It looks like #234 unintentionally regressed tests on 32-bit platforms where usize is (usually) the same size as u32. Removing the panic!ing code path for this case made the tests regress on 32-bit platforms since they no longer panic (logs from a build on i686-unknown-linux-gnu):
failures:
---- deny_max_capacity_arrayvec_value_const stdout ----
note: test did not panic as expected at tests/tests.rs:769:4
---- deny_max_capacity_arrayvec_value stdout ----
note: test did not panic as expected at tests/tests.rs:761:4
failures:
deny_max_capacity_arrayvec_value
deny_max_capacity_arrayvec_value_const
It looks like #234 unintentionally regressed tests on 32-bit platforms where
usizeis (usually) the same size asu32. Removing thepanic!ing code path for this case made the tests regress on 32-bit platforms since they no longer panic (logs from a build on i686-unknown-linux-gnu):