This package implements generic CRC calculations up to 64 bits wide. It aims to be fairly fast and fairly complete, allowing users to match pretty much any CRC algorithm used in the wild by choosing appropriate Parameters. This obviously includes all popular CRC algorithms, such as CRC64-ISO, CRC64-ECMA, CRC32, CRC32C, CRC16, CCITT, XMODEM and many others. See http://reveng.sourceforge.net/crc-catalogue/ for a good list of CRC algorithms and their parameters.
This package has been largely inspired by Ross Williams' 1993 paper "A Painless Guide to CRC Error Detection Algorithms".
To install, simply execute:
go get github.com/snksoft/crc
or:
go get gopkg.in/snksoft/crc.v1
Using crc
is easy. Here is an example of calculating a CCITT crc.
package main
import (
"fmt"
"github.com/snksoft/crc"
)
func main() {
data := "123456789"
ccittCrc := crc.CalculateCRC(crc.CCITT, []byte(data))
fmt.Printf("CRC is 0xX\n", ccittCrc) // prints "CRC is 0x29B1"
}
For larger data, table driven implementation is faster. Note that crc.Hash
implements hash.Hash
interface, so you can use it instead if you want.
Here is how to use it:
package main
import (
"fmt"
"github.com/snksoft/crc"
)
func main() {
data := "123456789"
hash := crc.NewHash(crc.XMODEM)
xmodemCrc := hash.CalculateCRC([]byte(data))
fmt.Printf("CRC is 0xX\n", xmodemCrc) // prints "CRC is 0x31C3"
// You can also reuse hash instance for another crc calculation
// And if data is too big, you may feed it in chunks
hash.Reset() // Discard crc data accumulated so far
hash.Update([]byte("123456789")) // feed first chunk
hash.Update([]byte("01234567890")) // feed next chunk
xmodemCrc2 := hash.CRC() // gets CRC of whole data "12345678901234567890"
fmt.Printf("CRC is 0xX\n", xmodemCrc2) // prints "CRC is 0x2C89"
}
In this version I have separated actual CRC caclulations and Hash interface implementation. New Table
type incorporates table based implementation which can be used without creating a Hash
instance. The main difference is that Table
instances are essentially immutable once initialized. This greatly simplifies concurrent use as Table
instances can be safely used in concurrent applications without tricky copying or synchronization. The downside is, however, that feeding data in multiple chunks becomes a bit more verbose (as you essentially maintain intermediate crc in your code and keep feeding it back to subsequent calls). So, you might prefer one or the other depending on situation at hand and personal preferences. You even can ask a Hash
instance for a Table
instance it uses internally and then use both in parallel without recalculating the crc table.
Anyway, here is how to use a Table
directly.
package main
import (
"fmt"
"github.com/snksoft/crc"
)
func main() {
data := []byte("123456789")
// create a Table
crcTable := crc.NewTable(crc.XMODEM)
// Simple calculation all in one go
xmodemCrc := crcTable.CalculateCRC(data)
fmt.Printf("CRC is 0xX\n", xmodemCrc) // prints "CRC is 0x31C3"
// You can also reuse same Table for another crc calculation
// or even calculate multiple crc in parallel using same Table
crc1 := crcTable.InitCrc()
crc1 = crcTable.UpdateCrc(crc1, []byte("1234567890")) // feed first chunk to first crc
crc2 := crcTable.InitCrc()
crc2 = crcTable.UpdateCrc(crc2, data) // feed first chunk to second crc
crc1 = crcTable.UpdateCrc(crc1, []byte("1234567890")) // feed second chunk to first crc
// Now finish calcuation for both
crc1 = crcTable.CRC(crc1)
crc2 = crcTable.CRC(crc2)
fmt.Printf("CRC is 0xX\n", crc1) // prints "CRC is 0x2C89"
fmt.Printf("CRC is 0xX\n", crc2) // prints "CRC is 0x31C3"
}
Beware that Hash
instance is not thread safe. If you want to do parallel CRC calculations (and actually need it to be Hash
, not Table
), then either use NewHash()
to create multiple Hash instances or simply make a copy of Hash whehever you need it. Latter option avoids recalculating CRC table, but keep in mind that NewHash()
returns a pointer, so simple assignement will point to the same instance.
Use either
hash2 := &crc.Hash{}
*hash2 = *hash
or simply
var hash2 = *hash