gemini - kennedy.gemi.dev

💾 Archived View for gemini.ctrl-c.club › ~phoebos › logs › kisslinux-2023-07-30.txt captured on 2024-03-21 at 15:48:37.
-=-=-=-=-=-=-
[2023-07-30T02:44:26Z] <iceman[m]> Hi
[2023-07-30T02:45:28Z] <iceman[m]> <sad_plan> "could use palemoon for now..." <- I used palemoon on fossapup (puppy linux)
[2023-07-30T02:45:28Z] <iceman[m]> it was decent enough
[2023-07-30T02:45:50Z] <iceman[m]> But it's still nice to keep going with upstream
[2023-07-30T03:23:49Z] <sewn> hi
[2023-07-30T04:16:06Z] <xdream8[m]> i am thinking of introducing threading to kiss-rs
[2023-07-30T04:16:22Z] <sewn> if it's easy, yes that is a great benefit
[2023-07-30T04:16:36Z] <sewn> but if it requires great rewrite of the code it's not reallyw orth it
[2023-07-30T04:17:43Z] <xdream8[m]> it is pretty easy, i just need to look at the code and change parts that uses for loops and add threading
[2023-07-30T04:23:42Z] <xdream8[m]> i will also add a flag, which it will force kiss-rs to use only 1 thread.
[2023-07-30T04:28:40Z] <sewn> it is july 30
[2023-07-30T04:55:36Z] <sad_plan> hi
[2023-07-30T04:56:33Z] <sad_plan> iceman[m]: sure. upstream is always best initially. atleast from a security perspective
[2023-07-30T05:03:02Z] <iceman[m]> <xdream8[m]> "it is pretty easy, i just need..." <- is it really beneficial?
[2023-07-30T05:04:01Z] <iceman[m]> the time taken for context switch for threads, time taken for setting up other stuff for the threads
[2023-07-30T05:04:01Z] <iceman[m]> all taken to consideration
[2023-07-30T05:16:04Z] <xdream8[m]> it is very useful. for example when kiss-rs creates a manifest file, it searches pkg directory and takes  found files as a vector of paths(you can think of vectors like arrays) and removes prefix($XDG_CACHE_HOME/kiss/proc/<pid>/pkg/<pkg_name>) from all items in the vector. adding threading would shorten the time we spend on this. this is mostly beneficial for big packages as you might have thought.
[2023-07-30T07:17:36Z] * xdream8[m] uploaded an image: (73KiB) < https://libera.ems.host/_matrix/media/v3/download/matrix.org/etaZraOOgFVbQQPnpiwvPclE/20230730_10h11m30s_grim.png >
[2023-07-30T07:17:37Z] <xdream8[m]> ^ this is single-threaded
[2023-07-30T07:17:57Z] * xdream8[m] uploaded an image: (64KiB) < https://libera.ems.host/_matrix/media/v3/download/matrix.org/cvyOpsZiESDxPAAdUShuuply/20230730_10h10m57s_grim.png >
[2023-07-30T07:18:01Z] <xdream8[m]> ^ this is multi-threaded
[2023-07-30T07:18:09Z] <sewn> i dont get how literally listing files and getting information from two files in them in a loop could be so much faster wth
[2023-07-30T07:18:29Z] <sewn> xdream8[m]: such a small difference
[2023-07-30T07:18:37Z] <sewn> your multi-thread will make difference when checking for conflicts
[2023-07-30T07:18:40Z] <sewn> a big one.
[2023-07-30T07:18:58Z] <xdream8[m]> i had to wait total 6-8 minutes for 'kiss list' test
[2023-07-30T07:19:19Z] <sewn> you set --min-runs to 1000 thats why
[2023-07-30T07:19:51Z] <xdream8[m]> yeah i want accurate tests
[2023-07-30T07:20:04Z] <sewn> :v
[2023-07-30T09:28:17Z] <xdream8[m]> should i parallelize source downloads and git checkouts?
