@david_chisnall @dantheclamman That last paragraph reminds me of another Onion classic: https://theonion.com/new-delicious-species-discovered-1819567839/

@cliffle @rubiline Since I don't see any explicit indicator, is that...Doritos-flavored vodka? Doritos-flavored rum? Or does the "uncategorized" stamp mean basically "we're not even trying to pretend this is any actual identifiable kind of liquor"?

@slava @steve @rk And of course an option for ENOENT: https://thedailywtf.com/articles/What_Is_Truth_0x3f_

@arj @acsawdey heh, nice. Aren't append-only caches with no eviction policy great?

@mxshift Servers really are an underappreciated food preparation tool: https://honk.bewilderbeest.net/u/zev/h/qB336jds7593W438TB

@petersanchez My initial thought is that it's a brief trend that will probably subside once the initial hype wears off.

I had similar thoughts about the rise of emoji though, so I'm probably wrong.

[click goes the shittiness ratchet...]

@cliffle I have to say, in this oh-so-hypothetical scenario, that based on past experience with similar phenomena that later became entries in the errata lists of an ARM core or two, I might take the opposite stance on your "probably" there.

Might it by any chance involve explicit clean or invalidate operations?

@tubetime @wren6991 @lofty @mei Alternately (or perhaps as a second level of bumper bowling), use __attribute__((constructor)) to install a SIGSEGV handler to just quietly "make the problem go away" a la https://people.csail.mit.edu/rinard/paper/osdi04.pdf (might also need to shim signal(2) and sigaction(2) to make sure it stays in place in case the Serious Enterprise Software tries to install its own crash handlers).

@mxshift Also wait, I seem to recall you mentioning an older Audi previously -- whence the switch? (Or am I misremembering...)

@mxshift Any fun decorations planned?

@cks @wingo Very tempted to implement this and then combine with https://github.com/zevweiss/bashc to see how many language roundtrips a simple program can make and still work.

@rk Step 1: #define CHAR_BIT 32768

@arj @wyldtom [raises glass] There's no staying power like that of a gross, ad-hoc hack.

@rk @mathew @mcc @darius I think that might just scale up the playback speed by about 6%...I await the "C♯ is faster than C!" blog posts.

@mxshift The amount of murder that can be gotten away with in software is one of the most depressing aspects of computing as a whole.

@mxshift

$ head -n6 src/sources/avstream/capture.c 
////////////////////////////////////////////////////////////////////////////////
// CONFIDENTIAL and PROPRIETARY software of Magewell Electronics Co., Ltd.
// Copyright (c) 2011-2024 Magewell Electronics Co., Ltd. (Nanjing)
// All rights reserved.
// This copyright notice MUST be reproduced on all authorized copies.
////////////////////////////////////////////////////////////////////////////////
$ tail -n3 src/sources/avstream/capture.c | head -n1
MODULE_LICENSE("GPL");

Nice.

The vendored PNG library, in-kernel file access, misunderstanding of the linux/list.h API, and passing kernel pointers to and from userspace are all very confidence-inspiring too.

@th @dashdsrdash @gvwilson I had the exact same thought, and figured this one could do with the graphical treatment as well.

@rk Likewise for right-clicking.

And scrolling.

And text selection & copy/paste.

(Okay, maybe github specifically isn't guilty of doing all of these, but I really think browsers should have a stricter, more aggressive demarcation of what is native browser UI that will be consistent no matter what jervascrapt fuckery is going on.)

@cliffle @kbob

(...omg I should try to modulate the data _onto the fans_)

I came very close to implementing that at my last job: https://honk.bewilderbeest.net/u/zev/h/HW2GcrgzthGts7DXR8

@jzb "If it bleeds, it leads" applied to niche nerd news.

@regehr

No system has bit addressing

Sounds like someone's never programmed an 8051.

@kevinrothrock It's probably these same teens from 2014 now of drinking age: https://theonion.com/teens-migrating-from-facebook-to-comments-section-of-sl-1819595619/

@arj @acsawdey Hmm, actually on further investigation it looks like that actually sorts via gnulib's filevercmp, which apparently has some additional smarts to make it slightly more DWIM-y. So, perhaps subtly different, but similar in spirit.

@acsawdey @arj Yep, -v:

$ ls -l
total 0
-rw------- 1 zev zev 0 Aug  4 01:58 10
-rw------- 1 zev zev 0 Aug  4 01:58 9
$ ls -lv
total 0
-rw------- 1 zev zev 0 Aug  4 01:58 9
-rw------- 1 zev zev 0 Aug  4 01:58 10

@corbet @lwn LWN is indeed an absolute gem; and while all the other topics the site covers are generally quite well done too, I would point to your kernel articles in particular as an exemplary job of combining both technical depth and remarkable accessibility despite the nature of the material (and yes, likewise something my own career has also benefited enormously from).

Bravo, and I will remain doggedly subscribed even as the scraper bots and LLM slop are busily trying to suffocate the last bits of useful life on the web.

@rk @RikerGoogling Little do we realize it, but as we enter commands at our shell prompts, we are also in the background composing...the greatest shell script in history.

@arj crates.io links at the end of the post look a bit broken, just FYI.

It still happens; I still love it.