**** BEGIN LOGGING AT Tue Oct 06 02:59:58 2015 Oct 06 03:29:40 build #98 of kirkwood is complete: Success [build successful] Build details are at http://buildbot.openwrt.org:8010/builders/kirkwood/builds/98 Oct 06 04:17:16 build #105 of imx6 is complete: Success [build successful] Build details are at http://buildbot.openwrt.org:8010/builders/imx6/builds/105 Oct 06 04:33:57 build #104 of ppc40x is complete: Failure [failed shell_10] Build details are at http://buildbot.openwrt.org:8010/builders/ppc40x/builds/104 Oct 06 06:52:39 build #96 of ar71xx is complete: Success [build successful] Build details are at http://buildbot.openwrt.org:8010/builders/ar71xx/builds/96 Oct 06 06:54:27 eigma: call make with IGNORE_ERRORS=m (like on buildbots) and with logging enabled your logs should cover like 90 percent of errors Oct 06 06:55:11 plntyk: thank you, I'll try that next time Oct 06 07:39:24 build #97 of lantiq.xrx200 is complete: Failure [failed compile_8] Build details are at http://buildbot.openwrt.org:8010/builders/lantiq.xrx200/builds/97 Oct 06 11:31:37 nbd: seems like all languages have something like (un)pack for creating blobs in memory... https://docs.python.org/2/library/struct.html Oct 06 11:31:45 lua has also that unpack feature Oct 06 11:32:39 lua unpack isn't the same as struct unpack, Oct 06 11:33:09 you need extra lua modules to get "struct" (un)pack Oct 06 11:33:24 karlp: not the same functionality? Oct 06 11:34:08 that's what I said, yes. Oct 06 11:34:16 lua unpack is not at all the same as python struct.unpack. Oct 06 11:34:45 to get something like python struct unpack or perl unpack, you need an extra module for lua providing that functionality Oct 06 11:35:45 http://www.lua.org/manual/5.1/manual.html and search for unpack, the first result will show you what it actually does. Oct 06 11:37:08 karlp: mine was more an observation that the feature itself has been implemented in many languages, although maybe because they didn't have access to proper structs Oct 06 11:38:27 right, it's a common enough desire that most languages end up having some way of doign something like it, Oct 06 11:38:43 but it's not available in openwrt's builds of lua out of the box, and not in any core packages at present. Oct 06 12:22:52 nbd: my bad, setsockopt worked correctly. Problem was I had an unitialized one Oct 06 12:23:08 ah Oct 06 12:24:50 yeah, I did a const int one; instead of const int one=1; Oct 06 12:25:19 compilers should warn when you use a const that hasn't been initialized... Oct 06 12:34:02 nbd r47146 trunk/target/linux/ramips/mt7620/ target.mk profiles/xiaomi.mk Oct 06 12:34:02 ramips: enable mt76 by default for the mt7620 subtarget (for devices with mt7620 + mt7612) Oct 06 12:34:27 nbd r47147 trunk/package/devel/oprofile/ patches/200-add_mips_1004kc.patch Makefile * oprofile: drop kmod-oprofile dependency, it uses the perf API now Oct 06 12:34:31 nbd r47148 trunk/package/kernel/linux/modules/other.mk * kernel: drop the now unused kmod-oprofile package Oct 06 12:34:38 nbd r47149 trunk/package/network/services/hostapd/patches/470-wait-for-nullfunc-longer.patch * hostapd: wait longer for inactive client probe (empty data frame) Oct 06 12:35:06 nbd r47150 branches/chaos_calmer/package/network/services/hostapd/patches/470-wait-for-nullfunc-longer.patch * hostapd: wait longer for inactive client probe (empty data frame) Oct 06 13:07:27 holy shit bonnie++ on a usb sd card takes for....evahhhh.... Oct 06 13:35:50 nbd: can you tell me how to enable squashfs+ubifs overlay image? I want to take a look at using that for ventana Oct 06 13:38:49 enable where? Oct 06 13:39:02 in target/imx6/images/Makefile Oct 06 13:39:36 well, it needs some code to be written to