#qi-hardware IRC log for Tuesday, 2012-04-03

qi-bot[commit] David Kühling: alfilesel: minor fixes (master) http://qi-hw.com/p/openwrt-packages/c3c191c00:02
DocScrutinizerviric: mhm, nice05:22
mirkolarsc: ping09:33
mirkowolfspraul: reviewing patches now09:36
mirkocurrently looking at http://pb.nanl.de/1061/34458611/ - this is mostly ben-specific kernel modules09:38
mirkoand they will only work when the xburst kernel patches are applied (implies the xburst target is selected)09:38
mirkoso it doesn't make sense to put them into package/kernel/modules/ but should go into own packages which depend on the xburst target09:39
wolfspraulmirko: that's the wpan stuff - ben specific?09:53
wolfspraulwhere do you want to have it in the tree?09:53
wolfspraulthe 802.15.4 stuff is meant to go into kernel.org, not ben-specific I think09:53
wolfspraulyou could plug an atusb into any router with usb...09:54
wolfspraulok I think you want this as an individual package, same as the other apps10:00
wolfspraulcan it still be compiled into the kernel statically then? as long as that's the case of course we can move it over to the qi packages10:01
wolfspraulso your choice: kernel module or qi package. I think kernel module makes more sense, especially given what the wpan project aims for and is working on.10:02
Action: lindi- maintains the qi package10:02
wolfspraulthe bootloader, wow10:07
wolfspraullindi-: do you have a feeling/guess/knowledge who is using it?10:08
lindi-wpwrak: I am :)10:15
cladamwwpwrak, i just manually fixed *.lib :-)10:30
cladamwwpwrak, it lost one line. :-)10:31
lindi-wolfspraul: I just got sick of having wiki.debian.org point people to random websites to download bootloader binaries10:32
wolfspraulsure, it's good10:33
cladamwwpwrak, I created two empty but only have rectangle multi-parts and then compared two multi-parts library http://dpaste.com/726195/  the line "S -1150 2700 1250 -2800 1 1 0 N" was lost. :-)10:34
wolfspraulI'm a little worn out with Debian packages, although I very much recognize the enormously important work happening there10:34
wolfspraulit took 1.5 years or so to get xburst-tools in, and I think it will be removed again due to cross-compilation issues10:35
wolfspraulwell then10:35
wolfspraulthe one thing I looked for recently and could not find was a debian package for openscad10:35
cladamwwpwrak, it makes me remembered that deleted the rectangle block in FIRST sub-part. then eeschema confuses/or can't handle such deleted act in FIRST sub-part, then no matter whatever I tried to draw new rectangle in which sub-part will ALWAYS add into all else sub-part. It's a BUG in eechema. ;-)10:51
kristianpaultoo early, but at least there is now a opencsg 10:52
kristianpaulabout openscad10:52
mirkowolfspraul: well, the wpan-stuff depends on TARGET_xburst - at elast what xiangfu sent me10:53
mirkospi_atben definitely is ben specific10:54
mirkoWHEN it is not xburst-specific, the kernel patches need to go into the 'generic patches' directory10:55
mirkootherwise they only get applied, when the xburst target is selected10:55
mirkothis way they won't be available for other platforms than xburst10:55
mirkoso, either it's xburst-specific: patches need to be target-specific, packages should be usual packages (not package/kernel/modules)10:59
mirkoor the patches are generic (applied for all targets) and do not depend on a particular target/architecture, then they could go into package/kernel/modules11:00
mirkoright now - looking at the patch - it's mixed up11:00
mirkoi've no problem committing the wpan-stuff for all targets11:02
qi-bot[commit] Adam Wang: 1. renamed wm9707scft.* to wolfson.* (master) http://qi-hw.com/p/kicad-libs/ff0787711:05
mirkowhich implies generic WPAN stuff needs to be extracted from the xburst-specific patches and converted into generic ones (referring to drivers/ieee802154/* in particular)11:06
larscmirko: pong11:44
wolfspraulmirko: perfect, I got it. so mostly you care that the builds and dependencies are consistent12:02
