a trash can upside down ... on top of what? a shallow mercury bath?

a passing wave, even a distant plane wave, will couple to the EM surface waves of the can surface.

also novices always home in and hyperfocus on reflectivity and transmission attenuation lengths, which works when done properly, but is a complete PITA without using absorbers.

simply placing an unopen plastic water bottle inside the cavity means that any EM energy that does make its way in, would be absorbed, instead of repeatedly reflecting from your perfect metallic reflector walls...

Consider a small 2D patch of metal sheet with some resistivity: any attenuated tails of EM waves will induce currents in the metal, and those currents can flow in all directions in the sheet.

Now consider the same metal sheet with a very fine unintentional slit, while currents can still run parallel to the slit, the metal doesn't admit significant current (ignore electron tunneling across the slit) across the slit, this means a current can circulate around the slit, this is why some of the commenters here point out the issues with slits. Simply placing mating metal lids / doors / ... against metal containers does not guarantee identical conductivity across the formed slit: imperfect planarity (think placing the Alps upside down on the Pyrenees), different metallic phases (annealing etc) at the surface, the spontaneous formation of an oxide layer (aluminum has a thin corundum layer, iron/steels/metals oxidize,...), protective coating layers, ... all contribute to the failure to text-book-reflect an EM waved with a perfect isotropic conductor.

Suppose you wish to protect your empire from an overrun by the barbarians: proper border control is important, but in practice imperfect, so you need a way to absorb or apprehend any barbarians that made their way in.

Thats why you want to place a wideband absorber within your protected cavity. This increases the demand on EMP power to disturb any sensitive devices.