CIRCLE WITH A DOT

@evan @funcrunch yep, the tech and the content is the easy part - it's pulling along the userbase that would be the really hard part. forks only tend to survive when there's sufficient momentum in community following imo

@funcrunch @evan Like I said, froma development perspective. Actually forking the contributor base would be a lot harder, although if there was a sufficiently bad enough incident I am sure many contributors would willingly join an alternative project. whether that'd be sufficient for critical mass is hard to say however

@cwebber garfield af

@cR0w @da_667 I may or may not have used it to snag some database dumps from a certain game project when I was a kid, and I may or may not have cracked those hashes to get an admins account and unban myself from their forums

@cR0w @da_667 does sf still dump all of the projects into the same filesystem and fail to prevent different project accounts from touching each others files? project account homedirs being 777 by default was certainly an amusing choice

@evan I'm really worried about wiki, but there's still a chance of recovery; it would also be significantly easier to fork wiki compared to forking Firefox (from a development aspect at least... legal questions in licensing not so much)

@foone it's a good reminder that CF cards are essentially just miniaturised 16 bit ISA devices

@da_667 is this their classic fiber infra, or part of their recent centurylink/quantum acquisition? I've been using the latter for a few years now and it's quite good when it works, but the support is horrendous. $80/m for ostensibly 1Gbps synchronous, slightly less than that if they install a G.fast extension from your demarc but still quite good.

I deobfuscated the #copyfail exploit and fixed up the payload to run under Alpine: https://astr.al/notes/2026-04-29_copyfail/

@q3k working around the broken execve is trivial enough like you said; https://social.treehouse.systems/@astraleureka/116490148181953204
it's pretty amusing seeing the trodden pagecache results persist afterwards

@ska wow, to be honest I expected you to be running Linux in a considerably customised fashion, but now I suspect we might have very similar feelings

@ska please excuse the impromptu blagpost but i developed a sudden case of historical waxing combined with an urge to rant about an immediate dilemma

the last time I daily drove Linux on a laptop was in 2017 - system76 oryp3 (i7-7700hq+16gb+nv 1060) running Fedora rawhide with Xorg+openbox+tint2+dmenu. that config was comfy & familiar and I had run some variation of it since fedora core 6 (dang.. i don't miss that pentium 4 though)
however 2017 had other plans and I found myself repeatedly having to patch the kernel by hand (because nvidia of course:))). nothing wild, just little struct ordering adjustments and swapping out EXPORT_SYMBOL_GPL because lol proprietary drivers.
unfortunately the recurring feeling of 80x25 on a 17" screen quickly became grating; I lasted all of 2 months before reaching the end of my rope. I flashed the OEM (clevo) UEFI image, installed windows 10 and never looked back

that same OS install is still present. the machine still works even though the battery holds about 4 minutes of charge (when new, it held about 50 minutes!!!), the heatsink no longer conducts heat effectively, and the CPU & GPU hit 95°C immediately after boot. despite all that clevo makes a fine machine and even under max load it only throttles about 10% from max clocks at the expense of sounding like the concorde taking off. on the other hand, 16gb for windows + hyper-v is very cramped these days; for single-tasking it is just barely sufficient. suitable for working in wsl2, browsing the web, running discord, playing a game - but no more than one at a time

i finally broke down last spring and got a new laptop after 8 years for a much needed hardware refresh, another system76 - addw4, i9-14900hx+64gb+nv 4060. it shipped with pop_os 22.04 which is essentially ubuntu + system76 hardware integrations + supposedly "plug and play" nvidia drivers.
my initial thought was: "i'll try and feel out today's mainstream linux ecosystem" fully expecting to become fed up, wipe the OS nvme and install alpine within a few weeks

instead, wtmp begins Wed Feb 12 13:06:50 2025. i am still using the stock gnome environment, albeit on wayland rather than the stock Xorg setup. wayland was necessary due to complex interplay between Xorg, the nvidia kmod, i915 & DP MST, and my USB dock. any time an EDID I2C transaction was invoked, it caused i915 to spin on a lock for 10-15 seconds whilst waiting for acks on the DP aux channel. this, of course, made the rest of the kernel (and Xorg) very unhappy. EDID probes happen surprisingly often, too - multiple times during boot and logon, while changing VTs, even just launching winecfg led to a hard-stalled display for upwards of 30 seconds. despite kthread stacks and DRM debug looking nearly identical when querying available displays & resolutions, wayland does not exhibit the same behavior

surprisingly, the dock (dell WD19) is otherwise really good. all the peripherals Just Work; fwupd supports it natively; it drives 2x1440p@170Hz along with 1GbE and multiple USB3 ports without breaking a sweat. it's my first experience with a USB-C dock. all that throughput over one dinky little cable still feels like magic

gnome is so goddamn bad; every single day I nearly reach breaking point before my mind is flooded with deep, dark thoughts about how large my local aports tree would be (although that's why i opted for a ridiculous core count), interspersed with a sense of dread around trying to hammer the nvidia userland blobs into a musl-compatible shape.
I am at a loss for what distro to run: what project is trustworthy enough while simultaneously capable of being driven without thought on occasion, like a windows machine? i'm fine with building some stuff, hell i'll even maintain personal patchsets (already am for freeRDP and others b/c ubuntu repos are so horribly outdated), but i do not want to be spending hours mucking about for basic functionality.

i feel like i'm kind of asking for an unrealistic set of conditions given the current state of F/OSS desktops. there's a lot of pretty decent newbie-friendly stuff out there like bazzite and cachyOS. problem is I don't want the extra layers of hand-holding that they're built around, while wanting the common-case ease of use that they're designed to provide. either I have rose-tinted glasses about how much of a headache classic fedora & arch used to be, or there really is a significant degradation when it comes to the extreme power-user (but lazy) niche which used to be somewhat well provided

@ska every year that goes by I have more and more difficulty distinguishing Windows behaviour from the behaviour of mainstream Linux distros

@atax1a that was of course the first thing i did, with essentially zero improvement in knowledge of going on. just a bunch of socket comms between each other (some very short frames of a few bytes each). there was nothing in any of their log streams.
anyway, i had a Hunch, brought up the rdp session and copied 1 character from notepad on my windows machine. the cpu immediately calmed down

i went downstairs to make some dinner, came back to my workstation and found my cpu fan at full tilt, with gnome-shell, xfreerdp and xwayland all ripping 100% of a core each

@wren6991 frankly, this is quite as dumb as you made it sound. "in-band signalling is bad" was a lesson learned by telecom developers more than 40 years ago before the advent of Modern Development as we know it. developers of all walks have understood this very basic concept after much blood, sweat and tears, and yet here we are with ANTHROPIC_MAGIC_STRING_TRIGGER_REFUSAL_1FAEFB6177B4672DEE07F9D3AFC62588CCD2631EDCF22E8CCC1FB35B501C9C86 and second-person prompts begging the model to Please Don't Do The Dangerous Thing as if it was self-aware

@wren6991 this absurdity is on-par with the "claude re-emits vaguely json-shaped output repeatedly until the linter says it's valid" discovery from the recent source leak

@thephd i believe you have invented the world's best compression algorithm. "haha" - the amount of pain contained within those 32 bits is astronomical

@zzt _Bool will_halt(void *prog) { return true; }

@wdormann Just finished reading through the source, thank you. Much simpler than the first version, and makes the root issue a lot more clear. This is pretty nasty, MS has some serious egg on their faces over this one