|
well a lot of android phones do indeed give you full control of your device, but all of the power management stuff lives in out-of-tree kernel hacks instead of system firmware, because the system firmware is a first-stage bootloader and nothing else beyond that. you would need a few other things on top of that, like a pci bus and nvme storage. and i suppose some sort of xhci extension that allows for the device to operate as a usb gadget. but that gives you enough firmware services and generic device classes to boot up a generic operating system image and suspend/resume it. arm macs are ipads with an unlocked bootloader. well, an unlocked bootloader on the application processor, but not on the dozen other cpus located inside the machine, the totality of which all run macos. so you can't really run alternative operating systems on it, you can dismember macos (which you cannot legally redistribute in whole or in part, by the way) and then reanalyze and reintegrate the undocumented interfaces every single time somebody in cupertino cuts a build. could those accessory cpus each run a generic firmware that provides a stable firmware interface? sure, but why would they, it's not supposed to be a hardware platform, it's supposed to be an appliance. easier to just drop support after five years and turn millions of machines into ewaste.
|
#
¿
Feb 27, 2024 15:24
|
|
|
|
| # ¿ May 16, 2024 21:10 |
|
|
come now friends, there is no need to get angry about Linux of all things
|
|
#
¿
Feb 28, 2024 22:52
|
|
|
https://www.collabora.com/news-and-blog/news-and-events/nvk-is-now-ready-for-prime-time.html NVK, the open source vulkan driver for 2080-series and newer nvidia hardware, is now Vulkan 1.3 conformant and will ship as a production-ready in the next Mesa release. DX11 over DXVK works, DX12 over VKD3D is a work in progress, OpenGL 4.6 over Zink is a work in progress.
|
|
#
¿
Feb 28, 2024 23:50
|
|
|
set the sleep mode from windows to linux in the firmware setup and it should sleep
|
|
#
¿
Feb 29, 2024 15:30
|
|
|
how is H264 still in patent anyway, this poo poo came out 20 years ago
|
|
#
¿
Mar 1, 2024 16:54
|
|
|
sounds to me like somebody caught a regression in the beta and it was promptly fixed unless i'm missing something
|
|
#
¿
Mar 7, 2024 17:08
|
|
|
shackleford posted:now to crack open the preferences and hose down the AI, cloudflare, identity protection / people search, VPN, etc. funk install Librewolf and disable fingerprinting protection. You don't even need to install ublock origin because it is built in.
|
|
#
¿
Mar 25, 2024 22:02
|
|
|
oh yeah that's right, Librewolf deletes all cookies on close by default, which is a little bit insane. you have to disable that. the important thing is that LW makes prompt decrapified Firefox releases any time there is an upstream release, and they are also trustworthy, erring on the side of being too aggressive about blocking ads and tracking than being too lax.
|
|
#
¿
Mar 26, 2024 09:43
|
|
|
mystes posted:I mean it just barely got caught by chance because someone was profiling sshd or something and it had gotten pretty far so it is a little concerning but I don't know what the solution is in this particular case, reduce the number of nooks in which some shady poo poo can hide. there was something lurking in the tarball that wasn't in the git repository. this sort of story is the way in which engineering processes improve in practice, by writing the rules in blood.
|
|
#
¿
Mar 31, 2024 17:23
|
|
|
a dude slipped a backdoor into a crack between what's in git and what's in a tarball. the solution to prevent this particular exploit from happening again is to have a stronger chain of custody from code reviews to source control to distribution build scripts, i.e. by tying build scripts to particular source control commit ids instead of a "dude trust me bro" tarball that is just assumed to reflect a particular git tag. this isn't new and we knew that we should be doing that to begin with, but nonetheless a bunch of distros (almost) got caught sleeping. this rule is part of the movement towards reproducible builds in general, but this rule is now written and underlined in blood. you learn from specific catastrophic or near-catastrophic incidents and devise specific rules to prevent them from happening again, that's how this sort of thing works in practice.
