This is a particularly salient point. Programs like web browsers are near as well constantly having previously existing zero-days reported, and each of these will translate into keylogging access on any X11 server the program is connected to. (Why X11 still permits full keyboard keystate polling when a client's particular window isn't focused is a conspiracy-worthy mystery on its own.)
Encryption is only as good as the user's opsec. If a password is being typed through the X server, it may be assumed that it has already been transmitted to an interested party of choice. For example: how many of us don't practice the boot partition passphrase a few tens of times in order to commit it to memory? There's even a --test-passphrase option in cryptsetup which enables just that. How many of us switch to the console, perhaps also kill the X server first, before doing that?
Thirdly, assuming keyloggers, the French security apparatus can just as easily present materials as supposedly haxxed from their target's computers when said materials were actually captured as they were typed. The custody chain is generally not very solid, nor its judicial requirements very strict, in "political" cases.
no subject
Encryption is only as good as the user's opsec. If a password is being typed through the X server, it may be assumed that it has already been transmitted to an interested party of choice. For example: how many of us don't practice the boot partition passphrase a few tens of times in order to commit it to memory? There's even a --test-passphrase option in cryptsetup which enables just that. How many of us switch to the console, perhaps also kill the X server first, before doing that?
Thirdly, assuming keyloggers, the French security apparatus can just as easily present materials as supposedly haxxed from their target's computers when said materials were actually captured as they were typed. The custody chain is generally not very solid, nor its judicial requirements very strict, in "political" cases.