wolfspraulI will find out whether we see wpan more ben-specific or more kernel-generic, and then you should get an updated patchset tomorrow12:02
wolfspraulthanks for feedback!12:02
wpwrakthe atben drivers are definitely ben-specific. atusb and the rest of the stack (from linux-zigbee) aren't.12:04
wpwrakwell, driver(s). one variant replaced the other.12:06
wpwrakcladamw: (lines in the wrong part) are you sure the part/body switch was set correctly ? i think, when you restart the component editor, it'll reset12:33
wpwrakcladamw: (wm9707scf rename) is that the only part wolfson produce ? :)12:34
cladamwwpwrak, yes, i was sure 'locked' switch was checked. ;-)12:36
cladamwwpwrak, once you deleted FIRST rectangle in first sub-part, you run into this problem.12:36
wpwrak(locked)  i mean "Edit pins per part"12:37
cladamwwpwrak, i'd prefer to use wolfson. ;-)12:37
cladamwyes, it's locked.12:37
cladamwyou can use Editor to delete it first then try it. ;-)12:38
cladamwwpwrak, i can easily reproduce the bug now. ;-)12:40
wpwrak(wolfson) what i mean is that things get messy if more wolfson chips get added since git can't track changes inside a .lib file. so all the changes to one part will appear to affect all the wolfson chips12:41
cladamwonce you deleted rectangle in first sub-part.12:41
cladamwno, i noticed that git can track changed codes inside *.lib , are you sure ?12:43
wpwrak(rectangle) seems that the flag doens't affect drawings12:43
mirkowolfspraul: welcome :)12:43
wpwrak(.lib) what i mean is that you can't search for changes of, say, only the wm9707scft in a .lib if there are other things in there 12:44
cladamwyes, flag doesn't affect it. but if you switch to else sub-part, you can see your all changes go to else sub-part. :-)12:44
wpwrakcladamw: if you keep one component per lib, then you can  git log wm9707scft.lib  and see the revision history of that part12:45
cladamwokay. have not used  git log  though. i try it. :)12:45
wpwrak(git) similar with other commands12:45
wpwrak(rectangle) if you double-click on it, you get a pop-up with a "Shared by all parts in component" checkbox. that one seems to do the trick.12:47
cladamwoah..yeah...that one i knew. :-)12:49
wpwrakso this solves the problem ? or do deletes still affect other parts if you un-check "Shared ..." ?12:50
cladamwbut i meant once you deleted rectangle in first sub-part, then draw a new rectangle still in first sub-part, then switch to else sub-part, you will see the new rectangle is all existed there, which is not the need we want. no matter what you checked  "Edit pins per part" or not. :-)12:52
cladamwyes, i added one line after comparing those two empty *.lib, then try a while, now it's okay.12:53
cladamwinserted one line in lib.12:53
wpwraki'm confused :)12:53
wpwrakdoes the "one line" refer to the rectangle problem ? or are you talking about something else ?12:54
cladamwlet me git push the new spartan-6 then later you can follow my steps above, you will see. :-)12:54
qi-bot[commit] Adam Wang: added Xilinx XC6SLX45-2FGG484C (master) http://qi-hw.com/p/kicad-libs/e2393f312:56
wpwrakyou have "Shared by all parts in component" set13:00
cladamwwpwrak, 1 - make you that's checked  2 - delete rectagnle in part A  3 - re-draw a new rectangle 4 - switch to other part [B:F], see there's a new rectangle on there. It's wrong 5. added one line as i posted earlier then fixed.13:00
wpwrakmove the mouse over the edge of the rectangle, then press E13:01
wpwrakthen you'll get a dialog with a checkbox "Shared by all parts in component"13:01
wpwrakif there's an [X], the rectangle will appear in all parts. if you click to change it to [ ], it won't be shared (as far as i can tell)13:02
cladamwyou meant that "Apply changes to all parts in components" ?13:03
wpwrakwhich verion of kicad are you using ?13:05
cladamwarh ? your dialog msg isn't same as mine ?13:05
wpwraki don't know :)13:05
cladamwBuild: (2010-08-11 BZR 2448)-unstable13:05
wpwraki have 2012-03-20 BZR 348213:06
wpwrakokay, so some things may be different13:06
cladamwnice ! yours are super new though. :-)13:06
