**** BEGIN LOGGING AT Thu Mar 25 02:59:56 2021
Mar 25 04:35:58 <mangix> fascinating.
Mar 25 04:36:08 <mangix> 2GB is not enough for my NAS anymore
Mar 25 04:46:11 <mangix> wow using a usb flash drive as swap makes a huge difference
Mar 25 05:09:26 <aparcar[m]> OutBackDingo: sure
Mar 25 05:09:47 <aparcar[m]> mangix: ideas on how I fix this CI depended issue?
Mar 25 05:10:24 <mangix> aparcar[m]: remove it </sarcasm>
Mar 25 05:10:44 <mangix> there should be a way
Mar 25 05:16:35 <aparcar[m]> mangix: mind thinking a bit about it?
Mar 25 05:26:37 <mangix> hmm no idea how to map a package name to a directory entry
Mar 25 05:27:17 <mangix> doing a git grep and matching based on that seems error prone.
Mar 25 05:33:45 <aparcar[m]> yup also because e.g. vim contains xxd which is hart to match
Mar 25 05:37:14 <mangix> I think were's gonna need jow's help on this. I don't know enough about the build system
Mar 25 06:51:11 <owrt-snap-builds> Build [#15](https://buildbot.openwrt.org/master/images/#builders/64/builds/15) of `realtek/generic` failed.
Mar 25 06:58:25 <OutBackDingo> aparcar[m]: openwrt c openssl
Mar 25 07:13:34 <mangix> OutBackDingo: you tried mailing list or forum?
Mar 25 07:16:55 <OutBackDingo> mangix: well... no
Mar 25 07:17:07 <OutBackDingo> figured irec would be faster
Mar 25 07:18:08 <mangix> nah IRC is usually dead
Mar 25 07:45:36 <owrt-snap-builds> Build [#28](https://buildbot.openwrt.org/master/images/#builders/26/builds/28) of `apm821xx/sata` failed.
Mar 25 08:21:56 <rsalvaterra> mangix: zram, no?
Mar 25 08:23:20 <rsalvaterra> mangix: zram? :)
Mar 25 08:25:25 <mangix> rsalvaterra: already enabled
Mar 25 08:25:52 <mangix> doesn't really help. I added my super fast USB flash drive as swap
Mar 25 08:25:56 <mangix> solved all my problems
Mar 25 08:26:53 <rsalvaterra> Hm… so you're really pushing it, then. What's the filesystem?
Mar 25 08:27:11 <mangix> what the hell. when i lost internet it seems my switch failed.
Mar 25 08:27:20 <mangix> [68198.102160] TCP: request_sock_TCP: Possible SYN flooding on port 5000. Sending cookies.  Check SNMP counters.
Mar 25 08:27:41 <mangix> rsalvaterra: fs for the swap?
Mar 25 08:27:55 <rsalvaterra> No, for the NAS.
Mar 25 08:28:41 <mangix> the OS is installed on an sd card running ext4. The hard drives are all on a btrfs volume. The USB flash drive is using a swap partition
Mar 25 08:29:33 <mangix> why would that matter?
Mar 25 08:29:33 <rsalvaterra> You said 2 GiB weren't enough RAM for you NAS anymore, right? :)
Mar 25 08:29:58 <rsalvaterra> And you're using btrfs, so that explains it.
Mar 25 08:30:00 <mangix> yeah. I have several docker containers running while compiling rust programs
Mar 25 08:30:19 <rsalvaterra> Oh. That's not a NAS, then. :P
Mar 25 08:30:23 <mangix> also, node is memory hungry like chrome
Mar 25 08:31:02 <mangix> anyone know what's up here? https://gist.github.com/neheb/1b15bf88e390496e8beb74b7645add90
Mar 25 08:31:16 <mangix> I have never had this happen.
Mar 25 08:32:23 <mangix> rsalvaterra: I mean, most of the containers are dedicated to multimedia stuff
Mar 25 08:32:56 <rsalvaterra> You mean just the flood, or the links going up/down too?
