[Graphs] g-snap vs. g-box and graph equality

Just trying to get a handle on people's expectations around named graphs as
g-snaps vs. g-boxes as it relates to graph equality (and inequality).

Given the two notional TriG/Qurtle fragments which we would like to combine:

file-1.trig:
:G1 { :a :b :c } .
:G2 { :d :e :f } .

file-2.trig:
:G1 { :a :b :d } .
:G3 { :d :e :f } .

My impression so far is that some people want to treat named graphs as
g-snaps, and some as g-boxes.  Suppose for a second that we treat them as
naming g-snaps.  Without expressing an opinion one way or another, I ask:

1.  Is the fact that G1 is mapped to two different g-snaps an inconsistency?

2.  From the fact that G2 and G3 are mapped to the same g-snap, can we
conclude that G2 and G3 are in fact the same resource?

Regards,
Alex