💾 Archived View for gemini.ctrl-c.club › ~phoebos › logs › freenode-kisslinux-2020-10-26.txt captured on 2024-05-10 at 14:32:13.

View Raw

More Information

⬅️ Previous capture (2021-12-17)

-=-=-=-=-=-=-

2020-10-26T00:05:17 #kisslinux <onodera> im using cock.li lol
2020-10-26T00:05:36 #kisslinux <onodera> as my primary email which is probably kinda retarded since they are down 60% of the time
2020-10-26T00:12:31 #kisslinux <dilynm> midfavila: I think KISS-kde has packaged shared-mime-info? And it might be the case that it was packaged on KISS long, long ago...
2020-10-26T00:17:15 #kisslinux <muevoid> E5ten you were the one that helped with the previous linker error on llvm/clang
2020-10-26T00:17:47 #kisslinux <E5ten> yeah? is there another issue?
2020-10-26T00:18:48 #kisslinux <muevoid> Different project now trying to get a sandbox game called the powder toy to compile
2020-10-26T00:19:04 #kisslinux <muevoid> just was wondering if you could give some insite
2020-10-26T00:19:07 #kisslinux <muevoid> insight*
2020-10-26T00:19:28 #kisslinux <E5ten> I can definitely try, what's the problem?
2020-10-26T00:19:45 #kisslinux <muevoid> Fails with undefined symbol for a libary called fftw which I installed from community
2020-10-26T00:20:16 #kisslinux <E5ten> can you send a link to the build log?
2020-10-26T00:20:25 #kisslinux <muevoid> https://termbin.com/czct
2020-10-26T00:20:49 #kisslinux <muevoid> Oh hold on
2020-10-26T00:20:57 #kisslinux <muevoid> Didn't show the linker errors for some reason
2020-10-26T00:21:29 #kisslinux <muevoid> there https://termbin.com/vdt2
2020-10-26T00:29:17 #kisslinux <muevoid> Right now trying the CXXFLAGS="$CXXFLAGS -fPIC" and CFLAGS="$CFLAGS -fPIC"
2020-10-26T00:29:40 #kisslinux <muevoid> Didn't seem to help
2020-10-26T00:31:37 #kisslinux <E5ten> it seems like fftw provides different symbols for different precisions based on the flags it's built with?
2020-10-26T00:31:55 #kisslinux <E5ten> like --enable-single (or --enable-float which is the same thing) will make fftwf_* symbols
2020-10-26T00:32:03 #kisslinux <muevoid> Gotcha
2020-10-26T00:32:31 #kisslinux <E5ten>  looking at the package for archlinux, they build 4 different copies with different precisions (single, double which seems to be default, long double, and quad)
2020-10-26T00:32:51 #kisslinux <muevoid> So maybe I should add --enable-double to the kiss one?
2020-10-26T00:33:26 #kisslinux <E5ten> I think you'd need --enable-single for this to work, --enable-double is default
2020-10-26T00:34:09 #kisslinux <E5ten> but I think doing --enable-single will disable the symbols that are provided from --enable-double? unless arch is doing 4 builds unnecessarily, I think to get all the symbols you'd need 4 libs each compiled with the different precision options
2020-10-26T00:34:14 #kisslinux <E5ten> ridiculous lol
2020-10-26T00:34:24 #kisslinux <muevoid> lol
2020-10-26T00:34:26 #kisslinux <E5ten> but I googled powder toy and fftw and it seems you might be able to disable fftw?
2020-10-26T00:34:36 #kisslinux <muevoid> Hmmm really
2020-10-26T00:35:25 #kisslinux <E5ten> try using the option "nofft"
2020-10-26T00:35:40 #kisslinux <E5ten> like for the build system
2020-10-26T00:35:57 #kisslinux <muevoid> Trying that now
2020-10-26T00:36:08 #kisslinux <muevoid> I didn't even notice that if so cause I looked there for a luajit option
2020-10-26T00:38:00 #kisslinux <muevoid> That worked thank you!
2020-10-26T00:39:02 #kisslinux <E5ten> no problem lol, glad it didn't end up being a linker issue cuz that would probably have been a much bigger pain in the ass
2020-10-26T00:39:09 #kisslinux <muevoid> Yeah lmao
2020-10-26T00:50:52 #kisslinux <dilynm> fftw is a bitch of a package
2020-10-26T00:51:01 #kisslinux <muevoid> What even is it?