Mar 25 08:32:56 <mangix> both
Mar 25 08:33:16 <rsalvaterra> Never seen that happen, to be honest. :/
Mar 25 08:35:43 <mangix> I lost internet at that time. my NAS output: https://gist.github.com/neheb/1729842c7a4b660bf37c2bbe080a69e9
Mar 25 08:36:10 <mangix> which is uninteresting
Mar 25 08:37:43 <rsalvaterra> Speaking of storage, I'm now doing my builds on this crazy thing: https://www.photofast.com/home-2/products/ssd-expandable/CR9000/
Mar 25 08:38:08 <rsalvaterra> I've had it for years, but never used it. :P
Mar 25 08:40:13 <mangix> oh god that looks freaky
Mar 25 08:40:46 <rsalvaterra> I can work with "freaky", I thought you'd say "insane". :P
Mar 25 08:41:37 <mangix> even worse version: https://www.amazon.com/1Pack-Micro-Memory-Adapter-Converter/dp/B01MU59ACL/ref=cm_cr_arp_d_bdcrb_top?ie=UTF8
Mar 25 08:42:23 <mangix> the question I have is who needs this?
Mar 25 08:42:28 <rsalvaterra> When you fill it up with 32 GiB A1 cards, it's surprisingly usable. Of course, when it came out (over 10 years ago), there were no A1 cards.
Mar 25 08:43:12 <mangix> can the controller even handle A1?
Mar 25 08:43:45 <rsalvaterra> mangix: Nobody needs that. I got this CR9000 because it was dirt-cheap and seemed so crazy to me I just had to try it.
Mar 25 08:44:01 <mangix> haha
Mar 25 08:44:39 <mangix> I recently spent way too much money buying two double sided USB adapters
Mar 25 08:44:52 <mangix> I hope I'll have a use for them
Mar 25 08:46:56 <rsalvaterra> So, I filled it up with 32 GiB A1 cards, put the thing in a UAS USB 3.1 enclosure and copied over my tree. It's working just fine. :)
Mar 25 08:48:19 <rsalvaterra> However, I formatted it first in f2fs and it ate my data. Reformatted in ext4 and it's all good.
Mar 25 08:49:09 <mangix> how did it eat your data?
Mar 25 08:49:27 <mangix> f2fs is not a journaled fs AFAIK
Mar 25 08:50:03 <rsalvaterra> mangix: Unmounted, rebooted, data lost. Weird.
Mar 25 08:51:02 <mangix> strange
Mar 25 08:58:24 <mangix> hmm running 5.4.94
Mar 25 08:58:29 <mangix> wonder if i should update
Mar 25 08:58:55 <rsalvaterra> mangix: Linux mallard 5.12.0-rc4+ #31 SMP Wed Mar 24 13:08:04 WET 2021 x86_64 x86_64 x86_64 GNU/Linux
Mar 25 08:59:29 <rsalvaterra> 5.4 is a bit long in the tooth… :P
Mar 25 09:00:10 <mangix> it's on my router
Mar 25 09:00:40 <mangix> I think you meant 4.4
Mar 25 09:00:47 <mangix> that's super LTS
Mar 25 09:01:25 <rsalvaterra> Ah, OpenWrt. In that case, sure, bump it to 5.10. What router is it?
Mar 25 09:01:53 <rsalvaterra> With an Atheros switch, probably ath79, no?
Mar 25 09:02:09 <mangix> yeah it is
Mar 25 09:02:49 <rsalvaterra> I have 5.10 on the TL-WDR3600, Archer C6 v2 and AirGrid M2, without noticeable issues.
Mar 25 09:08:08 <ynezz> nbd: can I get your Reviewed-by for https://patchwork.ozlabs.org/project/openwrt/patch/20210324122715.149933-1-arnout@mind.be/ (ubusd: convert tx_queue to linked list) ? I've tested it extensively and I am happy with it so I would merge it, but it changes current ubusd behaviour so second pair of eyes would be welcome
Mar 25 09:11:29 <ynezz> 21.02.0-rc1 seems to be around corner so it would be nice to have that fix in there as well
Mar 25 09:12:17 <nbd> will take a look
Mar 25 09:30:05 <mangix> 104 packages left. whoo
Mar 25 09:31:48 <aparcar[m]> mangix: for what? Ninja?
Mar 25 09:32:15 <aparcar[m]> OutBackDingo: try mailing list first
Mar 25 09:33:02 <mangix> aparcar[m]: yep
Mar 25 09:44:22 <Tapper> Hi people is kernel 5.10 ready for testing on the r7800 yet?
Mar 25 09:44:46 <mangix> AFAIK no
Mar 25 09:45:02 <Tapper> OK thanks
Mar 25 09:45:40 <Tapper> mangix BTW I fixt my ubuntu on wsl. I removed it and started from scratch.
