gemini - kennedy.gemi.dev

💾 Archived View for gemini.ctrl-c.club › ~phoebos › logs › kisslinux-2022-02-21.txt captured on 2024-05-26 at 16:10:08.
-=-=-=-=-=-=-
[2022-02-21T00:28:55Z] <dilyn> @illiliti: http://ix.io/3QeO
[2022-02-21T00:28:56Z] <dilyn>  womp
[2022-02-21T00:29:07Z] <dilyn> cloning kiss-libdemi and building libinput
[2022-02-21T01:29:45Z] <illiliti> dilyn: clang i suppose?
[2022-02-21T01:29:50Z] <dilyn> yessir
[2022-02-21T01:33:38Z] <illiliti> fixed
[2022-02-21T01:34:40Z] <dilyn> <3
[2022-02-21T01:40:54Z] <illiliti> noocsharp: this is bad
[2022-02-21T01:41:01Z] <illiliti> at least there's immediate dereference which will crash the program
[2022-02-21T01:41:21Z] <illiliti> still bad though
[2022-02-21T01:49:30Z] <noocsharp> i've been working on removing all the unnecessary allocations, but there's a lot of them
[2022-02-21T01:50:28Z] <noocsharp> and a bunch of other weird stuff
[2022-02-21T01:50:48Z] <noocsharp> it's weird that wl-clipboard is as widely used as it is given the code quality
[2022-02-21T01:51:29Z] <dilyn> because it's the only implementation of a clipboard for wayland basically :v
[2022-02-21T01:52:16Z] <illiliti> code quality is super bad
[2022-02-21T01:52:16Z] <illiliti> https://github.com/bugaevc/wl-clipboard/blob/master/src/util/files.c#L221-L222
[2022-02-21T01:52:19Z] <illiliti> look at this
[2022-02-21T01:52:42Z] <noocsharp> fixed: https://github.com/noocsharp/wl-clipboard
[2022-02-21T01:53:29Z] <illiliti> https://github.com/bugaevc/wl-clipboard/blob/master/src/util/files.c#L206
[2022-02-21T01:53:33Z] <illiliti> hardcoded /tmp
[2022-02-21T01:53:46Z] <illiliti> use TMPDIR first
[2022-02-21T01:54:10Z] <noocsharp> i haven't fixed that yet, but i'll put it on the list
[2022-02-21T01:56:37Z] <noocsharp> maybe it would be better to just write a better one from scratch
[2022-02-21T01:57:26Z] <illiliti> indeed
[2022-02-21T01:57:38Z] <illiliti> the code is too bad
[2022-02-21T01:58:22Z] <illiliti> https://github.com/bugaevc/wl-clipboard/blob/master/src/util/files.c#L169-L171
[2022-02-21T01:58:25Z] <illiliti> gosh
[2022-02-21T01:58:36Z] <noocsharp> and the maintainer is gone to help r invade u probably
[2022-02-21T01:59:07Z] <illiliti> let distro maintainer decide which path should be used, stop doing this nonsense
[2022-02-21T02:01:04Z] <noocsharp> https://github.com/bugaevc/wl-clipboard/blob/master/src/wl-paste.c#L103-L129
[2022-02-21T02:01:14Z] <noocsharp> lol
[2022-02-21T02:01:33Z] <dilyn> I should learn go faster to write a better version :V :V :V
[2022-02-21T02:01:44Z] <noocsharp> it has to be in c or nobody will use it
[2022-02-21T02:02:00Z] <dilyn> :|
[2022-02-21T02:02:01Z] <illiliti> ^
[2022-02-21T02:02:41Z] <dilyn> but golang so kewl
[2022-02-21T02:03:04Z] <noocsharp> it has to be in c or illiliti won't use it
[2022-02-21T02:03:36Z] <dilyn> weeeeeellllll
[2022-02-21T02:04:09Z] <dilyn> illiliti is your wlroots fork based on -.15x or 0.14?
