feat: Add bit_width for unsigned NonZero<T>

[sorairolake]

• Tracking issue: Tracking Issue for uint_bit_width #142326

This pull request adds a method to the unsigned NonZero<T> that return the minimum number of bits required to represent a value.

This can be achieved by using the get method and the methods added in #142328, but I think adding the NonZero::bit_width method is useful because it accomplishes the same thing a little more succinctly.

[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

[sorairolake]

@rustbot label +T-libs-api -T-libs

r? libs-api

[hanna-kruppe]

Can’t this return NonZeroU32? That would be a good argument for a dedicated NonZero method instead of delegating to get().

[scottmcm]

but I think adding the NonZero::bit_width method is useful because it accomplishes the same thing a little more succinctly.

Note that we generally don't add things to NonZero just to avoid a .get(). You'll note, for example, that NonZero::count_ones exists -- because it can return NonZero<u32> -- but NonZero::count_zeroes doesn't, despite it existing on the underlying types, since there's nothing special for NonZero to do on it.

If, as @hanna-kruppe mentions, this can return NonZero, that might be a reason to justify its existence.

@rustbot author

[rustbot]

Reminder, once the PR becomes ready for a review, use @rustbot ready.

[sorairolake]

@scottmcm In safe Rust, the minimum number of bits required to represent unsigned NonZero<T> is guaranteed to be greater than or equal to one. Therefore, this method can return NonZero<u32>.

Based on the comments, I changed the return type of this method from u32 to NonZero<u32>. This type is suitable to indicate that this method can never return zero. Also, other NonZero<T> methods may be easier to use with this type than with u32.

[sorairolake]

@rustbot ready

[scottmcm]

Sure; for unstable seems close enough to fit under the existing unstable library feature.

@bors r+ rollup

[bors]

📌 Commit a25950d has been approved by scottmcm

It is now in the queue for this repository.

[sorairolake]

• Tracking issue: Tracking Issue for uint_bit_width #142326

This pull request adds a method to the unsigned NonZero<T> that return the minimum number of bits required to represent a value.

This can be achieved by using the get method and the methods added in #142328, but I think adding the NonZero::bit_width method is useful because it accomplishes the same thing a little more succinctly.

[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

[sorairolake]

@rustbot label +T-libs-api -T-libs

r? libs-api

[hanna-kruppe]

Can’t this return NonZeroU32? That would be a good argument for a dedicated NonZero method instead of delegating to get().

[scottmcm]

but I think adding the NonZero::bit_width method is useful because it accomplishes the same thing a little more succinctly.

Note that we generally don't add things to NonZero just to avoid a .get(). You'll note, for example, that NonZero::count_ones exists -- because it can return NonZero<u32> -- but NonZero::count_zeroes doesn't, despite it existing on the underlying types, since there's nothing special for NonZero to do on it.

If, as @hanna-kruppe mentions, this can return NonZero, that might be a reason to justify its existence.

@rustbot author

[rustbot]

Reminder, once the PR becomes ready for a review, use @rustbot ready.

[sorairolake]

@scottmcm In safe Rust, the minimum number of bits required to represent unsigned NonZero<T> is guaranteed to be greater than or equal to one. Therefore, this method can return NonZero<u32>.

Based on the comments, I changed the return type of this method from u32 to NonZero<u32>. This type is suitable to indicate that this method can never return zero. Also, other NonZero<T> methods may be easier to use with this type than with u32.

[sorairolake]

@rustbot ready

[scottmcm]

Sure; for unstable seems close enough to fit under the existing unstable library feature.

@bors r+ rollup

[bors]

📌 Commit a25950d has been approved by scottmcm

It is now in the queue for this repository.

[sorairolake]

• Tracking issue: Tracking Issue for uint_bit_width #142326

This pull request adds a method to the unsigned NonZero<T> that return the minimum number of bits required to represent a value.

This can be achieved by using the get method and the methods added in #142328, but I think adding the NonZero::bit_width method is useful because it accomplishes the same thing a little more succinctly.

[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

[sorairolake]

@rustbot label +T-libs-api -T-libs

r? libs-api

[hanna-kruppe]

Can’t this return NonZeroU32? That would be a good argument for a dedicated NonZero method instead of delegating to get().

[scottmcm]

but I think adding the NonZero::bit_width method is useful because it accomplishes the same thing a little more succinctly.

Note that we generally don't add things to NonZero just to avoid a .get(). You'll note, for example, that NonZero::count_ones exists -- because it can return NonZero<u32> -- but NonZero::count_zeroes doesn't, despite it existing on the underlying types, since there's nothing special for NonZero to do on it.

If, as @hanna-kruppe mentions, this can return NonZero, that might be a reason to justify its existence.

@rustbot author

[rustbot]

Reminder, once the PR becomes ready for a review, use @rustbot ready.

