**** BEGIN LOGGING AT Thu Aug 11 02:59:58 2016 Aug 11 07:31:07 so on a krogoth setup, can i just switch my MACHINE and rebuild without issues ? or do i want to setup a different build dir ? Aug 11 07:33:43 is jethro still officially maintained, or is it abandonned ? (jethro can't be built when the host uses gcc-6 as gcc, even when setting the varions BUILD_* variables. I have a fix, but i'd like to discuss it first, not sure if anybody is still interested) Aug 11 07:54:28 general question... is it intended that rm_work removes stuff from work-shared ? should it ? Aug 11 07:57:20 * boucman_work has problem when rm_work is active which erases the kernel .config that other recipes try to use... Aug 11 08:16:10 rob_w: usually just rebuilding should be no problem Aug 11 08:16:18 thx Aug 11 08:16:55 rob_w: it will in fact do an almost complete rebuild then anyways, as automatic detection should avoid most of the exiting sstate in such a case. Aug 11 08:17:08 rob_w: if you have similar enough machines it can even re-use the cross-compiler Aug 11 08:18:49 jubr: yeah thats probably the main part where things could be recycled. Aug 11 08:20:18 and libc Aug 11 09:21:52 Hello all, while compiling my next package (streamhtmlparser) I get this error: /usr/include/python2.7/pyconfig.h:13:54: fatal error: arm-linux-gnueabihf/python2.7/pyconfig.h: No such file or directory Aug 11 09:21:56 # include Aug 11 09:22:28 I think I need a package like 'libpython' or something under debian the package is called: 'libpython2.7-dev' Aug 11 09:23:04 you're probably missing a dependency Aug 11 09:23:12 or it has raced Aug 11 09:23:16 How can I find this package in this big pool on recpies Aug 11 09:23:27 CTtpollard: its a new package from me, and yes I need a dependency Aug 11 09:23:35 But how is the target package called? Aug 11 09:23:50 HyP3r: is streamhtmlparser a python module? Aug 11 09:24:15 i'm suprised something specifically an arm-only include... Aug 11 09:24:24 *needs* Aug 11 09:24:48 rburton: no not really, Its totaly written in cython and python yes but its not a module Aug 11 09:25:16 did you depend on python? Aug 11 09:25:19 Oh sorry, no its not a python module it a c library (*.so) and its compiled with autotools Aug 11 09:25:28 yep Aug 11 09:25:49 Here is my package http://pastebin.com/aCL7TwP9 Aug 11 09:25:57 note that its looking for target paths (/usr/include/python2.7/pyconfig.h) which suggests that streamhtmlparser is broken Aug 11 09:26:34 you may need to inherit python-dir to get some more variables set Aug 11 09:27:44 This is slightly OT, but I find it difficult to retrace git checkouts against the official repos. E.g. fsl-mx6 use a repo manifest to check out meta-openembbeded on a specific commit. However when I clone meta-openembedded myself I have a hard time finding that commit chain which repo checked out. I'm not too used to git yet, do you have any pointers in how to deal with this? Aug 11 09:27:50 The file 'py_streamhtmlparser.c' inside the streamhtmlparser package depends only to '#include "Python.h"' and then... the Python.h has this weird include Aug 11 09:28:09 rburton: ok Aug 11 09:28:10 its the configure script that is likely doing the wrong thing Aug 11 09:28:20 yes I think so Aug 11 09:32:31 rburton: is jethro still officially alive ? Aug 11 09:32:42 rburton: thats inside the Makefile.am PYTHON_CFLAGS = -I/usr/include/python$(PYTHON_VERSION) -fpic Aug 11 09:32:50 Are there any predefined best pratices wrt layer priority? Aug 11 09:32:54 Static Path to the root of my system Aug 11 09:33:16 How can I fix this? Aug 11 09:34:05 E.g. I find that I constantly have to patch some recipes in meta-openembedded, and I think I'd rather copy them to my layer and give them a higher priority. That is possible, right? Aug 11 09:40:59 HyP3r: yeah that would be wrong :) autoconf-archive has some proper macros to find the headers using python-config if you want to fix upstream