**** BEGIN LOGGING AT Wed Jul 06 02:59:58 2016 Jul 06 07:26:34 morning Jul 06 09:29:11 php 5.6.16 and mariadb 5.5.46 doesnt work, php-cgi generates this error: php-cgi: relocation error: php-cgi: symbol mysql_server_init, version libmysqlclient_16 not defined in file libmysqlclient.so.18 with link time reference Jul 06 09:29:19 what could be the problem? Jul 06 09:30:07 it worked in jethro but not in krogoth Jul 06 09:31:27 * php 5.6.18 Jul 06 11:44:24 i have an issue with php and mariadb on krogoth, php-cgi generate this error on execute: php-cgi: relocation error: php-cgi: symbol mysql_server_init, version libmysqlclient_16 not defined in file libmysqlclient.so.18 with link time reference Jul 06 12:18:11 ndec_: who do i have to beg to get my linaro cloud instance request reviewed? (DC-28) Jul 06 12:18:50 Hi. python-core has a circular dependency on python-lang. Is it a known issue? Jul 06 12:19:45 mario-goulart: sounds like -lang should just be merged into -core, but is that actually an issue? Jul 06 12:21:49 rburton: not that I could observe in practice. I'm just doing some package analysis with data from buildhistory and realized that circular dependency. Jul 06 12:22:27 easy enough to merge them :) Jul 06 12:23:04 Are circular dependencies considered an issue in general? Jul 06 12:23:13 runtime deps, not massively Jul 06 12:23:27 they'll both be installed Jul 06 12:23:39 ok. Good. Thanks. Jul 06 12:33:50 rburton: i will try to see.. i got mine processed today.. Jul 06 12:34:25 rburton: we can assume the approval process might be simpler for someone with @linaro vs @intel address ;-) Jul 06 12:35:23 :) Jul 06 12:35:26 HE'S A SPY! Jul 06 12:43:57 morning Jul 06 12:49:23 I need to use shadow but the filesystem in normal operation is read only so the post_install script is not working, is there any way I can get around this? Jul 06 13:02:45 how is a postinstall running if the file system is read only? Jul 06 13:03:00 oh, a postinst on first boot? Jul 06 13:05:42 rburton: it's working because in development mode the rfs is not read only, I need to find a way to convert the files at build tinme Jul 06 13:06:15 chouimat: can you not do the postinst at rootfs time instead of on the target Jul 06 13:08:09 ok. do you know where /etc/passwd is coming from? I think I will go to plan b: don't install "non shadow passwd" and install the shadow one instead Jul 06 13:16:56 rburton: so you request is being processed, just hitting some technical issues, so nothing to do with your email address, that's cool ;-) Jul 06 13:17:05 heh Jul 06 13:17:11 thanks Jul 06 15:44:54 Greetings. Trying to use Poky with https://github.com/ARM-software/meta-mali to build core-image-minimal. And failing ... Jul 06 15:45:33 ~/odroid/openembedded-core/build $ bitbake core-image-minimal Jul 06 15:45:33 ERROR: Failure expanding variable MACHINE_ARCH, expression was ${@[d.getVar('TUNE_PKGARCH', True), d.getVar('MACHINE', True)][bool(d.getVar('MACHINE', True))].replace('-', '_')} which triggered exception AttributeError: 'NoneType' object has no attribute 'replace' Jul 06 15:46:18 Building for either default x86 or Beaglebone targets works as expected. Jul 06 15:46:47 local.conf has MACHINE ?= "odroid-xu3" Jul 06 15:47:56 bblayers.conf has BBPATH = " /home/user/src/meta-mali " Jul 06 15:48:05 any thoughts welcomed. Jul 06 15:56:30 that's not how you add a layer to your build. Jul 06 15:56:35 you add layers to BBLAYERS, not BBPATH Jul 06 15:58:08 Hi. OK, that makes a great deal of sense ... Jul 06 15:59:01 as in this: Jul 06 15:59:01 BBLAYERS ?