💾 Archived View for gemini.ctrl-c.club › ~phoebos › logs › freenode-kisslinux-2020-10-03.txt captured on 2024-05-12 at 16:07:29.
⬅️ Previous capture (2021-12-17)
-=-=-=-=-=-=-
2020-10-03T00:05:45 #kisslinux <muevoid> Ah it was just my bad 2020-10-03T02:31:59 #kisslinux <muevoid> Are atk, gdk-pixbuff, and pango supposed to clone glib when building? 2020-10-03T02:35:02 #kisslinux <sh4rm4^bnc> as in git clone? 2020-10-03T02:35:36 #kisslinux <muevoid> Not sure just starts to git clone glib when building 2020-10-03T02:35:49 #kisslinux <muevoid> clone* 2020-10-03T02:35:53 #kisslinux <muevoid> Says cloning glib 2020-10-03T02:36:59 #kisslinux <sh4rm4^bnc> lol, is that a feature of that new meson build system? certainly doesn't do that in the versions i use 2020-10-03T02:37:19 #kisslinux <muevoid> Possibly that is it 2020-10-03T02:37:22 #kisslinux <sh4rm4^bnc> in fact, sabotage even has a feature to supress network connections for package builds 2020-10-03T02:37:45 #kisslinux <sh4rm4^bnc> so we catch packages doing nasty stuff like that 2020-10-03T02:51:49 #kisslinux <muevoid> I'll be back later 2020-10-03T02:51:51 #kisslinux <muevoid> o/ 2020-10-03T03:01:37 #kisslinux <dylanaraps> Added sh to bash so 'kiss a bash /usr/bin/sh' can be used. 2020-10-03T03:15:08 #kisslinux <dylanaraps> muevoid: Send log for binutils. I cannot reproduce. 2020-10-03T03:15:41 #kisslinux <dylanaraps> muevoid: They clone the dependencies when they aren't found in the system. 2020-10-03T03:27:55 #kisslinux <sh4rm4^bnc> even if not in the expected version? 2020-10-03T03:35:22 #kisslinux <dylanaraps> Haven't tried 2020-10-03T03:38:44 #kisslinux <dylanaraps> I have an almost fully "static" kiss now ;) 2020-10-03T03:38:47 #kisslinux <sh4rm4^bnc> is that meson doing it? 2020-10-03T03:38:52 #kisslinux <dylanaraps> Yes 2020-10-03T03:38:58 #kisslinux <sh4rm4^bnc> can this be disabled? 2020-10-03T03:40:03 #kisslinux <micr0> anyone understand gobject-introspection? theres a build script for a fwupd plugin that checks for certain python libraries but i cannot figure out where they are in pip 2020-10-03T03:40:16 #kisslinux <micr0> cairo, and pangocairo gobject-introspection stuff 2020-10-03T03:41:26 #kisslinux <sh4rm4^bnc> yes, the introspection stuff is needed for python bindings to these libs 2020-10-03T03:41:37 #kisslinux <sh4rm4^bnc> in order to build it, python is required too 2020-10-03T03:42:57 #kisslinux <sh4rm4^bnc> basically glib has to be built with support for that turned on, and the gobject-introspection package 2020-10-03T03:43:17 #kisslinux <micr0> so i installed gobject-introspection, and have python as well as pygobject installed 2020-10-03T03:43:32 #kisslinux <micr0> but i am not sure if glib has gobject-introspection enabled; checking now 2020-10-03T03:49:40 #kisslinux <sh4rm4^bnc> hmm, can't see any introspection-related stuff in my glib build config. but it seems to automatically build dbus-introspection stuff when python is installed 2020-10-03T03:50:36 #kisslinux <sh4rm4^bnc> note that i went to great lengths to remove the python dependency of glib, e.g. by forward-porting the old mkenums.c program 2020-10-03T03:52:33 #kisslinux <sh4rm4^bnc> anyway after you have gobject-introspection, you gotta rebuild e.g. cairo or whatever that python plugin is trying to use 2020-10-03T03:53:39 #kisslinux <sh4rm4^bnc> cairo will then automatically create the bindings when it detects the toolkit 2020-10-03T04:17:12 #kisslinux <dylanaraps> I have a package locally called 'kiss-static' which creates a kiss binary including the package manager itself and whatever it needs from busybox. A "static" kiss in other words. 2020-10-03T04:18:40 #kisslinux <dylanaraps> See: https://termbin.com/ckmq 2020-10-03T04:19:10 #kisslinux <dylanaraps> This is configured to use gnugrep for reference. 2020-10-03T04:20:52 #kisslinux <dylanaraps> Size difference: https://termbin.com/e84rq 2020-10-03T04:48:34 #kisslinux <muevoid> after doing some testing gtk4 is not usable yet at least with webkit2gtk 2020-10-03T04:50:21 #kisslinux <muevoid> It also requires a few more packages 2020-10-03T04:51:35 #kisslinux <dylanaraps> libhandy probably? 