wpwrakah, ignore the date. that's just when i compiled it13:07
cladamwwell...nevermind now on my problems. let's ignore it now.13:07
cladamwwell....so how do i install yours ?13:07
wpwrakor wait .. no, that's actually from then. hmm13:07
wpwraklemme clean up my stuff :)13:08
wpwrakit should be BZR 3378. but that's still newer than what you have13:08
cladamwyeah...but it's safe on my site now. since I'll super carefully on editing multi-part. but no more such part needed to create now. haha...13:09
wpwrakseem that kicad's build process uses the information from the latest commit not from the one checked out.13:12
cladamw(electrical type of pin properties) wpwrak btw, although I git pushed lib, but may few of those pins are not set well though. It's quite a lot of check works. I think in DRC stage I'll meet err then. :-)13:13
cladamwthen I'll fix them somewhere. :-)13:14
wpwraksoon you'll be able to use this tool: http://downloads.qi-hardware.com/people/werner/tmp/out.pdf13:15
wpwrakmake sit much easier to check the pin types13:16
cladamwwow~ nice13:17
cladamw"make sit" ?13:17
wpwraks/make sit/makes it13:19
wpwrakrace condition between my fingers :)13:19
cladamwha :-)13:19
cladamwwpwrak, do you think that NC we should use "Unspecified" or "Not connected" ? 13:25
cladamwfrom your out.pdf you used an "Unspecified".13:26
wpwraki think NC didn't exist when i did these13:28
wpwraklemme check how it works ...13:28
wpwrakUnspecified for those that have no internal connection13:29
wpwrakNC for those that are "do not connect" or those where we don't know if it's safe to connect something or not13:29
cladamwthis is defined by yourself or KiCad ?13:29
wpwrakthis way, one can still use "NC" pins for routing13:30
wpwrakthis is what i suggest to use13:30
qi-botThe build was successful: http://fidelio.qi-hardware.com/~xiangfu/build-nanonote/openwrt-xburst.minimal-20120402-0952 13:50
whitequarkwolfspraul: hm, disassembled my microUSB-HDMI adapter15:39
whitequarkit only uses four pins15:39
whitequarkof which two are power15:39
whitequarkand the two rest seem to be just like USB data lanes15:39
whitequarkbut 480 mbit/s isn't enough for uncompressed full-HD video afaik, so how does it work?15:40
Aylax-It's enough for blu-ray16:17
mthwhitequark: 24bpp 1920*1080 at 60 Hz is almost 3 Gbit/s17:02
mth480 Mbit/s would be just enough for 640*48017:03
whitequarkmth: exactly17:03
whitequarkand I definitely remember a whole lot of problems with USB displays17:04
whitequarkthere's even a chip for that17:04
mthso it either supports lower resolution, lower bpp, lower frame rate or it does do compression17:04
whitequarkDisplayLink or DisplayPort or whatever17:04
qi-bot[commit] kyak: Revert "gmenu2x update to latest git commit date: 20120331" (master) http://qi-hw.com/p/openwrt-packages/cbd1c6117:04
whitequarkAylax-: blu-ray isn't uncompressed17:04
whitequarkmth: the problem with compression is, you can't compress screen contents fast and efficiently17:05
whitequarkyou need a good hardware encoder, a good hardware decoder and it still will lag on some patterns17:05
mthif you don't care about efficiency too much, you can compress it quickly though17:06
whitequarkbut you need to fit hd to 480mbit17:06
whitequarkthrough it probably does not do 60hz17:06
mthcompress every frame as JPEG image?17:06
whitequarknot to mention that jpeg is surprisingly costly for such a crappy algorithm17:07
whitequarkand the hw I have definitely does not do jpeg17:08
whitequarknor it does any other kind of lossy compression17:08
mthsince the PC side is USB, the compression has to be done by the driver, so it's likely not hardware accelerated, or at least not accelerated by this adapter (it could use the GPU)17:08
whitequarkit's not PC17:09
whitequarkit is a phone17:09
whitequarkhttp://en.wikipedia.org/wiki/Mobile_High-definition_Link with this17:09
whitequarklooks like it just uses USB data lanes for some kind of LVDS17:09
whitequarkI suppose you could fit 3gbit/s there if you try hard17:09
mthdoes it actually conform to the USB spec though?17:10
mthmaybe it cheats and uses the connector but not the protocol17:10
