**** BEGIN LOGGING AT Mon Nov 09 03:02:03 2020
Nov 09 03:09:52 <Grommish> damex_: https://github.com/openwrt/openwrt/tree/master/package/network/services/lldpd
Nov 09 03:10:35 <Grommish> still there, maybe it's erroring? or moved?
Nov 09 03:39:14 <swalker> updated openwrt/upstream, https://sdwalker.github.io/uscan/index.html
Nov 09 05:28:20 >KGB-0< https://tests.reproducible-builds.org/openwrt/openwrt_tegra.html has been updated. (0% images and 100.0% packages reproducible in our current test framework.)
Nov 09 05:47:12 <damex> Grommish: yeah, maybe? i wonder if there is logs for package builds
Nov 09 06:48:31 <novski__> in looking for somebody who can fix kmod-sound-mtk as paid service. is this the right place to ask?
Nov 09 06:52:22 <PaulFertser> It might be. But in any case you'll want some text explaining with all the necessary details the problem so that anyone can evaluate the potential resources required to fix it.
Nov 09 06:59:57 <novski__> ok.
Nov 09 06:59:57 <novski__> I'm looking for somebody who can fix an openwrt package.
Nov 09 06:59:58 <novski__> I have a device that can be used with a sound chip over i2s. Unfortunately the driver is broken.
Nov 09 06:59:58 <novski__> Device: Linkit Smart 7688
Nov 09 06:59:58 <novski__> Audiochip: WM8960
Nov 09 06:59:58 <novski__> driver: kmod-sound-mtk
Nov 09 06:59:58 <novski__> related:https://github.com/lede-project/source/commit/63aa7023d8e04fb40ed0dc4971e80e86330b7ac6
Nov 09 06:59:59 <novski__> Thanks for a feedback if you are able to solve such a problem.
Nov 09 07:00:00 <novski__> Price can be disscussed.
Nov 09 07:03:35 <PaulFertser> novski__: is there a ticket which gives details about "brokenness"?
Nov 09 07:04:17 <PaulFertser> Just doesn't compile but no known issues?
Nov 09 07:05:31 <PaulFertser> Probably it was fixed since then just not reenabled? :)
Nov 09 07:17:22 <novski__> how can i reenable and test? I have built own images in past
Nov 09 07:19:05 <Grommish> Anyone know the easy way to pass a -fcommon flag to the build system for a package?
Nov 09 07:25:29 <Grommish> TARGET_CFLAGS+=-fcommon maybe?
Nov 09 07:25:38 <mangix> yes
Nov 09 07:25:48 <Grommish> appreciated
Nov 09 07:26:03 <novski__> can i simply remove the @Broken tag in Line 108?
Nov 09 07:26:23 <mangix> I added that to some packages where fixing the actual issues was too difficult
Nov 09 07:26:53 <Grommish> mangix: I'm trying to get libunwind to work and there is a known issue with gcc10 and it
Nov 09 07:27:09 <Grommish> mangix: the solution is to pass the -fcommon *shrug* or they say :) we shall see
Nov 09 07:30:54 <mangix> haven't seen that issue
Nov 09 07:35:54 <Grommish> Well.. if it fixes it, i'll put up an update and a gcc check.. it's 2 versions behind anyway
Nov 09 07:36:06 <mangix> great
Nov 09 07:36:14 <mangix> my last attempt was denied
Nov 09 07:36:40 <Grommish> interesting.. any reason why other than 'cuz'?
Nov 09 07:37:07 <Grommish> Because I'll need to update the DEPENDS to include mips64 at the very least
Nov 09 07:37:59 <mangix> https://patchwork.ozlabs.org/project/openwrt/patch/20200728202524.11424-1-rosenp@gmail.com/
Nov 09 07:38:37 <mangix> he's completely wrong on the ARC change
Nov 09 07:42:54 <Grommish> That's almost like a language issue
Nov 09 07:43:42 <mangix> looks like libunwind 1.5 will be out soon
Nov 09 07:43:46 <mangix> might as well wait
Nov 09 07:43:52 <Grommish> I just need it to work at all
Nov 09 07:44:02 <Grommish> right now, I can't even select it without hackery
Nov 09 07:44:46 <mangix> anyway, https://patchwork.ozlabs.org/project/openwrt/patch/20200831071150.19577-1-rosenp@gmail.com/
Nov 09 07:46:34 <Grommish> Thanks :)
Nov 09 08:05:28 <novski__> i have feeds.buildinfo config.buildinfo and .config files from my last build. can somebody point me to a description how to apply those to a new build?