|
|
#
¿
Mar 31, 2024 17:36
|
|
|
if anything it is a far more specifically teachable moment than something like heartbleed or the log4j catastrofuck. the latter was particularly awful because that library literally had rce as its documented albeit unintended behavior and somehow nobody noticed.
|
|
#
¿
Mar 31, 2024 17:40
|
|
|
it would have been much harder because then a diff to introduce a backdoor would need to pass code review and a plausible good-faith justification would have to be presented. of course, this assumes that people are actually looking at and participating in a key dependency's development and it isn't just a one-man show that every major cloud company silently mooches off. the main factor saving us from that aspect of the attack is the fact that the bar for inclusion into the base packages for debian is way higher than it is for publishing something to the npm or pypi trash heap. a change of upstream maintainer for a key dependency like that should be cause for scrutiny. we already follow this rule elsewhere: in the browser add-on world a change in maintainership for popular a add-on is immediately treated with great suspicion because of a repeated history of attacks like this. so we have one human-factors takeaway and one technological takeaway from this incident.
|
|
#
¿
Mar 31, 2024 17:54
|
|
|
Beeftweeter posted:i mean, ultimately the exploit payload was within the lzma test files, and those are ultimately meant to be random junk data anyway you'd need to come up with a way for the build system to incorporate test data into the final executable artifacts. it is very difficult to come up with a plausible justification for doing that and it cuts against the grain of how every widely-used build system works. normally you first build the artifacts and then run the test suite against them. a compression algorithm could perhaps introduce a "new and improved" v2 bitstream format that incorporates some opaque tree-structured statistical models (derived from "real-world data", for instance), and it could access them in some complicated way such that the memory safety of those accesses is hard to prove, but at that point you're looking at a very different attack, and none of that changes the fact that openssh doesn't actually do any compression using this library to begin with, it was just the weakest link in /usr/sbin/sshd's ldd output that some nation-state actor could organize a bullying campaign against.
|
|
#
¿
Mar 31, 2024 18:07
|
|
|
fresh_cheese posted:i use a mac which is the best linux because the sound buttons all work as advertised right out of the box cool hey quick q how do i turn down the headphone volume on my usb-c dock then because macos goes  when i attempt to move the volume slider down from the maximum, tia
|
|
#
¿
Apr 7, 2024 15:32
|
|
|
AVC gets you shaken down by one mafia syndicate. not ideal but this situation has been around for decades and everybody knows how it works. HEVC gets you shaken down by the same mafia syndicate, but also several other mafia families at the same time. unlike actual mafia they don't protect you from rival mafia coming knocking for an extra piece of the pie. also i ask again why the gently caress is H264 still in patent, the standard is old enough to drink for christ's sake
|
|
#
¿
Apr 7, 2024 17:32
|
|
|
at this point i kind of hope that gnome sacks up and turns gnome os into an actual first-class distro instead of being a CI sewage pipe that spits out whatever is on git HEAD right this moment, because the patent-avoidance bullshit IBM keeps doing to fedora is becoming unbearable. application-wise this can be worked around by installing from flathub, but on the actual distro-provided desktop itself you still get blank thumbnails on video files encoded on a 20-year old codec because of softbank ip holding corp bullshit
|
|
#
¿
Apr 7, 2024 17:34
|
|
|
tiling window manager yeah ok lemme just open up a man page and a text editor and write a config file for my computer's user interface like some sort of caveman i wouldn't mind tiling wms if there was something like gnome that worked more or less ootb, but then again i seem to be the only person who thinks gnome's ootb defaults are mostly fine
|
|
#
¿
Apr 7, 2024 17:44
|
|
|