[sorairolake]

@scottmcm In safe Rust, the minimum number of bits required to represent unsigned NonZero<T> is guaranteed to be greater than or equal to one. Therefore, this method can return NonZero<u32>.

Based on the comments, I changed the return type of this method from u32 to NonZero<u32>. This type is suitable to indicate that this method can never return zero. Also, other NonZero<T> methods may be easier to use with this type than with u32.

[sorairolake]

@rustbot ready

[scottmcm]

Sure; for unstable seems close enough to fit under the existing unstable library feature.

@bors r+ rollup

[bors]

📌 Commit a25950d has been approved by scottmcm

It is now in the queue for this repository.

[sorairolake]

• Tracking issue: Tracking Issue for uint_bit_width #142326

This pull request adds a method to the unsigned NonZero<T> that return the minimum number of bits required to represent a value.

This can be achieved by using the get method and the methods added in #142328, but I think adding the NonZero::bit_width method is useful because it accomplishes the same thing a little more succinctly.

[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

[sorairolake]

@rustbot label +T-libs-api -T-libs

r? libs-api

[hanna-kruppe]

Can’t this return NonZeroU32? That would be a good argument for a dedicated NonZero method instead of delegating to get().

[scottmcm]

but I think adding the NonZero::bit_width method is useful because it accomplishes the same thing a little more succinctly.

Note that we generally don't add things to NonZero just to avoid a .get(). You'll note, for example, that NonZero::count_ones exists -- because it can return NonZero<u32> -- but NonZero::count_zeroes doesn't, despite it existing on the underlying types, since there's nothing special for NonZero to do on it.

If, as @hanna-kruppe mentions, this can return NonZero, that might be a reason to justify its existence.

@rustbot author

[rustbot]

Reminder, once the PR becomes ready for a review, use @rustbot ready.

[sorairolake]

@scottmcm In safe Rust, the minimum number of bits required to represent unsigned NonZero<T> is guaranteed to be greater than or equal to one. Therefore, this method can return NonZero<u32>.

Based on the comments, I changed the return type of this method from u32 to NonZero<u32>. This type is suitable to indicate that this method can never return zero. Also, other NonZero<T> methods may be easier to use with this type than with u32.

[sorairolake]

@rustbot ready

[scottmcm]

Sure; for unstable seems close enough to fit under the existing unstable library feature.

@bors r+ rollup

[bors]

📌 Commit a25950d has been approved by scottmcm

It is now in the queue for this repository.

[sorairolake]

• Tracking issue: Tracking Issue for uint_bit_width #142326

This pull request adds a method to the unsigned NonZero<T> that return the minimum number of bits required to represent a value.

This can be achieved by using the get method and the methods added in #142328, but I think adding the NonZero::bit_width method is useful because it accomplishes the same thing a little more succinctly.

[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

[sorairolake]

@rustbot label +T-libs-api -T-libs

r? libs-api

[hanna-kruppe]

Can’t this return NonZeroU32? That would be a good argument for a dedicated NonZero method instead of delegating to get().

[scottmcm]

but I think adding the NonZero::bit_width method is useful because it accomplishes the same thing a little more succinctly.

Note that we generally don't add things to NonZero just to avoid a .get(). You'll note, for example, that NonZero::count_ones exists -- because it can return NonZero<u32> -- but NonZero::count_zeroes doesn't, despite it existing on the underlying types, since there's nothing special for NonZero to do on it.

If, as @hanna-kruppe mentions, this can return NonZero, that might be a reason to justify its existence.

@rustbot author

[rustbot]

Reminder, once the PR becomes ready for a review, use @rustbot ready.

[sorairolake]

@scottmcm In safe Rust, the minimum number of bits required to represent unsigned NonZero<T> is guaranteed to be greater than or equal to one. Therefore, this method can return NonZero<u32>.

Based on the comments, I changed the return type of this method from u32 to NonZero<u32>. This type is suitable to indicate that this method can never return zero. Also, other NonZero<T> methods may be easier to use with this type than with u32.

[sorairolake]

@rustbot ready

[scottmcm]

Sure; for unstable seems close enough to fit under the existing unstable library feature.

@bors r+ rollup

[bors]

📌 Commit a25950d has been approved by scottmcm

It is now in the queue for this repository.

[sorairolake]

• Tracking issue: Tracking Issue for uint_bit_width #142326

This pull request adds a method to the unsigned NonZero<T> that return the minimum number of bits required to represent a value.

This can be achieved by using the get method and the methods added in #142328, but I think adding the NonZero::bit_width method is useful because it accomplishes the same thing a little more succinctly.

[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

[sorairolake]

@rustbot label +T-libs-api -T-libs

r? libs-api

[hanna-kruppe]

Can’t this return NonZeroU32? That would be a good argument for a dedicated NonZero method instead of delegating to get().

