**** BEGIN LOGGING AT Thu Sep 17 02:59:57 2020 **** BEGIN LOGGING AT Thu Sep 17 03:36:02 2020 Sep 17 04:26:05 Anyone know exactly what the make[2]: Nothing to be done for 'compile'. error means? Even if I override and define Build/Compile, it tells me that and I'm very confused by it Sep 17 04:26:43 it's a standard make/make install (PKG_INSTALL), and for whatever reason it isn't seeing it Sep 17 04:27:02 probably the dependencies haven't changed, so if you have blah.c, and then blah.o is newer, Make won't compile... Sep 17 04:27:18 I call a {clean, prepare,configure,compile} though Sep 17 04:27:30 I'll just kill all of it and try Sep 17 04:27:46 I'll try removing tmp/ Sep 17 04:28:20 I want to get colm and ragel installed, and then hyperscan Sep 17 04:28:31 it's being.. troublesome **** BEGIN LOGGING AT Thu Sep 17 04:54:00 2020 **** BEGIN LOGGING AT Thu Sep 17 05:04:53 2020 Sep 17 05:10:28 jow: morning, would you take a look at https://patchwork.ozlabs.org/project/openwrt/patch/20200915180511.22491-1-zajec5@gmail.com/ in a free moment? [PATCH] ubus: add ACL support for "subscribe" request **** BEGIN LOGGING AT Thu Sep 17 05:26:42 2020 **** BEGIN LOGGING AT Thu Sep 17 05:33:02 2020 **** BEGIN LOGGING AT Thu Sep 17 05:39:32 2020 Sep 17 05:57:24 mangix: ping Sep 17 07:49:36 aparcar[m]: pong Sep 17 07:53:41 mangix: do you have an idea why the installation of mpd fails? Sep 17 07:55:24 blogic: don't want to be a bore but 4488b260a02e169d6902df44debbe9b3c8e001cc contains no flashing instructions :( Sep 17 07:56:30 aparcar[m]: from what I see, the buildbot is trying to install conflicting packages Sep 17 07:56:40 libavahi-dbus and libavahi-nodbus Sep 17 07:56:52 same with pulseaudio and its avahi variant Sep 17 07:57:08 actually...let's see if I can select them Sep 17 07:57:20 I can Sep 17 07:57:24 that's a bug Sep 17 07:58:10 A bug in the CI? Sep 17 07:58:16 no, Makefile Sep 17 07:58:32 So CI found a bug? Sep 17 07:59:16 🍿 Sep 17 07:59:17 yep Sep 17 07:59:28 I'll mark that in my calendar Sep 17 07:59:37 lol Sep 17 08:02:30 I tried comparing the Makefile to one that works properly. I can't find the problem. Sep 17 08:07:01 mangix: sorry no idea, I just do bug detection Sep 17 08:19:27 rmilecki: LGTM Sep 17 08:19:35 jow: thanks Sep 17 08:20:11 I think I see what's happening. the avahi Makefile creates a libavahi package, which is totally bogus Sep 17 08:21:01 aparcar[m]: this problem is also present with dnsmasq **** BEGIN LOGGING AT Thu Sep 17 08:23:08 2020 Sep 17 08:23:49 you can select both dnsmasq and dnsmasq-full Sep 17 08:23:58 both cannot be installed Sep 17 08:25:31 hmmm there's nothing that relies on normal libavahi Sep 17 08:39:13 jow: Yay, thanks for the merge! :D Sep 17 08:40:50 jow: do you have any idea how development of new API works? i was trying to follow https://github.com/whatwg/html/issues/2177 Sep 17 08:41:21 jow: it seems devs don't have interest in spending time for adding API for setting custom headers (i.e. "Authorization") Sep 17 08:41:40 jow: i was wondering about developing patch for Fx or Chromium for that Sep 17 08:42:12 standard guys seems to not willing to document that because of lacking experimental implementation Sep 17 08:42:38 is that the right way? try to add experimental support for a browser, then have documentation guys document it? **** BEGIN LOGGING AT Thu Sep 17 08:51:23 2020 Sep 17 09:20:12 rmilecki: I think the development works like this: if it comes from google and is in chrome, then it ends up in the standard, else its shrugged away Sep 17 09:20:40 interesting ;) Sep 17 09:22:00 that's laregly my understanding of it as well. Sep 17 09:22:21 also, did you see that comment from a chrome dev about scale of usage? Sep 17 09:22:48 that's their metric, and that means that if it's not being used on facebook/twitter/gmail et al, it's _absolutely irrelevant_ Sep 17 09:26:57 lovely isn't it, that 'do no evil' company Sep 17 09:27:16 and thats why browser monopolies are bad Sep 17 09:27:21 but chrome is the new MSIE 6 Sep 17 09:27:42 not sure if there'll be another mozilla down the road Sep 17 09:28:27 just look at the https + selfsigned cert mess, that affects a way larger audience compared to credentials in eventsources Sep 17 09:28:37 it also is simply shrugged away Sep 17 09:30:08 yeah i really wonder how long firefox is gonna stick around the way mozilla is looking for money and axing jobs and side projects Sep 17 09:34:34 fighting self signed is part of their active "you must host all services with us, do not think you can do it yourself" worldview too though, so they're very against that. Sep 17 09:35:14 http is still far better user experience than self signed https. Sep 17 09:35:34 Google is working on deprecating it :) Sep 17 09:35:45 even on their roadmaps they admit that http can't be killed. Sep 17 09:35:59 you'll get an "insecure" icon in the url bar,but not the yellow full screen horror show Sep 17 09:53:45 nbd: https://bugs.openwrt.org/index.php?do=details&task_id=3343 Sep 17 09:57:28 jow: thanks, i'll take care of this **** BEGIN LOGGING AT Thu Sep 17 10:14:30 2020 Sep 17 10:15:11 rsalvaterra: I don't understand e9hack's comment about the zram commit Sep 17 10:15:18 so shall I revert it or not? Sep 17 10:15:43 I already regret touching it Sep 17 10:16:28 jow: Let me explain. Sep 17 10:19:12 The lzo compressor generates two modules (lzo and lzo-rle). I wasn't aware of this. My patch works around the issue by defaulting to lzo instead of lzo-rle, which isn't wrong, but it's only half of the fix. Sep 17 10:19:49 So, the full fix means appying e9hack's patch too. Sep 17 10:19:54 *applying Sep 17 10:20:35 What doesn't make sense (as I replied to e9hack) is the crypto layer advertising lzo-rle support without the module being present. Sep 17 10:21:54 Note that making zram default to lzo means it will always work, since lzo is a strict dependency of zram. We can be sure that it's always present. Sep 17 10:23:01 Feeling less regretful now? :) Sep 17 10:23:24 yes, sounds good. Also matches LuCI's behaviour Sep 17 10:28:43 blogic: hey, we have in OpenWrt that brcmfmac patch you accidentally pushed and planned to upstream Sep 17 10:28:59 blogic: can you try sending it, please? **** BEGIN LOGGING AT Thu Sep 17 11:52:03 2020 **** BEGIN LOGGING AT Thu Sep 17 12:13:31 2020