agreed, which is why you should use the cattle desktop os (fedora silverblue)
|
|
#
¿
Apr 8, 2024 17:40
|
|
|
or Debian so that you can actually play videos
|
|
#
¿
Apr 8, 2024 20:25
|
|
|
Jonny 290 posted:the only time i did was when i bought that pair of optane 58GB's on fire sale as a complete joke. they have like 1600 tbw which is just funny. i use them as my raid1 for /var/log. maybe after ten years i will hit 1% of that now optane i can get into
|
|
#
¿
Apr 9, 2024 00:30
|
|
|
i didn't realize osmocom was capable of bringing up anything more than a basic 2g network
|
|
#
¿
Apr 9, 2024 23:52
|
|
|
i've been running one continuous install of silverblue on my current laptop and desktop since i bought them in 2019 and 2021 respectively
|
|
#
¿
Apr 12, 2024 15:40
|
|
|
gnome people seem to be the only desktop people doing actual performance work on their poo poo, e.g. https://bxt.rs/blog/just-how-much-faster-are-the-gnome-46-terminals/ but they've been doing a bunch of analysis and picking of low-hanging fruit using their sysprof tool for the past few releases so yeah gnome-terminal is pretty deece
|
|
#
¿
Apr 12, 2024 19:28
|
|
|
I like firewalld but I can't get it to work with a wireguard server for the life of me. Traffic comes in over the tunnel but then fails to forward.
|
|
#
¿
Apr 14, 2024 13:57
|
|
|
Tankakern posted:Linux 6.10 To Merge NTSYNC Driver For Emulating Windows NT Synchronization Primitives looking forward to the 150% improvement in ... idk Bungholiomarks i presume phoronix is such a cesspit
|
|
#
¿
Apr 14, 2024 19:03
|
|
|
love to get 150% more performances per performance
|
|
#
¿
Apr 14, 2024 19:06
|
|
|
speaking of which i ran into some efi secure boot revocation thing recently because it turns out that silverblue/ostree/whatever doesn't update the actual bootloader in the efi partition like, ever, so i had to do it manually. only ended up at a grub command prompt one time even, not bad. they've had an action item to fix this for like three releases now although something like that can render your system unbootable if it fucks up so i do understand their trepidation edit: u gotta copy the contents of /usr/lib/ostree-boot/efi/ into your efi system partition, but you also have to make sure you preserve EFI\fedora\grub.cfg which is where i messed up like an absolute noob Sapozhnik fucked around with this message at 18:15 on Apr 16, 2024 |
|
#
¿
Apr 16, 2024 18:11
|
|
|
Tankakern posted:pretty sure it's the task of fwupd to update those uefi dbx files it is, but it is but it is not the task of fwupd to update the efi shim, because the efi shim is not firmware.
|
|
#
¿
Apr 16, 2024 22:10
|
|
|
if i wanted a giant pile of bloated poo poo on my desktop i would turn my monitor off
|
|
#
¿
Apr 17, 2024 23:12
|
|
|
https://www.youtube.com/watch?v=0yXlLOKqjpc NVK running Control at respectable framerates RTX OFF of course, and the driver isn't at the point where they're doing any serious perf work yet but like, it works. You can already play graphically intensive games with it.
|
|
#
¿
Apr 19, 2024 05:33
|
|
|
oh yeah I remember connecting some orinoco crap to a sharp zaurus back when i was a childe good poo poo
|
|
#
¿
Apr 19, 2024 21:53
|
|
|
i think about my filesystem a lot when i find myself using windows because that filesystem is slow as poo poo though technically that's because of windows io rather than any particular filesystem running on it
|
|
#
¿
Apr 22, 2024 05:44
|
|
|
I use gnome backup and it writes encrypted incremental backups to google drive it would be nice if they supported backblaze as well but their explicit goal is to only support "consumer" cloud storage and well yeah okay fair
|
|
#
¿
Apr 22, 2024 18:22
|
|
|
i think it's okay to underestimate it sometimes. every now and again. as a special treat
|
|
#
¿
Apr 27, 2024 13:14
|
|
|
well if the experimental label was removed and it ships with mesa by default then it should have some baseline functionality, surely? i seem to recall that it got certified as vulkan 1.0 compliant at least maybe it could have used another few months in the oven though, the videos of nvk running some fairly modern games via dxvk with decent framerates are still fairly new. anyway hopefully a good version makes it into freedesktop sdk 2024.08. not that i care though since i don't use nvidia hardware.
