Simon McVittie

> Does it really make sense [to] pass (all) of the hosts environment variables to Flatpak by default? FWIW, my experience from Steam's container runtime is that users generally expect...

> I have a useful use case where I point `LD_PRELOAD` to this adblock library I placed inside the app folder (`$HOME/.var/app/tld.domain.app`). > > I just hope that, when this...

> I'm still not convinced that a sandbox should pass through random env vars. Unfortunately, as I said, my experience is that users of Flatpak and similar app frameworks expect...

If someone wants to contribute it, I'd be happy to review a pull request for a `flatpak run --clear-env` option, or perhaps even a `flatpak override --clear-env` option, similar to...

Traditional package managers have the same behaviour: for example if you try to `apt install zsnes` (an x86-only package) on an ARM Debian/Ubuntu device, the error message you will get...

Runtime extensions are not apps, so there is no well-defined "latest version": they're more like plugins. If what you want is "the version of MangoHud that plugs into o.fd.Platform 22.08",...

> how exactly can XWayland be abused My understanding is that it can take screenshots of other X11 windows, and inject input into other X11 windows. For instance if you're...

If we install this into `profile.d`, can we be confident that it won't get sourced by bash and other POSIXy shells on any distribution?

Not yet tested (with AddressSanitizer or otherwise) but I wanted to get a version up so @TingPing could opine. What do you think about this as a reasonably non-ugly implementation?

In principle we could also have a `assert_cmpstr_free_rhs` (but there would be no uses yet, so I didn't add it) and/or a `assert_cmpstr_free_both` (but there would be only one use...