**** BEGIN LOGGING AT Tue Nov 05 02:59:58 2019 Nov 05 07:26:57 rr123: the minus sign idnciates whether whitespace before the start or after the end tag should be stripped Nov 05 07:44:02 jow: thanks for the review Nov 05 07:54:29 Tapper: ping Nov 05 07:55:14 jow pong Nov 05 07:55:50 jow: pretty success the whole weblate thing, I'm surprised so many people help out. Is the entire luci base now javascript? Nov 05 08:00:44 aparcar[m]: yes, quite a success. Nov 05 08:01:12 aparcar[m]: yeah, most of the base (luci-mod-status, luci-mod-network, luci-mod-system, luci-app-firewall and luci-app-opkg) use clietn side CBI views now Nov 05 08:01:29 only the routing and menu building happens on the server side yet, but that'll change soon as well Nov 05 08:01:57 I expect that luci base could be completely lua free in two to three months Nov 05 08:01:57 should we put some more power on the "scenario" based configuration? Nov 05 08:02:45 is that your goal for 19.07? Nov 05 08:03:10 Tapper: a while back you told me that the modal dialogs added in LuCI were not ideal for your screen reader, e.g. when you edit wireless settings, you need to jump to the end of the page to configure it etc. Nov 05 08:03:23 Tapper: would it be better to open the modal at the beginning of the page? Nov 05 08:04:27 aparcar[m]: yes on scenario conf, no on full js for 19.07. For 19.07 my goal was to to align the branch with master, to simplify support Nov 05 08:05:42 The thing is that after clicking on the button to change the settings the dialog does not grab the focus of the screen reader so unless you go looking around the page you would knot know that the dialog is there. Nov 05 08:05:57 hm, I understand Nov 05 08:06:27 What I do now is use controle down arrow to drop strate down to the dialog and then press enter then I can change what I need to. Nov 05 08:07:18 so you think it would be enough to focus on the first element of the dialog after it has been opened? Nov 05 08:07:44 and must that element be an input or form element or can it be any element, such as the dialog headline? Nov 05 08:09:10 jow: do you have something written up about this "scenario" based configuration? (even just an email) Nov 05 08:10:43 I will soon start mentoring a group of students to work on a more intuitive web interface Nov 05 08:11:05 jow I think it mite be ok being the dialog headline. Nov 05 08:11:50 I'm still hesitating on making them start a new project (easier but less maintainable) or integrate in an existing project (which raises the initial cost) Nov 05 08:12:15 zorun: one sec Nov 05 08:12:59 zorun: https://forum.openwrt.org/t/common-use-case-for-simplified-web-interface/38866 Nov 05 08:13:03 jow did you see this? https://www.w3.org/TR/wai-aria-practices/examples/dialog-modal/dialog.html Nov 05 08:13:13 jow: I'm not quite sure if this is a generic issue with luci-theme-material or specific to luci-app-wireguard, but the new'ish floating dialogue style configuration panes of luci-app-wireguard show a weird behaviour starting with the fourth peer config (the background of the overlay stopping earlier than its contents) Nov 05 08:13:26 jow: http://sl-h.de/tmp/Screenshot_2019-11-05%20nbg6817%20-%20Interfaces%20-%20LuCI.png that is firefox 72 on Debian/ unstable Nov 05 08:14:54 My thinking behind not dropping focus to a input field is when some one is using a keyboard they mite change that field without meaning to. Nov 05 08:15:22 that's a ~two days old snapshot, r11392-66ebca7103 with luci git-19.306.71062-aa5151e (feel free to disregard, if that's specific to wireguard) Nov 05 08:15:31 zorun: so is your plan to write something from scratch? if not I'd be interesting to mock up some rpcd backend stuff