**** BEGIN LOGGING AT Thu Jun 17 02:59:57 2010 Jun 17 07:00:09 morning Jun 17 07:00:23 Morning hrq Jun 17 07:00:26 hrq* Jun 17 07:00:30 Grrrrrrrrr Jun 17 07:00:35 hrw* Jun 17 07:00:40 :) Jun 17 07:00:56 Stupid fancy keyboard! Jun 17 07:01:01 ;D Jun 17 08:07:50 lag: can you boot the .34 panda kernel which was built by yourself? Jun 17 08:08:37 Yes Jun 17 08:08:48 :) Jun 17 08:08:53 I have been playing with it Jun 17 08:09:31 lag: excellent, man, good to hear that Jun 17 08:09:34 lag: thx Jun 17 08:09:52 cooloney: No problem Jun 17 08:10:01 Do you have HW yet? Jun 17 08:10:17 too bad, no Jun 17 08:10:24 i don't have omap3 and omap4 Jun 17 08:10:32 Are you going to get HW soon? Jun 17 08:11:20 igep v2 should be good omap3 testboard. compared to beagleboard (c4/xm) it is available Jun 17 08:11:48 hrw: igep v2 is from TI or other vendor? Jun 17 08:12:13 lag: no, i think there is no OMAP4 HW for me Jun 17 08:12:26 lag: and for omap3, no idea about that Jun 17 08:12:51 That sucks Jun 17 08:12:53 other Jun 17 08:12:56 lag: maybe i need to order some omap3 boards and ship to folks in US, and get the board during next sprint Jun 17 08:13:18 cooloney: spanish company Jun 17 08:13:28 hrw: ok, let me search Jun 17 08:13:32 hrw: thx Jun 17 08:13:36 I think I'd find it difficult working on an architecture without HW Jun 17 08:14:20 cooloney: Do you know if a console is configured on HDMI on the Panda board? Jun 17 08:15:21 lag: oh, not sure about that, let me check Jun 17 08:15:36 cooloney: http://shop.igep.es/index.php?main_page=product_info&cPath=1&products_id=38&zenid=j4novjt0a4t4k1j0o2igoldhm6 Jun 17 08:15:39 cooloney: we are trying to get some IGEPv2 boards for evalueation Jun 17 08:15:51 cooloney: 130€ for no wifi/bt, 145€ for wifi/bt Jun 17 08:16:11 they should work with our kernel, i've enabled it in the configs Jun 17 08:16:14 hrw: nice, ehe Jun 17 08:16:37 amitk: good to know that. hope i can get one of them to play with Jun 17 08:18:09 yeah, hope so too Jun 17 08:27:53 lag: i am not sure about the HDMI console. so i believe you are using HDMI display with panda, Jun 17 08:28:04 I want to Jun 17 08:28:11 lag: can you see anything on it? Jun 17 08:28:15 oh, you don't have Jun 17 08:28:19 No, nothing Jun 17 08:28:21 I do have Jun 17 08:28:49 oh, how about change some kernel cmdline to add 'console=tty0' or similar Jun 17 08:28:51 When I boot, I get 1,000's of these: omapdss DISPC error: SYNC_LOST_DIGIT Jun 17 08:29:16 That was my follow-up question - Which console ... Jun 17 08:30:05 ericm|ubuntu: I start merging your and mine patches now, OK? (You're not in #armlinux, so this is slightly off-topic) Jun 17 08:30:26 I've been meaning to install a full desktop environment on my new SD card Jun 17 08:30:28 I wonder how many working pandas are outside of TI Jun 17 08:30:37 ukleinek, i'm ok Jun 17 08:31:27 hrw: skip me, i don't have any Jun 17 08:31:30 hrw: ehe Jun 17 08:31:43 hrw: o/ Jun 17 08:31:51 lag: hmmm, actually, i have no idea about that. Jun 17 08:32:17 sebjan: do you not this console tty0 should be the right one for HDMI console? Jun 17 08:32:31 sebjan: i think lag is trying to output our UNR on it Jun 17 08:37:45 hrw: hi. not man panda outside of TI.. not many inside of TI anyways... Jun 17 08:37:53 lag: which kernel are you using? Jun 17 08:38:22 ndec: panda (the board) as endangered as panda (the animal) ? Jun 17 08:38:43 2.6.34-900-omap4 Jun 17 08:38:49 Maverick Jun 17 08:38:53 amitk: yep... we are trying to save the animal from disparation ;-) Jun 17 08:39:02 ndec: yep Jun 17 08:39:21 lag: do you have a HDMI screen or is it a DVI screen using the HDMI output of the board? Jun 17 08:41:07 It's the real deal Jun 17 08:41:21 HDMI->HDMI Jun 17 08:42:44 lag: ok. well this should work without any config change, assuming you have the right config in the kernel... but i have not tried on HDMI screen yet. Jun 17 08:43:42 Hmmm Jun 17 08:45:51 ndec: On which console? Jun 17 08:46:01 tty* Jun 