: ahh as you mean for reference material |
| 21:40.47 | CosmicPenguin | As opposed to all the other documentation that isn't stale? |
| 21:40.58 | prpplague | Mirell: i agreed most public wikis are badly maintained |
| 21:47.18 | prpplague | Mirell: i just find it easier to do the documentation on the wiki than in a journal |
| 21:48.33 | kergoth | what i hate about wikis is the lack of a definitive location for a given thing, so you can end up with 12 different pages all covering the same thing, and you don't know which is the most recent.. |
| 21:48.37 | kergoth | (amogn other things) |
| 21:48.56 | Mirell | That too |
| 21:50.02 | prpplague | wmat: http://www.elinux.org/index.php?title=Special:LonelyPages&limit=500&offset=0 |
| 21:50.29 | prpplague | wmat: with the exception of flamemans stuff, the remaining orphaned pages are all CELF related |
| 21:50.42 | kergoth | would rather have official docs (in general, anyway) be rst/sphinx in a git repository or something, or even docbook or latex, and let wiki do what its really good at.. brainstorming, gathering together disperate(sp?) bits of info, etc |
| 21:50.53 | Mirell | prpplague: You want the TimBirdFree Wiki |
| 21:51.04 | prpplague | kergoth: yea that problems is due to the lack of management |
| 21:51.11 | kergoth | nods |
| 21:51.16 | prpplague | Mirell: naw, just organization |
| 21:51.39 | prpplague | as kergoth is pointing out, if you can't find the information it might as well not exist |
| 21:51.41 | kergoth | something like elinux is particularly difficult, at least with a typical oss project there's usually a person or persons with sufficient Authority |
| 21:51.53 | *** join/#edev MonMotha (n=monmotha@cpe-65-29-71-209.indy.res.rr.com) |
| 21:52.10 | prpplague | kergoth: indeed, there are about 4 or 5 of us that work part time so to speak on keeping it organized |
| 21:52.42 | prpplague | just changes today - http://www.elinux.org/index.php?title=Special:RecentChanges&limit=500 |
| 21:54.55 | *** join/#edev MuscleNerd (i=eric@adsl-75-27-189-151.dsl.irvnca.sbcglobal.net) [NETSPLIT VICTIM] |
| 22:04.25 | wmat | yeah, changes are increasing of late |
| 22:04.42 | wmat | it's hard to keep up on a part time basis |
| 22:04.59 | chouimat | and looks like prpplague was busy today ;) |
| 22:16.06 | prpplague | chouimat: just trying to take a break on my project and do some docs |
| 22:16.26 | *** join/#edev likewise (n=likewise@82-171-51-231.ip.telfort.nl) |
| 22:16.30 | *** join/#edev Crofton (n=balister@12.238.63.177) |
| 22:16.33 | chouimat | prpplague: breaks are good :) |
| 22:16.39 | prpplague | kergoth / Mirell i use mediawiki to do all my docs on projects, along with keeping trying of reverse engineering info |
| 22:16.54 | Mirell | I have stared into the gaping maw of a build system, and won. |
| 22:17.32 | chouimat | Mirell: I spent the afternoon fighting opensolaris on sparc and lost ... |
| 22:18.54 | prpplague | anyone here own a toyota prius? |
| 22:25.38 | prpplague | wmat: nice we have the orphaned pages down to 116 |
| 22:28.30 | *** join/#edev chiques (n=working_@pool-71-189-111-152.lsanca.fios.verizon.net) |
| 22:30.00 | monkey_ | hey prpplague |
| 22:30.11 | prpplague | monkey_: greetings earthling |
| 22:30.25 | monkey_ | does this make any sense? Error accessing memory address 0x234: Unknown error 4294967295. |
| 22:30.34 | monkey_ | I'm technically an alien |
| 22:30.57 | monkey_ | i have a green card so Im' a resident alien breakpoints.c:73 breakpoint_add(): can't add hardware breakpoint, resource not available |
| 22:32.08 | roxfan | sounds like too many breakpoints |
| 22:32.15 | monkey_ | i only have 2 picked |
| 22:32.28 | roxfan | and trying to set a third? |
| 22:32.55 | monkey_ | nope |
| 22:33.01 | monkey_ | i just set the 2, and ran the program |
| 22:33.09 | roxfan | hmm |
| 22:33.22 | monkey_ | if i don't set a breakpoint for the 1st one it wont hit the 2nd one. let me verify this again. |
| 22:34.56 | monkey_ | gd it! |
| 22:35.11 | monkey_ | so, i only have one breakpoint, and I stepped after the first one and it gives me this breakpoints.c:73 breakpoint_add(): can't add hardware breakpoint, resource not available |
| 22:35.54 | roxfan | weird |
| 22:35.55 | monkey_ | some unknown error -308 as well |
| 22:36.26 | monkey_ | 2 breakpoints set now I get : Error: gdb_server.c:1024 gdb_error(): unexpected error -308 |
| 22:37.09 | monkey_ | Error: arm7_9_common.c:171 arm7_9_set_breakpoint(): BUG: no hardware comparator available |
| 22:39.02 | roxfan | i hear arm7s have only two breakpoints |
| 22:39.08 | monkey_ | hey Mirell I had to use flash magic in windows to erase the flash, and openocd started working again |
| 22:39.15 | Mirell | monkey_: Um, cool? |