[2022-02-21T02:04:22Z] <illiliti> latest master
[2022-02-21T02:04:31Z] <dilyn> hng
[2022-02-21T02:05:59Z] <illiliti> soliwilos said they was able to apply my patch to 0.15.0
[2022-02-21T02:06:30Z] <dilyn> my problem is that wayfire still uses 0.14.1
[2022-02-21T02:06:41Z] <illiliti> ah
[2022-02-21T02:12:51Z] <illiliti> noocsharp: https://github.com/noocsharp/wl-clipboard/blob/master/src/util/files.c#L150
[2022-02-21T02:12:54Z] <illiliti> ssize_t
[2022-02-21T02:17:03Z] <noocsharp> yeah, i saw that a few days ago
[2022-02-21T02:17:31Z] <noocsharp> forgot about it apparently
[2022-02-21T02:18:04Z] <noocsharp> it's inherited from master
[2022-02-21T02:22:35Z] <illiliti> https://github.com/noocsharp/wl-clipboard/blob/master/src/util/files.c#L152
[2022-02-21T02:23:00Z] <illiliti> i would put assert here or smth to make sure it's newline
[2022-02-21T02:24:49Z] <noocsharp> tbh i'm probably gonna abandon this and just write a new one
[2022-02-21T02:26:13Z] <illiliti> yeah
[2022-02-21T02:27:14Z] <noocsharp> i feel like it's much easier to get error checking right if you're diligent about it from the beginning than shoehorning it into a project that didn't pay attention to it
[2022-02-21T02:29:29Z] <illiliti> my thoughts exactly
[2022-02-21T02:33:04Z] <illiliti> no wonder why wl-clipboard is too buggy. i use fg/bg instead of tmux and sometimes if i copy something it can cause deadlock
[2022-02-21T02:33:20Z] <illiliti> i have no choice but to kill -9
[2022-02-21T02:39:38Z] <testuser[m]> Hi
[2022-02-21T02:42:53Z] <noocsharp> \o/
[2022-02-21T02:43:24Z] <testuser[m]> <illiliti> "ssize_t" <- Bruh
[2022-02-21T02:43:42Z] <testuser[m]> It's gonna explode if a signal is received 
[2022-02-21T02:51:10Z] <testuser[m]> typedef char * const *argv_t
[2022-02-21T02:51:10Z] <testuser[m]> Cursed
[2022-02-21T02:51:55Z] <illiliti> signals are bug by definition btw
[2022-02-21T02:52:06Z] <illiliti> we need signal-free versions of functions
[2022-02-21T02:52:16Z] <illiliti> see close() EINTR fallout
[2022-02-21T02:53:18Z] <testuser[m]> I write a wrapper for them with a loop to check errno for eintr 
[2022-02-21T02:53:48Z] <illiliti> never
[2022-02-21T02:54:01Z] <noocsharp> for close!?
[2022-02-21T02:55:56Z] <illiliti> never ever use close() in the loop
[2022-02-21T02:56:10Z] <testuser[m]> No
[2022-02-21T02:56:13Z] <testuser[m]> i meant write / read
[2022-02-21T02:56:18Z] <noocsharp> while (close() == -1);
[2022-02-21T02:56:24Z] <dilyn> http://ix.io/3Qf8 womp womp
[2022-02-21T02:56:24Z] <dilyn> the greatest shame is that hikari requires pango :v
[2022-02-21T02:56:59Z] <dilyn> hi testuser:)
[2022-02-21T02:57:36Z] <testuser[m]> Bruh illiliti something actually did that https://github.com/wahern/cqueues/issues/56
[2022-02-21T02:58:21Z] <noocsharp> unix is really simple when you don't look too hard
[2022-02-21T03:03:48Z] <illiliti> dilyn: what patch are you trying to port?
[2022-02-21T03:04:18Z] <dilyn> https://github.com/illiliti/wlroots/commit/c118c6d562fabbc2a456f47d0c1f94747de23d48.patch
[2022-02-21T03:04:26Z] <dilyn> latest on libdemi branch
[2022-02-21T03:04:38Z] <dilyn> oh harumph...
