Nathaniel Martin (v1.3.1)

What's Changed

This release fixes one bug in the DivMod method, which under certain types of aliased inputs would produce faulty output. It also adds a method IntoBig for alloc-free conversion into big.Int.

• conversion: introduce 0-alloc IntoBig method by @karalabe in #177
• divmod: fix aliasing error, add tests by @holiman in #180

Full Changelog: v1.3.0...v1.3.1

---

Nathaniel Martin is a character from Patrick O'Brian's Aubrey/Maturin - books, set in the 1800s, the age of fighting sail. Nathaniel Martin is a natural philosopher and is a kind of sidekick to Stephen Maturin.

Fitcher Penrose (v1.3.0)

The 1.3.0-release contains a lot of optimizations, contributed by @AaronChen0. Thanks for all the contributions! Other than that, a breaking change in the ssz-interface implementation and a bugfix.

What's Changed

• optimizations to Mul, squared, Exp, Set, DivMod, Byte, Lsh, Rsh, SRsh, ExtendSign, AddMod, MulOverflow, MulMod, MulDivOverflow, MulModWithReciprocal , Mod, Sqrt by @AaronChen0 in (#152, #153, #154, #158, #165, #166, #167, #168, #169, #172, #173, #174

• Fixes a bug in SetBytes29 (#157)

• Test-improvements, now using golang-native fuzzing instead of gofuzz-based fuzzing by @holiman (#163, #164)

• The method MarshalSSZTo was removed, and has been replaced by MarshalSSZAppend and MarshalSSZInto. This is an intentionally breaking change, forcing users to explicitly choose which implementation they desire. For more info, see #171

Full Changelog: v1.2.4...v1.2.5

---

The uint256 release naming-convention is "fictional mathematicians", later expanded into "fictional scientists", due to the dearth of fiction about mathematicians.

Fitcher Penrose appears in Amitav Ghosh's Ibis-trilogy, part II, "River of Smoke". He is a botanist collecting rare plants in Asia.

Stephen Maturin (v1.2.4)

What's Changed

• Add go 1.21, make go 1.19 minimum version by @holiman in #142
• Optimize Log10() by @fyfyrchik in #141
• Conversion: behavioural changes in String, MarshalText and MarshalJSON by @holiman in #144

OBS! #144 is a bit of a breaking change, in some situations, since it changes the output-formats when converting to string, in order to better mirror big.Int behaviour. Please read the PR-description to understand the effects of the change.

New Contributors

• @fyfyrchik made their first contribution in #141

Full Changelog: v1.2.3...v1.2.4

Kit Traverse (v1.2.3)

What's Changed

• Implemented Float64() float64 by @holiman in #132
• Implemented CmpBig by @holiman in #138
• Implemented Log10 and CmpUint64 by @holiman in #136
• Allow converting between nil big.Int and nil uint256.Int by @karalabe in #137
• tests: naming conventions for benchmarks, remove obsolete by @holiman in #133

New API Methods

// Float64 returns the float64 value nearest to x.
//
// Note: The `big.Float` version of `Float64` also returns an 'Accuracy', indicating
// whether the value was too small or too large to be represented by a
// `float64`. However, the `uint256` type is unable to represent values
// out of scope (|x| < math.SmallestNonzeroFloat64 or |x| > math.MaxFloat64),
// therefore this method does not return any accuracy.
func (z *Int) Float64() float64

// CmpBig compares z and x and returns:
//
//	-1 if z <  x
//	 0 if z == x
//	+1 if z >  x
func (z *Int) CmpBig(x *big.Int) (r int)

// CmpUint64 compares z and x and returns:
//
//	-1 if z <  x
//	 0 if z == x
//	+1 if z >  x
func (z *Int) CmpUint64(x uint64) int 


// Log10 returns the log in base 10, floored to nearest integer.
// **OBS** This method returns '0' for '0', not `-Inf`.
func (z *Int) Log10() uint

Full Changelog: v1.2.2...v1.2.3

Will Hunting (v1.2.2)

What's Changed

• Interfaces to work nicely with fastssz by @karalabe in #126
• String conversion from decimal by @elee1766 and @holiman in #122 and #127
• Nice constructions via MustFromBig, MustFromHex and MustFromDecimal by @karalabe and @holiman in #128, #131
• Implement PrettyDec and native Dec by @holiman in #130
• Optimize AddMod by @jwasinger in #120

New API-methods

Methods to create Ints

// FromDecimal is a convenience-constructor to create an Int from a
// decimal (base 10) string. Numbers larger than 256 bits are not accepted.
func FromDecimal(decimal string) (*Int, error) 

// MustFromBig is a convenience-constructor from big.Int.
// Returns a new Int and panics if overflow occurred.
func MustFromBig(b *big.Int) *Int

// MustFromHex is a convenience-constructor to create an Int from
// a hexadecimal string.
// Returns a new Int and panics if any error occurred.
func MustFromHex(hex string) *Int 

