TLDs are valid in emails, as are IP V6 addresses, so checking for a . is technically not correct. For example a@b and a@[IPv6:2001:db8::1] are both valid email addresses.

Oh I’m sure they can justify spending the time porting the game. I just can’t blame them for not doing so when Apple’s policy is to not give a shit.

Yes, but it’s highly unlikely they can reuse that effort. TF2 uses an older version of source, so you either end up having to update everything to a newer version or redo all the effort you’ve already put in to porting source, but for an older version. Now multiply that by the number of source games.

I can see how you might come to that conclusion, but porting to 64bit is way more involved than it may seem. x86-64 is different enough to x86 that you’re basically porting to a new architecture: all your assembly will break just like it would porting to arm. On top of that all your sizes have changed and caused all sorts of bugs, for instance: long is 4 bytes under 32 bit on all platforms, but it’s 8 bytes on macOS & Linux and 4 bytes on Windows under 64 bit.

It’s mind boggling that Apple would drop 32bit support. Adding 64 bit support can be a lot of effort, especially on older code bases, and there’s a whole bunch of games that’ll never receive updates again. You could’ve had arm32 support, with the ability to run any x86 games.

AMDGPU is open source: https://github.com/radeonopencompute/rock-kernel-driver/, it’s also upstreamed into Linux. The firmware is a binary blob though.

This is plainly false. Hash collisions aren’t more likely for longer passwords and there’s no guarantee there aren’t collisions for inputs smaller than the hash size. The way secure hashing algorithms avoid collisions is by making them astronomically unlikely and that doesn’t change for longer inputs.