17 08:47:01 lag: don't know. i had tried HDMI->DVI screen several weeks ago with the UNE UI Jun 17 08:47:52 I've just tested the monitor on my Revo running Karmic and it worked fine Jun 17 08:48:45 The DVI port on the Panda has an HDMI socket, so I can't test that Jun 17 08:50:35 lag: I am using the HDMI port (closest to USB). with the HDMI to DVI cable. I think we have problems with the DVI port. that one shouldn't work either... Jun 17 08:53:25 ndec: No, that doesn't work as far as I can tell Jun 17 08:53:35 But neither does the HDMI Jun 17 08:54:14 ndec: Have you seen this message: omapdss DISPC error: SYNC_LOST_DIGIT Jun 17 08:56:18 lag: yes... something wrong in DSS2/OMAP4 Jun 17 08:57:02 i want one Jun 17 08:57:06 Do you think it will be related to why I don't have HDMI? Jun 17 09:00:05 ericm|ubuntu: do you want to look into my patches 4 to 9 and comment/ack? Jun 17 09:00:29 ukleinek, OK Jun 17 09:01:14 ericm|ubuntu: well, my 7 = your 2 Jun 17 09:01:53 lag: yes. Jun 17 09:05:45 ndec: can i get a panda? :D Jun 17 09:06:11 armin76: why do you always ask me the same questions ;-) Jun 17 09:06:30 last time it wasn't a question :P Jun 17 09:06:40 armin76: good point... Jun 17 09:06:51 armin76: let us know polish it first... Jun 17 09:07:19 i want to polish it too :) Jun 17 09:16:03 hrw wants to polish it too! Jun 17 09:16:06 and zyga! Jun 17 09:16:25 (zyga and hrw are in Poland) Jun 17 09:17:22 Polish man meets English man and says: "I need to polish my English". English man replies: "No, your English is Polish enough". Jun 17 09:17:40 joke sounds better then looks Jun 17 09:22:30 hrw: nice... Jun 17 09:22:54 http://alt-tab.org/data/images/2010/06/angelamerkel.png Jun 17 09:38:43 lol :-) Jun 17 09:52:53 anyone knows if it is possible to tell debuild to generate both a -dbg package a striped binary version? I know that I need to set DEB_BUILD_OPTIONS=noopt,nostrip for debug but this will provide me only the debug binary. I'd like to generate both in one command if possible Jun 17 10:00:37 berco: Is this a package with -dbg packages in control? Jun 17 10:00:53 berco: What you can do, and what Ubuntu does on all packages, is using pkg-create-dbgsym to achieve that Jun 17 10:01:09 berco: You would build the package as usual, but the build will output -dbgsym packages along the other ones Jun 17 10:01:25 lool: I have not changed the control file for this Jun 17 10:01:31 berco: Just install the package and it should work (you might have to turn it on in /etc) Jun 17 10:01:45 lool: I was looking more for the general rule to achieve this :) Jun 17 10:02:14 lool: thanks. I will give it a try Jun 17 10:03:02 * armin76 steals lool's panda Jun 17 10:16:36 armin76: since when he has one? ;D Jun 17 10:17:07 he stole it too :D Jun 17 10:36:46 dyfet, how far are you with the gobject-introspection FTBFS ? Jun 17 13:16:24 npitre: ping Jun 17 13:17:18 npitre: do you care to answer to http://mid.gmane.org/20100611045507.GA10894@pengutronix.de Jun 17 13:17:22 ? Jun 17 13:21:33 lool: ping Jun 17 13:25:33 berco: pong Jun 17 13:25:49 berco: Ideally, don't ping but just ask something so that I dont context switch twice ;-) Jun 17 13:25:55 lool: thx for the dbg, seems to work. Jun 17 13:25:58 Cool Jun 17 13:26:03 lool: I now face another issue Jun 17 13:26:29 lool: I noticed in my source package some files lost the execution privilege Jun 17 13:26:40 lool: any idea how to prevent this? Jun 17 13:27:04 berco: Sorry, what source is this? Jun 17 13:27:32 lool: package that contains source code (apt-get source Jun 17 13:27:39 berco: Yes, which one? Jun 17 13:28:11 lool: well not public yet, it's for omap4, this is the "tiler" package Jun 17 13:28:37 berco: Ok; would be easier to suggest a fix for a public source package Jun 17 13:29:06 berco: So which file is losing the +x bit exactly and at which point? Jun 17 13:29:24 lool: code has never been package. it's available but not in deb format yet. Am working on