One of the rationales for cloud computing is it saves money by cracking up utilisation. Providers observe how much users "really" use and then provisioning that much.
True, sometimes you will leave boxes at low utilisation for various reasons, e.g. to deal with traffic spikes. But those reasons have not gone away. So now instead of heaving a predictable increase in CPU cost, you have an unpredictable increase in performance snafus.
The only good news is that the real performance hit will be less than 30% on many workloads. Especially once the providers start juggling and optimising.
The point is that gansterism will become a smaller part of the economy, have "less funding" if you want to think of it that way.
Gangsterism is funded by various econmic oppornities which for one reason or another are illegal. Some things like assasination and extortion need to be illegal by their very nature.
But by passing laws that make new things illegal, we increase the amount of resources available to fund criminalality.
Ok, I see distcc helping people with computer labs and server farms, but I thought mysterydip wanted to help the Gentoo community and other such. I.e. peer-to-peer build sharing.
So what kind of cryptographic guarantees would you need for that? And if you can only verify the build results by trusting signatures from upon high, then what is the point? Perhaps those builds could be turned into work in a proof-of-work blockchain. Do compilers contain any hard-to-do, but easy to verify steps?
Whole shelves full of useless PhDs thesis are just waiting to be written on this topic.
How happy have you been with Meson in complicated projects with multiple directories. Especially where things are complex and different options are used in different places. Make, in spite of all it's craziness would be a good tool if it any sane kind of support for this.
CMake tries hard to to do better, but then introduces its own layers of craziness. So it's fine as long as I am not doing anything unusual, but as soon as I need to understand what is going on, I find a dizzying array of barely working moving parts beneath me.
The point is that something seems to be actively looking at the user-agent string and then doing wrong things. Had they merely "not supported" Linux it would have worked fine.
Now it might still be a bug rather than malice, but it's a but that arises out of intentionally doing something stupid.
True, sometimes you will leave boxes at low utilisation for various reasons, e.g. to deal with traffic spikes. But those reasons have not gone away. So now instead of heaving a predictable increase in CPU cost, you have an unpredictable increase in performance snafus.
The only good news is that the real performance hit will be less than 30% on many workloads. Especially once the providers start juggling and optimising.