and give it a nice interface Nov 05 08:16:23 I do see the same visual behaviour on chromium 76.0.3809.100 (Debian/ unstable as well) Nov 05 08:24:47 Tapper: yeah, indeed makes sense. I'll add some code to auto-focus the dialog headline of the modal Nov 05 08:25:29 pkgadd: the screenshot looks like bootstrap ? Nov 05 08:25:56 jow: sorry, yes, of course luci-theme-bootstrap - git-19.306.71062-aa5151e-1 Nov 05 08:38:52 pkgadd: I remember that bug. There's some max-height set for the modal somewhere, will fix it Nov 05 08:39:06 jow: thanks, great Nov 05 08:46:33 aparcar[m]: they will only work on the frontend, I was planning to use uhttpd-mod-ubus for the backend Nov 05 08:46:39 https://openwrt.org/docs/techref/ubus#access_to_ubus_over_http Nov 05 08:47:05 zorun: should be fine too, they can fire multiple calls at once Nov 05 08:47:47 if I get it right, rpcd is a layer on top of ubus to make more complicated actions? Nov 05 08:48:28 that is, more complicated actions with a single client call, as opposed to several client calls to do the same with raw ubus access? Nov 05 08:49:24 zorun: rpcd mainly provides ubus methods Nov 05 08:49:51 zorun: each ubus method needs to be hosted by a service, e.g. the network.interface namespace is hosted by netifd, the system namespace is hosted by procd etc. Nov 05 08:50:18 rpcd is a deamon which provides further ubus namespaces which are not otherwise provided by services Nov 05 08:50:25 e.g. session, uci, file, iwinfo Nov 05 08:50:39 jow: I see, thanks Nov 05 08:50:59 it also provides the ability to load plugins for extending its functionality Nov 05 08:51:11 either through shared objects or cgi-like executables Nov 05 08:56:38 ok Nov 05 08:57:20 jow: so, LuCI is slowly moving towards full-JS, with no Lua backend? that would deprecate LuCI2? Nov 05 08:57:47 zorun: yes Nov 05 08:58:11 its becoming luci2 Nov 05 08:58:14 nice Nov 05 08:59:17 jow: btw, what are your thoughts on a simplified web interface with a set of "scenarios"? do you think it could be part of LuCI or should it be a different project? Nov 05 08:59:31 in both cases, LuCI is still needed for more complex use-cases Nov 05 08:59:37 zorun: aparcar[m] and me think that it would make sense as part of luci Nov 05 08:59:54 basically remove luci-mod-network, luci-mod-system (and maybe luci-mod-status) Nov 05 09:00:02 and ship with a luci-mod-simple instead Nov 05 09:01:00 Yeah, pointless to make competing stuff, if same framework allows alternates :) Nov 05 09:01:32 ok, it's harder to do but it makes sense long-term Nov 05 09:01:55 I will see if the students are up to the task (I meet them this week) Nov 05 09:02:08 most "vendor" interfaces allow a simple and a advanced settings menu. The simple one is likely enough for 90% of users only wanting to change passwords, wifi and maybe add static leases. Would be very awesome if you could work on that zorun Nov 05 09:02:15 anyway, their first task will be to determine what should go in this simplified interface :) Nov 05 09:03:08 aparcar[m]: yeah, let's keep in touch, especially regarding the backend stuff (I have never touched that part) Nov 05 09:03:37 zorun: regarding the style maybe https://docs.gl-inet.com/en/3/setup/mini_router/first-time_setup/ can be an inspiration Nov 05 09:04:01 zorun: sure let me know what you need. I'm happy to implement the bash part which can then later be rewritten in C Nov 05 09:04:23 thanks for the link Nov 05 09:04:31 aparcar[m]: ok, nice! Nov 05 09:04:56 bbl Nov 05 09:06:31 Ofcourse with openwrt one always has choice, but talking in more broader sense usually "simplified" transaltes to cripled too, because vendors nowadays doesn't want you to even change wifi settings (thanks Inteno