Mar 25 09:46:10 <Tapper> So I can do builds again now. all is good
Mar 25 09:47:12 <mangix> hrm this is strange. make is not producing a sysupgrade file. i'll look into it later.
Mar 25 09:51:42 <nbd> ynezz: sent my review
Mar 25 09:56:31 <ynezz> nbd: cool, thanks!
Mar 25 10:00:30 <rsalvaterra> mangix: I found the reason why I had data loss with f2fs (because I just had right now with ext4). USB auto-f***ing-suspend. *facepalm*
Mar 25 10:40:44 <karlp> your usb device shouldn't advertise supporting it if it doesn't....
Mar 25 10:40:59 <karlp> but hey, reality is allthat matters...
Mar 25 10:49:41 <ldir> rsalvaterra: https://git.openwrt.org/?p=openwrt/staging/ldir.git;a=commit;h=3fc007aaa1d8654673bad513fed1be88a0b7eed8
Mar 25 11:11:31 <rsalvaterra> karlp: Crappy JMicron controller…
Mar 25 11:12:00 <rsalvaterra> ldir: Bah. We need to organise ourselves. I did it already too. :P
Mar 25 11:12:28 <ldir> oh I haven't pushed it, but I have built it - and just about to press it into service.
Mar 25 11:12:40 <rsalvaterra> Likewise! :)
Mar 25 11:13:42 <ldir> it's a lot quicker at the moment 'cos not many targets at 5.10
Mar 25 11:14:23 <ldir> a 5.4 refresh is a much longer wait being essentially single threaded.
Mar 25 11:15:11 <rsalvaterra> ldir: I moved from the Pentium D 950 to an i7-4770R. Things are speedier now.
Mar 25 11:16:36 <ldir> imagine if it could be paralleled
Mar 25 11:16:43 <rsalvaterra> By the way, can you tell me a bit more about the your dnsmasq DSCP patch? Do you notice any difference in response times?
Mar 25 11:21:39 <ldir> ha - short answer, no.  It's there purely to exercise CAKE's tins - it scratches a personal itch of 'dns is important'
Mar 25 11:22:15 <rsalvaterra> Placebo, then. :)
Mar 25 11:22:30 <rsalvaterra> But I see a point in it, yes.
Mar 25 11:23:13 <ldir> brb
Mar 25 14:23:35 <owrt-1907-builds> Build [#6](https://buildbot.openwrt.org/openwrt-19.07/images/#builders/29/builds/6) of `oxnas/ox820` failed.
Mar 25 14:31:23 <ldir> openssl 1.1.1k - hmmmmmmm
Mar 25 14:32:09 <ldir> https://www.openssl.org/news/secadv/20210325.txt - hmmmmm
Mar 25 14:34:59 <olmari> There is always something (:
Mar 25 14:36:28 <rsalvaterra> Oh, c'est le CVE du jour, n'est-ce pas? :P
Mar 25 15:33:53 <Hauke> ldir: thanks
Mar 25 15:34:12 <Hauke> ustream-ssl and hostapd should not be affected by the first problem
Mar 25 15:34:25 <Hauke> I haven't seen any usage of X509_V_FLAG_X509_STRICT
Mar 25 15:35:03 <Hauke> could be that it is possile to crash ustream-ssl server with "NULL pointer deref in signature_algorithms processing (CVE-2021-3449)"
Mar 25 15:35:10 <Hauke> or even hostapd in EAP mode
Mar 25 15:35:42 <Hauke> I do not think they are a big problem
Mar 25 15:52:22 <plntyk> hm... I wonder if https://git.openwrt.org/?p=openwrt/openwrt.git;a=commit;h=af22991e03cae55f96b06996df2ff16752cec5d5 somehow broke valgrind compile
Mar 25 15:53:16 <plntyk> or if it was broken before ...
Mar 25 16:41:08 <ynezz> aparcar[m]: BTW do you know if it's possible to have browsable artifacts on GH?