= " \ Jul 06 15:59:01 /home/jevans/src/openembedded-core/meta \ Jul 06 15:59:01 " Jul 06 15:59:05 becomes .... Jul 06 16:00:11 BBLAYERS ?= " /home/user/src/openembedded-core/meta \ Jul 06 16:00:11 /home/user/src/meta-mali" Jul 06 16:00:16 ? Jul 06 16:15:24 g40 are you using daisy branch on Poky ? Jul 06 16:28:39 Hi armpit: daisy. yes. Jul 06 16:42:31 kergoth. _big_ thanks for spotting problem. things now appear to be working. Jul 06 16:45:29 np Jul 06 18:22:58 kergoth: in one case bitbake hangs on a git fetch on git ls-remote Jul 06 18:23:19 is there a simple way to insert a timeout for this wait Jul 06 19:21:30 khem: now that we're on python 3, i think subprocess may have a built in timeout option now. if so, it'd be trivial to add to the bitbake code Jul 06 19:21:46 that's what i'd suggest, rather than rolling our own subprocess timeouts Jul 06 19:22:23 kergoth: yes +100 to that Jul 06 19:22:34 we want less clones of subprocess, not more! Jul 06 19:23:00 the only clone i consider acceptable inside oe now is a copy of subprocess.run inside bitbake until we can depend on py3.5 Jul 06 19:23:13 and move everything over to that Jul 06 19:23:33 agreed. that should usually be the case, really. absolute worst case we backport, and only as a temporary measure, and always in one place Jul 06 19:25:59 the new run api looks tasty Jul 06 19:26:04 finally does what you want Jul 06 19:26:09 only took a decade! Jul 06 19:54:56 rburton, https://youtu.be/HqfFIWRYWoo Jul 06 19:55:06 local professor comments on #brexit Jul 06 20:21:01 rburton: my problem is coming from bb.process.run() which is tryig to run git ls-remote Jul 06 20:22:06 ndec_: I have signed up for linaro cloud but no answers yet Jul 06 20:35:05 clearly a conspiracy Jul 06 20:36:05 we should have a summer camp: http://lists.busybox.net/pipermail/buildroot/2016-July/167050.html Jul 06 20:51:21 generally such sprints can be helpful for backlogs Jul 06 21:04:23 khem: Linaro Cloud requests are being worked through Jul 06 21:07:39 XorA: OK. Jul 06 21:08:11 khem: slow rollout as its all a bit rought around the edges still Jul 06 21:08:20 understood Jul 06 21:08:59 hoping to get the instance and do some native clang work Jul 06 21:12:02 khem: hopefully Cello ships the end of the month too, then you can have one on your desk Jul 06 21:15:40 i have a linaro cloud account Jul 06 21:15:44 \o/ Jul 06 21:16:28 arse, core-image-sato fails to build on aarch64 Jul 06 21:16:48 ah well the point was "does gcc buid" and yes it does Jul 06 21:17:18 XorA: i discovered that you can't install rsync on the debian setup, as it hangs in postinst with systemd/password stuff Jul 06 21:17:57 rburton: file a bug please, although I think that has been fixed in the repo Jul 06 21:19:42 Im sure Dave Mandala hit a similar/the same bug during early testing Jul 06 21:20:39 fabo: ^^^ Jul 06 21:22:15 XorA: http://www.lenovator.com/product/103.html says Q2 and lists 299/- Jul 06 21:23:09 XorA: will I have to get RAM extra Jul 06 21:23:16 and hd too ? Jul 06 21:23:30 khem: I am not sure what its shipping kit is, HD you'll need to supply Jul 06 21:23:52 I have one and good news is for a 96borad is all support is currently in -next Jul 06 21:24:22 but mine has "issues" hence why its a month late Jul 06 21:26:25 but you do get an arm sock with a big ass heatsink and fan :-) Jul 06 21:33:57 heh ok Jul 06 21:34:34 96boards does not specify SOC arch. So someone can take mips or x86 SOCs and make 96boards too Jul 06 21:34:50 khem: dont even get started on that :-) Jul 06 21:34:57 with fans and heatsinks its easy Jul 06 21:35:20 since that know-how is plenty in market **** ENDING LOGGING AT Thu Jul 07 02:59:59 2016