once and for all, or the hack that might work would be to replace that hard-coded path with python-config —includes Aug 11 09:42:09 I thought that I simply patch this makefile? Aug 11 09:42:36 fixing it properly and getting that fix upstream would be Best, as then you never have to worry about it again. Aug 11 09:42:44 but yeah, patch the makefile is easy. Aug 11 09:43:12 yeah but then I have to do an pullrquest to this 7 ago repository https://github.com/google/streamhtmlparser Aug 11 09:43:31 But can you tell what I should write in this makefile at this line Aug 11 09:43:49 $(python-config --includes) will not work I think? Aug 11 09:43:50 ha Aug 11 09:43:58 HyP3r: sounds like you should fork it :) Aug 11 09:44:01 I'm not really good in this makefile.am syntax Aug 11 09:44:04 :S Aug 11 09:44:47 dare i ask why you want to use software that's been dead for 7 years? :) Aug 11 09:45:09 I don't know why https://github.com/swiperproxy/swiperproxy/releases depends on it Aug 11 09:45:13 \o/ Aug 11 09:46:12 This example looks good: http://stackoverflow.com/a/5099428/5232865 Aug 11 09:46:56 Or does some has a better stragey to load this python include paths Aug 11 09:48:22 yeah Aug 11 09:48:47 ? Aug 11 09:54:13 (the example looks good) Aug 11 10:03:37 There is no python-config inside my devshell enviroment Aug 11 10:07:05 Well this autotool component works well but inside my sysroots python-config is not inside the path enviroment its inside the folder /sysroots/x86_64-linux/usr/bin/python-native/python-config Aug 11 10:07:17 so if I generally call python-config I don't reach this file Aug 11 10:10:28 sweet , made hddimg but in its inird the busybox has no applets configured Aug 11 10:54:52 trying to build the glog package from meta-oe and it tries to link with /usr/lib/libgcc_s.so.1, is this something that pseudo is supposed to prevent/work around? Aug 11 10:55:48 This has worked for a long time and I don't know what's changed except I needed to patch pseudo to work with glibc 2.24 that came to my host distro recently. Aug 11 11:06:53 rburton: about that kernel.bbclass path fix Aug 11 11:07:13 the only commit I can find is a26ec548aab Aug 11 11:08:39 but that commit doesn't touch the same function as the patch I've sent Aug 11 11:22:03 hm Aug 11 11:24:58 aragua: oe-core d596286bc4cf19774447b136a27acf3c61aad2ff Aug 11 11:25:11 d596286bc4cf197 Aug 11 11:25:15 i found it Aug 11 11:25:23 ah almost at the same time Aug 11 11:25:33 I don't know why I'm missing that commit in my branch Aug 11 11:25:35 thanks Aug 11 11:30:49 so what happened is I had a commit on top of master which was reverting that change X| Aug 11 11:31:15 then I re-fixed it without knowing Aug 11 11:40:29 where can I find how the DISTRO_FEATURES vs DISTRO_FEATURES_DEFAULT logic is glued together? Aug 11 11:42:30 because I find examples of testing with ${@bb.utils.contains('DISTRO_FEATURES', while local.conf sets them with DISTRO_FEATURES_DEFAULT="... Aug 11 12:06:11 Is there a way to give a recpie some argument by command line or user input? It would be gerade if I can specify things like hostname and so on while connection Aug 11 12:06:22 *compilation not connnection \o/ Aug 11 12:09:29 HyP3r: use BB_ENV_EXTRAWHITE Aug 11 12:11:31 I don't find any documentation about this. Can you show me an example? Aug 11 12:11:37 Or how should I use this? Aug 11 12:14:25 http://stackoverflow.com/questions/17366984/is-it-possible-to-pass-in-command-line-variables-to-a-bitbake-build Aug 11 12:14:32 first google hit :) Aug 11 12:14:51 yeah I saw this article but I don't get it sorry Aug 11 12:15:34 Ok. Sorry now I get it. Sometimes its not good to do more than one thing at the same ... Aug 11 12:15:38 bitbake filters out ENV stuff before executing: thhis adds an exception, so you can use this var in your recipe's directly Aug 11 12:16:11 HyP3r: I always do too many things at the same time. King of the Context Switch is not