[2022-02-21T03:04:38Z] <dilyn> one sec
[2022-02-21T03:05:02Z] <illiliti> https://gitlab.freedesktop.org/wlroots/wlroots/-/merge_requests/3451.patch
[2022-02-21T03:05:34Z] <illiliti> i no longer update github branch
[2022-02-21T03:06:52Z] <dilyn> aha i see i see
[2022-02-21T03:07:06Z] <dilyn> we'll see :V
[2022-02-21T03:30:27Z] <illiliti> noocsharp: if you're going to (re)write wl-clipboard, consider taking a library approach
[2022-02-21T03:30:35Z] <illiliti> so people can use your API instead of shelling out to command
[2022-02-21T03:30:52Z] <illiliti> https://github.com/YaLTeR/wl-clipboard-rs
[2022-02-21T03:30:59Z] <illiliti> something like that, but in c
[2022-02-21T03:32:09Z] <noocsharp> yeah, let me get something working first and then i'll do that
[2022-02-21T03:33:39Z] <illiliti> yeah, no rush
[2022-02-21T03:34:32Z] <testuser[m]> Rich
[2022-02-21T03:35:19Z] <illiliti> yeah yeah yeah yeah
[2022-02-21T03:40:04Z] <dilyn> http://ix.io/3Qfi
[2022-02-21T03:40:27Z] <dilyn> if you're going to rewrite wl-clipboard consider using a nonGPL license<3
[2022-02-21T03:41:15Z] <noocsharp> i have never used gpl in my life
[2022-02-21T03:41:27Z] <dilyn> my man 8)
[2022-02-21T03:42:54Z] <noocsharp> although every time i start working on something i spend a day debating whether to make it gpl
[2022-02-21T03:43:18Z] <dilyn> include a Dilyn-exception
[2022-02-21T03:43:53Z] <illiliti> dilyn: make sure #include "backend/session/session.h" is present
[2022-02-21T03:44:07Z] <illiliti> in backend/session/session.c
[2022-02-21T03:44:17Z] <dilyn> it is :V
[2022-02-21T03:44:22Z] <dilyn> hence my confusion lmao
[2022-02-21T03:44:36Z] <dilyn> I thought I had bungled an ifdef but that isn't the case either
[2022-02-21T03:44:45Z] <dilyn> something just isn't being roped in somehow...
[2022-02-21T03:44:50Z] <testuser[m]> U mostly don't need to worry about corpos stealing ur stuff unless it's webscale 🚀 related
[2022-02-21T03:45:05Z] <testuser[m]> But i still use gpl
[2022-02-21T03:45:09Z] <illiliti> grep session_open_if_kms include/backend/session/session.h
[2022-02-21T03:46:07Z] <dilyn> ack
[2022-02-21T03:46:08Z] <dilyn> ain't there
[2022-02-21T03:46:11Z] <dilyn> only present in two files
[2022-02-21T03:46:26Z] <dilyn> oh wait hold on
[2022-02-21T03:47:39Z] <dilyn> it isn't present in that file;
[2022-02-21T03:47:40Z] <dilyn> ./backend/session/udev.c:			session_open_if_kms(session, udev_device_get_devnode(dev));
[2022-02-21T03:47:40Z] <dilyn> ./backend/session/demi.c.orig:	struct wlr_device *wlr_dev = session_open_if_kms(ctx->session, devnode);
[2022-02-21T03:47:41Z] <dilyn> ./backend/session/session.c:		struct wlr_device *wlr_dev = session_open_if_kms(session, devnode);
[2022-02-21T03:49:39Z] <illiliti> https://gitlab.freedesktop.org/wlroots/wlroots/-/blob/6da814b2c26509799553c6005ac4b62877129f01/backend/session/session.c#L256
[2022-02-21T03:50:29Z] <illiliti> dilyn: also, i recommend using git merge to merge patches
[2022-02-21T03:50:55Z] <illiliti> it's smarter than patch
[2022-02-21T03:51:59Z] <dilyn> hm yes that is very true :thinking: why didn't I think of that
[2022-02-21T03:52:42Z] <illiliti> re GPL, my rule: ISC for libs, 0BSD for algos/simple programs/build scripts, GPLv3 for everything else
[2022-02-21T03:52:53Z] <dilyn> interestingly that line exists in wlroots 0.14.1 but the struct isn't defined in include/backend/session/session.h in 0.14.1...
