style: Use binary literals instead of hex literals in doctests for highest_one and lowest_one

[sorairolake]

For example, I think it's easier to understand that the index of the highest bit set to one in 16 is 4 as 0b10000 than as 0x10.

assert_eq!(0x10_u64.highest_one(), Some(4));

Instead of:

assert_eq!(0b10000_u64.highest_one(), Some(4));

#145203

[rustbot]

r? @scottmcm

rustbot has assigned @scottmcm.
They will have a look at your PR within the next two weeks and either review your PR or reassign to another reviewer.

Use r? to explicitly pick a reviewer

[hkBst]

I was going to agree, but with more repeating ones the benefits of hex start to appear. This is also one of the test cases:

assert_eq!(0x1f_u64.highest_one(), Some(4));

Instead of:

assert_eq!(0b11111_u64.highest_one(), Some(4));

[sorairolake]

@hkBst When the value has many digits, hex literals are easier to understand:

assert_eq!(0x10000000_u64.highest_one(), Some(28));
assert_eq!(0b10000000000000000000000000000_u64.highest_one(), Some(28));

assert_eq!(0x1fffffff_u64.highest_one(), Some(28));
assert_eq!(0b11111111111111111111111111111_u64.highest_one(), Some(28));

However, the binary literals in the doctests are at most five digits long, so I think they are easier to understand than hex literals.

[tgross35]

I'd use underscores to split up the nibbles, e.g. 0b1_1111_u32

[hkBst]

@hkBst When the value has many digits, hex literals are easier to understand:
However, the binary literals in the doctests are at most five digits long, so I think they are easier to understand than hex literals.

For me 5 ones in a row is already over the limit of what I can easily glance.

tgross35's suggestion of adding underscores would help with that though...

[tgross35]

Squash please then r=me

[sorairolake]

Added underscores to binary literals longer than 4 digits:

• 0b10000 to 0b1_0000
• 0b11111 to 0b1_1111