whitequarkit has USB too17:10
mthit says "5-pin interface", while USB is 4-pin17:13
whitequarkmini/microUSB is 5-pin17:21
whitequarkwell, miniusb actually has a 4-pin variant, but no one ever uses it AFAIK17:21
whitequarknot sure about microusb, but they probably ditched the 4-pin variant17:22
whitequarkjust look at the connector.17:22
mthindeed, my cable has 5 pins too17:23
larscthe 5th pin is used to indicate which kind of device is on the other end17:24
whitequarkit's shorted to GND via a resistor17:33
whitequarkor just shorted to GND or left opened17:34
whitequarkone of the variants is for OTG and the other is for phone to detect a high-current charger17:34
larscthere are also some resistor values used for uart-via-usb and audio-via-usb17:35
larscin which case they just use the usb data lanes17:36
larscsee here for more: http://www.ulpi.org/ulpi_whitepaper_v2.pdf17:40
GNUtoolarsc, hi17:44
larschi GNUtoo 17:45
GNUtoolarsc, I sent a patch to linux-samsung from you about fixing boot of gta02, I got no responses, should I send to another list? the patch is linked from here: http://article.gmane.org/gmane.linux.kernel.samsung-soc/1035017:47
whitequarklarsc: huh, interesting, thanks for the link17:47
larscGNUtoo: try to ping the maintainer again17:50
larscpatches sometimes get overlooked17:50
whitequarkDocScrutinizer2: (qualcomm-atheros)17:55
whitequarkUPDATE: I have been contacted by Qualcomm PR regarding this expected presentation next week. The content of the Qualcomm Atheros developers was not approved by Qualcomm's legal department and the views to be expressed will be of their own personal beliefs.17:55
DocScrutinizerwhat presentation though?17:56
DocScrutinizerlarsc: actually the 5th pin just indicates whether this is an A or a B plug17:57
DocScrutinizerwhich arguably implies some general indication what the other end of cable might look like, and this which remote device17:58
DocScrutinizershort to GND = A, open = B, some resistor indicates fastcharger, or charging Y-cable setup for charging A-OTG host18:00
DocScrutinizersome ultraweird "standard" is using ID pin as *output* to control chargers18:01
DocScrutinizeror somesuch18:01
whitequarkDocScrutinizer: they [Atheros guys] presented their thoughts in a form of a presentation18:03
whitequarkmaybe want to speak at some conf18:03
DocScrutinizerand some very tricky battery dongles seem to use ID pin as anohter power-in or -out18:03
DocScrutinizerooh, those "nuke all proprietary drivers zealots?"18:05
whitequarkbut maybe it's the same as some Samsung employees giving away a lot of internal docs without any NDAs or similar shit18:06
DocScrutinizernow either they get twice the wages, or they are already fired18:06
whitequarkjust because they couldn't do it officially18:06
DocScrutinizerwhitequark: are those dudes relatives of me? ;-D18:07
whitequarkDocScrutinizer: no idea, I didn't talk to them, just heard on #replicant18:07
whitequarkIngenic folks tend to do the same18:07
whitequarke.g. they sent us a full minios source tree18:08
DocScrutinizerOM folks too ;-)18:08
whitequarkare they supposed to be proprietary?..18:08
DocScrutinizerhough there's not much closed stuff inside OM18:08
DocScrutinizersome crappy calypso GSM stack, 90% binary libs, already 'leaked' by somebody else18:09
DocScrutinizersome glamo docs that meanswhile are semi-public18:09
DocScrutinizerdunno shit about atheros wl6001 or whatever WLAN firmware18:10
DocScrutinizerbut that's no APE land blob anyway (neither is modem)18:10
whitequarkwell, given that some pre-android motorola actually had a probe in the wlan firmware...18:12
whitequarkI would care about this stuff18:12
DocScrutinizerand I think there actually were some accidentally publicly accessable schematics where the calypso modem pages didn't lack18:12
whitequarklike you cannot sandpaper the board and reverse-engineer that layout anyway18:13
whitequarkyou could ICMP it18:13
whitequarkand it would happily relay all traffic to whoever had sent the secret ICMP18:13
