00:05:52 <last1> how can I find which file names exist in two different directories ? ( and remove it from directory1 if it also exists in directory 2 )
00:09:27 <ngortheone> something along the lines of: for f in dir1/*; do test -f dir2/$ && rm dir1/$f; done
00:10:11 <jgh_> basename $f
00:10:12 <ngortheone> this will only remove file from dir1 if similarly named file exists in dir2
00:10:28 <jgh_> and just let the rm fail on nonexist
00:13:35 <ngortheone> jgh_ is right, you can do: for f in dir1/*; do test -f dir2/$(basename f$) && echo "file $(basename $f) exists in both dirs" && rm dir1/$(basename $f); done
00:13:59 <last1> nice, I had started using fdupes but I think it also compares file contents
00:14:03 <ek> You could toss an "if [ $? -gt 0 ]; then" in there to avoid the rm's that would fail.
00:14:05 <ngortheone> jgh_ given that you need to delete from dir1 how would you just let the rm fail on nonexist
00:14:08 <last1> jgh: is that tcsh or sh ?
00:14:15 <ngortheone> sh
00:14:49 <jgh_> just miss out the test ... &&
00:15:19 <jgh_> (I assume you don't really care about the echo)
00:15:34 <ngortheone> last1: there are many ways to skin this cat. names being equal doesn't mean files are the same. You might want to caclulate sha1 at least to be sure
00:15:41 <ngortheone> but it all depends on your usecase
00:15:49 <last1> I really just care about the names
00:16:07 <ngortheone> jgh_ is your solution then to iterate over dir2 and rm -f $file from dir1?
00:16:24 <jgh_> with basename, yup
00:16:58 <ngortheone> that would work too, rm -v to print out dupes, if accounting of deleted files is important
00:17:36 <ngortheone> last1: so the hivemind solution is
00:18:20 <ngortheone> for f in dir2/* ; do rm -v dir1/$(basename $f); done
00:20:18 <ngortheone> rm -vf to avoid error message
00:21:20 <jgh_> 'night y'all
00:26:48 <meena> rm -f never fails.
00:39:40 <ngortheone> -f not to avoid failure, but to avoid annyoing message "file does not exist"
08:51:22 <yuripv> it's both
09:47:48 <moon-child> I have a linux app which is running out of open fds.  I have raised the open file limit to a truly absurd 500m, but it still happens.  Is it possible I need to set the limit in a different way for linux apps?
09:48:25 <parv> (500m) lol💩
09:51:20 <parv> That point had come up recently (somewhere?). Only point I remember is that FreeBSD has too low of a default
09:51:32 <moon-child> if I run /compat/linux/bin/bash, then ulimit -n is a paltry 1024.  If I raise it to 500m from the linux bash, and then run the application again, I have the same problem.  But perhaps something in the linuxulator exec process breaks this?
09:52:08 <moon-child> do I need to trick the application into running setrlimit itself?
09:56:06 <parv> Points to that : https://www.cyberciti.biz/faq/linux-increase-the-maximum-number-of-open-files/
09:57:33 <moon-child> I don't that will be helpful for the linuxulator
09:58:47 <parv> If the application has the some setting to control the number of open files ...
09:59:41 <moon-child> I'm fine with its opening as many files as it wants.  I want to be able to raise the limits to allow that
10:14:39 <meena> the problem is that we're working thru layers
10:15:54 <moon-child> ah, seems perhaps I want compat.linux.default_openfiles
10:16:15 <meena> that a sysctl?
10:16:31 <moon-child> yeah
10:16:38 <meena> what's it say
10:18:02 <moon-child> nothing interesting.  'Default soft openfiles resource limit, or -1 for unlimited'
10:18:59 <moon-child> now it hangs on startup.  Improvement?
10:19:33 <meena> what did you set it to?
10:20:45 <parv> Is it possible to create a ZFS mirror out of 2 RAIDZ pools?
10:21:08 <moon-child> -1.  But an intermediate of 1000000 seems fine so far
10:21:32 <parv> OP (re ZFS mirror): https://mastodon.social/@shad⊙tl/109804803838302840
10:21:32 <VimDiesel> Title: Shad: "fellow techies, I have a zpool with a raidz2 of s…" - LGBTQIA+ Tech Mastodon
11:25:11 <meena> moon-child: there's every chance that Linux and FreeBSD have a different understanding of how big unlimited is, and do it's getting lost in translation (integer conversion)
11:27:39 <moon-child> this is a purely freebsd-specific thing, though
11:28:03 <meena> https://cathode.church/@meena/109769417124107787
11:28:05 <VimDiesel> Title: Mina: "c" - Cathode Church
11:28:48 <meena> moon-child: and you think it's completely tested and error free, too?
