**** BEGIN LOGGING AT Fri Sep 11 02:59:57 2020 **** BEGIN LOGGING AT Fri Sep 11 03:02:18 2020 **** BEGIN LOGGING AT Fri Sep 11 04:29:57 2020 **** BEGIN LOGGING AT Fri Sep 11 06:52:23 2020 Sep 11 06:52:29 I'm getting an ext4 image that differs in md5sum, but a tar file of the file system the md5sums are equal - which leads me to beleive that the issue is timestamps Sep 11 06:53:20 simonpe^^: works for everything up to the packages in core-image-sato and core-image-full-cmdline and we verify that with automated testing Sep 11 06:53:38 simonpe^^: sounds like something in the final ext4 image generation isn't quite right :( Sep 11 06:54:01 RP: I need binaries with the same md5sum for my "top" build system to determin a build is unnecessary Sep 11 06:55:10 I read something about reproducible_builds.bb but I don't know if thats something automagic or if I have to enable smth Sep 11 07:29:16 simonpe^^: it was made the default from dunfell Sep 11 07:29:44 packages in DEPENDS are pulling their own DEPENDS? Sep 11 07:32:22 RP: yeah I found it, I actually think the problem is with wic Sep 11 07:33:42 simonpe^^: patches very welcome to improve things, we've got so far but there are only a few people sending patches so its slow going Sep 11 07:33:47 looking at the rootfs image all the files are timestamped at time of creation Sep 11 07:34:02 Getting the tarballs working is a huge win in itself up to that point Sep 11 07:34:19 yeah, its super useful Sep 11 07:34:54 that's usually a counter argument for using linux in small systems, that you can't get reproducible builds Sep 11 07:35:56 And its nice to establish these things early in the project so it isn't "forgotten" by some eager project manager later on :P Sep 11 07:53:34 mihaix: initrdscripts mods did the trick, tks.. Now, I've to figure out how to supply a custom grub.cfg.. Sep 11 07:55:13 mihaix: yes Sep 11 07:58:05 PaowZ_: yw, you should also find some hints about grub config in those install scripts Sep 11 08:00:42 qschulz: ok, thanks Sep 11 08:02:27 is there a secret tool somewhere that can generate a "fuller" list of task-depends? Sep 11 08:03:13 hmm, oe-depends-dot Sep 11 08:07:12 ndec: re new docs, I wasn't able to find the mini scenario to reproduce the issue but it seems that the switch to/from singleindex.html isn't working all the time (sometimes adding /singleindex.html at the end of a url with anchors) Sep 11 08:07:41 ndec: also, when switching to/from singleindex.html repeatedly, more and more '/' are prefixing singleindex.html Sep 11 08:08:00 ah, thanks! at least the 2nd part should be easy to fix. Sep 11 08:09:13 it might be this leading '/' here: https://git.yoctoproject.org/cgit/cgit.cgi/yocto-docs/tree/documentation/sphinx-static/switchers.js?h=sphinx#n184 Sep 11 08:09:25 Hi! Yocto reboots into initramfs and there is a script which adds a missing sd card partition to /dev/mmcblk1, but after that fdisk operation I cannot see the newly created partition as e.g. /dev/mmcblk1p3. What is missing to have it in /dev ? Sep 11 08:11:12 mihaix: yes indeed, there's a bunch of leads in them, actually.. Sep 11 08:11:16 qschulz: right, this is the fix: Sep 11 08:11:20 https://www.irccloud.com/pastebin/jU3iJ02O/ Sep 11 08:11:33 The internet says I should call partprobe. But a program partprobe is not existing in my yocto image. If it's the right program, which package contains partprobe? Sep 11 08:11:54 ndec: is docroot guaranteed to end with a '/'? Sep 11 08:12:26 ndec: answering myself: yes :) Sep 11 08:12:30 maybe not, i need to check. Sep 11 08:12:34 https://git.yoctoproject.org/cgit/cgit.cgi/yocto-docs/tree/documentation/sphinx-static/switchers.js?h=sphinx#n140 Sep 11 08:12:36 oh, really? Sep 11 08:13:06 ah, perfect. Sep 11 08:16:49 qschulz: i should remove the '/' for the other case (var new_url = docroot + "/mega-manual/mega-manual.html";) Sep 11 08:18:40 it's slightly confusing that the mega manual link is singlehtml, but it's called 'mega' in all_doctypes. and the regular doc type is 'single' but it's not singlehtml ) Sep 11 08:21:27 qschulz: i reproduced the other issue, i