and a ISP) Nov 05 09:15:04 sure I'd always have both available Nov 05 09:36:07 mh, seems mit github username triggers bugs with github pages Nov 05 09:36:11 *my Nov 05 09:38:52 s/jow-/jow+/ Nov 05 09:39:38 thats probably an even worse idea :) Nov 05 09:39:49 https://jow-.github.io/luci/ Nov 05 09:40:05 lol Nov 05 09:40:13 konsole says invalid hostname, can't even open that link Nov 05 09:40:21 malformed url Nov 05 09:40:36 interesting, firefox opens it but fails at the certificate stuff Nov 05 09:40:53 jow: no nice things for you Nov 05 09:43:13 jow: https://aparcar.github.io/luci/ Nov 05 09:50:46 jow: I'm guessing you have to add some specific https://github.com/jekyll/jekyll/blob/master/docs/_config.yml but would be cool to have this! Nov 05 09:54:55 RFC3696: If the hyphen is used, it is not permitted to appear at either the beginning or end of a label. Nov 05 09:55:14 so that's to be expected… Nov 05 09:59:48 jow: re moving things to client side cbi, are there any notes or docs on doing this for other people? or just trolling the luci commits so far? Nov 05 10:07:36 fredrikhl: yeah I know, I guess my username just slipped the cracks of github validation :) Nov 05 10:07:46 karlp: I am in the process of writing documentaiton Nov 05 10:07:47 github validation? :P Nov 05 10:08:11 stintel: allegedly gh useranmes with multiple consecutive dashes, leading dashes or trailing dashes are not allowed Nov 05 10:08:16 yet I was able to register "jow-" Nov 05 10:08:40 maybe that limitation was added later because they planned to use usernames in dns labels Nov 05 10:11:25 aparcar[m]: you need to change your settings for gh pages to "use /docs from master branch" Nov 05 10:11:42 i did Nov 05 10:12:06 worked on the master repo: http://openwrt.github.io/luci/api/ Nov 05 10:12:10 I didn't yet add an index.html Nov 05 10:12:16 in docs/ Nov 05 10:13:44 strange don't know why it does not work... Nov 05 10:14:33 there's also http://openwrt.github.io/luci/jsapi/ now but the cdn still serves an incomplete version without the required assets Nov 05 10:14:34 anyway good that it work on master Nov 05 10:15:20 well overall quite cool! Nov 05 10:15:42 does it look correct to you? Nov 05 10:15:48 I mena with colors and stuff Nov 05 10:15:56 for me its just a blank unformatted html page Nov 05 10:16:00 no all oldschool Nov 05 10:16:07 jow: great thanks. I've got a bunch of things to tweak/fix/migrate from the variou schanges in tables to divs and want to get it all done and move to "the future" at the same time :) Nov 05 10:16:27 it should look like this: http://htmlpreview.github.io/?http://raw.githubusercontent.com/openwrt/luci/master/docs/jsapi/index.html Nov 05 10:18:45 very pretty Nov 05 11:26:18 greearb_: commits are created for the fw bump to 012. thanks! Nov 05 12:54:20 greearb_: I'm happy to confirm that the SBWA errors are gone. nice step forward! Nov 05 13:13:19 build #185 of arc770/generic is complete: Failure [failed pkgbuild] Build details are at http://buildbot.openwrt.org/master/images/builders/arc770%2Fgeneric/builds/185 blamelist: Adrian Schmutzler , Felix Fietkau Nov 05 13:44:36 05|14:43:03 < githubstatus> [bc] Currently investigating increase in queue backlogs affecting Pages builds and some webhooks deliveries. https://t.co/RM4zklcszC Nov 05 13:44:43 jow: did you break it with your username :P Nov 05 13:45:37 stintel: rofl Nov 05 13:48:08 closed loop automated testing, we need more of that ml/ai based cloud archestration Nov 05 13:48:27 *yawn* Nov 05 13:48:50 i am hungry but breakfast wont be for another 45 minutes, jet lag sucks Nov 05 13:59:51 username of jow- is certainly challenging all unix system still. alphanumeric plus _ symbol are the safe bet i