The lateral bracing of the trusses can be provided by the purlins PROVIDED THAT THEY ARE DESIGNED TO DO SO.  The purlins can then carry the loads back to the cross bracing which can then be designed to transfer the load down to the foundations.

The columns at the end of the structure are usually designed to carry vertical loads and only the wind load in the horizontal direction.  Structurally, it would be inefficient to design these columns to carry the lateral load when you have diagonal bracing which can do the job using much less steel and can do it with less deflection to the structure.

Thinking practically, on a 50m span one should expect the purlins to be spaced at approx 1-2m c/c along the top chord. This would give between 30 and 40 or so purlins along the length. Even on such a large building, the axial load on each purlin due to the wind load on will be reasonably small. It is reasonable to assume full restraint due to the cladding, therefore, the purlins should be checked mainly for bending with the nominal axial load due to wind. Bear in mind, of course, one would also expect the bottom chord of the truss to be restrained by bracing as well, therefore the wind load will also be shared with these elements.

Will P-delta analysis with a small lateral load applide do anything??