it... Jun 17 13:29:27 berco: There are two places where that can happen: in your source files, because of the source packaging format, or in the resulting .debs Jun 17 13:29:44 lool: this is bootstrap.sh file Jun 17 13:29:56 berco: are you saying that the +x is lost when you unpack the source package, or in the resulting .debs? Jun 17 13:30:33 lool: I need to check. I just did an "apt-get source" and noticed it's lost Jun 17 13:30:43 berco: Right, so while unpacking the source Jun 17 13:30:59 berco: So this is expected in the default source package format Jun 17 13:31:16 berco: you could of course change the source format, but there's something odd here Jun 17 13:31:29 berco: Usually, bootstrap.sh is part of the "upstream" tarball, the .orig.tar.gz Jun 17 13:31:37 berco: Files in this tarball wouldn't lose this bit Jun 17 13:31:55 berco: What might have happened here is that you intended to ship this file in the tarball, but it got shipped as part of the packaging instead Jun 17 13:32:04 lool: ok, so I need to check b'cos this file should be in my .orig.tar.gz Jun 17 13:32:05 berco: You really want this file to be in the upstream tarball I guess Jun 17 13:32:15 berco: Please do Jun 17 13:32:30 lool: thanks, that would make sense. I will check Jun 17 13:32:50 berco: Also, might be best to just ask the chan rather than me specifically -- will get you faster answers, spread the load, and also I dont actually work on omap4 stuff (yet?) but others here do Jun 17 13:33:32 lool: ok. Sorry I'm new to the Linux world so learning all the habits and stuff Jun 17 13:35:32 berco: This is fine; I hope you don't mind me telling you Jun 17 13:36:09 lool: not at all, just noticed you knew quite well the pcakaging topci Jun 17 13:48:34 :q Jun 17 13:58:42 when packaging from the .orig.tar.gz, I created a patch to move files but those that have +x attribute will lose it in the .deb source package. Anyone knows how that can be fixed. I want to keep the +x on these file. Jun 17 14:00:41 btw, I'm using cdbs-edit-patch for creating the patch Jun 17 14:02:35 berco: Why do you move the files? and why in a patch? Jun 17 14:03:26 lool: git tree has the source code in a subfolder Jun 17 14:03:57 lool: am not really allowed to change this and I created the orig.tar.gz with the simple "git archive" command Jun 17 14:03:57 berco: You could either tell your package to build in a subfolder rather than build from the top, or you could release tarballs from the proper level? Jun 17 14:04:49 lool: might be simpler in my case to build from the subfolder until the upstream maintainer modifies the git tree Jun 17 14:04:49 berco: Without moving files in the git repo, could you strip a component when creating the tarball? Jun 17 14:05:06 berco: Ok; if you're not upstream, then just go for the sub-folder build Jun 17 14:05:40 berco: Set DEB_SRCDIR near the top of your rules if you're using cdbs Jun 17 14:05:57 lool: yes, am using cdb Jun 17 14:06:12 The default is ".", you could set it to subfolder or to $(CURDIR)/subfolder if you need an absolute pathname Jun 17 14:06:17 lool: cdbs. I will try this, that sounds the most reasonable at this point Jun 17 15:48:02 anybody formilliar with the sharp netwalker? Jun 17 15:56:33 in-game: What's up? Jun 17 15:57:16 in-game: persia might be the one to ask, but he's on holiday. Jun 17 15:57:25 I believe he has one, though. Jun 17 15:57:27 ahh ok Jun 17 15:57:52 sweet device but the ubuntu running is 9.04 Jun 17 15:57:59 yep. Jun 17 15:58:10 so i am searching for a way to upgrade somehw Jun 17 15:58:34 now running lxdm to make it some faster Jun 17 15:58:35 in-game: unlikely you'll be able to upgrade Jun 17 15:58:50 I don't think it supports 10.04. Iirc, it was a hardware issue. Jun 17 15:58:52 the kernel hasn't been mainlined Jun 17 15:59:00 thought so Jun 17 15:59:47 pitty, the arm packages are some what lagging behind Jun 17 16:00:24 in-game: it is mostly kernel and any binary drivers if