// MustFromDecimal is a convenience-constructor to create an Int from a
// decimal (base 10) string.
// Returns a new Int and panics if any error occurred.
func MustFromDecimal(decimal string) *Int 

Methods to initialize Int instances

// SetFromDecimal sets z from the given string, interpreted as a decimal number.
// OBS! This method is _not_ strictly identical to the (*big.Int).SetString(..., 10) method.
// Notable differences:
// - This method does not accept underscore input, e.g. "100_000",
// - This method does not accept negative zero as valid, e.g "-0",
//   - (this method does not accept any negative input as valid))
func (z *Int) SetFromDecimal(s string) (err error)

// SetFromHex sets z from the given string, interpreted as a hexadecimal number.
// OBS! This method is _not_ strictly identical to the (*big.Int).SetString(..., 16) method.
// Notable differences:
// - This method _require_ "0x" or "0X" prefix.
// - This method does not accept zero-prefixed hex, e.g. "0x0001"
// - This method does not accept underscore input, e.g. "100_000",
// - This method does not accept negative zero as valid, e.g "-0x0",
//   - (this method does not accept any negative input as valid)
func (z *Int) SetFromHex(hex string) error

// Scan implements the database/sql Scanner interface.
// It decodes a string, because that is what postgres uses for its numeric type
func (dst *Int) Scan(src interface{}) error

Methods to output into strings

// MarshalJSON implements json.Marshaler.
func (z *Int) MarshalJSON() ([]byte, error)

// Value implements the database/sql/driver Valuer interface.
func (src *Int) Value() (driver.Value, error)

// Dec returns the decimal representation of z.
func (z *Int) Dec() string

// PrettyDec returns the decimal representation of z, with thousands-separators.
func (z *Int) PrettyDec(separator byte) string 

Methods related to ssz encoding

// MarshalSSZTo implements the fastssz.Marshaler interface and serializes the
// integer into an already pre-allocated buffer.
func (z *Int) MarshalSSZTo(dst []byte) ([]byte, error) 

// MarshalSSZ implements the fastssz.Marshaler interface and returns the integer
// marshalled into a newly allocated byte slice.
func (z *Int) MarshalSSZ() ([]byte, error) 

// SizeSSZ implements the fastssz.Marshaler interface and returns the byte size
// of the 256 bit int.
func (*Int) SizeSSZ() int 

// UnmarshalSSZ implements the fastssz.Unmarshaler interface and parses an encoded
// integer into the local struct.
func (z *Int) UnmarshalSSZ(buf []byte) error

// HashTreeRoot implements the fastssz.HashRoot interface's non-dependent part.
func (z *Int) HashTreeRoot() ([32]byte, error)

New Contributors

• @jwasinger made their first contribution in #120
• @elee1766 made their first contribution in #122

Full Changelog: v1.2.1...v1.2.2

Hari Seldon

The v1.2.1 release, "Hari Seldon", contains some new API-methods and a lot of improvements under the hood and in the surrounding infrastructure.

• OSS-fuzz integration has finally happened, as of (#98) and google/oss-fuzz#6497. The uint256 library is now continuously fuzzed by the fine folks at OSS-Fuzz.
• Optimised modular arithmetic targeting elliptic curve operations (#86). This was a major overhaul and complete rewrite of the modular arithmetic operations.
• Added DivMod() (#113)
• Added MulDivOverflow() (#110)
• Added Sqrt() (#104)
• Improvements ot Cmp() (#85)
• Test improvements (#112, #95, #93, #92, #91, #89, #88)

Thanks to all contributors to this release: @holiman, @chfast,@daosvik, @Planxnx and @fyrchik.

Enoch Root

The 1.2.0 release, "Enoch Root", contains an important fix to SubUint64, which was flaweed (see #82 ).
Other than that some API methods have been changed, to make uint256.Int closer resemble big.Int, to be easier to use as a drop-in replacement (also #82 ).

Daniel Waterhouse

This release adds various methods for marshalling from text/json, to make integration with go-ethereum easier

Lawrence Waterhouse

The 1.1.0 release of uint256 contains some improvements and new features.

• The function MulOverflow was added (#75 and #76)
• Squaring, multiplication and expontentiation was improved with 15-20% (#71)
• EncodeRLP method was added (#74)
• The SetBytes-method was improved by an order of magnitude, and the extremely fast specialized methods SetBytes1, SetBytes2 etc were added (#70).
• Compatibility with golang-1.12 was fixed (#72)
• Tests are now also run on big-endian architecture (#73)

Roger Mexico

This is the first official production release of uint256.

Most of the really great changes recently are due to @chfast, who has done tremendous work on:

• Ensuring a 100% test coverage,
• Ensuring allocation free operations,
• Improve the public API,
• Verify functionality on big endian architecture

And last but not least, further improve speeds everywhere.

Release-names are mathematicians in fiction