2020-10-26T00:51:03 #kisslinux <dilynm> Building four packages is no exaggeration
2020-10-26T00:51:05 #kisslinux <dilynm> https://github.com/dilyn-corner/KISS-kde/blob/master/extra/fftw/build
2020-10-26T00:51:21 #kisslinux <dilynm> It's just a bunch of math libs
2020-10-26T00:51:27 #kisslinux <dilynm> For fast courier transforms
2020-10-26T00:51:34 #kisslinux <dilynm> Fourier*
2020-10-26T00:51:46 #kisslinux <dilynm> Useful for things like audio stuff
2020-10-26T00:52:16 #kisslinux <muevoid> Gotcha
2020-10-26T00:52:55 #kisslinux <dilynm> But since different audio processors make use of the different precision's, you're left either making four fftw packages or building it four times for a main package :eyeroll:
2020-10-26T00:53:59 #kisslinux <soliwilos> muevoid: wayherb's performance is much better now. :-)
2020-10-26T00:54:23 #kisslinux <muevoid> That's good I was just being stupid
2020-10-26T01:01:14 #kisslinux <muevoid> Again if you ever would like to see something just lmk
2020-10-26T01:03:24 #kisslinux <soliwilos> Thank you, I will keep that in mind. :-)
2020-10-26T13:23:00 #kisslinux <ominous_anonymou> Can anyone think of some fun workarounds to Firefox not supporting ALSA any more?  Like I don't know, making a minimal container that runs firefox in a virtual X session with a pulseaudio server?
2020-10-26T13:23:50 #kisslinux <icy> That sounds too painful to even bother with.
2020-10-26T13:23:56 #kisslinux <ominous_anonymou> :'(
2020-10-26T13:24:11 #kisslinux <icy> Also woah, they don't support ALSA anymore?
2020-10-26T13:24:14 #kisslinux <icy> That's retarded.
2020-10-26T13:24:40 #kisslinux <icy> Mozilla and making bad decisions.  Name a better duo.
2020-10-26T13:25:59 #kisslinux <ominous_anonymou> yeah it's been over a year now since they "officially" stopped support.  people have still been able to compile with alsa enabled since they at least didn't _remove_ code, but it's been getting buggier and buggier
2020-10-26T13:26:09 #kisslinux <sh4rm4^bnc> i'd check the commit log, find the commit that removed the support and revert it
2020-10-26T13:26:11 #kisslinux <icy> That's rough.
2020-10-26T13:26:55 #kisslinux <icy> Audio on Linux is a shitshow, tbh.
2020-10-26T13:27:10 #kisslinux <sh4rm4^bnc> there's probably also patches flying around on the bugtracker or on other distro's repos
2020-10-26T13:27:36 #kisslinux <sh4rm4^bnc> icy, it works very well for me since i got rid of pulseaudio :)
2020-10-26T13:27:37 #kisslinux <ominous_anonymou> https://freenode.logbot.info/kisslinux/20200423#c3711184
2020-10-26T13:27:58 #kisslinux <ominous_anonymou> yeah i've had zero issues with alsa, and i don't have to deal with pulseaudio at all
2020-10-26T13:28:17 #kisslinux <onodera> there is always apulse
2020-10-26T13:28:29 #kisslinux <onodera> whichs translates the pulse calse to alsa ones or something
2020-10-26T13:28:32 #kisslinux <ominous_anonymou> i tried apulse, didn't work for me
2020-10-26T13:28:44 #kisslinux <onodera> though my firefox sound stopped working since a week, anyone else having this problem?
2020-10-26T13:28:58 #kisslinux <sh4rm4^bnc> does it inject itself too via LD_PRELOAD like pulseaudio does?
2020-10-26T13:29:05 #kisslinux <ominous_anonymou> yeah mine stopped too that's kind of why i was asking.  i figured they finally broke support
2020-10-26T13:29:24 #kisslinux <ominous_anonymou> https://github.com/i-rinat/apulse
2020-10-26T13:29:38 #kisslinux <soliwilos> Sound in Firefox works for me, with only alsa.
2020-10-26T13:29:38 #kisslinux <onodera> i think you can also compile firefox with openbsd's sndio as a last resort
2020-10-26T13:29:45 #kisslinux <icy> How does is stop working all of a sudden?  Unless you updated Firefox.
2020-10-26T13:29:52 #kisslinux <ominous_anonymou> what version firefox, soliwilos?
