[02:45] cbreak: you can remove the dock in default Ubuntu (it's just a Gnome shell extension) [02:46] also, not sure what's different about the virtual desktops? [02:47] possibly you want some extension to customize these too... :) [02:49] also the dock can autohide, if you prefer [04:31] good morning [04:46] ok i'm here [04:46] noel: its 6h46 here [04:46] 6 pm?.. [04:47] my country doesnt do am/pm [04:47] so 6 in the morning [04:48] so, what you will doing, hmm my buddy heh [04:48] im having coffee first, read some linux news [04:49] waw i like coffe + chocolate, are know about python language broo?. [04:50] im more a hardware guy noel [04:52] what's your name broo?..can you tell me?. hehe [04:53] lotuspsychje :p [04:54] haha i see in screen,it private right, sorry my bad english man [04:55] whats your native language noel [04:56] i mean my language, im not study eng language [04:56] Hello! [04:56] bakayaroo [04:56] haha [06:03] konoyaroo [08:57] JanC: I can assign programs to the virtual desktops somewhat easily. And when I log out and back in, the programs (at least some of them) get restarted and put into the right spot. [08:57] I can arange the virtual desktops in a grid, not just a line [08:57] there are keyboard shortcuts to get a grid overview of all desktops, and all windows on a desktop. And to move between desktops in various ways [08:58] im using the workspaces to dock extension, ill be curious to test gnomes new workspaces system [08:58] when I'm using then when I hold a window, the window moves to the new desktop too, which is nice [08:58] overall, the plasma thing seems a lot more customizable, and more like I'm used from MacOS [08:59] there's even a small extension that moves full-screen windows to new desktops, but I'm struggling with making this work properly [09:00] in addition to the virtual desktops, there are workspaces [09:00] !info gnome-shell-extension-workspaces-to-dock [09:00] Package gnome-shell-extension-workspaces-to-dock does not exist in impish [09:00] this is what im using ontop 20.04 [09:00] "activities" [09:00] nicely configurable [09:00] not sure what the point of those are yet [09:00] i never liked activities much [09:01] too much clicks to do what you want [09:01] it seems kind of functionally overlapping with virtual desktops [09:02] I already have tabs in many applications, windows, and virtual desktops, three levels of hierarchy are enough for me :) [09:02] maybe I find out what they're good for eventually [09:02] cbreak: https://imgur.com/a/rVxHbNb mine [09:02] anyway, plasma is nice [09:03] lotuspsychje: yeah, I know gnome had virtual desktops too. I used them a lot. They are nice. [09:04] the auto-creation is nice too [09:04] but not being able to bind applications to them is annoying slightly [09:04] im not using it default, installed the extension from repos [09:05] works differently then activities/dynamic workspaces [09:05] I have one virtual desktop for firefox, one for thunderbird, with split screen for mail and calendar [09:05] cool, sounds organised [09:05] one for my main kitty terminal, some more for additional terminals and VSCode [09:05] and the first for what ever other thing I do [09:06] I really liked dynamically creating them in gnome, but assigning apps that are restarted on login is more important [09:23] an other advantage of plasma over gnome is the file manager. Although it's only slightly better [09:27] the only useful file manager I found that has miller columns was ranger. It's nice, but has its own disadvantages :) [09:37] now I just have to find a way to transfer all my settings to my computer at work... :D [15:23] hello [15:23] anybody can help me with bluetooth dongle please [15:24] geomarcin: ask in #ubuntu not here [15:24] and please don't cross post [15:24] aske dthere they forward me here [15:24] thanks in advance [15:24] geomarcin: no one forwarded you here [15:24] it was for nikolam and his rant about firefox snap [15:24] wrote me with link to here [15:24] nope [15:25] same like here is #ubuntu [15:25] read again [15:25] discuss is not for support [15:25] so keep it in #ubuntu [15:26] and again nikolam refuses to join discuss... always the same [15:27] my biggest issue with snaps is there is no standard way to get the exact source used to build the binaries [15:27] how in hell is "snap reboot" a thing but not "snap source" ? [15:29] im totally fine with snaps for skype, discord or whatever [15:29] they have no source anyway [15:29] but a default browser should be part of the regular ubuntu archives [15:29] just my opinion [15:30] TJ-, most snaps ship their snapcraft.yaml ... thats the full build receipe and usually includes all links to all sources a snap is built from [15:30] but that is a massive regression to "apt-get source" and being confident of getting