assume, everything else is 'risky' Nov 05 14:00:47 Error: Cannot load https://raw.githubusercontent.com/openwrt/luci/master/documentation/api/index.html: 404 Not Found Nov 05 14:06:07 jow: trick question for template on luci, what does - mean in <%- and/or -%> Nov 05 14:06:13 s/trick/quick/ Nov 05 14:06:52 they already answered you? Nov 05 14:07:09 oops, let me re-check, did not find that Nov 05 14:07:26 about 7 hours ago by my clock. Nov 05 14:07:43 saw it, thanks Nov 05 14:22:39 nbd: can you port https://git.openwrt.org/?p=openwrt/openwrt.git;a=commitdiff;h=8b15e7f661b4263a7b9fb6a410a5f58263b72caa to https://git.openwrt.org/?p=project/iwinfo.git;a=blob;f=iwinfo_nl80211.c;h=4b6ef9124e66c6705b9ba356e8f2d8456b18f12a;hb=31dcef3169d29dc9bb11c6901fa7b2894438dc78#l236 as well? I'm too dense to follow that shell code Nov 05 14:33:40 i'll see if i find the time Nov 05 14:33:44 the basic principle is simple Nov 05 14:33:48 first phy of a device has the same path Nov 05 14:33:55 second phy has +1 appended to the path name Nov 05 14:33:56 etc. Nov 05 14:35:59 that numbering scheme relies on implicit ascii ordering of sysfs? Nov 05 14:36:16 yes Nov 05 14:36:50 ok Nov 05 14:47:59 https://github.com/openwrt/luci/wiki/Documentation works now after a page reload Nov 05 14:48:09 I updated the link Nov 05 14:48:12 ...s Nov 05 14:48:37 the documentaiton folder got renamed to docs/ in preparation for github pages which ... is still broken now Nov 05 14:48:45 cool, saw the new client-side thing, what's that Nov 05 14:49:02 the Lua cbi reimplemented in JS Nov 05 14:49:08 in a nutshell Nov 05 14:49:35 for luci-ng right? or luci also uses it? learning luci these days here Nov 05 14:49:46 luci uses it Nov 05 14:50:22 cool,thanks Nov 05 15:10:42 jow, still some links not working, the server-side-lua-apis and client-side-js-apis at https://htmlpreview.github.io/?https://github.com/openwrt/luci/blob/master/docs/jsapi/index.html Nov 05 15:15:08 works for me Nov 05 15:15:46 https://htmlpreview.github.io/?http://htmlpreview.github.io/?http://raw.githubusercontent.com/openwrt/luci/master/docs/api/index.html Nov 05 15:16:02 thats an invalid link Nov 05 15:16:31 that's what i got after click on https://htmlpreview.github.io/?https://github.com/openwrt/luci/blob/master/docs/jsapi/index.html Server side Lua APIs Nov 05 15:16:54 sigh, don't click there Nov 05 15:17:04 it will be solved with github pages Nov 05 15:17:15 but https://www.githubstatus.com/ Nov 05 15:18:13 ok Nov 05 15:42:13 jow: just looked at luci the master branch has lots of js files(18.06 is pure lua, 19.07 has a few js files only), looks like next release luci might get rid of lua-mvc fully, with all these new js files, will luck use some js-framework for the frontend, or you will write an openwrt-version of it? any doc on the forthcoming luci revamp? Nov 05 15:42:44 s/luck/openwrt/ Nov 05 15:43:41 I don't plan to use any frameworks Nov 05 15:44:03 and...what will be the openwrt backend look like, rest? luci-mod-rpc? uhttpd-mod-ubus? Nov 05 15:44:21 luci-mod-rpc, uhttpd-mod-ubus, rpcd Nov 05 15:44:57 frontend will probably be just bootstrap + jquey + all-those-js-files-with-ajax-to-rpc? Nov 05 15:45:13 no jquery Nov 05 15:45:28 vanilla JS Nov 05 15:46:35 what for css then? a cut-down bootstrap or writing your own css as well? a white and black theme will be nice Nov 05 15:46:50 its currently usinga bootstrap derived theme Nov 05 15:46:58 but its not using the bootstrap css framework Nov 05 15:49:32 building master branch now, this is big change, hope 2020 will see this in the official release,can't wait Nov 05 15:50:13 just spent last few days to better lua, now it seems outdated already :) Nov 05 15:51:16 i hope the api