[2023-07-30T09:28:24Z] <sewn> how
[2023-07-30T09:28:47Z] <sewn> btw i would reccomend you try out looking at a git library instead of executing git commands
[2023-07-30T09:28:56Z] <xdream8[m]> since they are functions that i wrote myself its pretty easy
[2023-07-30T09:29:07Z] <xdream8[m]> sewn: i already use git2 rust lib
[2023-07-30T09:30:20Z] <sad_plan> if you can, I would. but make it use a variable, so users can choose how many
[2023-07-30T09:30:50Z] <sad_plan> sabotage does this. really neat if you have a good connection
[2023-07-30T09:31:04Z] <xdream8[m]> sewn: kiss-rs does not depend on any system commands(except strip)
[2023-07-30T09:31:13Z] <xdream8[m]> sad_plan: it will have --jobs flag
[2023-07-30T09:31:18Z] <sad_plan> nice
[2023-07-30T09:55:21Z] <sewn> oh no
[2023-07-30T10:44:53Z] * xdream8[m] uploaded an image: (45KiB) < https://libera.ems.host/_matrix/media/v3/download/matrix.org/faVVcdBFcPRAwamIuokvGqPq/20230730_13h39m36s_grim.png >
[2023-07-30T10:44:55Z] <xdream8[m]> print order gets messed up
[2023-07-30T10:45:06Z] <xdream8[m]> maybe i should not print Reading sources
[2023-07-30T10:45:11Z] <sewn> there is NO way that is fast
[2023-07-30T10:45:46Z] <sewn> how on earth is that  1.5 seconds
[2023-07-30T10:46:28Z] <xdream8[m]> it did not download anything because they are already in cache
[2023-07-30T10:46:34Z] <sewn> yeah ik
[2023-07-30T10:46:41Z] <xdream8[m]> it only tried to fetch latest git commits
[2023-07-30T10:47:28Z] <xdream8[m]> so how should i handle printed output as it gets messed up?
[2023-07-30T10:47:43Z] <xdream8[m]> i should make it only print Reading sources... when debug is set
[2023-07-30T10:47:44Z] <sewn> is it multi threaded too
[2023-07-30T10:47:52Z] <xdream8[m]> yeah
[2023-07-30T10:47:58Z] <sewn> you're kidding me
[2023-07-30T10:48:46Z] <xdream8[m]> that is why print order gets messed up .d
[2023-07-30T10:49:03Z] <sewn> imo it should be one by one, this isnt a binary package manager
[2023-07-30T10:53:06Z] <xdream8[m]> this is better. i just need to improve my code
[2023-07-30T11:13:51Z] * xdream8[m] uploaded an image: (53KiB) < https://libera.ems.host/_matrix/media/v3/download/matrix.org/zdlXgizaCbdVkYHMAZwzTYwR/20230730_14h09m01s_grim.png >
[2023-07-30T11:13:54Z] <xdream8[m]> does it look good enough?
[2023-07-30T11:14:05Z] <sewn> whats wrong with -> pkg
[2023-07-30T11:14:34Z] <xdream8[m]> i think i removed them when illiliti was suggesting me design ideas
[2023-07-30T11:14:43Z] <sewn> looks about right
[2023-07-30T11:25:39Z] <xdream8[m]> i found a better way to do this. instead of creating threads for packages, we check/download sources in threads
[2023-07-30T13:37:25Z] <xdream8[m]> i am thinking of making some of the compression/decompression libraries in kiss-rs optional at compile-time as its just a lot of  dependency
[2023-07-30T14:09:39Z] * xdream8[m] uploaded an image: (119KiB) < https://libera.ems.host/_matrix/media/v3/download/matrix.org/vWWCNDqIsfOcloOInMRixZjk/20230730_17h04m25s_grim.png >
[2023-07-30T14:09:39Z] <xdream8[m]> ^ another test
[2023-07-30T14:10:09Z] <sewn> where can i see this kiss-rs thing
[2023-07-30T14:11:11Z] <xdream8[m]> its in codeberg but i havent pushed in a while
[2023-07-30T14:12:20Z] <sewn> maybe one day someone will write kiss-ha
[2023-07-30T14:13:53Z] * xdream8[m] uploaded an image: (66KiB) < https://libera.ems.host/_matrix/media/v3/download/matrix.org/UuFRiBeOtqmKfYxzJKaLNmzn/20230730_17h08m54s_grim.png >
[2023-07-30T14:14:10Z] <xdream8[m]> ^ ccache was used for precise results
[2023-07-30T14:14:13Z] <sewn> did you
[2023-07-30T14:14:14Z] <sewn> did you finish building
[2023-07-30T14:14:20Z] <sewn> how is it faster...
[2023-07-30T14:14:26Z] <sewn> shell sisters...
[2023-07-30T14:17:17Z] <xdream8[m]> pkg_conflicts has lots of bugs and is incomplete other than that i almost finished
[2023-07-30T14:17:46Z] <xdream8[m]> sewn: i feel like single-threaded version was faster
[2023-07-30T14:18:31Z] <sewn> statistics > feelings
[2023-07-30T14:18:34Z] <sewn> :^)
[2023-07-30T14:22:26Z] * xdream8[m] uploaded an image: (102KiB) < https://libera.ems.host/_matrix/media/v3/download/matrix.org/FTYsQZKVZbttzDnDrYLjtiSD/20230730_17h17m38s_grim.png >
[2023-07-30T14:25:03Z] * xdream8[m] uploaded an image: (92KiB) < https://libera.ems.host/_matrix/media/v3/download/matrix.org/XVGsnkBjKCZDGwNsoNShTrlr/20230730_17h20m14s_grim.png >
[2023-07-30T14:26:01Z] <sewn> whats so multi-threaded about generating checksums
[2023-07-30T14:26:17Z] * xdream8[m] uploaded an image: (105KiB) < https://libera.ems.host/_matrix/media/v3/download/matrix.org/EqLnzMCflQwTVLjwBZLBRFnq/20230730_17h21m27s_grim.png >
[2023-07-30T14:26:39Z] <sewn> looks like you will be using multi-threaded when its needed yay
[2023-07-30T14:28:52Z] <xdream8[m]> should i set number of jobs to 1 by default?