Nov 09 08:20:44 <Grommish> novski__: the .config file just sits in your build root
Nov 09 08:24:06 <novski__> but i noticed that make menuconfig picks up only some config i have made for my previous builds. seams like some packages are not configured in .config? like language->python3 is one of them
Nov 09 08:25:41 <Grommish> .config isn't removed from build to build.  What are you doing between builds?
Nov 09 08:34:07 <novski__> i blown the vm a few months ago to free diskspace. Before I stored my build image and the .config file so i can rebuild it someday but now it seams my config file that i droped in to the openwrt/ folder and then applyed feeds update and install -a did not take all my packages because make menuconfig has them not set...
Nov 09 08:41:30 <aparcar[m]> mangix: What do you think about the SDK issue? I can't reproduce it locally.
Nov 09 08:42:27 <aparcar[m]> the Docker images haven't been updated for a week because the CI tests fail: https://gitlab.com/openwrt/docker/-/jobs/837338425
Nov 09 08:42:54 <aparcar[m]> In that case it's actually x86/64
Nov 09 08:43:03 <aparcar[m]> so it's not target specific but CI specific
Nov 09 09:01:51 <ynezz> you can't reproduce it locally?
Nov 09 09:03:56 <ynezz> I didn't noticed it, because I'm not using x86/64 SDK, but I can reproduce it easily
Nov 09 09:04:26 <ynezz> it's caused by this commit of mine 4b7165722cb0e2ba62ab68e8b90d4d4b0909744c (scripts: bundle-libraries.sh: fix broken SDK compiler)
Nov 09 09:15:51 <PaulFertser> novski__: I think you can also enable a switch that allows to select broken packages in menuconfig.
Nov 09 09:16:29 <PaulFertser> novski__: I'll try to reproduce build failure now
Nov 09 09:27:00 <PaulFertser> novski__: sound-mtk seems to be an old driver, and currently sound-mt7620 is offered and used by default. What makes you think it's inadequate?
Nov 09 09:30:46 <novski__> aplay -l gives aplay: device_list:272: no soundcards found...  sound-mt7620 is not installed in my build i made self. does the selection of Linkit Smart 7688 realy select it by default?
Nov 09 09:31:35 <PaulFertser> novski__: hm, no it's not in the profile. I'll check the dts now to see if that's the right driver.
Nov 09 09:35:01 <PaulFertser> novski__: yes, it's indeed the driver that should handle i2s module in mt7628.
Nov 09 09:35:58 <PaulFertser> novski__: but you'll need a small DTS change to enabel it.
Nov 09 09:36:23 <novski__> well aso for mt7688AN? thats my chip... but i already seen some drivers work that are described as mt7628...
Nov 09 09:36:43 <PaulFertser> novski__: it's using mt7628an.dtsi so it should work
Nov 09 09:37:07 <novski__> ok. can you patch it in the next month? im in no hurry...
Nov 09 09:37:40 <PaulFertser> novski__: I can send a patch for you to test about right now
Nov 09 09:38:01 <novski__> ok
Nov 09 09:40:27 <PaulFertser> novski__: http://paste.debian.net/1170555/ to enable it in DTS, and just select sound-mt7620 in menuconfig when building the image
Nov 09 09:41:31 <PaulFertser> novski__: it won't be enough most probably, but I need to know more about your hardware to configure it properly. Because just i2s support is not a sound card.
Nov 09 09:41:51 <PaulFertser> novski__: I need to know if hardware is similar to e.g. mt7628an_duzun_dm06.dts
Nov 09 09:42:47 <PaulFertser> novski__: you say your board has WM8960 but is it using teh same I2C address as this other board? 0x1a?
Nov 09 09:43:21 <novski__> ah ok. i have a fresh git download and my old .config file but when i copy my old .config file to /env/.config and go in to make menuconfig not even the Terget Profile is set to my board. do i have to do something more than transfairing .config ?
Nov 09 09:44:07 <PaulFertser> novski__: .config is to be put in the root of the sources
Nov 09 09:45:15 <novski__> the root has a symlink to /env/.config becaus i used ./scripts/env new current
Nov 09 09:46:46 <novski__> it seams to pick some info but not all...the Target System is set to (medieTek Ralik MIPS) and the subtarget to MT7688 based boards but the Taget Profile is (Default  Profile) not Linkit Smart 7688
Nov 09 09:51:06 <PaulFertser> novski__: well, select it manually. Probably your old config is not fully compatible.
