Fighting the last war
One feels that a lot of the design ideas in selective availability came from cold war thinking rather than addressing the current threats. Not that this means that in the lifetime of the next generation of satellites the issues won't re-emerge. SA simply meant that a Soviet tank commander wasn't able to navigate usefully with the hatches shut, whilst a NATO tank could. P channel accuracy of 3 metres is useful. But as has been noted, differential GPS with only the CA channel is actually superior to the military only P channel. In general the specific question of using SA or not has been overtaken by technology.
Commercial GPS receiver chips are required to limit their use to essentially subsonic velocities. They are required to stop providing data if they detect their velocity is above a given threshold - specifically to prevent their use as a cheap missile guidance system. It is assumed that anyone with the technical competence to build a GPS receiver from scratch, one without such a limitation is already much too sophisticated for such restrictions to matter. Although this logic is now flawed.
Further, totally turning off the CA channel reduces the P channel's accuracy to that of the CA channel - it is the use of both together that provides the P channel's accuracy (by allowing for real time estimation of ionospheric propagation delay.)
Use of GPS signals abounds. For instance the protocol timing of GSM and other mobile phone networks is commonly provided by a fixed GPS receiver at each cell. Total loss of GPS would probably result in the failure of most mobile phone networks.
The issues are now clearly much wider, and strangely the reverse of what had been intended. The extraordinary and growing reliance on GPS in the modern world means that reduction of civilian availability probably presents a greater security threat than use of GPS by an enemy. Here is where thinking must move. For the Europeans the spectre of having security critical technology controlled by the US is not pleasing. Worse, it has becomes single point of failure for technological rather than political failures. Every country on the planet needs to recognise that the loss of GPS would represent a varying, but very real, economic and security risk. So, rather than reside in cold war thinking about the military control of GPS, thinking needs to move to the converse, the military's role in protecting a nation's security by ensuring GPS availability.
Eventually the only viable solution is an additional loosely tied second constellation. Not because of any mistrust of the US government, but simply from an argument of reliability of what is becoming a mission critical resource for the day to day life of a modern nation.