2020-10-26T13:30:00 #kisslinux <icy> Yeah, I'm on OpenBSD, using sndio.
2020-10-26T13:30:06 #kisslinux <icy> Is sndio ported to Linux?
2020-10-26T13:30:09 #kisslinux <ominous_anonymou> and yes icy i updated my system which included firefox
2020-10-26T13:30:11 #kisslinux <onodera> I don't know why, it didn't stop for all videos, only for most
2020-10-26T13:30:17 #kisslinux <icy> ominous_anonymou: Ah.
2020-10-26T13:30:22 #kisslinux <onodera> maybe I updated
2020-10-26T13:30:28 #kisslinux <soliwilos> Version 82.
2020-10-26T13:31:00 #kisslinux <sh4rm4^bnc> might also be an issue with a alsa-lib update
2020-10-26T13:31:16 #kisslinux <ominous_anonymou> hm that's the one i installed, i'll try reinstalling
2020-10-26T13:31:19 #kisslinux <icy> Y'all should try this: https://sndio.org/install.html
2020-10-26T13:31:37 #kisslinux <ominous_anonymou> shoot i see that amention above but didn't check whether mine got updated sh4rm4^bnc
2020-10-26T13:31:52 #kisslinux <sh4rm4^bnc> if you updated firefox AND alsa-lib, it's kinda hard to tell what is actually the issue...
2020-10-26T13:32:06 #kisslinux <icy> Have sndio as a dependency for Firefox, and compile against it.
2020-10-26T13:32:14 #kisslinux <icy> Should justwork.
2020-10-26T13:32:36 #kisslinux <ominous_anonymou> thanks icy, if reinstall and checking alsa-lib doesn't work i'll try that
2020-10-26T13:33:07 #kisslinux <onodera> btw firefox-bin is gone?
2020-10-26T13:33:29 #kisslinux <soliwilos> My alsa-lib is version 1.2.4
2020-10-26T13:34:32 #kisslinux <sh4rm4^bnc> mine is 1.1.8 and it works :)
2020-10-26T13:35:01 #kisslinux <sh4rm4^bnc> although with a patch: alsa-lib-1.1.2-avoid_forking.patch
2020-10-26T13:36:15 #kisslinux <ominous_anonymou> probably on me for blindly updating
2020-10-26T13:37:12 #kisslinux <konimex> onodera: aye, dylan doesn't provide it anymore
2020-10-26T13:37:34 #kisslinux <mcpcpc[m]> onodera: it appears so.  i was enjoying having a rust-free system. :P
2020-10-26T13:38:50 #kisslinux <onodera> yeah I don't care much about firefox but fucking rust
2020-10-26T13:39:15 #kisslinux <mcpcpc[m]> yup
2020-10-26T13:39:37 #kisslinux <ominous_anonymou> ok firefox-bin at 81.0 and alsa-lib at 1.2.4 work.  thanks guys
2020-10-26T13:40:12 #kisslinux <icy> Why is Rust so shit
2020-10-26T13:40:42 #kisslinux <icy> Something about the language and its absolutist community that puts me off.
2020-10-26T13:42:04 #kisslinux <konimex> honestly rust might be not so shit if its bootstrap system was more like Go
2020-10-26T13:42:39 #kisslinux <icy> Agreed.
2020-10-26T13:43:29 #kisslinux <onodera> seems like gentoo doesn't provide any alsa patches
2020-10-26T13:43:41 #kisslinux <onodera> though a lot of musl patches that we don't apply...
2020-10-26T13:45:00 #kisslinux <soliwilos> I'm only using the no-gtk2 patch for Firefox.
2020-10-26T13:55:25 #kisslinux <onodera> also this retarded rust shit failed after 40 minutes of compiling
2020-10-26T13:55:30 #kisslinux <onodera> ugggggg
2020-10-26T14:00:33 #kisslinux <icy> lol
2020-10-26T14:00:58 #kisslinux <icy> What's keeping you from using WebKit-based browsers?
2020-10-26T14:03:50 #kisslinux <onodera> not customizable enough
2020-10-26T14:06:15 #kisslinux <icy> I think codebase-wise, and autism-wise, WebKit is probably the best.
2020-10-26T14:07:05 #kisslinux <icy> You can relatively easily write a frontend wrapper around it, or use some existing ones like Epiphany or Midori.
2020-10-26T14:07:14 #kisslinux <icy> Or vimb etc.