Nov 09 09:52:24 <karlp> you might do a cp my-old-config .config; make olddefconfig ort something?
Nov 09 09:57:29 <PaulFertser> novski__: http://paste.debian.net/1170556/ builds for me. Assuming the I2C address is the same it should work.
Nov 09 09:57:44 <PaulFertser> novski__: and I selected the kmod needed
Nov 09 10:09:01 <enyc> I think this is a somewhat devel question!
Nov 09 10:09:08 <enyc> r.e.   https://git.openwrt.org/?p=openwrt/openwrt.git;a=commit;h=d3868f15f876507db54afacdef22a7059011a54e
Nov 09 10:09:33 <enyc> Looks like this technically affects ALL architectures,  but references especially affecting lantiq, I think ....
Nov 09 10:10:56 <enyc> (a) is that correct (b) is that 'fix' now only in master (c) is 20.x simply not forked yet ?
Nov 09 10:13:17 <PaulFertser> novski__: so did you try with http://paste.debian.net/1170556/ ?
Nov 09 10:14:30 <novski__> still adding all my special packages again. its a bit a pity that it does not take what i had already configured...
Nov 09 10:19:47 <PaulFertser> I checked the WM8960 datasheet, it's indeed on address 0x34, hardcoded.
Nov 09 10:20:06 <PaulFertser> They carefully avoid even mentioning I2C :)
Nov 09 10:22:06 <PaulFertser> 0x34 is 0x1a * 2, so it matches the config I added.
Nov 09 10:27:54 <novski__> perfect! thanks!
Nov 09 10:36:17 <Tapper> Hi people I have a .config file that I last used to build for a dveice more than 1 year ago.
Nov 09 10:37:21 <Tapper> Can I just inport that .config and kick off a new master build or is there a way to make shure that the config file is updated with any new flags that have bin introduced?
Nov 09 11:39:09 <PaulFertser> novski__: when booting the image please save whole dmesg, that'd be essential to look at if something goes wrong
Nov 09 11:41:23 <karlp> jow: I've got a view I'd like to use the "new js" style, L.view.extend(), but I want to use some external js in it, what do I need to do to make my external js available via 'require blah'; ? or would I be better off writing a stub lua templated .htm, that includes my external js, and then uses "ui.instantiateView(model)" ?
Nov 09 11:41:29 <karlp> or is there a third way?
Nov 09 12:48:55 <eduardas> Hello, anyone trying to hack on Xiaomi ax1800? Would like to collaborate if anyone is interested.
Nov 09 12:49:28 <eduardas> Trying to use mainline's minimal boot support for IPQ6018 to boot on this board, but failing so far
Nov 09 13:44:06 <russell--> eduardas: failing how?
Nov 09 14:03:01 <eduardas> russell--: no output after "Starting kernel ..."
Nov 09 14:32:52 <jow> karlp: can you change the external js?
Nov 09 14:33:14 <jow> karlp: or do you need to use it unmodified (e.g. jquery or some charting library)
Nov 09 14:33:38 <karlp> I ... can :) I'd rather not, (initially it's the paho mqtt js library, but in general looking for right approaches)
Nov 09 14:34:05 <karlp> sorry, been away at lunch, should have had some more experiments finished by now :)
Nov 09 14:34:36 <jow> hm, right. So personally I'd create script tags in view.load() and return a promise resolving on onload
Nov 09 14:37:48 <jow> something along these lines: https://pastebin.com/33m8mw5B
Nov 09 14:38:29 <karlp> wow, I would never even have thought of that sort of thing :)
Nov 09 14:38:50 <karlp> makes sense seeing it, but totally not me the not-really-js dev :)
Nov 09 14:39:49 <jow> making that external js `require` compatible would invole moving it to a specific location (/www/luci-static/resources) and modifying it to return a constructor (something inheriting baseclass and implementing a constructor function)
Nov 09 14:40:17 <jow> the approach above is likely simpler and less invasive
Nov 09 14:40:26 <karlp> yeah, I was looking at some of the existing ones, wondering how much of the baseclass bit was required, and how much was just "needs to return an object"
Nov 09 14:40:33 <karlp> but hadn't gotten very far experimenting.