whitequarkno "verifiable" (in terms of wikipedia) source, but I pretty much sure that there was18:14
Aylax-Somebody highlighted me here?18:14
DocScrutinizerwould show immediately on my monitored and logged WLAN AP18:14
whitequarkDocScrutinizer: and now you're saying you are not paranoid ;)18:14
DocScrutinizerI am just laughing at paranoids, they always are so helpless18:15
whitequarkmaybe paranoidness should be measured on a logarithmic scale18:15
whitequarkmost people are at -118:15
whitequarkI'm, dunno, maybe at 318:15
whitequarkthen you are at 8.18:15
whitequarkor measure it in decibels18:16
whitequark80 dBp!18:16
DocScrutinizerlook, the diff between me and a paranoid is: I'm enjoying it18:17
DocScrutinizerI'm eager to see some attack18:17
DocScrutinizerso I got something to laugh and to fight and to investigate18:17
whitequarkoh, I understand it18:18
whitequarkbut all this should eat a lot of time, no?18:18
DocScrutinizerand damn, if the dude is really really *really* good, he might take my browser history and I'm not scared at all, just respectful for a really great hacker18:18
DocScrutinizerand all that takes just as much time as I can spend on it18:19
DocScrutinizeroften less18:19
whitequarkthat's sole reason I'm not doing similar things. I do things which don't require constant attention from you, or require you to understand loads of bullshit you'll forget anyway and then screw up something important18:20
DocScrutinizerwell, probably you got a life18:20
DocScrutinizeri'm not that sure about me in that regard18:21
whitequarkso I have SSH and passphrases on _really_ important things (other ones just go in KeePassX), and even an encrypted VM for _really really_ important things18:21
whitequarkbut heck, to monitor AP logs?! no no no 18:21
whitequarkdaily logwatch from 2+ servers is more than enough18:21
DocScrutinizerapropos, have to check my own box' munin18:22
whitequark(got a life) probably not. as I recently understood, I wake up, stare at the screen, go to screen.18:22
whitequarksometimes even eat.18:22
whitequark*go to sleep.18:22
DocScrutinizergo to work, return to screen, switch screen, go to sleep18:23
DocScrutinizernot that "work" means anything but "stare atyet another screen"18:23
DocScrutinizerWITH WIN-XP!!!!!!18:24
whitequarknah, I can work at home18:24
whitequark(winxp) interestingly, I'm way more irritated by win7 or gnome3 or osx than winxp18:24
DocScrutinizerand now that new shit18:25
DocScrutinizercoming up18:25
whitequarkwin8, yes18:25
DocScrutinizerthe embedded desktop for desktops18:25
whitequarkfar worse than anything which happened before18:25
whitequarkI tried beta18:25
DocScrutinizerhope it will kill them18:25
whitequarkapart from the fact that it BSODs endlessly in VMs18:25
whitequarkit has a REALLY SHITTY interface18:25
whitequarkorders of magnitude more shitty than vista18:25
DocScrutinizermagnitudes shittier than even wind2.018:26
whitequarklike if someone at MS thought "tablets are cool, let's pretend desktops are tablets"18:26
whitequarkand put all those GIANT BUTTONS18:26
whitequarkand removed all the compat18:26
whitequarkfuck him18:26
DocScrutinizermonster idiots18:27
whitequarkand fuck tablets. there's no device humanity created that is less useful than a tablet18:27
DocScrutinizerwell, a good reason to refuse ever again touching that stuff, even for earning a living18:27
whitequark(touching) you have to touch it to toss it to garbage18:27
whitequarkor maybe you could use a hazmat suit18:28
DocScrutinizerfigure a tablet on desk of boss' secretary ;-P18:28
whitequarkyou need: 1 (one) hammer, 1 (one) nail (optional)18:28
DocScrutinizer"Ms Jones, could you please write that letter til 2'o?"18:28
whitequarkhow much hours did it take?18:28
whitequarkis she done yet?18:28
DocScrutinizerwith win8 M$ abandoned office usecase18:29
whitequarkwell I don't really understand18:30
whitequarkis win8 "the next windows"?18:30
DocScrutinizerthe ONLY market niche they were relevant in18:30
whitequarkare they pretending that everyone should now use that? no, seriously?18:30
DocScrutinizerwell, it's their next windows18:30