[2022-02-21T03:54:14Z] <dilyn> jk it exists as *open_if_kms* smh
[2022-02-21T03:54:26Z] <illiliti> hmm
[2022-02-21T03:55:33Z] <illiliti> i see
[2022-02-21T03:55:45Z] <illiliti> rename it to session_open_if_kms
[2022-02-21T03:58:24Z] <dilyn> a few more errors
[2022-02-21T03:58:39Z] <dilyn> specifically "static dec of "session_open_if_kms" follows non-static dec"
[2022-02-21T03:59:10Z] <dilyn> http://ix.io/3Qfk
[2022-02-21T03:59:45Z] <dilyn> oh shit no lol that's a bad thing; my mistake
[2022-02-21T04:00:16Z] <dilyn> fixed the patch; now it's the same error as before except open_if_kms is undef. so perhaps just rewrite open_if_kms as session_open_if_kms?
[2022-02-21T04:02:44Z] <dilyn> indeed, there was one lingering open_if_kms. changed it, and it built.
[2022-02-21T04:03:55Z] <dilyn> okay. if everything worked, i should be back in less than two minutes. If I'm not, it did *not* work :V
[2022-02-21T04:10:09Z] <noocsharp> lol
[2022-02-21T04:13:07Z] <illiliti> rip
[2022-02-21T04:18:29Z] <illiliti> https://termbin.com/nog8 cherry-picked
[2022-02-21T04:18:32Z] <illiliti> not tested
[2022-02-21T04:28:09Z] <illiliti> https://termbin.com/9poh
[2022-02-21T04:28:20Z] <illiliti> should work now
[2022-02-21T04:33:37Z] <dilyn> :v
[2022-02-21T04:33:40Z] <dilyn> will test soonish
[2022-02-21T04:53:24Z] <dilyn> it may theoretically be my problem
[2022-02-21T04:59:19Z] <illiliti> wdym?
[2022-02-21T05:02:31Z] <illiliti> can you post logs or something?
[2022-02-21T05:02:50Z] <illiliti> backtrace if it's segfault
[2022-02-21T05:11:21Z] <dilyn> no segfaults or anything; wayfire launches just fine
[2022-02-21T05:11:30Z] <dilyn> just no input, not a permission problem afaik
[2022-02-21T05:12:11Z] <illiliti> hmm
[2022-02-21T05:12:14Z] <dilyn> I can try to nab some logs from wayfire later to see what's up
[2022-02-21T05:12:21Z] <dilyn> but for now, pizza and reacher time :v
[2022-02-21T05:12:59Z] <dilyn> most of this work will probably be rendered irrelevant once wayfire is updated to latest wlroots anyways; wayfire itself doesn't actually use any libudev stuff
[2022-02-21T05:13:26Z] <dilyn> well that may or may not be true; I suppose I'd have to see how it's using wlroots
[2022-02-21T05:14:03Z] <illiliti> git grep udev
[2022-02-21T09:49:31Z] <phinxy> Why do you prefer ISC license for libs?
[2022-02-21T10:17:37Z] <illiliti> permissive, allows static link unlike gpl
[2022-02-21T15:28:58Z] <illiliti> noocsharp: https://emersion.fr/blog/2020/wayland-clipboard-drag-and-drop/
[2022-02-21T16:38:39Z] <dilyn> illiliti: `git grep udev` only reveals a single CI workflow :o hmhmhm
[2022-02-21T16:41:32Z] <illiliti> good
[2022-02-21T16:41:38Z] <illiliti> dilyn: does sway work with libdemi btw?
[2022-02-21T16:41:48Z] <dilyn> no idea, never tried!