the exact same source. [15:31] shouldn't have to look for a yaml file and hope [15:31] it just feels like getting the microsoft store forces on my like in windows [15:31] *forced [15:32] and i always hated that piece of software deeply [15:33] well, mozilla asked for it ... was not an ubuntu/canonical decision (though indeed nobody opposed to it simply because it frees 1-2 fulltime devs to actually fix desktop ssues instead of working on the deb) [15:34] cant wait for new ads in firefox that will be enabled by default in the new snaps : [15:34] too late to take back control about that then [15:35] well, its a joint package ... it isnt like ubuntu devs have no control [15:35] we will see. i have a little more time until 22.04 [15:36] make sure to file bugs if you hit things that dont work as expected 🙂 [15:36] im on 20.04. everything is perfectly fine in my world yet :P [17:33] ravage: im going early 22.04 as soon as we have an iso [17:39] lotuspsychje: helping during the dev cycle? [17:39] yeah Mekaneck always on LTS releases : ) [17:39] cool :) [17:39] blueskaj pinged me toolchain already launched [17:40] but maybe bit more patience for a daily iso [17:42] wow, things go fast [17:42] yeah indeed [17:42] i could try early 22.04 on my private notebook for sure [17:42] for work i prefer a stable system [17:42] the ubuntu jammy 22.04 LTS toolchain is already uploaded ...just switched over with 350 successful upgrades...sed'd my sources.list and changed /etc/hosts and hostname to reflect the change on my system. [17:43] will look like 21.10 a lot at this stage for a while [17:43] i know [17:43] but still the kind of package updates will flow in fast [17:55] yeah, the queue-ot i #ubuntu-release is pretty busy these days (announing uploads ad syncs to jammy) [18:02] cool ogra [18:20] geez ... "the queue-bot in" ... [18:20] * ogra glares at his fingers [22:16] Here's some impressive speed-up/size-reductions, for initramfs. https://paste.ubuntu.com/p/fJXmnyS9mr/ [22:17] /join #gandi [22:17] 102MiB > 27MiB (26.5s > 8.1s) [22:22] login required to view these days eh, boo [22:24] daftykins: https://termbin.com/wexh [22:25] According to Colin Watson, viewing was supposed to allowed anonymously if the post was from a logged-in account; but when using pastebinit that won't work so it is login + login [22:26] aww [22:28] this kernel patch I developed (4 lines) leads to 20MiB reduction in initrd size https://termbin.com/ebsp [22:30] wow, what's getting messed up? [22:30] how do you mean? [22:30] is it correcting a mistake? [22:31] no, PRUNE and FIRMWARE_LOADED allow the initramfs-tools scripts to tailor to the active system by leaving out modules and lots of firmware not required to boot [22:31] ah i see [22:32] Without the kernel patch there's no way to know which firmware is loaded. With it, and enabling FIRMWARE_LOADED=true the kernel log is grepped for "Firmware loaded:" lines and no other firmware is included [22:35] without that, PRUNE=true can still do quite well but needs shell functions tailored to each module. In my case the amdgpu module lists 474 firmware files but only uses about 12. My filter_amdgpu() function determines the AMD GPU codename and only includes firmware files with the codename in. [22:35] I've also got conditional code to disable the framebuffer hooks that include loads of VERY old legacy video drivers, and the plymouth hooks that pull in EVERY GPU and firmware files they require [22:36] I've never understood why this isn't standard because there's no need for everything in the initrd unless it is a portable install on a USB device or whatever [22:36] i recommend nicknaming this "speedstripes" [22:37] hehehe why stripes? [22:37] a bit like folks would do to their cars :) [22:37] I'm about to post my initramfs-tools patches to Debian [22:37] jluc, gandi ? you mean Ganbbu ? [22:37] oh! [22:38] actually Gganbu 간부 lol [22:38] I need to do some boot tests in qemu/kvm too, to see how long it takes to load/decompress the various sized initrds [22:38] I'm going to try and get the 'Firmware loaded:' patches into mainline kernel [22:39] I'll need to add patches to dracut too; not dived into that code yet [22:40] very neat, nice work! [22:40] i've often begun to think that there's a lot of cruft in Linux land [22:40] but i wouldn't want to go full gentoo ricer to trim things down [22:41] why u need to do all those things [22:41] the code is very minimal and non-invasive so shouldn't be too difficult to get included, with these kind of benefits [22:41] TJ-: lovely. but, you'll probably be asked to answer this question next: how will you support the "user rips out graphics card 1, installs graphics card 2" sceanrio, though? [22:42] tomreyn: It's up to the user; these options will be off by default [22:42] ok, you win ;) [22:43] tomreyn: that won't stop the regular vga driver from operating in initrd [22:43] or efifb or whatever [22:43] how about wireless :) [22:43] how about it? [22:44] if i have this usb wireless dongle which needs firmware a, then replace it by usb wireless dongle which needs firmware b [22:44] okay, that's a bit of a corner case, i guess [22:44] you're network booting off of wlan? [22:44] eh you'd be safe in the knowledge that you know what you've done with your own system :) [22:45] that's totally orthogonal really, unless it needs wifi in initrd, in which case you've got a lot more customisation needed to achieve that [22:45] oh right this is just about booting, ok ok [22:45] right [22:45] now that initrd.img sizes are exceeding 100MiB and hitting limits that break things [22:46] I've been using these patches for a while but decided to modularise and update them today [22:46] Bug #1931024 [22:46] Bug 1931024 in initramfs-tools (Ubuntu) "BOOT fails on UEFI with 100MB initrd.img [Dell XPS 9550]" [Undecided, Confirmed] https://launchpad.net/bugs/1931024 [22:46] will you submit them against debian? [22:46] ^^^ that's one I dealt with recently [22:47] tomreyn: yes [22:47] I want to get the kernel patch upstream [22:47] 😍 [22:48] thanks! [22:51] daftykins, uh, going from gentoo and trimming things down sounds like an awful direction, you should rather go from yocto and build things up with only the bits you need 😉 way more effective [22:51] those are not things i do, i think you misunderstood [22:52] i think i understood, just wanted to make a snarky comment 😉 [22:53] seems like it was wasted [22:54] yeah, thats my destiny 🙂 [22:54] now to look at the remaining files list and see what else I can leave out! See if I can get it down to 12MiB [22:54] TJ-, hah ... awesome [22:54] * ogra remembers times where we considered 2MB to big for the initrd [22:56] people seem to have forgotten that the initrd has exactly one purpose ... find your disk ... nothing more ... all the rest should be done from the rootfs [22:57] it is nice to have ssh access during said process for multiple reasons [22:57] all the time ? [22:58] just have a special debug initrd if you need one ... 99% of users would never need ssh access in the initrd [22:59] ow! looks like all the kernels ever installed haven't completely cleaned out their /usr/lib/modules/XXXX/ directories, and the remnants of every one have been included in the initrd! [22:59] lol [23:00] initramf-tools usually only pics the versioned dir ... did that change since i looked last ? [23:00] hmmmph - anyone else see this kind of thing? https://termbin.com/edf0 [23:01] (this is on the root-fs) [23:01] 34 dirs here ... [23:02] arggh, I was wrong, it's not in the initrd. I just prefixed a path with / after extracting the initrd and looked at the root-fs not the initrd content! [23:02] most likely the modules.dep and aliases files [23:02] there's no reason to keep those around is there? should be purged [23:02] if you do "dpkg -l linux-image" you'll see an rc entry for each of them i guess [23:03] kernels arent purged, just removed [23:03] i guess thats the issue [23:04] err dpkg -l | grep linux-image indeed 🙂 [23:04] i always purge old ones to keep the list nice and lean :) [23:04] really helps for those VPSs where you have a small virtual disk [23:05] i occasionally do that on my machines ... [23:40] what on earth is going on with apt-cache!? trying to grep output and fails to match. then tried manually "apt-cache policy nonexistentpackage" and it isn't on stdout nor stderr but redirecting either to /dev/null results in no output at all! [23:50] don't think I'll get the initrd any slimer via kernel; uncompressed /usr/lib/modules/ is now just 20M. /usr/lib/ is 53M - the rest is in the various libraries to support lvm cryptsetup and so on [23:52] TJ-: I see no issue with apt-cache here on 20.04 - got a particular I can test ? [23:54] yes " apt-cache policy linux-image-lowlatency111; echo next; apt-cache policy linux-image-lowlatency111 | cat " [23:54] the second invocation doesn't send anything to cat over stdout, but nothing appears from stderr either [23:54] I think it is directly writing to /dev/console [23:54] TJ-: doing [23:55] same thing if we redirect both stdout and stderr "!&" with " apt-cache policy linux-image-lowlatency111; echo next; apt-cache policy linux-image-lowlatency111 |& cat " [23:55] err "!&" should have been "|&" of course [23:56] redo :) .. now at >> N: Unable to locate package linux-image-lowlatency111 // next . [23:56] what's with those 3 additional l's? [23:57] to ensure the package isn't known [23:57] ah right, ok - not really paying attention here :D [23:57] so the output is "N: Unable to locate package ..." /but/ if piped to another process the output doesn't come through, from stdout nor stderr