Mar 25 16:41:16 <ynezz> aparcar[m]: something like this https://ynezz.gitlab.io/-/ucentral-wifi/-/jobs/1128365038/artifacts/build/scan/2021-03-25-163845-70-1/index.html
Mar 25 16:41:53 <ynezz> aparcar[m]: it's seems like I can only download the complete tarball https://github.com/ynezz/ucentral-wifi/suites/2344614205/artifacts/49512203
Mar 25 16:45:40 <dorf> you should be able to browse artifacts, ynezz
Mar 25 16:47:09 <dorf> pipelines -> select pipeline -> select build job -> hit browse
Mar 25 16:48:07 <dorf> that's from the gitlab.com UI, not gitlab.io
Mar 25 16:48:30 <ynezz> dorf: GH -> GitHub
Mar 25 16:48:50 <ynezz> it works fine on GitLab, just would like to have same feature on GitHub
Mar 25 16:49:05 <dorf> ah, right.
Mar 25 16:49:41 <ynezz> bummer https://github.com/actions/upload-artifact#zipped-artifact-downloads -> There is currently no way to download artifacts after a workflow run finishes in a format other than a zip or to download artifact contents individually.
Mar 25 16:50:15 <dorf> I do all my CI stuff on gitlab, don't know about github, but it should be possible to script something to place individual artifacts somewhere.
Mar 25 16:51:02 <ynezz> I prefer GL as well, but some folks and projects prefer GH
Mar 25 16:53:30 <dorf> well, if there are repos you're interested in on github that you'd like individual artifacts for, you could always import them into gitlab. there's that.
Mar 25 16:54:17 <dorf> possibly more hassle than it's worth, though, if all you're after is an index.html :)
Mar 25 16:54:37 <ynezz> in this case it's output from static code analyzer
Mar 25 16:55:10 <ynezz> actually pretty nice stuff as you can browse the codepaths and see the notes
Mar 25 16:55:11 <dorf> do you have privs on the repo in question?
Mar 25 16:55:34 <ynezz> in this example they're my repos
Mar 25 16:57:03 <dorf> I just wonder if you can't do a cp /www/ during the job run.
Mar 25 16:57:19 <dorf> or wherever you're hosting your .io files.
Mar 25 16:59:19 <ynezz> you don't need to do that, GL does it behind the scenes
Mar 25 17:00:09 <dorf> *GH :)
Mar 25 17:00:21 <ynezz> there is no such feature AFAIK
Mar 25 17:00:53 <dorf> oh well, time to migrate to GL then if they're your projects. unchain yourself!
Mar 25 17:00:56 <ynezz> you would probably need to use some custom storage like S3, GCS etc.
Mar 25 17:01:31 <dorf> feature? you mean storage for a website?
Mar 25 17:02:05 <dorf> like https://vituperative.github.io/i2pchat/ for example?
Mar 25 17:03:13 <ynezz> feature as in already available GH actions plugin
Mar 25 17:06:28 <ynezz> maybe there is peaceiris/actions-gh-pages
Mar 25 17:06:56 <dorf> yeah, I'd be surprised if someone hasn't figured it out tbh.
Mar 25 17:07:18 <dorf> seems like such an obvious feature.
Mar 25 17:14:34 <philipp64> Anyone around that (like Paul) that understands the CI/CD stuff?  I have a PR for strongswan (actually, several but I won't go into that) and the build checks are failing...  but for a completely incongruous reason.  https://github.com/openwrt/packages/pull/14711/checks?check_run_id=2189418878
Mar 25 17:15:40 <philipp64> pigeonhole fails to build, but pigeonhole is a standalone package! nothing depends on it! least of all strongswan!
Mar 25 17:15:53 <philipp64> WTAH is going on?
Mar 25 17:17:11 <ynezz> broken AI?
Mar 25 17:23:28 <ynezz> anyway to have more complete picture you need to read https://github.com/openwrt/packages/blob/master/.github/workflows/multi-arch-test-build.yml and https://github.com/openwrt/gh-action-sdk/blob/master/entrypoint.sh
Mar 25 17:24:54 <ynezz> I assume that logic somehow (wrongly?) computes that pigeonhole package as changed https://github.com/openwrt/packages/blob/master/.github/workflows/multi-arch-test-build.yml#L61
Mar 25 17:25:01 <ynezz> so that's why it's building it
Mar 25 17:34:45 <philipp64> ynezz: that's exactly right. look here... https://paste.ubuntu.com/p/CMtMrQGMcD/
Mar 25 17:40:45 <aparcar[m]> philipp64: please rebase your PR. Circleci is no longer used and was disabled
Mar 25 17:41:43 <stintel> philipp64: while at it, feel free to make a PR to take over maintainership of strongSwan from me. I don't have the energy to deal with the PR storm
**** BEGIN LOGGING AT Thu Mar 25 17:42:35 2021
Mar 25 17:48:32 <stintel> philipp64: reposting in case it was lost due to netsplit: while at it, feel free to make a PR to take over maintainership of strongSwan from me. I don't have the energy to deal with the PR storm
Mar 25 17:49:02 <stintel> blogic_: feel free to take over maintainership of lldpd and revert as you please
Mar 25 17:49:33 <philipp64> stintel: I'm not sure it's a "storm".  I'm happy to do co-ownership and deal with the UCI stuff...