2020-10-03T04:51:54 #kisslinux <muevoid> No libsass and sassc 2020-10-03T04:52:01 #kisslinux <dylanaraps> kek 2020-10-03T04:52:06 #kisslinux <muevoid> It's a css compile library :( 2020-10-03T04:52:09 #kisslinux <dylanaraps> Yes 2020-10-03T04:52:19 #kisslinux <dylanaraps> I've used it extensively 2020-10-03T04:52:26 #kisslinux <muevoid> Also graphene 2020-10-03T04:52:30 #kisslinux <muevoid> Is a new dep 2020-10-03T04:52:40 #kisslinux <dylanaraps> Just thank god that a C impl exists and you don't have to use ruby or nodejs for it =) 2020-10-03T04:52:46 #kisslinux <muevoid> That's true 2020-10-03T04:52:59 #kisslinux <muevoid> I hate that webkit2gtk needs ruby 2020-10-03T04:53:17 #kisslinux <dylanaraps> Just a build time dependency 2020-10-03T04:53:22 #kisslinux <muevoid> Yeah I know 2020-10-03T04:53:28 #kisslinux <dylanaraps> Not a big one either 2020-10-03T04:53:31 #kisslinux <muevoid> I wish I could ditch it completely though :P 2020-10-03T04:53:50 #kisslinux <dylanaraps> You could replace all ruby with shell/awk/whatever 2020-10-03T04:54:08 #kisslinux <muevoid> I don't have that big of an issue with ruby cause I don't know it well enough I just prefer to keep the number of languages on my laptop to a min 2020-10-03T04:54:30 #kisslinux <muevoid> I may look through what ruby is used for in webkit2gtk 2020-10-03T04:54:43 #kisslinux <dylanaraps> Build scripts iirc 2020-10-03T04:54:59 #kisslinux <muevoid> Ah 2020-10-03T04:55:57 #kisslinux <dylanaraps> I'm going to start preparing things for the new package manager release 2020-10-03T04:56:10 #kisslinux <muevoid> I'm trying to work on making webkit2gtk 'less' bloated 2020-10-03T04:56:19 #kisslinux <muevoid> Any major things? 2020-10-03T04:56:30 #kisslinux <dylanaraps> Yeah 2020-10-03T04:56:31 #kisslinux <dylanaraps> Lots 2020-10-03T04:56:51 #kisslinux <dylanaraps> Will all be in the changelog 2020-10-03T04:56:54 #kisslinux <muevoid> Any thing that packages will need to do different? 2020-10-03T04:56:59 #kisslinux <dylanaraps> Nope 2020-10-03T04:57:14 #kisslinux <muevoid> Excited to see it as always 2020-10-03T04:59:28 #kisslinux <muevoid> Whenever I do a reinstall I notice my pkg count go down and I don't know how 2020-10-03T04:59:33 #kisslinux <muevoid> 20 less packages 2020-10-03T04:59:38 #kisslinux <muevoid> But I have no clue what they were 2020-10-03T05:00:45 #kisslinux <sh4rm4^bnc> why you're reinstalling? 2020-10-03T05:01:32 #kisslinux <muevoid> Cause I mess with things and my system gets "messy" 2020-10-03T05:01:51 #kisslinux <muevoid> Plus I make it super easy on myself 2020-10-03T05:02:07 #kisslinux <sh4rm4^bnc> i've got my sys since 4 years, never reinstalled 2020-10-03T05:02:10 #kisslinux <muevoid> backup important etc files, leave /boot, and make a tar of my home folder 2020-10-03T05:02:17 #kisslinux <muevoid> I hope to do that this time 2020-10-03T05:25:17 #kisslinux <mcpcpc[m]> Can anyone on Wayland verify if this change to foot was effective? https://codeberg.org/dnkl/foot/pulls/153. i think you just need to `echo -e “x1b[?45h”` in a new terminal window, run `cat`, and see if you can type and delete multiple lines =S. 2020-10-03T05:27:59 #kisslinux <dylanaraps> Sure 2020-10-03T05:29:35 #kisslinux <dylanaraps> Works fine 2020-10-03T05:29:50 #kisslinux <dylanaraps> Shorter command for sequence: printf '033[?45h' 2020-10-03T05:30:44 #kisslinux <dylanaraps> One thing I can notice: Holding a key down to fill lines scrolls the terminal. Holding backspace to clear lines maintains new scroll position. 2020-10-03T05:30:50 #kisslinux <dylanaraps> Not sure if it should scroll back up or not. 2020-10-03T05:31:06 #kisslinux <dylanaraps> Either way, multiple lines can be deleted 2020-10-03T05:32:00 #kisslinux <dylanaraps> mcpcpc[m]: 2020-10-03T05:35:04 #kisslinux <mcpcpc[m]> dylanaraps: thanks! that is a “strange”. will send a note to the developer. 