necessarily a good title :) Aug 11 12:16:37 yeah but it generates headache if you do more than three things :( Aug 11 12:20:23 yup Aug 11 12:22:24 HyP3r: its usual to have a variable in your recipe that the user sets in local.conf Aug 11 12:22:28 instead of env vars Aug 11 12:30:32 rburton: true, that is better. I currently have a Jenkins Parametrized Job that writes a local.conf based on the parameters Aug 11 12:30:43 before kicking off the build Aug 11 12:32:50 rburton: I have a bug (and a workaround) on jethro, is there a bugtracker where I should submit, or a mail to the ML is enough ? Aug 11 12:34:43 bugzilla.yoctoproject.org for bugs, but if you have a patch then the list for master and backport requests for krogoth and jethro would be great Aug 11 12:35:29 I don't have a fix. I have a workaround, but I don't understand the gcc-cross recipe to be sure it's the correct way to fix it Aug 11 12:36:37 nor can I test it easily in master, since gcc supports gcc-6 so it won't trigger the bug (the bug is that BUILD_CC is not used correctly, so cross-gcc fails if gcc is an alias to gcc-c despite the $BUILD_CC variable) Aug 11 12:37:22 ah Aug 11 12:37:37 the list then i guess Aug 11 12:37:48 ok, thx Aug 11 12:37:51 we don't really support jethro with gcc6 Aug 11 12:38:15 (and point to the 2.0 buildtools tarball for anyone who wants jethro on a gcc6 host) Aug 11 12:40:13 rburton: I didn't expect gcc-6 to be supported, but it's a bug in BUILD_CC, which doesn't work correctly, so I can't use the gcc-5 I have, I think it's a real bug worth fixing Aug 11 12:40:35 anyway, mail posted. Thx Aug 11 13:02:22 Is there a way to configure yocto to generate a script which should be executed for the first startup? Aug 11 13:03:52 HyP3r: on first startupe, I don't really know (I think systemd can do that i'm not sure) but you can do something similar by having a post-install script that will run only on the target Aug 11 13:04:29 Well I thought I create a new recpie caled ' Aug 11 13:05:16 'firstrun' and other packages can depend on this and can write scripts to /etc/firstrun.d/ and the recpie firstrun registers itself by systemd and exectues this scripts in this directory Aug 11 13:05:49 And then deletes everything. Itself and the directory firstrun.d Aug 11 13:05:55 But what is post-install? Aug 11 13:08:07 HyP3r: https://wiki.yoctoproject.org/wiki/Run_postinst_during_rootfs_generation Aug 11 13:08:31 it also explains type-2 Aug 11 13:08:46 @firstboot Aug 11 13:09:43 a postinst that refuses to run at rootfs time and waits for first boot is the usual solution Aug 11 13:09:56 systemd has support for first boot units apparently if you are using systemd Aug 11 13:12:23 https://www.freedesktop.org/software/systemd/man/systemd-firstboot.html Aug 11 13:20:26 nice thats what I searched Aug 11 13:25:54 Is there a way to create users while package runtime? e.g. while do_install() { }? Aug 11 13:31:59 Something like that will be extecuted while startup? http://git.yoctoproject.org/cgit/cgit.cgi/poky/tree/meta/recipes-extended/sudo/sudo.inc?id=d18aba9c1cb5b2cf77cfb2dff150e9003b2e63ef#n33 Aug 11 13:50:00 this netsplit removed some people who might know the answer :D Aug 11 14:05:46 Is there an option/tool that can show why a recipe is being rebuilt? Aug 11 14:06:38 My fsl-imx repo rebuilds qt5 on *every* build, and it's quite tedious (read: minimum 1hrs waiting game) Aug 11 14:13:01 Is the dependencies of the extensible SDK installer documented somewhere? Aug 11 14:47:50 what's the current way of modifying PR from a .bbappend ? Aug 11 14:48:52 same way you'd modify any other variable Aug 11 14:49:01 but if you're using a somewhat modern release you don't need to touch PR anymore Aug 11 14:51:03 RP: did you have time to discuss my unset syntax ? Aug 11 14:52:43 rburton: what do you mean by "don't need" ? the ${PV}-${PR} stays the same as with the original recipe Aug 11 14:53:06 (and well, I was using += instead of _append, silly me) Aug 11 14:53:12 Come again: W: Unable to read /build/yocto/var-som-mx6-yocto-jethro/build/tmp/work/var_som_mx6-poky-linux-gnueabi/fsl-image-qt5/1.0-r0/apt/preferences.d/ - DirectoryExists (2: No such file or directory) Aug 11 14:54:24 Why does a build go haywire like this if I add DISTRO_FEATURES_DEFAULT += "myfeature"? It failes on completely different things that what that feature unlocks Aug 11 14:54:25 yann: rebuilds happen without touching PR, and if you want a package feed to work (so need PR to increase) then turn on the PR service Aug 11 14:54:30 Is there a way to configure a package this way that everytime you compile packages which depend on this recpie this recpie should fetch _new_ and compile _new_ Aug 11 14:54:47 sveinse: is DISTRO_FEATURES_DEFAULT assigned to with ?= Aug 11 14:54:48 I have configured the REVSION (SRCREV) to AUTOREV but thats not all Aug 11 14:54:54 SRCREV = "${AUTOREV}" Aug 11 14:55:06 rburton, += atm Aug 11 14:55:36 sveinse: i meant in oe-core. meta/conf/distro/include/default-distrovars.inc:DISTRO_FEATURES_DEFAULT ?= "alsa argp bluetooth ext2 irda largefile pcmcia usbgadget us Aug 11 14:55:40 rburton: I mostly wanted to see it changed so we know those packages come from recipes we modified Aug 11 14:56:07 sveinse: ?= foo then += bar results in "bar" as ?= is lazy Aug 11 14:57:29 rburton: what is the distinction between using DISTRO_FEATURES_DEFAULT and DISTRO_FEATURES ? Aug 11 14:57:47 sveinse: if you want to fiddle distro features from your local.conf, just do DISTRO_FEATURES_append Aug 11 14:59:12 rburton: ok. Point is, I have a recipe which tests on a DISTRO_FEATURE, and I need somehow to add it Aug 11 14:59:39 rburton: Just tested with DISTRO_FEATURES_DEFAULT ?=. That failed. Trying _append now Aug 11 14:59:53 sveinse: so if you have your own distro then set it in there. if you don't DISTRO_FEATURES_append in local.conf is best. Aug 11 15:00:01 rburton: This is a custom distro (fsl-imx-fb), yes Aug 11 15:00:44 every distro is a custom distro, i meant is it one you own or one you extend Aug 11 15:01:16 eg for testing and stuff i have DISTRO=poky and a pile of appends and removes in my local.conf Aug 11 15:01:34 but if i was doing a product id have a custom distro conf file that assigned to DISTRO_FEATURES directly Aug 11 15:01:54 rburton: I guess extend. For me I don't care if the distro is poky or anything else. For me both of them are third party stuff Aug 11 15:02:20 rburton: Point is, this is the codebase for the imx6 that I'm going to run our software one Aug 11 15:02:41 i recommend making your own distro that pulls in fsl-imx-fb and customises it Aug 11 15:02:57 (but for now, local.conf and append) Aug 11 15:03:16 DISTRO_FEATURES_append = " myfeature" failed even more miserable... unbuildable tasks were found and LOTS of loops Aug 11 15:03:32 such fun Aug 11 15:04:22 uncomment the entry, and the image is perfectly fine :p Aug 11 15:05:13 what feature is this? Aug 11 15:07:00 yes, isn't it. I am working with three trees: our jethro poky-based custom HW tree, a vanilla poky kergoth tree and a vericite/freescale tree. I am only able to build our qt5 app on the first.... Aug 11 15:07:41 rburton: its "lm-qt5", and grepping for it, reveals that this is only referred to from my own recipe Aug 11 15:08:29 I seriously doubt a name collision, but it's apparent that it garbles DISTRO_FEATURES somehow Aug 11 15:08:45 are you remembering to add whitespace? Aug 11 15:08:57 and why are you using a distro feature for a singlle recipe Aug 11 15:09:11 you can set package-level variables from distro or local configs Aug 11 15:11:22 rburton: This is ultimately the patch (to kergoth) that I want to include: http://paste.ubuntu.com/23030126/ Aug 11 15:11:41 please stop typoing krogoth as kergoth, it keeps pinging me :) Aug 11 15:12:49 hehe Aug 11 15:12:51 kergoth: very sorry. Both names are very similar to me, so I mix them up. My bad. Aug 11 15:13:03 np Aug 11 15:13:15 you're not the only one :) Aug 11 15:13:28 I can tell... Aug 11 15:14:22 sveinse: thats not a distro feature. have a look at the PACKAGECONFIG block in avahi.in, add a new one for qt5, then do PACKAGECONFIG_append_pn-avahi= " qt" in your config Aug 11 15:15:02 +FILES_libavahi-core = "${libdir}/libavahi-core.so.