ended up with https://docs.yoctoproject.org//index.html#document-kernel-dev//singleindex.html Sep 11 08:21:54 i think we need to trim the anchors when going back and forth. RP, you did that, what do you think? Sep 11 08:23:52 the scenario is go to docs.yp.org , switch to mega manual, open a doc using the nav bar, then switch to 'single' , then switch back to 'mega'. things keep beeing added to the url Sep 11 08:24:58 i am pushing a fix for the leading '/' for now. Sep 11 08:27:30 ndec: thx! with the proper scenario at least we have a way to debug it :) Sep 11 08:31:31 ndec: that sounds reasonable, thanks for fixing Sep 11 08:31:51 i fixed 50% of the reported issues only ;) Sep 11 08:32:08 ndec: its progress :) Sep 11 08:32:50 and it's Friday, let's keep the optimism :) Sep 11 08:41:14 RP: qschulz : there are more '/' issues.. docs.yp.org -> select 3.1 version (e.g. pre-sphinx) -> select mega manual, and we have https://docs.yoctoproject.org/3.1.2//mega-manual/mega-manual.html Sep 11 08:48:10 ok.. it was a deployment issue.. when rebuilding the docs it didn't update the .js file... Sep 11 08:52:46 ndec: phew :) Sep 11 08:53:40 RP: actually.. it's not that.. I need to update the 'transition' branch with the same fix! Sep 11 09:09:40 ndec: ah, right. I wondered about that Sep 11 09:09:50 it's done now. Sep 11 09:13:04 ;/buffer 13 Sep 11 09:13:07 oopsie Sep 11 09:15:17 RP: docs.python.org has the same problem with anchors (without surprise). I can end up here: https://docs.python.org/3.5/library/intro.html#notes-on-availability which doesn't exist! Sep 11 09:17:36 ndec: I haven't followed exactly what you've been discussing but I can imagine ways it breaks :/ Sep 11 09:19:42 when you switch version, we only replace the in the url, so if the user was in a page somewhere down in the hierarchy, we might end up on a page on that does not exist in . and if the current docroot has anchors (#xxx) we might end up on a page that exist, but not the anchors.. Sep 11 09:20:31 we might want to decide that switching version brings us back to docs.yp.org// and trim the url. Sep 11 09:21:40 ndec: I think it's okay if we keep the full url with the anchor? if the anchor does not exist, you'll just land at the beginning of the page Sep 11 09:21:53 IIRC that's what's happening currently with yp.org/docs? Sep 11 09:21:54 ndec: I think the feeling was that if there are anchors its helpful and if there aren't, well, you're no worse off Sep 11 09:22:49 right. i agree it's not a 'big issue'. until we introduce a new html page that exist in one version and not the others. Sep 11 09:23:07 ndec: also, in mega-manuals, the anchors are named the same as in "normal"/"signle" pages. Which means if it happens that there are two anchors with the same name, either it breaks or you fall back to the first anchor in the page/ Sep 11 09:23:09 ? Sep 11 09:23:28 ndec: there is in theory "fallback" code with iterates back through urls checking if they exist Sep 11 09:24:16 ok.. that said, they aren't the biggest problems we have with docs right now :-) Sep 11 09:24:36 ndec: you beat me to it. indeed, there's room for improvement but nothing critical :) Sep 11 09:24:56 we are over engineering a little bit here ;) Sep 11 09:25:07 ndec: yes, worry about it later I think Sep 11 09:26:25 ndec: btw thx for the migration to Sphinx, just the search is going to make it much easier for people to get started IMO :) Sep 11 09:26:55 you mean the search box? Sep 11 09:27:48 ndec: yup: https://docs.yoctoproject.org/search.html?q=IMAGE_INSTALL Sep 11 09:28:53 cool. Sep 11 09:29:23 i think there are various ways to improve/replace the builtin search engine. we can look into that later as well Sep 11 09:39:54 RP: what's the rationale for "bitbake: fetch2: Change git fetcher not to destroy old references" patches? does that fix known issue, or upcoming potential issues? **** BEGIN LOGGING AT Fri Sep 11 09:47:18 2020 Sep 11 09:51:19 ndec: it would have averted one problem already Sep 11 09:51:51 ndec: if upstreams change "master" -> "main" the fetcher was destroying the old branches, breaking