hash in test time_zones::names::test_compute_bcp47_ids_hash is not stable over architectures

This test seems to give different result on different machine architectures, on i386 and armhf: ``` 1013s 1013s thread 'time_zones::names::test_compute_bcp47_ids_hash' (19032) panicked at src/time_zones/names.rs:146:5: 1013s assertion `left == right` failed 1013s left: 7305464005179878320 1013s right: 2080308884639987833 1013s stack backtrace: 1013s 0: __rustc::rust_begin_unwind 1013s at /usr/src/rustc-1.95.0/library/std/src/panicking.rs:689:5 1013s 1: core::panicking::panic_fmt 1013s at /usr/src/rustc-1.95.0/library/core/src/panicking.rs:80:14 1013s 2: core::panicking::assert_failed_inner 1013s 3: core::panicking::assert_failed::<u64, u64> 1013s at /usr/src/rustc-1.95.0/library/core/src/panicking.rs:394:5 1013s 4: icu_provider_source::time_zones::names::test_compute_bcp47_ids_hash 1013s at /usr/share/cargo/registry/icu-provider-source-2.2.0/src/time_zones/names.rs:146:5 1013s 5: icu_provider_source::time_zones::names::test_compute_bcp47_ids_hash::{{closure}} 1013s at /usr/share/cargo/registry/icu-provider-source-2.2.0/src/time_zones/names.rs:135:33 1013s 6: core::ops::function::FnOnce::call_once 1013s at /usr/src/rustc-1.95.0/library/core/src/ops/function.rs:250:5 1013s 7: <fn() -> core::result::Result<(), alloc::string::String> as core::ops::function::FnOnce<()>>::call_once 1013s at /usr/src/rustc-1.95.0/library/core/src/ops/function.rs:250:5 1013s note: Some details are omitted, run with `RUST_BACKTRACE=full` for a verbose backtrace. ``` full log: https://ci.debian.net/data/autopkgtest/testing/i386/r/rust-icu-provider-source/74267524/log.gz Typically these kind of problems happens because there is an usize as input to the hashing function, and those have different lengths on 32-bit platforms.