Mar 25 17:51:53 <stintel> I'm counting 12 PRs related to strongswan in this year alone. it's supposed to be a hobby, but it's starting to make me feel bad, so I'm done with it
Mar 25 17:53:36 <philipp64> Some of them you asked to be restructured, so I did.  I've closed the obsolete ones, I thought... yeah, right now there are only 5 open.
Mar 25 17:53:52 <philipp64> let me see if I can recruit Noel.
Mar 25 18:01:03 <aparcar[m]> philipp64: I'm not familiar with strongswan but isn't wire guard kinda doing it's job?
Mar 25 18:02:27 <philipp64> there are times when Strongswan is what you need (compatibility with other systems, especially front-end systems that generate profiles, etc.). plus it's a known quantity when interoperating with Cisco IOS, MacOS, etc.
Mar 25 19:36:28 <philipp64> how do I invite someone (on github) to join the packagers community?
Mar 25 19:37:29 <aparcar[m]> philipp64: what kind of invitation?
Mar 25 19:37:29 <aparcar[m]> nbd_: ping
Mar 25 19:39:12 <philipp64> aparcar[m]: Noel Kuntze (Thermi) has agreed to co-own strongswan with me if stintel still wants to re-parent it.
Mar 25 19:39:51 <philipp64> so, wanted to bring him into the Contributors group.
Mar 25 19:39:51 <aparcar[m]> I'm assuming you want to give him commit access? Else that person can just create PRs like everyone lese
Mar 25 19:39:57 <aparcar[m]> *else
Mar 25 19:40:13 <philipp64> yes, as a co-owner, he'd be pulling the trigger on merges.
Mar 25 19:41:18 <aparcar[m]> I think the usual process is that somebody does a few quality contributions and then receives commit access on request
Mar 25 19:41:43 <aparcar[m]> but there are other people at packages.git, not so much here at IRC, who have better opinions on that
Mar 25 19:43:23 <philipp64> well, he's a regular contributor on strongswan/strongswan.git ...
Mar 25 19:43:31 <philipp64> if that matters.
Mar 25 19:44:10 <aparcar[m]> I'll open an issue for that because people keep asking for commit access and I don't have a reproducible answer to that
Mar 25 19:53:23 <aparcar[m]> philipp64: https://github.com/openwrt/packages/issues/15257
Mar 25 19:53:27 <aparcar[m]> mangix: -^
Mar 25 20:03:05 <philipp64> LGTM
Mar 25 20:03:55 <Thermi> Salut
Mar 25 20:04:03 <Thermi> philipp64: ^
Mar 25 20:05:02 <philipp64> aparcar[m]: This is Noel/Thermi... he's very active in the strongswan.git repo and I think a good downstream maintainer candidate for openwrt...
Mar 25 20:05:49 <Thermi> aparcar[m]: After Tobias and Andreas, who are the main developers left, I am doing most of the support
Mar 25 20:06:35 <Thermi> aparcar[m]: Using strongSwan is part of my work and I've been working with it since 2013, so I got "enough" experience. Previously, before it was packaged in the official repos, I was maintaining the strongSwan AUR package for Arch Linux.
Mar 25 20:07:00 <Thermi> aparcar[m]: I also work support contracts for other companies
Mar 25 20:09:09 <Thermi> I think that's about it.
Mar 25 20:09:37 <Thermi> Errr, okay, one more thing. I am working on getting some Windows related stuff upstream, but that's not relevant for OpenWRT.
Mar 25 20:10:48 <philipp64> ... and we won't hold it against you.
Mar 25 20:12:20 <philipp64> Thermi: how many of your customers use openwrt (that you know of)... keeping in mind that it's rebranded by Linksys and Netgear and TPLink and...
