ErgoScript release v3.1.0 for Ergo mainnet

• Merged core modules from scalan/special into sigmastate repository (closes #457)
• Necessary code is copied from Special and added to this repo into
  common, core, library-api, library-impl, library modules. (see #566)
• The core code from Special is minimised and optimised for performance
• The copied code is kept in sync with Special (which allow to use Special tests which cannot be copied)
• the content of sigma-api, sigma-impl, sigma-library modules is minimised and optimised for performance
• the changes preserve equivalence of costing, complexity and script results
• fixed Pay2SHAddress construction from ErgoTree (#596)
• Release publishing (#619)

ErgoScript release v3.0.3 for Ergo mainnet

• DataJsonEncoder for Json serialisation of ErgoScript data values

ErgoScript release v3.0.2 for Ergo mainnet

• Json encoders for Context data
• refactoring and improvement in Context representation.

ErgoScript release v3.0.1 for Ergo mainnet

• Reduced memory footprint (especially in (de-)serialization)
• PR review policy (see pr-review-policy.md)

ErgoScript release v3.0.0 for Ergo mainnet

• improved reliability of cost estimation
• more tests added for better coverage

ErgoScript Release v2.2

• soft-forkability for cost estimation (#503)
• optimization of cost estimation rules (#523)
• new Context parameters (initCost, costLimit)
• implemented fast complexity measure of box propositions
  (to fail-fast on too complex scripts #537)
• implemented test cases to check fast script rejections of
  of either oversized of too costly scripts

ErgoScript Release v2.1.2

• more tests added

ErgoScript Release v2.1.1

• Fix for running on JRE8

Soft-forkability, ErgoTree specification, ICO & LETS examples

• soft-forkability for language evolution (add types, operations) #500
• ErgoTree language specification (abstract syntax, typing rules, semantics, serialization format) #495
• ErgoTree IR extended with metadata to generate specification appendixes.
• more operations implemented (zip #498, element-wise xor for collections #494, flatMap #493,
  plusModQ, minusModQ #488, logical xor #478, filter #471, Option.map #469, groupGenerator #440)
• security improvements (like limiting ErgoTree depth during deserialization #459 #482)
• ICO and LETS examples #450
• final version of ErgoScript white paper (#410)
• improvements in scorex-util serialization
• various fixes code cleanup and better test coverage (#504, #508, #515)

Improved costing, new operations, new context properties

• new and more general costing algorithm
• optimized execution of cost formulas
• new elliptic curve for GroupElement operations
• new context properties (see special.sigma.Context trait)
  • dataInputs - A collection of inputs of the current transaction that will not be spent
  • headers - a fixed number of last block headers in descending order (first header is the newest one)
  • preHeader - fields of block header with the current spending transaction, that can be predicted
    by a miner before it's formation
• new operations for authenticated trees (see special.sigma.AvlTree trait), including chained compositions
• proveDlog, proveDHTuple can accept any expression as arguments
• def f(x: Int): Int declarations of functions
• various internal improvements in type checker, compiler, byte code interpreter etc.