[2022-05-26T02:16:35Z] I'm going to be shoving a small computer into a freezer in the near future, and would like to load it with a distribution in which very little can go wrong. I'd like to remove everything non-essential up to and including the package manager. Is kiss a good starting point, or should I look elsewhere? [2022-05-26T02:33:59Z] something like nix is probably more reliable, but less understandable [2022-05-26T02:35:50Z] openwrt [2022-05-26T02:35:59Z] alpine [2022-05-26T02:37:28Z] kiss is acceptable too tbh, but you would have to compile everything from source, including kernel [2022-05-26T02:57:01Z] Hi [2022-05-26T03:00:11Z] testuser: I noticed the latest chromium binary(v100) doesn't crash like v96-99 (aw, snap..etc). But the pdf viewer is smoother on the previous ones...do you know why? [2022-05-26T03:01:17Z] ashin: maybe oasis linux [2022-05-26T03:05:01Z] humaid_: https://gitlab.alpinelinux.org/alpine/aports/-/issues/13579 [2022-05-26T03:05:37Z] no idea about pdf viewer though [2022-05-26T04:22:39Z] ashln: go wrong in what sense? [2022-05-26T04:41:25Z] Hi [2022-05-26T04:45:50Z] hi [2022-05-26T04:48:42Z] o/ [2022-05-26T05:12:25Z] illiliti: perhaps we should make a meson symlink for muon, like how there's a ninja symlink for samurai. [2022-05-26T05:15:26Z] https://git.sr.ht/~lattis/muon/#non-features [2022-05-26T05:15:40Z] > cli compatibility with meson. muon has different flags, subcommands, etc., and should not be renamed/symlinked to meson. [2022-05-26T05:17:16Z] if the build files are written to adhere to a strict (i.e. muon-compliant) meson spec, I don't see how it matters. [2022-05-26T05:19:42Z] I suppose in use outside of the repository [2022-05-26T05:21:01Z] cli interface isn't in meson spec [2022-05-26T05:21:24Z] there's no standard to adhere in this case [2022-05-26T05:21:45Z] ah [2022-05-26T05:22:24Z] meson has more stuff like coverage and pgo support so it wouldn't make sense for general use [2022-05-26T05:25:32Z] even if we do symlink, it would not work anyway. for example, meson provides install target in build.ninja file, while muon rightfully don't and so on [2022-05-26T05:43:12Z] * illiliti need to resume work on muon... [2022-05-26T05:43:41Z] there were some problems with subprojects in sway* packages before i postponed my work [2022-05-26T05:45:59Z] also another problem is that mesa depends on llvm which doesn't have pkg-config since 2011 [2022-05-26T05:46:15Z] https://github.com/llvm/llvm-project/issues/9777 [2022-05-26T05:47:09Z] meson has special handling of llvm in order to fix this [2022-05-26T05:47:33Z] which is insane [2022-05-26T05:52:43Z] illiliti: couldn't we ship our own pc file for lolvm [2022-05-26T05:53:48Z] yeah, i think we can [2022-05-26T05:59:37Z] still i would like to reduce the amount of patches and workarounds [2022-05-26T06:01:38Z] there are not too many of them atm, most of them related to convoluted gnome module [2022-05-26T12:51:20Z] acheam: https://git.bvnf.space/kisscommunity/commit/72141f385be0db947188b3121dc2009f172adb6a.html [2022-05-26T13:14:59Z] not sure "tag support" is the right phrase [2022-05-26T13:55:31Z] https://drewdevault.com/2022/05/25/Google-has-been-DDoSing-sourcehut.html [2022-05-26T13:57:42Z] acheam: noocsharp: correct, nobody asked me :v [2022-05-26T13:57:47Z] I've been a :ghost: [2022-05-26T14:11:16Z] I like that drew frames it all as a targeted attack on him and sourcehut [2022-05-26T15:22:12Z] i don't think he's saying its targetted, but just that google is being negligent with the amount of requests that they send [2022-05-26T15:50:21Z] It is the wording he chooses to use which makes