Mar 25 20:12:37 <Thermi> Paying customers 0
Mar 25 20:12:44 <Thermi> Well
Mar 25 20:12:45 <Thermi> 1
Mar 25 20:13:06 <Thermi> TPLink brand, some enterprise switch/router. But I don't know if it's openwrt.
Mar 25 20:13:18 <Thermi> People coming to me for help sometimes use OpenWRT.
Mar 25 20:13:30 <aparcar[m]> all right let's not get into more details why you're qualified because else other maintainers feel intimidated
Mar 25 20:13:38 <aparcar[m]> what's you github handle
Mar 25 20:13:41 <Thermi> Thermi
Mar 25 20:15:01 <aparcar[m]> 2FA enabled?
Mar 25 20:15:45 <aparcar[m]> philipp64: please create a PR which adds Thermi as maintainer. Once that's done I'll add him
Mar 25 20:15:51 <Thermi> Yes
Mar 25 20:16:08 <Thermi> And my front door is quite sturdy.
Mar 25 20:16:39 <aparcar[m]> Thanks that's usually my second question
Mar 25 20:17:58 <aparcar[m]> philipp64: actually please create a PR which adds you and Thermi as maintainer, removing stintel
Mar 25 20:21:19 <aparcar[m]> sorry for the extra steps, I'm just trying to step on nobodies toes here
Mar 25 20:27:49 <Pepe> 🤔
Mar 25 20:29:44 <aparcar[m]> Pepe: share your thoughts\
Mar 25 20:31:18 <Pepe> Yeah, I will do that in the PR, where you mentioned me, but I dont want to offend anyone here, but what happened during one hour it surprised me how it easy to get a commit access to packages repository when I am looking for example on jeffery/commodo who are maintainers of Python and they dont have commit access. In my opinion it is not fair, though.
Mar 25 20:34:57 <aparcar[m]> Pepe: yea that's my thought. I then asked around and got as response that "whoever is a maintainer in packages.git should have commit access"
Mar 25 20:35:45 <aparcar[m]> since stintel approved philipp64 to become maintainer and he declared Thermi as co-maintainer....
Mar 25 20:36:06 <aparcar[m]> But well, please respond to the issue I opened and start a discussion
Mar 25 20:36:32 <Pepe> Will do, I need to sumarise it and I should not use any hateful speech, right? :D
Mar 25 20:38:29 <aparcar[m]> Ideally
Mar 25 20:39:06 <zorun> whoever is a maintainer in packages.git does not have commit access
Mar 25 20:39:09 <zorun> it's easy to check
Mar 25 20:42:06 <aparcar[m]> zorun: I know but that doesn't mean they can't have it
Mar 25 20:42:38 <aparcar[m]> anyway, point is there is no clear rule and I'm happy to follow whatever people agree on
Mar 25 20:48:33 <dl12345> commit 512229ce4967814a09ce202855e6b632d2e97a3c for include/image.mk has a syntax error - a spurious close parenthesis at the end of the line
Mar 25 20:48:40 <dl12345> @@ -474,15 +476,49 @@ endef
Mar 25 20:48:41 <dl12345>  ifndef IB
Mar 25 20:48:41 <dl12345>  define Device/Build/initramfs
Mar 25 20:48:42 <dl12345>    $(call Device/Export,$(KDIR)/tmp/$$(KERNEL_INITRAMFS_IMAGE),$(1))
Mar 25 20:48:42 <dl12345> -  $$(_TARGET): $$(if $$(KERNEL_INITRAMFS),$(BIN_DIR)/$$(KERNEL_INITRAMFS_IMAGE))
Mar 25 20:48:43 <dl12345> +  $$(_TARGET): $$(if $$(KERNEL_INITRAMFS),$(BIN_DIR)/$$(KERNEL_INITRAMFS_IMAGE) \
Mar 25 20:48:43 <dl12345> +         $$(if $$(CONFIG_JSON_OVERVIEW_IMAGE_INFO), $(BUILD_DIR)/json_info_files/$$(KERNEL_INITRAMFS_IMAGE).json,))
Mar 25 20:48:54 <dl12345> ian@ubuntu:/opt/openwrt $ make download V-s
Mar 25 20:48:55 <dl12345> make[1]: Entering directory '/opt/openwrt/scripts/config'
Mar 25 20:48:55 <dl12345> make[1]: 'conf' is up to date.