is well defined to accommodate other frontends, though for me bootstrap4 is good enough Nov 05 16:47:58 https://github.com/rosywrt/luci-theme-rosy looks nice Nov 05 17:21:21 xback, thanks for the testing and handling the patches Nov 05 17:55:23 jow: i assume the new luci changes are not enabled in master yet, just built it and still the default luci is there. I did notice if I enable rose theme it will be used directly in trunk overriding the real default theme Nov 05 18:02:43 Hi all, since Fedora 31, python now points to python3, openwrt master is fine but I cannot build 19.07, my questions are: 1) is it on someone's todo list ? 2) if not do you prefer a patch that just replace python with python2 or cherry-pick everything needed from master ? Nov 05 18:03:41 champtar: probably cherry pick, if reasonably small Nov 05 18:04:09 install venv with python2 and build with venv activated? Nov 05 18:05:55 For now i've patched all the python that I could find to python2, and in master it seems everything is migrated to python3 Nov 05 18:07:23 python2's EOL is planned in less than 2 months, so we should try to not require it in an upcoming release Nov 05 18:07:29 So cherry-pick might not be that small, but I like the idea of removing python2 dependency for 19.07 Nov 05 18:08:21 so let me rephrase that, as long as it is not unreasonably huge, do the backport ;) Nov 05 18:12:14 ./scripts/diffconfig.sh Nov 05 18:12:22 tmp/.config-package.in:716:warning: ignoring unsupported character '@' Nov 05 18:12:32 new thing for me Nov 05 18:15:17 https://github.com/openwrt/packages/issues/9283, what triggers this is i switch between master and 18.06 build Nov 05 18:52:23 tmp/.config-package.in:100123:warning: ignoring unsupported character '@' is happening in 18.06 Nov 05 18:55:33 rr123: the issue is that the switch doesn't rebuild *config; after a make -C scripts/config clean it should go away Nov 05 19:01:03 KanjiMonster: i did make distclean; ./scripts/feeds update/install; make defconfig; make; but still see it Nov 05 19:01:55 unless my old config file used for defconfig is causing this, which I never saw that in the past using the same config Nov 05 19:04:44 rr123: are you sure you did a distclean and not a dirclean? Nov 05 19:06:06 i'm sure, in fact i forgot to relink dl and the new build spent a lot of time re-downloading stuff Nov 05 19:08:26 jumping between master/19/18.06 with different archs for a while and I lost the confidence of the sanity of the build system so I just cleaned up all for good and rebuild them Nov 05 19:19:10 rr123: they are Nov 05 19:19:19 rr123: not sure why you expect something different Nov 05 19:19:30 its stil lthe same luci with the same design, just js instead of lua Nov 05 19:21:01 jow: thanks, got it now. Nov 05 19:21:14 rr123: have you tried the command I wrote? Nov 05 19:21:45 KanjiMonster: not yet, in the middle of building, will try after it's done Nov 05 19:29:20 KanjiMonster: make -C scripts/config clean && make defconfig -- about 30+ lines of "tmp/.config-package.in:715:warning: ignoring unsupported character '@'" still persist Nov 05 19:30:04 make sure your feeds are current Nov 05 19:30:24 also if you copied feeds.conf.defualt to feeds.conf at some point, makre sure that this one points to the proper branches Nov 05 19:30:38 did update/install -a all feeds Nov 05 19:31:53 feeds.conf unchanged for a long while, never saw this "@" message before until I sync to master and build it, then i can never get rid of it. anyway it's not really an error message Nov 05 19:32:47 rr123: if you can't get rid of it then it sounds like you are trying to build 18.06 feeds with a master base Nov 05 19:34:59 the @ error happens when you pull in