* ${libdir}/girepository-1.0/AvahiCore*.typelib ${libdir}/libavahi-core.so.*" <— why? Aug 11 15:15:17 +RDEPENDS_${PN}-dev = "avahi-daemon (= ${EXTENDPKGV}) libavahi-core (= ${EXTENDPKGV}) libavahi-client (= ${EXTENDPKGV}) libavahi-qt5 (= ${EXTENDPKGV})" <— breaks if qt is disabled Aug 11 15:15:30 rburton: All right. (My excuse is that not all patches/fixes I read from other, such as the fsl image, is done by the book. It's often difficult to know what is the right approach.) Aug 11 15:15:40 Hello Everyone. Aug 11 15:16:10 I'm having issue with linux-mainline customization. (I hope, it's a right place to ask) Aug 11 15:16:26 I understand the process : Aug 11 15:16:50 1/ bitbake linux-mainline -c clean Aug 11 15:17:04 2/ bitbake linux-mainline -c menuconfig Aug 11 15:17:23 3/ Check/uncheck in the menu config Aug 11 15:17:50 4/ bitbake linux-mainline -c diffconfig in order to get the fragment of modified values Aug 11 15:18:19 5/ Add to recipes-kernel/linux in my layer Aug 11 15:18:25 rburton: The first +: the add is superflous agreed. The second +: I don't know why it's like this. I didn't write this code, but I'm left with it Aug 11 15:18:56 The issue is simple, all lines in my fragment are commented. Such as "# CONFIG_MEDIA_CAMERA_SUPPORT is not set" Aug 11 15:19:06 "# CONFIG_SOUND is not set" Aug 11 15:19:12 "# CONFIG_LOGO is not set" Aug 11 15:19:14 etc Aug 11 15:19:39 Is it normal? Or do you know where it could come from? Aug 11 15:21:13 (It only does this for boxes I uncheck) Aug 11 15:21:34 (the other ones are not commented : "CONFIG_MEMSTICK=y") Aug 11 15:22:10 Should I modify directly the line ? Such as ""CONFIG_LOGO=y" ? Aug 11 15:23:28 rburton: Is there a standardized method to set PACKAGECONFIGs which can be resolved by the inclusion of other packages? E.g. for avahi.inc, it manually sets the "dbus" packageconfig, but it does not detect if dbus otherwise have been selected Aug 11 15:23:49 no Aug 11 15:23:55 ok Aug 11 15:24:03 each recipe makes a stab at what a 'norma' configuration is Aug 11 15:24:09 normal, even Aug 11 15:24:15 so avahi without dbus is 99% crippled Aug 11 15:24:26 qt5 is a separate layer so has to be disabled out of the box. Aug 11 15:24:49 but as i show above, make it a packageconfig and you can enable it from distro or local conf with one line Aug 11 15:25:03 yeah, that's perfectly fine Aug 11 15:29:09 My bad, it seems the '#' are not comments; when I insert my fragments in my recipe, a clean followed by menuconfig shows the modification are taken in account. Aug 11 15:29:21 Good to know Aug 11 15:30:35 Sorry to question now again, but if I read the source and the small parts of the documenation correct: pkg_postinst will not be extecuted while startup, its executed while do_rootfs. Aug 11 15:30:49 Is this correct? Aug 11 15:33:40 HyP3r: executed at rootfs, if it fails it is ran again on first boot Aug 11 15:34:11 so postinsts which want to be ran at first boot for whatever reason just do exit 1 if $D is set Aug 11 15:37:12 rburton: ah thats the trick \o/ Aug 11 15:40:14 rburton: next thing is configuring those mass of conf files (smb.conf, avahi.conf, httpd.conf) what I did is for each package a bbappend recpie (e.g. avahi_0.6.31.bappend) added a patch and a bit do_install_append like that http://pastebin.com/HpeFfsAX Aug 11 15:40:41 the var ${RACECHIP_HOSTNAME} is injected with this stragey http://stackoverflow.com/questions/17366984/is-it-possible-to-pass-in-command-line-variables-to-a-bitbake-build Aug 11 15:40:47 Is this a good practise? Aug 11 15:40:51 I'm not really sure Aug 11 