|
|
#
¿
Apr 29, 2024 21:32
|
|
|
ostree is the gold standard for os upgrades and it is bizarre that none of the other desktop oses have anything that comes close. mobile things have a/b partitions which work similarly and are technically more secure in some rather insane threat scenarios i guess. but yeah if silverblue wants to update itself it builds a deduplicated second /usr in the background from the bits it downloads and then it prompts you to reboot into it at time convenient for you. if the reboot fails then the kernel and /usr are rolled back. and each new build of silverblue is built from scratch on fedora's build servers so you don't have several years of god-knows-what crap being dragged around.
|
|
#
¿
May 4, 2024 03:39
|
|
|
i ain't reading all that looks like the guys who caused the term cadt to be coined are doing some cadt poo poo that breaks compatibility, again. is that about the size of it?
|
|
#
¿
May 6, 2024 02:32
|
|
|
reminds me of that time not too long ago when somebody politely asked for fractional dpi rendering and some gnome dev left a pissy note suggesting that they implement it themselves, then the gnome devs quietly added it over the course of the next release or two
|
|
#
¿
May 6, 2024 17:40
|
|
|
the dpi problem has a few aspects to it. one is that television is the spawn of satan and its standards continue to plague the decent respectable world of computing. there's hdmi ("it's just like displayport but it has patents on it!") and there's also the godawful claustrophobic 1920x1080 logical pixel standard, which then got doubled in both dimensions to give us 4K. on a reasonable computer screen (i.e. 27", the objectively correct desktop computer screen diagonal) this gives user interface components that are way too big or way too small. the objectively correct logical pixel standard for computer monitors is 2560x1440. an ideal high-resolution computer display would simply 2x2 this standard and call it a day. unfortunately this does not exist as a mass market product, so for all practical purposes you cannot buy a display with a 5120x2880 physical pixel grid. i've been waiting for this to come to market for almost 10 years and it still does not exist. it's not as simple as doing "just vectors", graphic designers will tell you that it is more complicated than that (i am not a graphic designer btw just to be clear). artists produce subtly different icons for each standard set of pixel dimensions that a logical "icon" can be rendered in. they will distort and emphasize different aspects of the icon image depending on the size at which the icon is being displayed. you can't "just" have a universal svg for this purpose, because if it was that simple then we would already be doing it; every linux gui toolkit has supported svg icons for over a decade now. there are two platforms that are not shackled by the tyranny of 96dpi: web, and mobile. on web the range of possible dpis is too great to depend on any particular pixel grid, which is why websites these days don't have many/any bitmapped user interface components and don't use icons much either, even though that was very much the fashion on turn-of-the-century websites. compare gamespot circa 2005 with gamespot today, but on a desktop because your mobile phone is going to have a hard time with it. on mobile, modern devices have a stupidly high density like 300 dpi or 250dpi or whatever, and at that level of oversampling it isn't too hard to size things a little bit bigger or a little bit smaller in order to get an integer multiple of 96 dpi. but a 4k screen's ~150 dpi absolutely can not. and it doesn't even matter because mobile apps all look like websites anyway, and so does electron hellworld for that matter. Sapozhnik fucked around with this message at 19:17 on May 6, 2024 |
|
#
¿
May 6, 2024 19:14
|
|
|
|
| # ¿ May 16, 2024 21:10 |
|
|
dudun dun dudun dudun dun dudun dudun dun dudun dudun dun dudun
|
|
#
¿
May 14, 2024 20:29
|
|