**** BEGIN LOGGING AT Sat Feb 29 02:59:59 2020 Feb 29 07:35:02 build #277 of mvebu/cortexa53 is complete: Failure [failed pkgbuild] Build details are at http://buildbot.openwrt.org/master/images/builders/mvebu%2Fcortexa53/builds/277 blamelist: Mark-MC Lee (???) , John Crispin , Koen Vandeputte , David Bauer , Christian Feb 29 07:35:02 Lamparter , Steven Lin , Adrian Schmutzler , Kevin Darbyshire-Bryant , ?lvaro Fern?ndez Rojas , Eneas U de Queiroz , Hauke Mehrtens Feb 29 11:15:46 build #75 of x86/64 is complete: Failure [failed kmods] Build details are at http://buildbot.openwrt.org/openwrt-18.06/images/builders/x86%2F64/builds/75 blamelist: Koen Vandeputte , Jo-Philipp Wich , Petr ?tetiar Feb 29 12:49:20 build #73 of arc770/generic is complete: Failure [failed] Build details are at http://buildbot.openwrt.org/openwrt-18.06/images/builders/arc770%2Fgeneric/builds/73 blamelist: Koen Vandeputte , Jo-Philipp Wich , Petr ?tetiar Feb 29 12:50:17 build #73 of ath25/generic is complete: Failure [failed] Build details are at http://buildbot.openwrt.org/openwrt-18.06/images/builders/ath25%2Fgeneric/builds/73 blamelist: Koen Vandeputte , Jo-Philipp Wich , Petr ?tetiar Feb 29 12:51:15 build #75 of pistachio/generic is complete: Failure [failed] Build details are at http://buildbot.openwrt.org/openwrt-18.06/images/builders/pistachio%2Fgeneric/builds/75 blamelist: Koen Vandeputte , Jo-Philipp Wich , Petr ?tetiar Feb 29 12:51:25 build #99 of armvirt/64 is complete: Failure [failed updatefeeds] Build details are at http://buildbot.openwrt.org/openwrt-19.07/images/builders/armvirt%2F64/builds/99 Feb 29 12:52:13 build #74 of mediatek/mt7623 is complete: Failure [failed] Build details are at http://buildbot.openwrt.org/openwrt-18.06/images/builders/mediatek%2Fmt7623/builds/74 blamelist: Koen Vandeputte , Jo-Philipp Wich , Petr ?tetiar Feb 29 13:55:33 build #100 of armvirt/64 is complete: Success [build successful] Build details are at http://buildbot.openwrt.org/openwrt-19.07/images/builders/armvirt%2F64/builds/100 Feb 29 15:49:26 build #249 of mpc85xx/p1020 is complete: Success [build successful] Build details are at http://buildbot.openwrt.org/master/images/builders/mpc85xx%2Fp1020/builds/249 Feb 29 17:50:21 How can i switch the buildsystem to kernel 5.4? Feb 29 17:59:50 nvm.. i guess its CONFIG_TESTING_KERNEL Feb 29 18:57:12 build #69 of x86/legacy is complete: Failure [failed kmods] Build details are at http://buildbot.openwrt.org/openwrt-18.06/images/builders/x86%2Flegacy/builds/69 blamelist: Koen Vandeputte , Jo-Philipp Wich , Petr ?tetiar Feb 29 19:49:43 Quick sanity check on https://git.openwrt.org/?p=openwrt/staging/ldir.git;a=commit;h=fc89320141d88f689e8aa2c23e63e2d67838f76c please Feb 29 19:56:52 ldir: if I got that error message, I wouldn't even know where to start looking for an 'xargs' that supports that without any explanation what '-r' does (single letter stuff is a PITA to google for) Feb 29 19:57:54 it's equally fun when you build on a system without a suitable xargs. Feb 29 19:58:59 ldir: the interesting question if you actually need it - the *BSD xargs does the -r behaviour by default, and "supports" -r only in the way it doesn't error when being passed, but the argument does nothing there Feb 29 20:00:51 ha, so having just tested darwin it assumes '-r' by default BUT errors as an 'unknown option' Feb 29 20:02:21 so maybe check if xargs supports -r and only pass it then (because otherwise it isn't needed), and no need to install any gxargs Feb 29 20:03:48 probably easier to check if it supports it than if it needs it Feb 29 20:05:58 hmm, problem with that is catching all the instances of xargs. this way it's guaranteed that the xargs linked to in staging understands '-r', whether it used $XARGS or not. Feb 29 20:08:07 let me expand on the '-r' requirement description Feb 29 20:12:35 oh and gxargs comes 'for free' with homebrew's findutils package which is needed anyway. Feb 29 20:13:06 then maybe use a setup host command macro? Feb 29 20:13:31 err, isn't that what I've done in the commit ? Feb 29 20:16:25 oh, I didn't actually look at the link :) Looks fine Feb 29 20:18:02 lol - good stuff - courtesy of ynezz encouragement of gitlab-ci things I'm falling over a few macos builder gotchas and I intend to update the 'build on mac' wiki page to have up to date & correct