Nov 09 14:40:48 <karlp> thanks for this, this should get me a long way,
Nov 09 14:41:45 <jow> yw
Nov 09 14:42:36 <jow> if you need to load multiple files you can extend that approach to return a promise.all() aggregate
Nov 09 14:43:57 <jow> https://pastebin.com/Jr8i1AQc
Nov 09 14:47:13 <jow> or more concise https://pastebin.com/mq2cHq76
Nov 09 14:47:22 <jow> can be shortened further by using arrow functions
Nov 09 14:50:36 <karlp> neh, still want to support ie, been avoiding arrows so far :)
Nov 09 15:18:58 <Zero_Chaos> PaulFertser: hey man, did that mac address issue I was having ever go anywhere/
Nov 09 15:21:27 <PaulFertser> Zero_Chaos: I guess it's pending review from nbd
Nov 09 15:21:50 <Zero_Chaos> PaulFertser: got a link to something I can bump?
Nov 09 15:22:10 <PaulFertser> Zero_Chaos: hm, it's a mailing list message you can reply to
Nov 09 15:23:33 <Zero_Chaos> https://lists.openwrt.org/pipermail/openwrt-devel/2020-October/031781.html ?
Nov 09 15:23:49 <PaulFertser> Zero_Chaos: yes
Nov 09 15:24:55 <Zero_Chaos> nbd: I have tested PaulFertser's patch (and was the original bug reporter) for fixing mac address allocation.  Can you please approve the patch and maybe drop it in the stable tag as well?
Nov 09 15:24:59 <Zero_Chaos> nbd: https://lists.openwrt.org/pipermail/openwrt-devel/2020-October/031781.html
Nov 09 17:17:34 >KGB-0< https://tests.reproducible-builds.org/openwrt/openwrt_kirkwood.html has been updated. (100.0% images and 98.8% packages reproducible in our current test framework.)
Nov 09 17:34:30 <jow> ynezz: I don't understand your mail
Nov 09 17:35:48 <ynezz> :)
Nov 09 17:36:11 <ynezz> jow: https://gitlab.com/openwrt/docker/-/jobs/837338425#L229
Nov 09 17:36:33 <ynezz> fakeroot segfaults
Nov 09 17:36:52 <jow> and how is that related to my change? did it work before?
Nov 09 17:36:57 <ynezz> yes
Nov 09 17:37:22 <ynezz> that CI job is run every day
Nov 09 17:38:40 <ynezz> last succesfull build https://gitlab.com/openwrt/docker/-/pipelines/209536830 was on 30.10. at 5:28am
Nov 09 17:40:27 <jow> okay, well feel free to revert the change
Nov 09 17:41:19 <ynezz> I'm not sure if that's the correct fix
Nov 09 17:41:51 <jow> probably not because before it uses hardcoded paths
Nov 09 17:45:20 <jow> ynezz: what is the actual command sequence to reproduce?
Nov 09 17:45:26 <jow> that "uname" reproducer is invalid
Nov 09 17:45:43 <jow> as in, that it is expected that invoking host binaries with bundled libs crashes
Nov 09 17:46:04 <ynezz> it happens with fakeroot
Nov 09 17:46:17 <ynezz> so perhaps faked?
Nov 09 17:46:40 <ynezz> /home/build/openwrt/staging_dir/host/bin/fakeroot: line 185:  5536 Segmentation fault      (core dumped)
Nov 09 17:46:44 <jow> I mena if I spawn an interactive shell in that container I am root
Nov 09 17:46:57 <jow> what do I need to do to make it compile busybox and crash?
Nov 09 17:47:12 <jow> somehow doing sudo?
Nov 09 17:48:24 <ynezz> it works even without container
Nov 09 17:48:31 <ynezz> just download the latest SDK
Nov 09 17:48:58 <ynezz> ./scripts/feeds update base; ./scripts/feeds install libubox; make defconfig; make package/libubox/compile
Nov 09 17:49:24 <jow> thanks
Nov 09 17:50:53 <ynezz> docker run --rm -it registry.gitlab.com/openwrt/docker:sdk-snapshot-master /bin/sh -c "./scripts/feeds update base; ./scripts/feeds install libubox; make defconfig; make package/libubox/compile V=sc"
Nov 09 17:54:54 <jow> ynezz: yeah, so the addition of the LD_LIBRARY_PATH seems to be the culprit
Nov 09 17:56:48 <jow> and before that change it worked by accident because the expaneded @libdir@ points to /builder/shared-workdir/build/staging_dir/host/lib which normally does not exist on systems of SDK users
Nov 09 17:57:19 <jow> after my change it is built relative to STAGING_DIR and starts to fail
Nov 09 18:01:30 <novski__> i got this file from @PaulFertser https://paste.debian.net/1170556/ but this does not work:
Nov 09 18:01:30 <novski__> ```root@openwrt:~/openwrt# patch -p1 < sound.patch
Nov 09 18:01:30 <novski__> can't find file to patch at input line 5
Nov 09 18:01:30 <novski__> Perhaps you used the wrong -p or --strip option?