[scottmcm]

but I think adding the NonZero::bit_width method is useful because it accomplishes the same thing a little more succinctly.

Note that we generally don't add things to NonZero just to avoid a .get(). You'll note, for example, that NonZero::count_ones exists -- because it can return NonZero<u32> -- but NonZero::count_zeroes doesn't, despite it existing on the underlying types, since there's nothing special for NonZero to do on it.

If, as @hanna-kruppe mentions, this can return NonZero, that might be a reason to justify its existence.

@rustbot author

[rustbot]

Reminder, once the PR becomes ready for a review, use @rustbot ready.

[sorairolake]

@scottmcm In safe Rust, the minimum number of bits required to represent unsigned NonZero<T> is guaranteed to be greater than or equal to one. Therefore, this method can return NonZero<u32>.

Based on the comments, I changed the return type of this method from u32 to NonZero<u32>. This type is suitable to indicate that this method can never return zero. Also, other NonZero<T> methods may be easier to use with this type than with u32.

[sorairolake]

@rustbot ready

[scottmcm]

Sure; for unstable seems close enough to fit under the existing unstable library feature.

@bors r+ rollup

[bors]

📌 Commit a25950d has been approved by scottmcm

It is now in the queue for this repository.

[sorairolake]

• Tracking issue: Tracking Issue for uint_bit_width #142326

This pull request adds a method to the unsigned NonZero<T> that return the minimum number of bits required to represent a value.

This can be achieved by using the get method and the methods added in #142328, but I think adding the NonZero::bit_width method is useful because it accomplishes the same thing a little more succinctly.

[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

[sorairolake]

@rustbot label +T-libs-api -T-libs

r? libs-api

[hanna-kruppe]

Can’t this return NonZeroU32? That would be a good argument for a dedicated NonZero method instead of delegating to get().

[scottmcm]

but I think adding the NonZero::bit_width method is useful because it accomplishes the same thing a little more succinctly.

Note that we generally don't add things to NonZero just to avoid a .get(). You'll note, for example, that NonZero::count_ones exists -- because it can return NonZero<u32> -- but NonZero::count_zeroes doesn't, despite it existing on the underlying types, since there's nothing special for NonZero to do on it.

If, as @hanna-kruppe mentions, this can return NonZero, that might be a reason to justify its existence.

@rustbot author

[rustbot]

Reminder, once the PR becomes ready for a review, use @rustbot ready.

[sorairolake]

@scottmcm In safe Rust, the minimum number of bits required to represent unsigned NonZero<T> is guaranteed to be greater than or equal to one. Therefore, this method can return NonZero<u32>.

Based on the comments, I changed the return type of this method from u32 to NonZero<u32>. This type is suitable to indicate that this method can never return zero. Also, other NonZero<T> methods may be easier to use with this type than with u32.

[sorairolake]

@rustbot ready

[scottmcm]

Sure; for unstable seems close enough to fit under the existing unstable library feature.

@bors r+ rollup

[bors]

📌 Commit a25950d has been approved by scottmcm

It is now in the queue for this repository.

[sorairolake]

• Tracking issue: Tracking Issue for uint_bit_width #142326

This pull request adds a method to the unsigned NonZero<T> that return the minimum number of bits required to represent a value.

This can be achieved by using the get method and the methods added in #142328, but I think adding the NonZero::bit_width method is useful because it accomplishes the same thing a little more succinctly.

[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

[sorairolake]

@rustbot label +T-libs-api -T-libs

r? libs-api

[hanna-kruppe]

Can’t this return NonZeroU32? That would be a good argument for a dedicated NonZero method instead of delegating to get().

[scottmcm]

but I think adding the NonZero::bit_width method is useful because it accomplishes the same thing a little more succinctly.

Note that we generally don't add things to NonZero just to avoid a .get(). You'll note, for example, that NonZero::count_ones exists -- because it can return NonZero<u32> -- but NonZero::count_zeroes doesn't, despite it existing on the underlying types, since there's nothing special for NonZero to do on it.

If, as @hanna-kruppe mentions, this can return NonZero, that might be a reason to justify its existence.

@rustbot author

[rustbot]

Reminder, once the PR becomes ready for a review, use @rustbot ready.

[sorairolake]

@scottmcm In safe Rust, the minimum number of bits required to represent unsigned NonZero<T> is guaranteed to be greater than or equal to one. Therefore, this method can return NonZero<u32>.

Based on the comments, I changed the return type of this method from u32 to NonZero<u32>. This type is suitable to indicate that this method can never return zero. Also, other NonZero<T> methods may be easier to use with this type than with u32.

[sorairolake]

@rustbot ready

[scottmcm]

Sure; for unstable seems close enough to fit under the existing unstable library feature.

@bors r+ rollup

[bors]

📌 Commit a25950d has been approved by scottmcm

It is now in the queue for this repository.