15:41:21 HyP3r: better to use a variable in bitbake.conf Aug 11 15:41:35 and just make users set it in local.conf Aug 11 15:42:09 then you can pre-set it in site.conf, local.conf, or your distro.conf as relevant Aug 11 15:44:04 rburton: ok but this 'sed' thing is a good way to solve that? Aug 11 15:45:23 rburton: I ended up with something like this: http://paste.ubuntu.com/23033128/ Aug 11 15:47:28 I admit this being far simpler, so thanks! Aug 11 15:58:02 HyP3r: sed Aug 11 15:58:03 ? Aug 11 16:00:29 take a look at my recpie, thats how I insert the current hostname (now from the site.conf) Aug 11 16:02:10 HyP3r: you'd probably be able to remove the patch with a well crafted sed but yeah Aug 11 16:02:41 I'm not able to replace a string with a patch? Aug 11 16:03:32 i'm guessing your patch is to add RACECHIP_HOSTNAME that you then sed out Aug 11 16:03:42 if thats teh only change then you can do that with just a sed Aug 11 16:03:44 rburton: yes thats the stragegy Aug 11 16:04:02 rburton: but the patch is also enabling some other things ... Aug 11 16:04:05 fair enough Aug 11 16:04:19 ok, I want to realize that in best practise, so ok Aug 11 16:35:19 Is there a wy to force yocto to build a recpie everytime its touched? Aug 11 16:36:01 I have here some git and hg repostiroy code, and my coworkers create everyday a new version \o/. Atm. I do everytime just a 'clean' Aug 11 16:39:03 HyP3r: set SRCREV to ${AUTOREV} and add ${SRCPV} to PV Aug 11 16:43:12 kergoth: I allready set SRCREV to AUTOREV Aug 11 16:43:34 that would be one of the two things, yes Aug 11 16:44:25 and the PV thing I don't get: the recpie file name is tes.bb so PV seems to be empty. Why should I set SRCPV to PV? Aug 11 17:43:54 Hi all...a rookie question Aug 11 17:44:38 I have a bbclass that needs to know what architecture it is running on...for example x86, x86-64, ARM or MIPS Aug 11 17:45:07 Is there a variable that I can look at to get that info Aug 11 17:46:28 hahmed, http://www.yoctoproject.org/docs/1.8/ref-manual/ref-manual.html see TARGET_ARCH Aug 11 17:46:53 there are various variables that have related meaning, scan through the *ARCH* vars for more info Aug 11 17:47:12 Thanks....wow can't believe I missed that... Aug 11 17:47:19 thats exactly what I am looking for Aug 11 18:06:23 HyP3r: 'add' :) Aug 11 21:02:40 Hi, I have some recipes I've created that use external source code via EXTERNALSRC = "...". This works fine the first time I build my project however afterwards if I make changes to this source it isn't rebuilt on subsequent builds. Is there a way to do that? Aug 11 21:19:43 dmiess: which version of the build system are you using? Aug 11 21:28:38 I'm using Yocto 1.7, is that what you mean? Aug 11 21:29:08 BitBake Build Tool Core version 1.24.0, bitbake version 1.24.0 Aug 11 21:39:38 dmiess: ah... how soon we forget Aug 11 21:39:56 in 1.7 we didn't have anything set by default to ensure it rebuilt Aug 11 21:40:27 you can set do_compile[nostamp] = "1" (just in the specific recipe, nowhere else!) Aug 11 21:41:23 the result will be that it will build every time regardless of whether it needs to, but then that should be offset by whatever build system the software your recipe is using (e.g. make) which should figure out what needs to be done and what doesn't Aug 11 21:41:40 bluelightning: Great, thanks! I'll give that a try. Aug 11 21:41:58 btw, in 2.1 we have made it so that it actually determines if files have changed and only rebuilds if they have Aug 11 21:42:22 it's also worth noting that 1.7 is very much unsupported these days, so if you can I would strongly suggest upgrading Aug 11 21:43:03 you're likely missing quite a few security fixes in the metadata if nothing else, assuming you haven't backported them all yourself Aug 11 21:43:17 At the moment upgrading isn't really a possibility for me but I'll be sure to keep that in mind for the future. **** ENDING LOGGING AT Fri Aug 12 02:59:58 2016