[2023-07-30T14:29:09Z] <sewn> you should hard-set it automatically when needed
[2023-07-30T14:29:17Z] <sewn> for example in big downloads, 4 jobs, and with small ones one job
[2023-07-30T14:29:22Z] <sewn> same goes for checksums, etc
[2023-07-30T14:29:38Z] <sewn> like if its small disable multi-threading
[2023-07-30T14:31:45Z] <xdream8[m]> i can directly number of jobs to 1 by default, so it will be disabled by default
[2023-07-30T14:31:55Z] <xdream8[m]> * can directly set number of
[2023-07-30T14:32:02Z] <sewn> sure
[2023-07-30T16:38:28Z] <sewn> i wonder why the irc bridge hasnt gone down yet
[2023-07-30T17:18:58Z] <mako2> chromium -> build failed https://termbin.com/7odu
[2023-07-30T17:20:45Z] <mako2> :/
[2023-07-30T17:28:40Z] <illiliti> log is truncated
[2023-07-30T17:29:26Z] <sewn> ^
[2023-07-30T17:41:36Z] <mako3> can i just shorten the log
[2023-07-30T17:42:16Z] <mako3> just from the beginning and in the end? i can't upload the file with the full log
[2023-07-30T17:42:48Z] <illiliti> use better pastebin
[2023-07-30T17:43:06Z] <illiliti> pastes.sh, envs.sh, 0x0.st
[2023-07-30T17:46:15Z] <mako3> http://0x0.st/H2ML.txt
[2023-07-30T17:47:37Z] <illiliti> gcc/clang is up-to-date?
[2023-07-30T17:48:07Z] <mako3> I updated gcc along with nss yesterday
[2023-07-30T17:48:22Z] <mako3> I don't think I've heard an update with clang
[2023-07-30T17:56:24Z] <illiliti> weird
[2023-07-30T17:57:30Z] <illiliti> https://chromium.googlesource.com/v8/v8/+/620388b1a6445f88fbe7c19ce4041c70eb948d21%5E%21/
[2023-07-30T17:58:06Z] <illiliti> not sure why this fix is still not included
[2023-07-30T17:59:24Z] <illiliti> you can apply it and check if it works
[2023-07-30T17:59:46Z] <illiliti> or you can build chromium with clang
[2023-07-30T18:00:47Z] <mako3> does clang work most of the time?
[2023-07-30T18:02:42Z] <illiliti> yeah it should bacause google devs use it to build chromium
[2023-07-30T18:06:38Z] <mako3> i'll just build it with clang this time, will switch back to gcc if something happens
[2023-07-30T18:06:52Z] <illiliti> btw are you sure your computer is capable of building chromium?
[2023-07-30T18:07:23Z] <mako3> well, my pc is a r5 5600 and rx 6600 with 32gb
[2023-07-30T18:07:39Z] <illiliti> still not enough
[2023-07-30T18:07:43Z] <illiliti> get ready for OOM
[2023-07-30T18:07:48Z] <mako3> :/
[2023-07-30T18:09:17Z] <illiliti> yeah linking step will kill your computer
[2023-07-30T18:12:02Z] <mako3> my temps go from 56 to 60's when compiling chromium
[2023-07-30T18:13:23Z] <mako3> i just cat /sys/class/hwmon/hwmon0/temp1_input and a sed command to check my temps
[2023-07-30T18:13:53Z] <mako3> since there isnt a gotop package in any of the kiss repos
[2023-07-30T18:14:14Z] <illiliti> if you live in a cold climate that's probably advantage
[2023-07-30T18:14:34Z] <mako3> i wish
[2023-07-30T18:15:23Z] <mako3> i live where the sun makes me uncomfortable and itchy, typhoons are frequent
[2023-07-30T18:25:31Z] <illiliti> that sucks
[2023-07-30T18:43:12Z] <mako3> btw, can i compile the linux kernel without using gcc?
[2023-07-30T18:57:09Z] <illiliti> i believe you can
[2023-07-30T18:57:26Z] <illiliti> with clang
[2023-07-30T18:58:24Z] <mako3> ah i see