Mar 25 20:48:56 <dl12345> make[1]: Leaving directory '/opt/openwrt/scripts/config'
Mar 25 20:48:56 <dl12345> make[1]: Entering directory '/opt/openwrt'
Mar 25 20:48:57 <dl12345> make[2]: Entering directory '/opt/openwrt'
Mar 25 20:48:57 <dl12345> make[3]: Entering directory '/opt/openwrt/target/linux'
Mar 25 20:48:58 <dl12345> make[4]: Entering directory '/opt/openwrt/target/linux/x86'
Mar 25 20:48:58 <dl12345> make[5]: Entering directory '/opt/openwrt/target/linux/x86/image'
Mar 25 20:48:59 <dl12345> Makefile:150: *** missing separator.  Stop.
Mar 25 22:16:26 <philipp64> doh!
Mar 25 22:58:24 <mangix> aparcar[m]: the kernel cannot be built with GCC7 anymore. Should be removed.
Mar 25 23:02:36 <aparcar[m]> mangix: tell me more
Mar 25 23:04:16 <mangix> https://gist.github.com/neheb/51367f34883ddca87a7c0ddefed55cb2
Mar 25 23:07:08 <mangix> hmmm something's wrong with my setup
Mar 25 23:10:14 <plntyk> mangix maybe your CC is clang ? https://curl.se/mail/lib-2012-06/0401.html
Mar 25 23:10:41 <plntyk> same "did you mean"
Mar 25 23:12:24 <mangix> plntyk: I removed it
Mar 25 23:12:40 <mangix> still errors
Mar 25 23:16:01 <plntyk> make --trace should show the make / CC cmdline iirc
Mar 25 23:43:14 <mangix> hmm builds after completely wiping build/staging_dir
Mar 26 01:36:30 <aparcar[m]> mangix: so we keep gcc7
Mar 26 01:36:30 <aparcar[m]> ?
Mar 26 01:41:18 <mangix> aparcar[m]: I mean, it should be removed eventually.
Mar 26 01:41:29 <mangix> GCC7 can't compile gerbera for example
Mar 26 01:42:16 <mangix> std::filesystem came with GCC8
Mar 26 01:43:11 <mangix> for an unrelated reason, I tried to move it to boost::filesystem but that's not equivalent surprisingly.
Mar 26 01:44:42 <aparcar[m]> gcc8 is default anyway right?
Mar 26 01:46:14 <mangix> yeah
Mar 26 02:08:00 <mangix> https://github.com/openwrt/openwrt/pull/4028 <-- surprising
Mar 26 02:11:31 <pkgadd> aside from the incorrect signed-off-by, there is no indication that this change really has an effect on the archer c2600. ipq806x is using cortex a15 since day one (so roughly 2016), it would (imho) need a whole lot more of a justification to change that after half a decade
Mar 26 02:16:00 <mangix> pkgadd: huh that's funny. look at target/linux/ipq806x/patches-5.4/0065-arm-override-compiler-flags.patch
Mar 26 02:16:00 <mangix> upstream kernel.org seems to disagree
Mar 26 02:20:02 <pkgadd> more verbose commit messages would certainly help, but that equally (respectively even more-) applies to that new PR
Mar 26 02:24:07 <pkgadd> it would be interesting what lineageos is doing as well
Mar 26 02:40:38 <mangix> pkgadd: fewer toolchains would be nice
Mar 26 02:40:48 <mangix> even if there is no real problem with a15
Mar 26 02:45:15 <pkgadd> mangix: armvirt is also using cortex-a15
Mar 26 02:49:06 <pkgadd> sure, reducing the set of architectures would always be nice, but there should still be a justification in the commit message - that may very well be that cortex-a7 works well enough, but not that it "might eventually" fix the c2600 without clarifying if it really does
Mar 26 02:53:07 <pkgadd> (I wouldn't assume that there'd be a significant performance difference for typical routing tasks either)
Mar 26 02:54:01 <mangix> pkgadd: that armvirt seems to be wrong
Mar 26 02:54:15 <mangix> given that it is not true with most devices
Mar 26 02:56:03 <mangix> i would think qemu targets should be using something close to hardware
Mar 26 02:58:16 <pkgadd> well, so far both targets have the justification that there's the other one also using cortex-a15 ;)  I've never used armvirt and have no idea if using cortex-a15 might be faster there
**** ENDING LOGGING AT Fri Mar 26 02:59:56 2021