create that kind of image Oct 06 13:39:50 the ventana still pulls kernels from a filesystem on ubi, right? Oct 06 13:40:14 right and it uses dt so there are 6 or so dtbs Oct 06 13:40:36 so it seems best to keep a 'boot' ubifs for kernel, dtbs, bootscript Oct 06 13:41:04 accessible from the bootloader (std ubi/ubifs) but for a root fs I would like to move to overlay with squash+ubifs Oct 06 13:41:08 yeah Oct 06 13:41:34 to implement this, i'd suggest ripping out all the old ventana specific image building code and creating something new based on the new image building code Oct 06 13:41:41 you can probably use the bcm53xx target as reference Oct 06 13:41:52 obviously I'll need to package up a sysupgrade image that has both boot+root but Its been so long I can't recall how to simply build the overlay rootfs Oct 06 13:42:50 here's how it should work: Oct 06 13:43:19 you create a Build/kernel-fs template, which runs some commands to build the ubifs image for the kernel files, dtb and boot scripts Oct 06 13:44:06 you then create a Build/ubinize template which creates the ubi image from the kernel ubifs and the squashfs image Oct 06 13:44:15 with a ventana specific ubinize.cfg Oct 06 13:44:56 you can use the bcm53xx ubinize.cfg as reference Oct 06 13:45:19 it declares a rootfs_data partition with no image and the autoresize flag Oct 06 13:46:23 once you've made the image for initial flashing, i can also integrate some code that you can use to build a proper sysupgrade image that works with the generic nand sysupgrade code Oct 06 13:47:03 or maybe you don't need that Oct 06 13:47:35 depending on whether you want to go with a tarball of block images, or rather use ubiformat to flash an entire new .ubi image Oct 06 13:47:53 bcm53xx uses the ubiformat way Oct 06 13:48:04 (it actually extracts the ubi data from a .trx image) Oct 06 13:49:54 so selecting BCM53xx target to see what it builds me - any subtarget? Oct 06 13:50:10 any profile? Oct 06 13:50:12 no need to build it, just look at the image building code Oct 06 13:50:25 here are no subtargets Oct 06 13:51:21 its the squashfs part that I'm confused about, I have a handle on creating ubi's Oct 06 13:51:55 but I see hw bcm53xx Build/mkfs/squashfs uses ubinize? Oct 06 13:52:10 yes Oct 06 13:56:39 i'm pretty lost in owrt's image build process - you say something has been changed and there is a new process? Oct 06 13:58:06 does bcm53xx Build/mkfs/squashfs which creates a ubi happen 'before' something else wraps the ubi in a squashfs? Oct 06 14:10:47 tharvey: Build/mkfs/squashfs is called at filesystem build time, but you should do it in a different way Oct 06 14:15:01 <_trine> This looks interesting >> https://github.com/pepe2k/u-boot_mod Oct 06 14:16:40 once you've got it there... Oct 06 15:19:28 build #98 of avr32 is complete: Failure [failed shell_10] Build details are at http://buildbot.openwrt.org:8010/builders/avr32/builds/98 Oct 06 18:27:52 hi, do images contain a hash or something to check its integrity? Oct 06 18:31:49 mwarning: some Oct 06 18:32:37 sometimes images are just renamed tar files, sometimes its a custom format with embedded md5sum, sometimes its a vendor format with crc32, sometimes its just a raw disk image Oct 06 18:32:43 one could verify if squashfs or jffs are valid meaning uncompressing them works Oct 06 18:32:53 etc. Oct 06 18:33:30 jow_laptop: ah, thanks. does this apply to both factory and sysupgrade images? Oct 06 18:39:40 sorry, connection went down Oct 06 18:42:09 did I miss something? Oct 06 18:43:09 mwarning: Nothing between your last 'ah, thanks' and now. Oct 06 18:43:41 k Oct 06 19:07:19 mwarning: yes, it applieas to both kinds of images **** ENDING LOGGING AT Wed Oct 07 02:59:58 2015