whitequarkis it even compatible with old apps at all?18:31
DocScrutinizerthere's no other next M$-OS18:31
DocScrutinizerare ribbons in M$-office any compatible to anybody used to menus?18:31
DocScrutinizerI curse it every single day, at work18:31
whitequarkthey put MORE RIBBONS to it18:31
whitequarkto Explorer18:31
whitequarkdid they accidentally hire someone to destroy them from inside?..18:32
DocScrutinizerhonestly since ribbons even excel shit takes 5 min for the real work, plus 2h for searching how to do that work18:32
DocScrutinizermaybe Nokia's secret revenge18:33
whitequarkthe funniest thing is, ribbon is absolutely unusable on tablets18:34
whitequarkbecause of small buttons18:34
whitequarkand other screwups18:34
whitequarkjust read the points18:35
whitequarkthey're incredibly hilarious18:35
whitequark2. No more digging around for options through menus and submenus  There are approximately 200 commands on the new Windows 8 Explorer UI, even hidden features that would otherwise sit unused. 18:35
whitequarkdoes he understand what he says _at all_?18:36
DocScrutinizernah, the first picture already made my eyes bleed18:36
whitequarkI should have pixelized it18:36
whitequarkyou know, like in porn18:36
whitequarkor maybe the other way around. anyway, you got the idea18:37
DocScrutinizera few days ago I wondered how T major F you do conditional formatting in excel2007. Found out there's a way to add a "menu" to window bar, via "system"-"preferences"18:38
DocScrutinizerDIE!!!  M$ DIE!!!18:39
larscwe should get you a bill gates voodoo doll ;)18:43
whitequarkso much hate :D18:43
larscGNUtoo: btw. you should send the patch to the maintainer with the list in cc18:46
GNUtoothat's what I did18:47
GNUtooI guess it just doesn't show up in the list18:47
GNUtooI sent TO Ben Dooks18:48
larscok. i only had one mail from you in my inbox which only had the mailinglist not the maintainer18:48
GNUtooand CC the list18:48
GNUtoo+ other people18:48
GNUtooyes that's why I [RESENT] it18:48
GNUtooIt was early the morning18:48
GNUtooand I wasn't awake18:48
larscben dooks has been quite quiet over the last few weeks18:48
GNUtooand I forgott to do ./script/get-maintainer.pl18:49
GNUtooso I resent18:49
GNUtooso who should I send to?18:49
larsctry to ping Kukjin Kim18:49
larsche is usually responsive18:49
GNUtoook thanks a lot!!!!18:50
GNUtooaltough there is a problem in get-maintainer.pl or MAINTAINERS...since it always give me ben dooks18:50
whitequarkDocScrutinizer: http://wiki.freebsd.org/AdrianChadd/AtherosTxAgg18:52
whitequark>Qualcomm Atheros, Inc. have also provided source code and documentation under an NDA which allows for contribution back to an open source project.18:53
larscGNUtoo: scripts/get_maintainer.pl -f arch/arm/mach-s3c244018:53
larscBen Dooks <ben-linux@fluff.org> (maintainer:ARM/SAMSUNG ARM A...)18:53
larscKukjin Kim <kgene.kim@samsung.com> (maintainer:ARM/SAMSUNG ARM A...)18:53
larscwhitequark: which presentation are you guys talking about?19:00
whitequarklarsc: http://www.phoronix.com/scan.php?page=article&item=qualcomm_kill_blobs19:01
larscsounds nice19:08
larscit's always good when development overrules PR ;)19:08
GNUtoowhitequark, what about the firmwares (ath9k_htc)19:45
whitequarkGNUtoo: no idea21:01
GNUtoobecause the status of free firmwares is even worse than the free drivers21:05
GNUtoofor wifi we have openfwwf for old b43 cards and carl something for an atheros card21:05
GNUtooand I think that's all21:05
GNUtooso no free firmwares for embedded devices21:06
GNUtoo(and for graphic card I think nouveau also does the firmwares so it's free for nouveau if I remember well)21:06
GNUtooI hope lima comes up with something 21:07
GNUtoofor embedded devices21:07
wpwrakhmm, we really ought to have some pre-built toolchains for the ben. the setup process is rather nightmareish. and jlime only has an i386 version which doesn't work on ubuntu onerous21:24
shevekDoesn't the one from emdebian work?21:24
GNUtoooops wrong channel about alsa21:25
wpwrakshevek: with the target system running openwrt or jlime ?21:26