11:29:18 <moon-child> I never said that
11:29:57 <moon-child> regardless, my problem now is no audio.  Much less fun
11:32:21 <moon-child> perhaps I'll compile my own sdl and tell it to use oss
11:33:05 <meena> fun.
11:34:02 <moon-child> not really.  Computers always suck at audio, ime
11:34:30 <moon-child> and now I have to work out how to set up a linux toolchain, too...
11:35:14 <meena> yeah, I had to share rooms with my patents, later with my brother where my computer was during school, so installing Linux on it, fixed the computer is making sound problems back in the day
11:35:38 <meena> would always catch me out, when it actually worked, and i had to mute it!
11:36:52 <meena> even today, none of my devices make sound, unless i want to / have headphones
11:36:57 <yuripv> sharing a room with my patents is nice, as i have none
11:38:51 <meena> yuripv: see, these days I'm a patent myself, but i don't have to share my room with my… accomplishment, because we're not that poor / live 300 years ago
11:39:54 <yuripv> sorry, just making a joke from your typo
11:40:04 <meena> the amount of times this stupid keyboard interprets patent when I'm trying to write parent is infuriating, and i don't understand why i haven't removed it from predictions yet
11:40:17 <yuripv> :)
11:43:00 <meena> I imagine somebody at Microsoft sitting at day at a dictionary and giving weights to words that have the same swype pattern, and going, yes, patent is definitely written way more than parent
11:43:50 * moon-child grumbles
11:44:05 <moon-child> apparently cc _does_ work in a linuxulator chroot.  But linux-c7-devtools does not come with make
11:45:33 <moon-child> and apparently gnu make wants to be built with gnu make.  Hilarious
11:47:43 <AmyMalik> wind turbine
11:47:47 <moon-child> oh, no, it does come with a shellscript
12:13:34 <meena> moon-child: which Software is this
12:17:29 <moon-child> celesgte
12:17:31 <moon-child> celeste*
12:24:34 * meena recommends getting a PlayStation, which is FreeBSD based… 
12:29:21 <moon-child> playstation is expensive and can't do much
12:31:49 <moon-child> really, I ought to be able to do this without linuxulator, as it is a c# app.  But it relies on a closedsource c lib (fmod).  If I had more patience I'd userspace-shim it
15:20:32 <debdrup> jbo: I hope my reply didn't come too late.
19:50:08 <PredatorONormies> Any reason why temperature readings on 1 FreeBSD install would suddendly go away? I think they were there in the past
19:50:48 <PredatorONormies> loading amd_temp or whatever it's called works though. Any idea why it's not loaded by default, or perhaps it faulted somehow and went offline?
19:51:31 <PredatorONormies> amdtemp
20:38:57 <scoobybejesus> does anyone know ocaml? when i run `opam install mad`, it will fail because `The compilation of conf-mad.1 failed at "pkg-config --exists mad".` which returns `1`. Of course `pkg-config --exists libmad` returns `0`. This used to work with 4.12.0. Same problem with 14.4.1
21:51:05 <Serpent7776> scoobybejesus: Are you sure it used to work? It seems conf-mad lacks support for FreeBSD https://github.com/ocaml/opam-repository/blob/master/packages/conf-mad/conf-mad.1/opam
21:51:06 <VimDiesel> Title: opam-repository/opam at master · ocaml/opam-repository · GitHub
21:52:39 <Serpent7776> You might want to ask on #ocaml
21:53:07 <scoobybejesus> thank you for the help. I think it didn't used to be a dependency of the opam mad package
21:54:15 <Serpent7776> Ah, it might be the case that it didn't depend on the system package previously
23:06:18 <thumbs> 31/
23:06:23 <thumbs> hrm, sorry
23:07:29 <mason> thumbs: Shockingly, you're not in my /31 at all.
23:09:46 <thumbs> mason: this is /30
23:10:28 <mason> /11 here
23:10:37 <mason> alt alt-q
23:10:39 <mason> or*