template<RangeUnit Unit, Endian Order>
struct HalfOpenRange< Unit, Order >
A half-open range of bits or bytes - [lo, hi) - specified in terms of a specific endian order. Half-open ranges include lo but do not include hi, so HalfOpenRange(3, 5) contains 3 and 4 but not 5.
Use a half-open range when you want to allow for the possibility that the range may be empty, which may be represented by setting lo and hi to the same value. Using a half-open range may also make some algorithms more natural to express, and it may make working with external code easier since half-open ranges are idiomatic in C++.
Note that there are many ways to represent an empty range - (1, 1) is empty, for example, but so is (2, 2). Many operations on HalfOpenRanges will canonicalize the empty range representation, so don't rely on it - just call empty() to determine if a range is empty.
XXX: Currently, for backwards compatibility, it's possible to construct ranges where lo is greater than hi. We should enforce that ranges are consistent; we'll add the necessary checks after the existing code has been audited.
XXX: We should also add checks to avoid integer overflow.
template<RangeUnit Unit, Endian Order>
template<Endian DestOrder>
Convert this range to a range with the specified endian order.
Making this conversion requires specifying the size of the larger space that this range is embedded in. That's because the place we start numbering indices from (the origin, in other words) is changing. For example, given a space with 10 elements, a range of 3 indices within that space may be numbered in two ways as follows: Space: [o o o o o o o o o o] Big endian range: --> [2 3 4] Little endian range: [7 6 5] <-----— We couldn't switch between those numberings without knowing that the space has 10 elements.
Note that this operation may result in integer overflow when dealing with very large ranges. It's generally safe for ZeroToMax as long as the space size is not also near INT_MAX. It's never safe for MinToMax.
- Template Parameters
-
| DestOrder | The endian order to convert to. |
- Parameters
-
| spaceSize | The size of the space this range is embedded in. |
- Returns
- this range, but converted to the specified endian ordering.
template<RangeUnit Unit, Endian Order>
template<RangeUnit DestUnit>
Convert this range to the smallest enclosing range with the specified unit.
Conversion from bytes to bits is exact, but conversion from bits to bytes is lossy and may cause the size of the range to grow if the bits aren't byte-aligned. If that would be problematic, use isLoAligned() and isHiAligned() to check before converting, or just check that converting the result back to bits yields the original range.
This operation will result in integer overflow when dealing with ZeroToMax or MinToMax-sized ranges.
- Template Parameters
-
| DestUnit | The unit to convert to. |
- Returns
- this range, but converted to the specified unit.
1.11.0