2020-10-26T14:12:25 #kisslinux <onodera> i tried most browsers probably
2020-10-26T14:12:31 #kisslinux <onodera> i always end up returning to firefox
2020-10-26T14:15:00 #kisslinux <sh4rm4^bnc> you tried palemoon too ?
2020-10-26T14:28:16 #kisslinux <uhrenmacher> join/join #emacs
2020-10-26T14:52:59 #kisslinux <jedavies> Work in progress for anyone interested in KISS with clang/llvm on different architectures: https://www.glasnost.org/
2020-10-26T15:05:33 #kisslinux <ominous_anonymou> oo thanks jedavies !
2020-10-26T15:35:44 #kisslinux <micr0> ominous_anonymou: my firefox 82 youtube sound is working fine
2020-10-26T15:36:45 #kisslinux <micr0> as is peertube
2020-10-26T15:40:10 #kisslinux <micr0> onodera care to share the (anki?) cli app to practice japanese in your screenshot? https://i.redd.it/puxe57tylhs41.png
2020-10-26T15:43:00 #kisslinux <mcpcpc[m]> giving up on firefox.. been building src+dependancies for 2hrs+.  back to surf. sad to see *-bin gone. :’(.
2020-10-26T15:54:09 #kisslinux <micr0> mcpcpc[m] are you on X or wayland?
2020-10-26T15:54:43 #kisslinux <mcpcpc[m]> micr0: X. been meaning to test out wayland.
2020-10-26T15:57:01 #kisslinux <micr0> mcpcpc[m]: I just did `kiss export firefox` and uploaded it to http://0x0.st/ik_G.gz
2020-10-26T15:57:26 #kisslinux <micr0> sha256sum: 4c27624754d4e12549ec08a4afdef60282c1a2e09ddf523f4fb85450f1551532  firefox#82.0-1.tar.gz
2020-10-26T15:57:50 #kisslinux <micr0> I have only tested it on wayland
2020-10-26T15:59:42 #kisslinux <mcpcpc[m]> micr0: will test it out! tyty!
2020-10-26T16:53:08 #kisslinux <onodera> micro what wm do you use on wayland
2020-10-26T16:53:14 #kisslinux <onodera> or compositor or whatever it is called
2020-10-26T17:15:06 #kisslinux <micr0> onodera sway
2020-10-26T17:15:18 #kisslinux <micr0> i miss subtle though, that was a nice wm back in the day
2020-10-26T17:15:24 #kisslinux <micr0> hikari looks okay to
2020-10-26T17:15:57 #kisslinux <dilynm> It's the best
2020-10-26T17:26:27 #kisslinux <micr0> yeah i have like a never-ending list of improvements to workflow etc, trying out hikari is kinda lower on that list
2020-10-26T17:27:12 #kisslinux <micr0> by the time evening rolls around i just wanna play videogames
2020-10-26T17:34:07 #kisslinux <midfavila> Anyone have luck getting LXAppearance running? I can configure and compile it just fine, but when running (even a debug-enabled copy) there's no output and the program appears to "exit" successfully without anything happening
2020-10-26T17:51:51 #kisslinux <onodera> is there an st like terminal for wayland
2020-10-26T17:55:25 #kisslinux <jedavies> Have been using foot recently instead of alacritty on wayland. At least it doesn't require rust to build!
2020-10-26T17:55:55 #kisslinux <midfavila> I wonder if they've started including Rust in the kernel yet...
2020-10-26T17:56:04 #kisslinux <midfavila> that would be annoying
2020-10-26T17:59:40 #kisslinux <dilynm> Isn't it in next?
2020-10-26T17:59:50 #kisslinux <dilynm> Wayst is an st for Wayland
2020-10-26T18:00:03 #kisslinux <midfavila> I haven't checked. I normally run stable kernels
2020-10-26T18:00:46 #kisslinux <midfavila> if they __do__ introduce Rust, I'm going to seriously consider HyperbolaBSD for my main PC once it releases
2020-10-26T18:05:44 #kisslinux <soliwilos> I quite like wayst.
2020-10-26T18:14:26 #kisslinux <micr0> onodera i would second foot being a good alternative to alacritty
2020-10-26T18:15:26 #kisslinux <micr0> I don't care what language the developers of my kernel use, as I'm not really planning on contributing to said kernel in my lifetime
2020-10-26T18:18:47 #kisslinux <ominous_anonymou> i would care if it made the compile time a lot longer
2020-10-26T18:19:50 #kisslinux <ominous_anonymou> building new versions of rust is already horrendously long and forces me to up my KISS VM's RAM to 4GB+ or it shits the bed.  i'd lose my mind if the kernel did the same.