archive tarballs for older releases **** BEGIN LOGGING AT Fri Sep 11 10:04:50 2020 **** BEGIN LOGGING AT Fri Sep 11 11:13:42 2020 Sep 11 12:12:08 zeddii: got this again Sep 11 12:12:10 https://www.irccloud.com/pastebin/zhenB51e/ **** BEGIN LOGGING AT Fri Sep 11 13:01:33 2020 Sep 11 14:28:53 rburton: odd. I'll have another look at it! Sep 11 14:49:06 i wonder if linaro kernel is doing something dumb Sep 11 15:00:11 Morning folks Sep 11 15:01:36 morning sgw Sep 11 15:01:45 how's the fire around your way? Sep 11 15:03:25 Havent finished all the checking yet, but still very smokey and ashy here, about 400 AQI (Hazardous) where 20-30 is normal. Sep 11 15:05:13 rburton: https://inciweb.nwcg.gov/incident/7170/ This will not be updated until about 10:00am here (another 2 hours). My guess is it's grown in size again. Sep 11 15:09:46 urgh Sep 11 15:09:48 good luck Sep 11 15:35:24 ugh scm mirror tarballs have no many annoyances since there's no telling what is or isn't in one. that's actually one positive to the use of generated shallow tarballs, much more specific Sep 11 15:35:32 course it has its own tradeoffs Sep 11 15:57:34 kanavin_home: question: http://git.yoctoproject.org/cgit/cgit.cgi/poky/tree/meta/classes/package_rpm.bbclass#n687 does set xz compression. but rpm-native does not DEPEND on xz, so rpmbuild seems to use bz2 or gz as fallback. intentional ? **** BEGIN LOGGING AT Fri Sep 11 15:59:31 2020 Sep 11 16:46:23 bitbake: file foo in package barr requires a package that doesn't exist Sep 11 16:46:32 me: alright don't need it, let's delete the file Sep 11 16:46:47 bitbake on the next run: file foo in package barr requires a package that doesn't exist Sep 11 16:46:56 =# **** BEGIN LOGGING AT Fri Sep 11 16:50:32 2020 Sep 11 17:06:40 any layerindex admin online? I've tried to reset my password and after a while it failed with internal server error and now says my account is locked and that I should contact admin (e-mail with PW reset didn't arrive) Sep 11 17:09:00 bluelightning: otavio: ^? Sep 11 17:10:58 it shows "Your account has been locked out. Please contact the admin." even when I clear all cookies or use different browser, when I click on "Log in" so it's probably broken for everybody not just my account Sep 11 17:11:25 JaMa: halstead may be around? Sep 11 17:12:38 JaMa: I am here but I am not sure I can unlock your accound Sep 11 17:12:44 JaMa: I am here but I am not sure I can unlock your accoundt Sep 11 17:13:25 Mine is locked as well: " Sep 11 17:13:25 Your account has been locked out. Please contact the admin. Sep 11 17:13:45 otavio: ok, then we really need someone with more powers like halstead :) Sep 11 17:19:15 :-) indeed Sep 11 17:19:28 we are poor humans ;) Sep 11 17:36:47 Sabertooth Tiger Battle Zord! Sep 11 18:45:57 JaMa: argh I really wish I had not implemented that :/ Sep 11 18:46:33 let me see if I can unblock it Sep 11 18:47:45 hmm, seems the access attempts list is empty - JaMa / Otavio are you still locked out ? Sep 11 18:47:47 dl9pf: can you provide a link to the rpm recipe you're looking at? Sep 11 18:48:05 one in latest master does depend on xz Sep 11 18:53:52 bluelightning: it doesn't show the message about locked account anymore, but still cannot login, should I try to reset password again? Sep 11 19:04:47 bluelightning: logged in Sep 11 19:05:28 otavio: great, thanks - I think JaMa's access is fixed now as well hopefully Sep 11 19:05:47 though there is still something wrong with the pw reset functionality, maybe with the email sending **** BEGIN LOGGING AT Fri Sep 11 19:35:39 2020 Sep 11 20:51:54 yes, works for me again, thanks to bluelightning **** BEGIN LOGGING AT Fri Sep 11 21:16:24 2020 Sep 11 21:58:52 win 2 Sep 11 21:58:54 gah **** BEGIN LOGGING AT Fri Sep 11 22:40:24 2020 Sep 12 02:42:54 Is there a way to have a U-Boot script run on device startup and then continue normal boot? Sep 12 02:43:24 I guess I can resume boot from the script with the last command being `boot` Sep 12 02:43:46 But how do I get it to load automatically in the beginning? **** ENDING LOGGING AT Sat Sep 12 02:59:57 2020