it sound like SourceHut is the only one being affected by this -- "Google has been DDoSing SourceHut", "a year later I am still being DDoSed by Google", "having our little-fish service DoS’d by a 1.38 trillion dollar company" [2022-05-26T16:30:40Z] Maybe it’s because crypto actually needs the advanced features and correctness that Rust provides, while language doesn’t matter for most of your basic CRUD jobs so people use JS or whatever [2022-05-26T18:53:22Z] man [2022-05-26T18:53:38Z] is just me or kiss and CPU not work well together [2022-05-26T18:53:59Z] i tried do install the same package with pip in kiss and fedora [2022-05-26T18:54:17Z] and fedora is much faster than kiss [2022-05-26T19:05:55Z] has anyone had a problem building firefox, #include_next not found [2022-05-26T19:06:05Z] i think i've had this before but can't remember what i did [2022-05-26T19:06:45Z] rohan: I don't think pip is very CPU-dependent [2022-05-26T19:07:02Z] phoebos: not lately. log? [2022-05-26T19:07:31Z] ioraff: so who maybe is the problem for this? [2022-05-26T19:08:09Z] are you saying pip is faster or the package you installed with pip is faster? [2022-05-26T19:08:29Z] the package [2022-05-26T19:08:34Z] what package? [2022-05-26T19:08:48Z] is line 15 minutes with matplotlib to install [2022-05-26T19:08:57Z] and fedora is like 2 minutes [2022-05-26T19:09:45Z] i using a r7 3700u [2022-05-26T19:14:23Z] I would guess fedora makes use of some multithreading, which our pip apparently does not. in any case, have a look at matplotlib+numpy in my personal repository (old/ folder), which link against a kiss-suitable openblas [2022-05-26T19:17:32Z] so kiss not handle multithreading? [2022-05-26T19:20:30Z] you should try to rule out differences in how python is compiled and installed first [2022-05-26T19:21:31Z] are you sure the difference isn't from network caching or something [2022-05-26T19:31:07Z] "KISS" isn't what doesn't support multithreading -- if you enabled it in your kernel, you have SMT [2022-05-26T19:31:22Z] rohan: what's your CPU clocked at in KISS? [2022-05-26T19:31:26Z] you might be using a bad governor [2022-05-26T19:40:26Z] my governor is performance [2022-05-26T19:46:48Z] fedora may just be installing prebuilt wheels, whereas our pip needs to do some compilation [2022-05-26T19:47:45Z] yeah [2022-05-26T19:48:15Z] thats a resonable anwser [2022-05-26T19:50:07Z] this line: "Using legacy 'setup.py install' for matplotlib, since package 'wheel' is not installed." would seem to indicate that [2022-05-26T20:14:36Z] yeah thats why too long [2022-05-26T20:14:44Z] and it seems to compile [2022-05-26T20:14:57Z] when starts compilation oguri seems slow [2022-05-26T20:16:44Z] try installing wheel first [2022-05-26T21:35:22Z] illiliti: I see. pdf viewer is usable but sluggish a bit. v100 chromium doesn't snap, so I'm okay with it. [2022-05-26T21:36:02Z] (because we don't have intel core i9999999 with 9999999gb ram to compile chromium all day [2022-05-26T21:37:08Z] ^ haha. [2022-05-26T21:43:24Z] rohan: yeah there are some optimisations that should be done in kiss [2022-05-26T21:44:14Z] A couple of months ago I rewrote the dependency fixing [2022-05-26T21:44:23Z] Makes a huge difference [2022-05-26T21:45:07Z] In terms of both CPU usage and the time it requires to fix dependencies [2022-05-26T21:46:21Z] https://fossil.carbslinux.org/cpt/info/5dfbf828f84f5a29 [2022-05-26T21:51:41Z] I now re-read the conversation, and realised that it was not related to the package manager :) [2022-05-26T21:51:59Z] But still putting it out there, heh [2022-05-26T22:12:21Z] cem! :D [2022-05-26T22:14:19Z] Hey dilyn!