2020-10-26T18:21:52 #kisslinux <midfavila> :^)
2020-10-26T18:22:58 #kisslinux <micr0> yeah if you are upgrading your kernel a lot and don't like to wait that would be annoying. I also doubt kernel devs would incorporate rust if build times swelled to even 2x what it is now, let alone what I would expect to be more.
2020-10-26T18:28:07 #kisslinux <eudaldgr> Rust into the kernel? what a shit
2020-10-26T18:30:32 #kisslinux <midfavila> just wait until systemd integrates Linux as a module
2020-10-26T18:30:37 #kisslinux <midfavila> systemd-linuxd
2020-10-26T18:31:32 #kisslinux <dilynm> Luckily the kernel devs have evaded that so far
2020-10-26T18:31:41 #kisslinux <midfavila> i'm calling it rn
2020-10-26T18:40:19 #kisslinux <micr0> any reason my laptop would not respond to ping by default?
2020-10-26T18:47:19 #kisslinux <onodera> eh can't be bothered to learn all this new stuff
2020-10-26T18:47:24 #kisslinux <onodera> ill stick to X for the next few year
2020-10-26T18:48:59 #kisslinux <muevoid> I'm considering switching back to vim but one thing I really like from emacs is the subscript and superscript characters when there are multiple of them they actually look like one number for example 22 but in vim it will look like two seperate twos next to each other.
2020-10-26T19:43:18 #kisslinux <soliwilos> micr0: ping replies can be disabled by sysctl, if net.ipv4.icmp_echo_ignore_all is set to 1, but I think the default should be 0.
2020-10-26T19:51:55 #kisslinux <mcpcpc[m]> soliwilos: oh. this is a useful tidbit.
2020-10-26T21:08:29 #kisslinux <dilynm> The Makefiles automake generates are disgusting
2020-10-26T22:31:17 #kisslinux <E5ten> @freenode_dylanaraps:matrix.org: for replacing nftw with something fully POSIX, I've got this http://ix.io/2C98, it's probably got some problems, and definitely needs some cleanup because I adapted it from something else I used (e.g., the return values are pretty useless in present state), but it's a start I think
2020-10-26T22:34:16 #kisslinux <E5ten> might be better to use scandir instead of readdir so it's done alphabetically
2020-10-26T22:35:22 #kisslinux <mcf> E5ten: one thing you could do is use openat and fdopendir to avoid path allocation and construction
2020-10-26T22:35:25 #kisslinux <E5ten> and probably you could do something like passing dirfd's instead of pathnames, and using fdopendir instead of opendir, fstatat instead of lstat, unlinkat instead of remove, to avoid having to allocate and construct a full path for each time
2020-10-26T22:35:37 #kisslinux <mcf> heh
2020-10-26T22:35:42 #kisslinux <E5ten> lol
2020-10-26T22:38:16 #kisslinux <E5ten> oh there is an issue though, apparently you can't safely use the fd you pass to fdopendir after the fdopendir call? so you wouldn't be able to use that fd for fstatat and unlinkat?
2020-10-26T22:39:32 #kisslinux <E5ten> or actually, the dirfd man page says you can use the internal fd used by the DIR for stuff that doesn't depend on/alter the file position, like fstat, so maybe you can still use the fd?
2020-10-26T22:42:05 #kisslinux <mcf> yeah, i would assume that using the fd for *at functions do not "modify the state of the associated description"
2020-10-26T22:44:00 #kisslinux <E5ten> yeah but I just don't know if being allowed to do that for dirfd's return means you're also allowed to do it with the fd you pass to fdopendir
2020-10-26T22:45:33 #kisslinux <mcf> fdopendir only prohibits closing the fd or modifying the state of the state of the associated description (other than functions that act on the DIR)
2020-10-26T22:45:46 #kisslinux <E5ten> ah ok, but looking at the POSIX man page instead of the linux one, I think it seems like you can do that, cuz that man page doesn't say you're not allowed to do anything with the fd, just not allowed to modify (like you just said)
2020-10-26T22:46:08 #kisslinux <E5ten> the "but" at the beginning there is in response to my last message not yours to be clear
2020-10-26T22:53:56 #kisslinux <E5ten> mcf: what would I use instead of "remove(path)" at the end of the function?