makefiles with broken DEPENDS:= sysntax Nov 05 19:36:29 which we fixed for master (and I guess 19.07), but only master (and 19.07?) warn about it, so only the combo master + pre-fix feeds can cause the warning, not e.g. 18.06 base + 18.06 feeds Nov 05 19:37:10 i'm on 18.06 for sure, I notice all those warning lines have something like "@TARGET_brcm47xx, or @arm" Nov 05 19:37:53 rr123: what the top rev of git log scripts/config ? Nov 05 19:38:34 commit 9d401013fc61b0ca51cd557b9922052b2c3cce26 ustream-ssl: backport fix for CVE-2019-5101, CVE-2019-5102 Nov 05 19:40:36 rr123: that commit doesn't touch scripts/config, that can't be the top one Nov 05 19:41:21 oops, thought it's for the whole git, anyway it is: commit 02a1a8af955180439be6316049ba8a70cf5d1b1c build: cleanup leftover qconf files Nov 05 19:44:05 rr123: ah, I think I know what's up - you need do the distclean on master *before* the switch to 18.06; 18.06 doesn't do a proper clean scripts/config Nov 05 19:44:19 see 0096a1cf0015e483b99e51c74f2f0bbae7247342 Nov 05 19:45:19 thanks, trying Nov 05 19:46:55 I think that one would be a safe cherry-pick to 18.06 Nov 05 19:48:03 git checkout master; make -C scripts/config clean && git checkout openwrt-18.06 && make defconfig, looked clean, and fixed it Nov 05 20:21:41 is there a way nowadays to override packages from a certain feed (e.g. with low priotity) with packages from another higher prioritised feed? Nov 05 20:29:16 jow: isn't the 19.07-rc0 almost in the oven? Nov 05 20:34:16 mirko: i do: ./scripts/feeds install -a -p my-high-priority-feed && ./scripts/feeds install -a to make sure my higher priority packages are overriding the rest Nov 05 20:43:55 Guys i am ripping my hair off, I have not updated my routers for the better part of a year due to them just working. Finally got around to update and everything works great except my Sony TV just hates my AP. It will see it, connect to it then disconnect. Rinse and repeat. Works fine with an Ethernet Nov 05 20:44:28 rr123: hm :) Nov 05 20:44:45 fair enough though :) Nov 05 20:46:56 ynezz: I got distracted by verious stuff today. I could make the tag *now* and immediately leave the PC afterwards or I'll do it first thing tomorrow morning Nov 05 20:47:09 I prefer the latter Nov 05 20:50:14 thagabe: If you want help you should give version before, after, AP version, maybe you have some warnings in the logs Nov 05 20:56:08 jow: sure, no problem, just wondering about that python3 backports, that's all about it Nov 05 20:56:32 ynezz: from my pov we could take them after rc0 Nov 05 20:56:57 ok, sounds like a plan Nov 05 20:59:19 rr123: when you say "built the master and the new luci isn't ther" what did you mean? Nov 05 20:59:36 the changes to use js instead of backend lua are not visible for the most part, Nov 05 21:01:06 ynezz: do you know if everything now build with scons & python3 in master ? Nov 05 21:03:05 (else i'll check tomorrow) Nov 05 21:03:23 champtar: well, master is easy, nothing depends on scons, packages might be different story Nov 05 21:03:39 that's why I've asked mangix for his view in that PR, he knows better Nov 05 21:06:46 Ok thanks Nov 05 21:17:33 mirko: https://github.com/openwrt/openwrt/pull/2481 ( Nov 05 21:18:07 "automake: Fix "make V=1" behaviour in automake packages") is it ok to add that Acked-by as suggested? Nov 05 21:18:53 ynezz: yes, as already confirmed in https://github.com/openwrt/openwrt/pull/2481#issuecomment-541683807 Nov 05 21:19:30 mirko: I'm just unsure about the mail address Nov 05 21:19:47 ah, mirko-openwrt@nanl.de is fine Nov 05 21:20:12 ok, thanks! Nov 05 21:20:26 i'm still curious what's going to break :) Nov 05 21:20:51 subtle things, in dark corners, beneath