“There is a cost to silence and a cost to using your voice and every day I wake up and decide which bill I’m going to pay.”
Holy crap, yes.
(stolen from https://twitter.com/AshaRangappa_/status/1577697756319076359 )
re: Actually valid complaint about rust in Linux
@vertigo @Azure @tek_dmn @nytpu Note that overcommit is not the only reason allocation-failure recovery is problematic. It's also that standard coverage testing and fault injection is grossly inadequate to find all the bugs in the error recovery paths, because you can easily have a bug that only manifests with a particular pattern of allocation failures *and successes*, and the number of such patterns is exponential in the number of allocation sites.
Type systems can stop you from *leaving out* an error recovery path but they can't stop you from having a logic error in a recovery path because a *previous* recovery cycle left your day structure in a state you thought was impossible.
Tumblr users know when some asshole buys your website you STAY and be even more of an asshole until they realize their mistake
When you think about it, everything's driven by online review scores today, and scores less than 90% can get managers fired, which can cripple a company.
So, giving any product by a major company an online review of less than 90% is really an act of Industrial Sabotage. And if that company is also a defense contractor, well, sabotaging an ongoing war effort is Treason, isn't it? And should be punished accordingly.
In conclusion, I believe Rings of Power is a perfect 10/10 show. Must watch TV.
@alcinnz I've always held the attitude that forks are good as long as appropriate deference is given to the project it was forked from.
It doesn't even have to be fancy. "We decided to fork project X over these features, which we feel distinguishes our project from theirs," followed by a list of unique features or concerns.
As followers of my own projects know, not only have I stated forthrightly that I would not support some types of features, but I've actively encouraged those who do want those features to fork my projects and I'd be grateful to link to their projects on my own websites.
I'm a big supporter of forking, as long as the relationships it creates remain amicable.
"Subhealth" basically means "yeah you're sick but we don't know why" and it's a real diagnosis. So we (usually) have less "Your tests are fine therefore you're lying" than the West. It's a really good framework for treating stuff like Long-COVID.
Once again, a bug affecting a core piece of infrastructure at work, that has defeated the concerted efforts of a dozen people over the years, appears to have a two-line fix.
I just had to think about it for 3 years and then look at the code in the right frame of mind, and a "well, what if you just ..." springs to mind, unbidden.
Tell me again about how ticket estimates, sprint velocity, and SLOC are a useful way to measure developer productivity...
On the one hand: container, kernel, hypervisor, Secure Enclave, TPM, Management Engine, Endpoint Security Coordinator, I do get what you're all trying to do, you all want to be the good cop that reports on the antihero cop that reports on the bad cop, but on the other hand, what if we just STOPPED trying to make bad things good by stacking up hilariously higher and higher Jenga piles of top secret bad things? And also, what if we put just ONE CPU in our CPUs, and that one CPU WAS the CPU?
A newer server operated by the Mastodon gGmbH non-profit