| 22:39.22 | Mirell | monkey_: Think you were talking to landley about that, not me. |
| 22:39.32 | roxfan | if you can run the code from ram that should work better |
| 22:39.35 | monkey_ | hey landley I had to use flash magic in windows to erase the flash, and openocd started working again |
| 22:39.45 | monkey_ | :) |
| 22:39.48 | Mirell | monkey_: landley is also not here, heh. |
| 22:40.01 | monkey_ | roxfan: you're right. there are only 2 hardware breakpoints, but it shouldn't be giving me this error |
| 22:40.10 | monkey_ | :( |
| 22:40.40 | roxfan | well, if there's no hardware step then to step from a breakpoint the debugger needs to put a temp breakpoint on next instruction |
| 22:40.43 | monkey_ | roxfan: the only purpose of this code, is from the teacher, we don't modify it or anything, we just poke around the dissassembly and debugging |
| 22:40.49 | monkey_ | ok |
| 22:41.27 | monkey_ | also, the first time i ran this, it worked |
| 22:41.33 | roxfan | hm |
| 22:41.50 | roxfan | maybe the old breakpoints are still there then |
| 22:41.50 | monkey_ | the first assembly instruction shoudl've been a comp |
| 22:42.03 | roxfan | try a power-off reset? |
| 22:42.24 | monkey_ | roxfan: just hit the reset button while powering on? |
| 22:42.47 | roxfan | i mean, cut power to the chip |
| 22:42.53 | roxfan | not just press reset button |
| 22:43.10 | roxfan | i think warm reset doesn't clear debug stuff |
| 22:43.53 | monkey_ | do i need to pull the battery too? |
| 22:44.16 | roxfan | maybe... depends on how it's connected |
| 22:44.30 | roxfan | if it's just for rtc then probably not |
| 22:44.51 | monkey_ | there is a small circular battery directly on the board, and it's using usb power too |
| 22:44.51 | monkey_ | connected to a arm-usb-ocd thingie |
| 22:45.41 | roxfan | coin battery is usually for rtc |
| 22:45.48 | monkey_ | ok |
| 22:45.50 | roxfan | so shouldn't be necessary |
| 22:46.43 | monkey_ | ok |
| 22:46.47 | monkey_ | let me try that and see what happens |
| 22:47.47 | *** join/#edev archae0pteryx (n=snewman@207.47.42.130.static.nextweb.net) |
| 22:48.39 | prpplague | monkey_: how long did it take you to get a visa and what kind is it? |
| 22:49.38 | prpplague | monkey_: what board are you working on? |
| 22:49.53 | prpplague | wmat: ping |
| 22:49.58 | monkey_ | prpplague: lpc2378 olimex board |
| 22:50.22 | monkey_ | monkey_: i came here as a kid with a soldier who married my mom, so i got a green card pretty easily |
| 22:50.40 | prpplague | monkey_: must have been over 10 years ago |
| 22:50.49 | prpplague | monkey_: i'm guessing over 15 |
| 22:51.27 | prpplague | monkey_: because all the immigration laws changed around the mid 90's |
| 22:51.46 | monkey_ | yeah |
| 22:51.51 | monkey_ | it was in the 10s of years ago |
| 22:51.57 | prpplague | my wife is a LPR |
| 22:51.59 | monkey_ | 20 or so |
| 22:52.02 | monkey_ | ah |
| 22:52.35 | prpplague | took us 3 years 8 months and 23 days to get her a visa |
| 22:54.13 | prpplague | @!#$ my office is cold |
| 22:54.29 | monkey_ | is that from barbados right? |
| 22:55.02 | prpplague | yea |
| 22:55.47 | monkey_ | ok, after the reset it didn't give me more breakpoint messages, but the assembly is still not right |
| 22:57.40 | monkey_ | the first time i ran this, went to cmp r1, #0, but now it's a mov |
| 22:57.52 | monkey_ | the instructions say it sould be cmp r1,#0 |
| 23:01.43 | roxfan | is the address correct? |
| 23:01.56 | monkey_ | well, it's the first instruction in the READ8() |
| 23:02.26 | monkey_ | after the first time i ran it , i guess i had a bad flash, as openocd wasn't working after that |
| 23:02.39 | monkey_ | i had to use flash magic to clear the flash and then I was able to reprogram it |
| 23:02.43 | monkey_ | and it's been different ever since |
| 23:03.17 | SpeedEvil | checks the inside temp. |
| 23:03.20 | SpeedEvil | 7.5C |
| 23:03.26 | SpeedEvil | Electric blankets ++ |
| 23:04.32 | *** join/#edev davidc__ (n=davidc__@user-387g3od.cable.mindspring.com) |
| 23:05.06 | chouimat | sudo |
| 23:08.46 | monkey_ | roxfan: oh lol. i missed the point about disabling the breakpoint once it stops on READ8(), now I'm where I need to be. |
| 23:10.55 | *** join/#edev likewise (n=likewise@82-171-51-231.ip.telfort.nl) |
| 23:17.55 | monkey_ | some of the values are still different though |
| 23:28.54 | monkey_ | roxfan: are there any reasons why the register values would be different? |
| 23:29.16 | roxfan | umm, different from what? |
| 23:30.48 | monkey_ | the teacher had a list of registers to check and values they should be |
| 23:32.51 | roxfan | don't know really... there can be a million of reasons |
| 23:33.25 | monkey_ | ok |