reqd. Jun 17 16:00:34 There were some hardware changes between 9.04 & 10.04. The newer kernels in 10.04 were never made to support the older hardware rev. Jun 17 16:00:48 ukleinek: sorry Jun 17 16:01:03 If you are kernel savvy, you could probably forward port the changes though. Jun 17 16:01:05 in-game: the rest of the arm packages have been carried forward just fine Jun 17 16:02:26 ok, yeah i wanted to update lxdm because of the network xplore function Jun 17 16:03:58 Unfortunately 9.04 packages are built for ARMv5 and 10.04 were built for ARMv7. It would be like running x86_64 code on an Intel Pentium (P5). Jun 17 16:04:08 in-game: you could enable the lucid repo in /etc/apt/sources.list and only upgrade certain packages Jun 17 16:04:23 GrueMaster: the processor is a ARMv7 Jun 17 16:04:41 amitk: The processor may be, but the kernel doesn't have support for it. Jun 17 16:04:55 amitk tried that, but it deleted my x packages somehow Jun 17 16:05:14 tried it twice now Jun 17 16:06:17 GrueMaster: in-game does NOT want to upgrade the kernel, just the desktop. He can do that if the dependencies were carried over correctly Jun 17 16:07:26 in-game: it is quite possible that the new version of the desktop brings in a new version of X which ofcourse doesn't have new version of some display driver or similar. That could kill your X Jun 17 16:09:08 yups. Jun 17 16:10:10 so hm first uprade the x Jun 17 16:10:36 can add 10.04 to the repro and upgrade just without the kernel? Jun 17 16:11:14 in-game: are you trying a 9.04->10.04 upgrade direct? Jun 17 16:11:28 hmm Jun 17 16:11:34 yeps Jun 17 16:11:39 maye first 9.10 Jun 17 16:11:39 that is probably untested. Have you tried a 9.04->9.10 ? Jun 17 16:11:47 not yet Jun 17 16:12:07 heh that is pretty good one :) Jun 17 16:12:58 amitk: in-game: this is probably a bad idea. Netwalker kernel is very dated, and doesn't have proper thumb2 handling support which is required for lucid (you might get karmic to work) Jun 17 16:13:41 NCommander: that is what I just suggested ^^ :) Jun 17 16:13:53 yep :) Jun 17 16:13:59 * NCommander brain farts, and goes to hide in a corner Jun 17 16:15:42 no need for hiding, I tried it twice.... >< Jun 17 16:16:33 hehe Jun 17 16:21:10 Well lets try a update! Jun 17 16:21:20 thanks for the info people Jun 17 16:21:42 Speak to you later, if it works I will be back soon Jun 17 19:52:48 Lucid isn't built with neon enabled by default, right? Jun 17 19:53:59 neon in what? Jun 17 19:54:27 libpixman includes neon routines for instance Jun 17 19:54:44 but as far as I know, gcc never outputs neon instructions automatically Jun 17 19:54:52 huh. that won't run on Dove then... Jun 17 19:55:31 various packages include neon-optimized routines as alternatives but they're expected to work on non-neon systems too Jun 17 19:55:34 I'll check, but I'd expect there to be guard clauses Jun 17 19:55:41 exactly Jun 17 19:55:47 ojn: its not Jun 17 19:56:14 ok, thanks. Jun 17 19:56:17 ojn: last time i checked ubntu had an old version of pixman which didn't had the neon routines Jun 17 19:57:34 iirc pixman compiled automatically with neon, but it gets disabled/enabled at runtime Jun 17 19:57:59 ssvb: is the master here :D Jun 17 19:59:14 armin76, pixman-0.16.4 has some neon fastpaths Jun 17 19:59:25 I'm looking at apt-get source libpixman-1-0 Jun 17 20:00:29 doesn't have the more recent work done in january though Jun 17 20:01:13 cwillu_at_work: as i said, ssvb should know for sure :) Jun 17 20:01:33 armin76, argument screens off authority :p Jun 17 20:01:35 iirc the neon stuff was reworked on >0.16 by ssvb itself :) Jun 17 20:01:41 i.e., I'm looking at the source code :) Jun 17 21:44:34 ojn, armin76, cwillu_at_work: lucid's and maverick's pixman have some NEON bits which are detected at runtime (via /proc/self/auxv) Jun 17 23:54:37 pixman in maverick is muuch faster on arm, I merged that as soon as I could :) **** ENDING LOGGING AT Fri Jun 18 02:59:57 2010