five layer of m4sh code Nov 05 21:21:42 well, it's a revert of a fix and i'm *sure* i fixed sth. - only nobody knows what exactly Nov 05 21:21:58 I know that situation Nov 05 21:22:11 I love the things that only work due to a combination of different bugs Nov 05 21:22:22 though, i'm really happy about how elaborate my commit messages were back then - it helps a lot Nov 05 21:22:32 unfortunately i forgot the part what it actually fixes Nov 05 21:25:52 lets see :) Nov 05 21:27:09 karlp: i thought it's a luci revamp allover, it's more of replacing luci-lua-cbi with js-promises-cbi and the theme etc are not affected Nov 05 21:33:49 switch arch and build openwrt is safe sharing one git tree as long as they're on the same code base(e.g. 18.06), switch between 18.06, 19.07, master versions under one git checkout seems not safe at all, so to have 3 builds for each version(18.06,19.07.master) do I have to use three git-checkouts? in my above case, build with master then switch to 18.06 caused trouble, for example Nov 05 21:34:51 i want to have 18.06, 19.07 and master 3 builds in parallel basically, probably i need 3 git-clone sharing one dl/ then Nov 05 21:37:18 rr123: i am using separate folders per release / master as well. that being said, i keep multiple checkouts (19.07 and master) in my 19.07 folder e.g. to be able to cherry-pick stuff from master into my local 19.07 tree, but actually compiling master, that's done in a different folder (and a separate tree) altogether Nov 05 21:37:45 i don't know what interferes with one another but if memory serves me right then yes there seems to be stuff that gets messed up. Nov 05 21:38:19 i tend to actively prune the staging and build dirs for the separate trees Nov 05 21:40:26 i will use 3 git for now, buildroot went nuts once a while and the only sane way out of it is to make distclean(though keep dl) Nov 05 21:51:26 Is Roger D (rogerdammit@gmail.com) here? (maintainer micropython) Nov 05 22:03:15 rr123: yes, this is a reasonable approach, I also have 3 build trees, and a symlink to share dl/ Nov 05 22:05:46 rr123: you can share them, but youll need to doo way more cleaning than you'd like to do. it's better to just have multiple trees :) Nov 05 22:40:15 there is also a .config option to aim downloads (dl) directory, but i use a symlink too ;-) Nov 05 22:40:56 CONFIG_DOWNLOAD_FOLDER Nov 05 22:50:43 aparcar[m]: bash: GZ_SUFFIX: command not found Nov 05 22:50:43 Makefile:86: recipe for target '/opt/devel/openwrt/openwrt.git/bin/targets/ath79/generic/openwrt-ath79-generic-tplink_archer-c7-v5-squashfs-sysupgrade.bin' failed Nov 05 22:51:02 reverting 881ed09ee6e23f6c224184bb7493253c4624fb9f fixes it :p Nov 05 22:51:18 ynezz: yea I saw that too, did not figure out yet what it is Nov 05 22:52:19 ynezz: well kindly help me figure out hat the problem is, I'd be sad if the JSON patch is gone for that very reason Nov 05 22:52:53 hm # CONFIG_JSON_ADD_IMAGE_INFO is not set Nov 05 22:53:32 if it is not set shouldn't it then just ignore the json_add_image_info.py part? Nov 05 22:54:57 it's make, whitespace sensitive minefield Nov 05 22:56:05 it's quite interesting, that it's fine on the bots Nov 05 22:56:15 buildbots I mean Nov 05 22:56:41 ah, it's because they've this option enabled Nov 05 22:56:58 anyway, it's caused by the missing new line you've removed in that patch Nov 05 22:57:01 I'll fix it Nov 05 22:57:48 thanks ynezz Nov 05 23:14:01 jow: do you manage the mailing list openwrt-devel? Nov 06 02:12:02 tested material and rosy theme, definitely not as good as the default bootstrap, luci-mod-failsafe will mess up the login page, pages not displayed correctly on cellphones,etc **** ENDING LOGGING AT Wed Nov 06 03:00:06 2019