repr(x) == repr(y) <=> x == y AND eval(repr(x)) == x

[Tim Peters]

[Thorsten Kampe]
> ...
> This relies on two assumptions:
>
> 1. repr(x) == repr(y) <=> x == y

You'll be deluged with counterexamples soon enough <wink>, so I'll just give
you a surprising one:  this can fail if x and y are ordinary dicts, with x
== y.  The order in which keys are listed isn't defined, and if the the same
set of key+value pairs is inserted into two dicts but in different orders,
internal key collisions can cause different orders internally too (which
repr() then reveals).  It isn't possible to give a specific example that
will fail across all versions of Python, though, because nothing is defined
about this internal order, and it can (and does!) vary across releases.