2020-10-26T22:54:42 #kisslinux <E5ten> actually that's kind of a problem, because there's probably issues with removing the dir before doing closedir, but after closedir, the dirfd is invalid
2020-10-26T22:55:47 #kisslinux <E5ten> would unlinkat with the dirfd, "." as pathname, and AT_REMOVEDIR work?
2020-10-26T22:56:07 #kisslinux <E5ten> I think probably not
2020-10-26T23:02:35 #kisslinux <mcf> oh, hmm. this reminds me of a musl thread a while ago: https://inbox.vuxu.org/musl/20191121175418.GI4262⊙me/T/
2020-10-26T23:05:13 #kisslinux <E5ten> hmmm
2020-10-26T23:05:44 #kisslinux <E5ten> wouldn't a kernel issue like that probably also affect nftw (which I'm guessing probably uses the dirent stuff)?
2020-10-26T23:06:27 #kisslinux <mcf> well, i think this was just an issue for buggy filesystems
2020-10-26T23:06:35 #kisslinux <E5ten> I see
2020-10-26T23:10:19 #kisslinux <E5ten> I assume given that you maintain sbase and recommended the dirfd approach that sbase also uses dirfds and *at functions, how does it deal with deleting the top-level dir?
2020-10-26T23:11:10 #kisslinux <E5ten> for subdirs, I think since I have the dirfd and the file name, I can just unlinkat them after calling rm_rf on them (within rm_rf), but I'm not really sure how to delete the top-level dir
2020-10-26T23:11:58 #kisslinux <mcf> i don't quite understand how the subdirs is different from the top-level dir. could you paste what you have so far?
2020-10-26T23:13:18 #kisslinux <E5ten> in a second yeah, but what I mean is that for deleting subdirs, I can do it outside the rm_rf on that subdir, like higher level rm_rf calls rm_rf on a subdir, and then deletes the subdir (so the rm_rf on the subdir itself doesn't delete it). what I'd really want is the rm_rf call to delete the subdir itself, and if it did that there'd be no difference from subdirs and top dir
2020-10-26T23:14:50 #kisslinux <E5ten> http://ix.io/2C9n I haven't tested this one yet (I did with the non-dirfd one), and it won't even compile because there's still remove(path) at the end, because I don't know how to delete the dir after going through it, so it's a placeholder
2020-10-26T23:15:25 #kisslinux <E5ten> also the openat call should probably have O_DIRECTORY or something
2020-10-26T23:16:05 #kisslinux <mcf> it could also have O_SEARCH
2020-10-26T23:16:06 #kisslinux <E5ten> actually, maybe that'd mean I don't need the fstatat, because if I openat with O_DIRECTORY, and errno is whatever value means "not a directory" I treat it as a file, and if not I call rm_rf on it
2020-10-26T23:16:32 #kisslinux <E5ten> I don't see that in the open man page?
2020-10-26T23:18:47 #kisslinux <mcf> i'd move the openat to start of rm_rf, and make it take a *parent* dir fd and a name of a subdirectory to remove. then at the end you do unlinkat(parentfd, name)
2020-10-26T23:19:26 #kisslinux <E5ten> so I'd use fdopendir on the result of the openat at the beginning?
2020-10-26T23:19:34 #kisslinux <mcf> yeah
2020-10-26T23:20:55 #kisslinux <E5ten> and I guess for top-level I'd pass AT_FDCWD as the parent dir fd? or maybe to make the API simpler and not reliant on AT_* it could take a negative value for the parent dir fd, and if that happens use AT_FDCWD for the openat
2020-10-26T23:21:37 #kisslinux <mcf> yeah, whatever you prefer
2020-10-26T23:38:23 #kisslinux <E5ten> http://ix.io/2C9s haven't tested it very thoroughly, but it does seem to work
2020-10-26T23:39:21 #kisslinux <E5ten> very likely needs at least more cleanup, if not important fixes/changes
2020-10-26T23:42:35 #kisslinux <E5ten> possible thing that could need fixing, since this is supposed to be like rm -rf, could need to fchmodat on files without write perms
2020-10-26T23:48:18 #kisslinux <E5ten> hmm and since the use of nftw in k uses FTW_MOUNT, I might want to save the .st_dev of the top-level dir somehow and make sure it's the same for each file I fstatat, and if it's different bail or skip