Nov 09 18:01:30 <novski__> The text leading up to this was:
Nov 09 18:01:31 <novski__> --------------------------
Nov 09 18:01:31 <novski__> |diff --git a/target/linux/ramips/dts/mt7628an_mediatek_linkit-smart-7688.dts b/target/linux/ramips/dts/mt7628an_mediatek_linkit-smart-7688.dts
Nov 09 18:01:32 <novski__> |index cbec8c9d40..b9f254244d 100644
Nov 09 18:01:32 <novski__> |--- a/target/linux/ramips/dts/mt7628an_mediatek_linkit-smart-7688.dts
Nov 09 18:01:33 <novski__> |+++ b/target/linux/ramips/dts/mt7628an_mediatek_linkit-smart-7688.dts
Nov 09 18:01:33 <novski__> --------------------------
Nov 09 18:01:34 <novski__> File to patch:
Nov 09 18:01:34 <novski__> ```
Nov 09 18:01:35 <novski__> What am i doing wrong?
Nov 09 18:21:14 <PaulFertser> novski__: please do not paste that many lines to a public channel
Nov 09 18:21:48 <novski__> whats the reason?
Nov 09 18:22:23 <stintel> there's pastebin for that
Nov 09 18:22:38 <PaulFertser> novski__: spamming the channel. When pasting many lines essential conversation goes up too fast.
Nov 09 18:24:17 <PaulFertser> novski__: see pm
Nov 09 18:48:50 <jow> ynezz: can you CI test a change without me pushing it first?
Nov 09 19:22:20 <aparcar[m]> jow: if you talk about the docker CI, it just downloads an pre-build SDK. You could run the CI on your own fork and change $FILE_HOST to point to a self build SDK. Looking at the code it would still require a valid usign or gpg signature
Nov 09 19:23:15 <aparcar[m]> do you have a fix in place? I can try it with the method suggested by ynezz
Nov 09 20:42:34 <ynezz> jow: yes, I would simply patch manually the current broken SDK and see if it helps
Nov 09 20:43:19 <ynezz> but it depends on the content of the fix itself
Nov 09 21:51:27 <jow> ynezz: https://git.openwrt.org/?p=openwrt/staging/jow.git;a=commit;h=2d1b3b554b8e0c1bf318a8f652d3aeaa2fc1653e
Nov 09 21:51:38 <jow> this solves the issue for me (was able to reproduce as well)
Nov 09 22:23:37 <thagabe> I've dumped the rootfs (jffs2) for my router and now im trying to find the wifi's firmware (which was updated) anyone know what directory it is located? (not on /lib/firmware) This is a Linksys router
Nov 09 22:30:47 <thagabe> nevermind, jefferson needed cstuct 1.0 to work
Nov 10 00:23:38 <Grommish> What is the best way to run a complex shell command from a Makefile?  $(shell xxx)?
**** ENDING LOGGING AT Tue Nov 10 00:42:46 2020

**** BEGIN LOGGING AT Tue Nov 10 00:48:06 2020
Nov 10 01:50:08 <aparcar[m]> Grommish: I'd call a script
Nov 10 01:50:47 <aparcar[m]> https://github.com/openwrt/openwrt/blob/master/include/package-ipkg.mk#L263
Nov 10 01:50:57 <Grommish> aparcar[m]: https://forum.openwrt.org/t/pkg-devs-complex-shell-commands-from-makefile/78858 to example of what I'm needing to do
Nov 10 01:51:52 <Grommish> My biggest issue right now is that it takes 2 hours to compile again once it fails at untar'ing everything
Nov 10 01:52:24 <Grommish> But yes, I'll look at the script way
Nov 10 01:54:28 <Grommish> aparcar[m]: Thanks :)
Nov 10 02:29:28 >KGB-0< https://tests.reproducible-builds.org/openwrt/openwrt_ar71xx.html has been updated. (99.1% images and 100.0% packages reproducible in our current test framework.)
**** ENDING LOGGING AT Tue Nov 10 02:59:56 2020