2020-10-03T05:36:04 #kisslinux <dylanaraps> Yeah. Not sure if it should do that or not. 2020-10-03T05:36:47 #kisslinux <dylanaraps> Let me test a few more things (resize, etc) 2020-10-03T05:37:52 #kisslinux <dylanaraps> Resize works fine. 2020-10-03T05:38:48 #kisslinux <dylanaraps> Another thing. If I resize sometimes there's a case where cursor is at end of line. Character is inserted, but cursor does not move to next line. 2020-10-03T05:38:51 #kisslinux <dylanaraps> If that makes sense. 2020-10-03T05:39:14 #kisslinux <dylanaraps> The cursor ceases to be /in front of/ the inputted char for a single key press 2020-10-03T05:39:51 #kisslinux <dylanaraps> This happens /after/ resize. Works fine prior. 2020-10-03T05:40:53 #kisslinux <dylanaraps> Actually. I can "randomly" reproduce this. At some sizes it seems to work and at others it doesn't(?) 2020-10-03T05:42:39 #kisslinux <dylanaraps> Also. When line wrapping is disabled in the terminal, you can input only up to end of line (as intended). However, foot allows you to backspace N number of keypresses backwards. 2020-10-03T05:43:12 #kisslinux <dylanaraps> Example: Line is 20 chars long. Line wrapping is disabled. I input 40 chars. Cursor stops at EOL. I can now backspace 40 chars even though only 20 were displayed 2020-10-03T05:44:02 #kisslinux <dylanaraps> I don't know how much of this is intended/normal behavior for reference. 2020-10-03T05:44:28 #kisslinux <mcpcpc[m]> dylanaraps: wow. ok. the reverse-wraparound terminfo flag is strange. i imagine that it has something to do with what flags foot has implemented (as opposed to xterm, and what they set as default). 2020-10-03T05:44:54 #kisslinux <mcpcpc[m]> xterm or urxvt* 2020-10-03T05:45:05 #kisslinux <mcpcpc[m]> this is all very new to me. lol. 2020-10-03T05:45:54 #kisslinux <dylanaraps> OK. With normal foot (master), backspace + resize is really funky. 2020-10-03T05:47:37 #kisslinux <dylanaraps> Behavior of backspace while line wrapping disabled works fine except for the fact that the /last/ char on the line is not cleared by backspace 2020-10-03T05:47:45 #kisslinux <dylanaraps> (using normal master) 2020-10-03T05:56:17 #kisslinux <mcpcpc[m]> i believe that’s expected behavior in foot for normal master. st is slightly different (also does not have reverse-linewrap implemented). 2020-10-03T06:04:01 #kisslinux <mcpcpc[m]> my take away from all of this... don’t force “normal” text editing in canonical mode. no solution will be portable enough due to the variation in default flags from TE to TE. just use raw mode and be done with it... that was a ~2 week hard lesson learned. lol 2020-10-03T06:04:47 #kisslinux <mcpcpc[m]> (a.k.a should have listened to E5ten) 2020-10-03T07:43:31 #kisslinux <E5ten> mcpcpc: so you're gonna switch to just using raw mode the whole time instead of switching in and out? 2020-10-03T11:21:00 #kisslinux <mcpcpc[m]> E5ten: I’m strongly considering it. it was pointed out to me that even a terminal emulator that strictly complies with standards will implicitly behave different cursor behavior (see ECMA-48 6.1.6). also the low adoption of the <bw> flag in terminal emulators is only ~30%, 2020-10-03T11:22:37 #kisslinux <mcpcpc[m]> * E5ten: I’m strongly considering it. it was pointed out to me that even a terminal emulator that strictly complies with standards will implicitly behave different cursor behavior (see ECMA-48 6.1.6). also the low adoption of the “bw” flag in terminal emulators is only ~30%, so that’s A LOT of VTEs that kirc just wont play nicely in. 2020-10-03T11:27:52 #kisslinux <mcpcpc[m]> err sorry for poor “english” (if you can call it that). 2020-10-03T11:29:25 #kisslinux <soliwilos> mcpcpc[m]: I only tried kirc on wayst for a little bit, but it seemed to work nicely. Will be trying that out more later to get a bit beyond just a first impression of "works". 2020-10-03T11:46:53 #kisslinux <mcpcpc[m]> soliwilos: nice! glad to here. let me know if you see anything weird. so far, urxvt and xterm (with “bw” flag set) seem to work the best. foot will also behave like urxvt once the PR is merged. 