## Expand description

## §fletcher

A dependency free implementation of the Fletcher’s checksum algorithm

Fletcher’s Checksum was developed to provide nearly the same error checking capability as a CRC but with a faster software implementation. This is not a cryptographically secure checksum, it’s only meant to be used for checking the integrity of data NOT the authenticity.

### §Algorithm Pros ✅

This algorithm is faster to run in software than most CRCs. This is because the CRC algorithm was originally designed to be simple to implement in hardware, but not neccesarily in software. The Fletcher Checksum was designed specifically to be suited for implementation in software.

### §Algorithm Cons ❌

This checksum algorithm does suffer from not being able to distinguish `0xFF`

from `0x00`

. Meaning a block of data with all bits set to 1 will have the exact
same the same checksum as a block of data with all bits set to 0. This comes
from the fact that the algorithm uses one’s complement math.

Fletcher’s checksum isn’t quite as good at detecting bit errors in data as a CRC with a well choosen polynomial.

### §How To Use

If you have an entire block of data you want to get the checksum of you can
use the calc functions (`calc_fletcher16`

, `calc_fletcher32`

, `calc_fletcher64`

)
to get the checksum in a single function call.

If you are getting the data in chunks you can make a `Fletcher`

object
(`Fletcher16`

, `Fletcher32`

, `Fletcher64`

) to manage the intermediate
state between chunks of data. The checksum objects take in slices of data to
process. There is no minimum length required of the slices, all of the provided
data will be processed to completion. The type of the input data is dictated by
the size of the checksum value. i.e. a 64-bit checksum operates on 32-bit wide
values.

The checksum object can be queried for it’s current checksum value as any
time with the `Fletcher::value()`

function.

### §Example

```
let data: [u8; 6] = [0xC1, 0x77, 0xE9, 0xC0, 0xAB, 0x1E];
assert_eq!(fletcher::calc_fletcher16(&data), 0x3FAD);
// Or if you want to work on smaller chunks of data
let mut checksum = fletcher::Fletcher16::new();
checksum.update(&data[0..3]);
checksum.update(&data[3..]);
assert_eq!(checksum.value(), 0x3FAD);
```

## Structs§

- Type to hold the state for calculating a fletcher checksum.

## Traits§

- Base set of values and operations needed for our implementation

## Functions§

- Get the 16-bit checksum in one shot
- Get the 32-bit checksum in one shot
- Get the 64-bit checksum in one shot

## Type Aliases§

- Produces a 16-bit checksum from a stream of 8-bit data.
- Produces a 32-bit checksum from a stream of 16-bit data.
- Produces a 64-bit checksum from a stream of 32-bit data.