whitequarkwpwrak: crosstool-ng sets up some pretty statically built toolchains21:26
whitequarkthey're filesystem relocatable21:26
whitequarkthat is, do not depend on LD_LIBRARY_PATH being set correctly or other shit21:26
wpwrakbut they still need to match the target system's libc21:26
viricnixpkgs toolchains... nanonixos...21:26
wpwraki knew you'd say that ;-)21:27
Action: kristianpaul going to Ecuador to give talk about Copyleft Hardware21:27
viricsalute Correa :)21:28
viricwpwrak: bwahahaha ;)21:29
viricwpwrak: I bet I could crossbuilt for the ben in a mips.21:29
kristianpauli insisted for workshop... but they said no?21:29
kristianpauli'll confirm again21:30
whitequarkwpwrak: how often do you change libcs?21:30
virichow often change libc?21:31
viricthere is one new every half a year21:31
viric(for glibc)21:31
viricI use to change once a year21:33
viricwell, by change, I mean *start using the newer*. I still use some old glibc for programs I have not updated21:34
wpwrakwhitequark: on the ben ? almost never. on the desktop, it depends on what ubuntu does.21:34
viricwpwrak: what one you use now?21:35
viricdid ubuntu start using 2.14?21:35
viricthey made that mess about RPC...21:36
wpwrakGNU C Library (Ubuntu EGLIBC 2.13-20ubuntu5.1) stable release version 2.1321:39
viricah even eglibc. ok21:41
wpwrakhmm, ftp://ftp.belnet.be/mirror/ftp.gnu.org/gnu/gcc/gcc-4.6-linaro/ (et al.) no longer exists21:50
wpwrakthis breaks owrt toolchain compilation21:51
wpwrakphew. found an old owrt toolchain in my backups. and it doesn't even mind the new ubuntu :)21:56
whitequarkwpwrak: ah, the host libc22:09
whitequarkwell, it theoretically requires the same libc22:09
whitequarkpractically you can build it once on debian stable and forget about it for ~3 yrs22:09
wpwrakyeah, host libc usually isn't a big issue. luckily :)22:12
mthwe're distributing the OpenDingux toolchain as i386 binary and I haven't heard of any problems running it22:38
mthit is not relocatable though22:38
mthand if you enable LTO on GCC 4.5, you get a dependency on libelf, which is not installed everywhere22:39
mthand is hard or impossible to install on 64-bit Debian22:39
mthbut with 4.6 and 4.7 that dependency is gone22:39
mthit is also gone if you don't enable LTO22:39
mthwe haven't had much success with LTO anyway, so that's not a big sacrifice22:40
wpwrakmth: ubuntu onerous doesn't install i386 support on amd64 by default. ia32-libs and ia32-libs-multiarch both have unresolved dependencies. aptitude flat out refuses to install them. apt-get valiantly tries but hits a problem later on.23:02
wpwrakon any case, the gcc support libs (libmfpr or whatever it's called) don't seem to get included23:03
wpwraktrying to install the i386 version manually results in apt-get and aptitude happily announcing that they going to uninstall pretty much all of the rest of the system. i was too chicken to try and see what would actually happen if i let it.23:04
dvdklatest mplayer in openwrt-packages is broken!?23:41
qi-bot[commit] David Kühling: add new icons for mplayer, nupdf that start the app via 'alfilesel' (master) http://qi-hw.com/p/gmenu2x/7ab0e1623:45
qi-bot[commit] David Kühling: alfilesel: cleanup, fixes, add support for directly exec'ing applications (master) http://qi-hw.com/p/openwrt-packages/2a118e023:47
qi-bot[commit] David Kühling: config.full_system: enable alfilesel file selector and stuff needed to build it (master) http://qi-hw.com/p/openwrt-packages/befba9a23:47
wolfsprauldvdk: really? haven't tried yet but will notice right when I upgrade23:53
dvdkwolfspraul: didn't work due to some library depenency problem (libfaad.so not present), but if i recompile myself, i only get a black picture, not even sound.  maybe the video driver deadlocks23:55
wolfspraulouch, got it23:57
dvdkso, just nobody realized till now, as masked by the dependency problem.23:58
dvdkor maybe my local compilation just has problems, who knows.23:58
--- Wed Apr 4 201200:00

Generated by irclog2html.py 2.9.2 by Marius Gedminas - find it at mg.pov.lt!