2020-10-03T11:53:36 #kisslinux <periish> Hello! 2020-10-03T11:53:48 #kisslinux <mcpcpc[m]> o/ 2020-10-03T12:02:59 #kisslinux <mcpcpc[m]> btw, this is a “nifty” tool and summary of the “line-wrapping” issue: https://github.com/mattiase/wraptest 2020-10-03T14:11:39 #kisslinux <noemata> ok. you people know something. is this a vulnerability, or not. #gnome, #fedora #gentoo, nobody cares: https://web.archive.org/web/20201002095820/https://bugs.gentoo.org/show_bug.cgi?id=685722 2020-10-03T14:12:32 #kisslinux <noemata> "it can be used as exploit" . not "clamav "thinks" it can be used as an exploit" 2020-10-03T14:27:19 #kisslinux <dylanaraps> noemata: From what I understand, gdk-pixbuf includes a test which can be exploited (and is why clamav picks up on it). It's a false-positive and intentional. Regardless... we don't run the test suite. 2020-10-03T14:29:52 #kisslinux <noemata> hey dylan, the master itself. 1st thx, 2nd, i do not care (anymore) about flaptpak & co, but i was a long time on something like silverblue. you say "it can be exploited" - why is it then a false postive. i just want to understand this strangeness 2020-10-03T14:40:08 #kisslinux <dylanaraps> noemata: I'll look deeper into this over the next couple of hours. 2020-10-03T14:40:26 #kisslinux <dylanaraps> I really just gave you a summary of the link you posted here. 2020-10-03T14:42:32 #kisslinux <noemata> thx dylan, in fact, after 2 hours on 4 linux channels, you're the only one who cares. i find this fuzzy logic resp. fragile superposition behind all this strange, and i do not understand, why you can't get any answers from gnome or bluehat devs 2020-10-03T15:36:55 #kisslinux <dylanaraps> noemata: gdk-pixbuf had buffer overflow with crafted input image and the loader would crash. This was found by a fuzzer. They fixed the issue and added some tests to enure that this doesn't regress. 2020-10-03T15:37:03 #kisslinux <dylanaraps> The source code includes said crafted input image for use by the tests. It doesn't make its way into gdk-pixbuf itself and is only used when running the test suite. 2020-10-03T15:43:00 #kisslinux <noemata> @dylanaraps big thx for your research. therfore, it is not some "janus"-thing with 2 faces. it is designed to fix/close something malicious, but at the cost of open another malicious path at the same time? sry for my en 2020-10-03T15:44:14 #kisslinux <noemata> "malicious" is not the proper term, but i hope ypu know what i mean.. 2020-10-03T15:45:48 #kisslinux <noemata> maybe i' 2020-10-03T15:46:12 #kisslinux <dylanaraps> If you're worried about it, just don't run the test suite (or disable the relevant tests) 2020-10-03T15:47:35 #kisslinux <noemata> maybe i'm paranoid, because what i saw these 2 years with redhat & co. but i was right about ca-certificates too (thankfully this is not on kiss) 2020-10-03T15:49:34 #kisslinux <noemata> i will block anything flatpak related on my git-project, without explanation, anyway. but thx for your research, i'm not that concerened about this particular issue anymore 2020-10-03T15:50:49 #kisslinux <noemata> @dylanaraps i wish you the best, stay healthy and go on with kiss. big thx again! 2020-10-03T15:50:55 #kisslinux <dylanaraps> No problem 2020-10-03T17:05:49 #kisslinux <Rio6> LC_* on musl doesn't work yet right? 2020-10-03T17:09:01 #kisslinux <periish> Ello boo 2020-10-03T17:09:16 #kisslinux <periish> > LC_* on musl doesn't work yet right? 2020-10-03T17:09:21 #kisslinux <periish> No locales so I assume not 2020-10-03T17:09:39 #kisslinux <Rio6> ye ok 2020-10-03T18:23:36 #kisslinux <kciN> Trying to install kiss, couldn't find Dylan's key on keys.gnupg.net 2020-10-03T20:31:39 #kisslinux <soliwilos> What image viewers are people on wayland using? imv, mpv or something else? 2020-10-03T21:26:27 #kisslinux <onodera> i don't use wayland, but I use mpv with a profile that makes shortcuts and such like and imageviewer 2020-10-03T21:35:43 #kisslinux <soliwilos> Sticking to mpv for now, will look at making a profile for image-viewing.