[2022-02-21T16:41:59Z] <dilyn> I am not a super big fan of tiling wms nowadays :\
[2022-02-21T16:42:48Z] <illiliti> no input devices means there's a possible bug in my libinput patch or libdemi
[2022-02-21T16:43:09Z] <illiliti> you can help me catch one
[2022-02-21T16:43:11Z] <dilyn> i agree
[2022-02-21T16:43:25Z] <dilyn> wayfire has several debugging options; I'll collect them independently and send 'em your way
[2022-02-21T16:43:51Z] <illiliti> that would be awesome
[2022-02-21T16:43:55Z] <dilyn> so I'll be back after a few reboots haha
[2022-02-21T16:51:39Z] <noocsharp> illiliti: thanks, that's much easier to read than the wayland headers
[2022-02-21T16:57:20Z] <noocsharp> seems like command line copy and paste isn't possible with the core wayland protocol, since only the window with keyboard focus can copy things
[2022-02-21T16:57:24Z] <noocsharp> so wlr-data-control it is
[2022-02-21T17:06:03Z] <dilyn> broken input: http://ix.io/3QiT
[2022-02-21T17:06:04Z] <dilyn> working input: http://ix.io/3QiU
[2022-02-21T17:06:04Z] <dilyn> Biggest difference between first and second is device enumeration actually happens when it works; libinput is 'initialized' in the first, but no enumeration happens. I don't think there's a way to increase verbosity
[2022-02-21T17:06:40Z] <dilyn> places of interest in wayfire code may be src/core/seat/seat.cpp
[2022-02-21T17:11:04Z] <illiliti> it happends randomly, right?
[2022-02-21T17:11:19Z] <dilyn> jk that code isn't actually that interesting
[2022-02-21T17:11:23Z] <dilyn> what do you mean it happens randomly?
[2022-02-21T17:11:32Z] <dilyn> it happens every time I start wayfire. no input, regardless of how long I wait
[2022-02-21T17:11:56Z] <illiliti> how did you get working input then
[2022-02-21T17:12:04Z] <dilyn> oh i rebuilt everything against libudev-zero lmao
[2022-02-21T17:12:11Z] <illiliti> ah
[2022-02-21T17:12:33Z] <illiliti> sounds like a bug tbh
[2022-02-21T17:13:05Z] <dilyn> mhm
[2022-02-21T17:14:31Z] <illiliti> post output of: ls -la /dev/input
[2022-02-21T17:14:40Z] <dilyn> it's what you would expect
[2022-02-21T17:14:43Z] <dilyn> root:input 660
[2022-02-21T17:14:55Z] <dilyn> all the devices i would expect to exist are there
[2022-02-21T17:15:17Z] <illiliti> post anyway
[2022-02-21T17:15:57Z] <dilyn> http://ix.io/3Qj0
[2022-02-21T17:19:43Z] <illiliti> revert this: https://github.com/illiliti/libinput/commit/75959303e3ee629fcdc1e4a4dff152942052bf45
[2022-02-21T17:26:34Z] <illiliti> dilyn: did you know that kiss caches git repos?
[2022-02-21T17:26:41Z] <illiliti> that's might be a problem
[2022-02-21T17:26:44Z] <dilyn> I did indeed, it's a new thing
[2022-02-21T17:26:54Z] <dilyn> but I'm not using git+ i'm using archive/commitnum.tar.gz
[2022-02-21T17:27:13Z] <illiliti> ah, that's ok then
[2022-02-21T17:27:23Z] <dilyn> it's on my long pursuit to kiss r git :v
[2022-02-21T17:27:58Z] <illiliti> kiss b got
[2022-02-21T17:30:17Z] <dilyn> :O  don't do this to me illiliti
[2022-02-21T17:36:41Z] <illiliti> pretty sure the bug somewhere in the 0.14.1 wlroots patch
[2022-02-21T17:37:45Z] <illiliti> need confirmation that bug is present on latest sway/wlroots
[2022-02-21T17:40:32Z] <dilyn> http://ix.io/3Qj8
[2022-02-21T17:40:32Z] <dilyn> I suppose I can test latest wlroots/sway for you
[2022-02-21T17:40:49Z] <dilyn> anything to put off learning about io.Reader in golang I suppose
[2022-02-21T17:44:17Z] <illiliti> io.Reader is easy if you know interfaces and type assertion
[2022-02-21T17:44:26Z] <dilyn> exactly
[2022-02-21T17:44:31Z] <dilyn> (:
[2022-02-21T17:44:45Z] <testuser[m]> Soon
[2022-02-21T17:49:23Z] <illiliti> dilyn: https://libera.irclog.whitequark.org/kisslinux/2022-02-11#31751015 << what do you think?