details. Feb 29 20:20:15 this xargs thing was a real surprise, never seen it before. Related to CONFIG_AUTOREMOVE, which I've never set. Feb 29 20:21:49 ldir: the main source for xargs -r is rules.mk's definition of XARGS:=xargs -r which is used in a few places, makes me wonder why you haven't seen it before Feb 29 20:22:16 ^^^ Feb 29 20:22:23 AUTOREMOVE Feb 29 20:25:03 ldir: scripts/gen-dependencies.sh uses xargs -r, and is used for every package Feb 29 20:28:56 I don't understand either - I've been building openwrt for 4+ years, the last 2 of those under macos, never, ever seen it complain about lack of 'xargs -r' before. Feb 29 20:33:19 ha! right - it's 'cos the build system builds it for host. Feb 29 20:33:42 but it's a pre-req for the build IF you have AUTOREMOVE set. Feb 29 20:34:55 or that's my guess anyway Feb 29 20:42:58 so I guess there is a catch 22 there, can't use the hostbuild xargs for autoremove until its built, but you would need it to autoremove the stuff from findutils's dependencies (i.e. bison) Feb 29 20:45:01 it's a bit of a corner case - trust me to find it. Feb 29 20:47:17 if I will try to build on macos I guess I deserve what I get Feb 29 20:59:45 jow: hi, got my email? :) Feb 29 20:59:53 (buildbot patch) Feb 29 21:00:52 f00b4r0: yes, thanks Feb 29 21:00:56 np, yw Feb 29 21:01:12 jow: so is it normal to upload the kmods twice? Feb 29 21:01:21 yes Feb 29 21:01:24 k Feb 29 21:02:31 different topic: no new hw support patch accepted in ar71xx, correct? I'm looking at a piece of mikrotik kit that should be a dozen lines of code to add support for in the generic mikrotik image Feb 29 21:04:30 i guess that's a question i should ask whoever handles ath79/ar71xx :) Feb 29 21:06:15 very unlikely Feb 29 21:08:38 ynezz: got cut at "very unlikely", I assume that was a reply to my question? Feb 29 21:11:15 I don't see any T-Bone's question in my backlog, I was refering to ar71xx Feb 29 21:12:16 * T-Bone curses unstable link Feb 29 21:12:29 ynezz: T-Bone / f00b4r0 identity crisis :) Feb 29 21:12:35 *sigh* Feb 29 21:13:51 what I'll do I guess is I'll propose a quick patch to add support via ar71xx, hopefully that'll stir some comments as to the 'right'(TM) way to move the mikrotik nor machines to at79 Feb 29 21:14:26 specifically I'm curious whether there's any interest in keeping the single "one image to rule them all" approach (which I personally like) Feb 29 22:19:58 Hauke: mkresin: FYI: https://github.com/xdarklight/openwrt/commits/lantiq-linux-5.4-20190229 - it compiles for me, no further testing done yet Feb 29 22:20:59 Hauke: how about switching to the DSA driver for 5.4? I don't want to mess with the GPHY loader and the out-of-tree Ethernet driver Feb 29 22:32:36 xdarklight: I would like to swithc to the DSA driver Feb 29 22:32:47 but we probably need some migration path Feb 29 22:36:55 Hauke: maybe we can share a TODO-list somewhere so we can write down everything that comes to our mind and then work on it one-by-one Feb 29 22:51:43 Hauke: I think it makes sense to switch to dsa without migration Feb 29 22:59:17 the wish for migration scripts exists for at least two years (on mvebu and ipq806x) - and it's a good idea in principle, but /etc/config/network is complex - and may be very complex at times (and lantiq with the VLAN tagging necessary for multiple xDSL ISPs is not an exception to that rule), so making migration support a requirement probably ends up in a self-fullfilling prophecy to never switch to Feb 29 22:59:23 DSA (just because writing these scripts in a way that makes them remotely safe to use is somewhere between very hard and close to impossible) Feb 29 23:30:30 we do need to some enhance the existing for defining vlan membership, pvids etc with vlan_aware bridges Feb 29 23:34:09 yes Feb 29 23:34:26 and this time we should define how the config should look like, then implement it Feb 29 23:35:00 to avoid inconsistencies and poor config layout like with swconfig in the past Mar 01 01:41:20 build #281 of mvebu/cortexa72 is complete: Success [build successful] Build details are at http://buildbot.openwrt.org/master/images/builders/mvebu%2Fcortexa72/builds/281 Mar 01 02:31:37 Hauke: I suggest switching to DSA without migration as well. The biggest problem is the one CPU port issue. Needs a patch. **** ENDING LOGGING AT Sun Mar 01 02:59:57 2020