Trait DigitsIter

Source

pub trait DigitsIter<'a>: Iterator<Item = &'a u8> + Iter<'a> {
Show 13 methods    // Required methods
    fn increment_count(&mut self);
    fn peek(&mut self) -> Option<Self::Item>;
    fn is_digit(&self, value: u8) -> bool;

    // Provided methods
    fn is_consumed(&mut self) -> bool { ... }
    fn try_read(&mut self) -> Option<Self::Item> { ... }
    fn peek_is_cased(&mut self, value: u8) -> bool { ... }
    fn peek_is_uncased(&mut self, value: u8) -> bool { ... }
    fn peek_is(&mut self, value: u8, is_cased: bool) -> bool { ... }
    fn read_if<Pred: FnOnce(u8) -> bool>(&mut self, pred: Pred) -> Option<u8> { ... }
    fn read_if_value_cased(&mut self, value: u8) -> Option<u8> { ... }
    fn read_if_value_uncased(&mut self, value: u8) -> Option<u8> { ... }
    fn read_if_value(&mut self, value: u8, is_cased: bool) -> Option<u8> { ... }
    fn skip_zeros(&mut self) -> usize { ... }
}

Expand description

Iterator over a contiguous block of bytes.

This allows us to convert to-and-from-slices, raw pointers, and peek/query the data from either end cheaply.

A default implementation is provided for slice iterators. This trait should never return null from as_ptr, or be implemented for non-contiguous data.

Required Methods§

Source

fn increment_count(&mut self)

Increment the number of digits that have been returned by the iterator.

For contiguous iterators, this is a no-op. For non-contiguous iterators, this increments the count by 1.

Source

fn peek(&mut self) -> Option<Self::Item>

Peek the next value of the iterator, without consuming it.

Note that this can modify the internal state, by skipping digits for iterators that find the first non-zero value, etc. We optimize this for the case where we have contiguous iterators, since non-contiguous iterators already have a major performance penalty.

Source

fn is_digit(&self, value: u8) -> bool

Determine if the character is a digit.

Provided Methods§

Source

fn is_consumed(&mut self) -> bool

Get if the iterator cannot return any more elements.

This may advance the internal iterator state, but not modify the next returned value.

If this is an iterator, this is based on the number of items left to be returned. We do not necessarly know the length of the buffer. If this is a non-contiguous iterator, this MUST advance the state until it knows a value can be returned.

Any incorrect implementations of this affect all safety invariants for the rest of the trait. For contiguous iterators, this can be as simple as checking if self.cursor >= self.slc.len(), but for non-contiguous iterators you MUST advance to the next element to be returned, then check to see if a value exists. The safest implementation is always to check if self.peek().is_none() and ensure [peek] is always safe.

If you would like to see if the cursor is at the end of the buffer, see [is_buffer_empty] instead.

Source