[2022-02-21T17:50:25Z] <dilyn> do you still want to maintain it, just under the umbrella of the org?
[2022-02-21T17:50:35Z] <dilyn> I think to transfer the repo you just need to be a member and that can be arranged
[2022-02-21T17:56:32Z] <noocsharp> i wrote a 100 line program that lets you paste a fixed string in wlroots
[2022-02-21T17:56:47Z] <noocsharp> how is wl-clipboard possibly 1800 lines
[2022-02-21T17:57:08Z] <testuser[m]> Show
[2022-02-21T17:57:16Z] <illiliti> under umbrella, indeed. i also want to revisit and modernize it a bit
[2022-02-21T17:57:28Z] <dilyn> for sure. I'll send the invite now, feel free to transfer it!
[2022-02-21T17:58:42Z] <dilyn> built latest libdemi/libinput/wlroots/sway from kiss-libdemi; will reboot and see. Though I've rarely had luck running sway with input being detected at all, so I might not be the best tester lmao
[2022-02-21T18:01:32Z] <noocsharp> https://github.com/noocsharp/wayclip
[2022-02-21T18:02:26Z] <illiliti> nice!
[2022-02-21T18:05:07Z] <illiliti> dilyn: thanks. will do something today and probably tomorrow
[2022-02-21T18:05:17Z] <illiliti> s/and/or/
[2022-02-21T18:15:11Z] <dilyn> can't even get sway to connect to my outputs :V :V :V
[2022-02-21T18:15:11Z] <dilyn> http://ix.io/3Qjg
[2022-02-21T18:15:12Z] <dilyn> But it seems that libinput also fails to enumerate devices, can try with reverting that commit if you'd like
[2022-02-21T18:18:54Z] <illiliti> WLR_DRM_NO_MODIFIERS=1
[2022-02-21T18:24:24Z] <dilyn> you're making me try too hard lmao
[2022-02-21T18:25:08Z] <illiliti> xd
[2022-02-21T18:32:25Z] <dilyn> off we go again :X
[2022-02-21T18:54:21Z] <dilyn> illiliti: no changes
[2022-02-21T19:04:33Z] <illiliti> ok. we can debug this if you want to continue
[2022-02-21T19:04:43Z] <illiliti> https://termbin.com/3y33
[2022-02-21T19:13:59Z] <dilyn> do you want me to stick with sway or go back to wayfire?
[2022-02-21T19:14:11Z] <dilyn> because i'll have to first get sway working lmfao
[2022-02-21T19:16:05Z] <illiliti> doesn't matter
[2022-02-21T19:16:55Z] <dilyn> perfecto. more than happy to help!
[2022-02-21T19:17:15Z] <illiliti> thanks!
[2022-02-21T20:32:13Z] <rohan> someone here using kde in kiss?
[2022-02-21T20:37:18Z] <rohan> dilyn: kiss KDE is working great?
[2022-02-21T20:39:51Z] <dilyn> last I checked, not so much :\
[2022-02-21T20:40:06Z] <dilyn> issue with plasmashell crashing or simply not doing what it should when using kwin
[2022-02-21T20:40:28Z] <dilyn> but if you set KDE_WM=sowm or some other wm it works quite well, you just won't have the traditional plasma desktop
[2022-02-21T22:44:34Z] <rohan> sad kde will not work
[2022-02-21T22:44:47Z] <rohan> then i will give up my project
[2022-02-21T22:58:48Z] <dilyn> @illiliti: this is interesting; http://ix.io/3QkK
[2022-02-21T22:59:13Z] <dilyn> we can also move this to github if you'd rather track the issue there