ChangeLog.org

v0.2.8

Newest to oldest additions in each section.

New

• destructure/pattern match in let bindings, j14159
• alpaca_printer, type formatting and printing for compiler feedback, lepoetemaudit
• build for OTP 20, getong
• logical operators (and, or, xor), Licenser
• explicit type signatures for top-level bindings, e.g. val add: fn int int -> int, lepoetemaudit
• nicely formatted and humane compiler output with colours, Licenser
• local/inner functions can be curried, lepoetemaudit
• able to specify default types and functions to be imported in a project, lepoetemaudit
• infix functions starting with < are now right-associative, lepoetemaudit
• generate a module’s dependencies from source to support partial compilation, lepoetemaudit
• support loading of pre-compiled BEAM files rather than requiring every module in source form, lepoetemaudit
• quoted atoms, lepoetemaudit
• German error message translation, ypaq
• Spanish error message translation, arpunk and yurrriq
• symbol, integer, and float AST nodes defined in Alpaca itself, j14159
• enable internationalization and formatting of error messages, danabr

Fixes

• wildcard variables (the `_` variable) in function head matches are now renamed properly, j14159
• partial record matching works, no more cycles in reference cells, j14159
• type inferencer reference cells in ETS, no more processes, j14159
• unsized binaries work correctly when constructing binaries, lepoetemaudit
• simple expressions don’t require parens, e.g. [1 + 2, 3 + 4], lepoetemaudit
• lambdas inside match expressions now work correctly, lepoetemaudit
• don’t re-type modules for every module-qualified function call, lepoetemaudit
• correctly rewrite the internals of unbound lambdas, j14159
• don’t perform exhaustiveness checks on already-compiled modules, lepoetemaudit
• fix unification failure for nested ADTs, j14159
• fn as a reserved word in tests, Licenser
• purge warnings removed for cleaner test output, Licenser
• substantial test and warnings cleanup, Licenser
• fixed variable lookup failure in exhaustiveness checks, j14159
• broken BIF fixes (!=, /), lepoetemaudit
• fix lambdas in records, j14159
• properly escape \ sequences in strings, lepoetemaudit
• usage example correction, monkeygroover
• rewrite imported functions in tests, j14159
• rewrite lambdas in tests, j14159
• allow forward references to top-level bindings, j14159
• parameterized pid type fix when using is_pid, j14159

v0.2.7

New

• parser and AST generation errors report in a uniform manner and are specified, https://github.com/danabr
• lambdas (anonymous functions), https://github.com/j14159
• function type specs have same syntax as lambdas, https://github.com/danabr
• generation test/0 for EUnit, https://github.com/j14159

Fixes

• correct arity for curried versions of functions, https://github.com/lepoetemaudit
• built-in types are no longer reserved words, https://github.com/danabr
• type arity errors unified, https://github.com/danabr
• bad variables in errors (throw, error, exit) are errors, https://github.com/j14159
• all parser and AST generation errors include line numbers, https://github.com/danabr
• line numbers for module definitions, https://github.com/danabr
• line numbers for unexported functions, https://github.com/danabr
• all clauses can lead with a |, https://github.com/j14159

v0.2.6

New

• top-level let deprecates all significant whitespace, https://github.com/lepoetemaudit
• same function name allowed for different arity, https://github.com/j14159
• Alpaca’s compiled modules prefixed with alpaca_, https://github.com/lepoetemaudit
• zero-arity functions constrained to literals for use as values, https://github.com/lepoetemaudit
• property-based testing with PropEr, generates full modules for testing, https://github.com/ypaq
• import functions from other modules, https://github.com/j14159
• exhaustiveness checking for top-level functions, https://github.com/danabr
• functions as ADT members, https://github.com/danabr
• automatic currying of top-level functions, https://github.com/lepoetemaudit
• type aliases for arrow types, https://github.com/danabr
• record transformations, https://github.com/j14159

Fixes

• apply arguments to an expression that evaluates to a function, https://github.com/j14159
• many whitespace fixes, https://github.com/lepoetemaudit and https://github.com/ypaq
• unused code and warnings cleanup, https://github.com/jkakar
• FFI calls must refer to missing variables, https://github.com/danabr
• unit value code generation, https://github.com/danabr
• reduced number of reserved words, https://github.com/danabr
• module-qualified constructors and types default to private, https://github.com/j14159
• more obvious typing of higher order types, https://github.com/danabr
• unified internal types for lists and maps, https://github.com/danabr
• referring to nonexistent types and type variables are errors, https://github.com/j14159
• unified error handling in AST generation, https://github.com/danabr

v0.2.5

• early infix function support (operators, e.g. |>) from https://github.com/lepoetemaudit
• multiple instances of _ permitted in patterns
• throw/1, exit1/, and error/1 are now available
• patterns in function arguments, e.g. g f Some x = Some (f x) or get_x {x=x} = x
• types default to private in their defining module with explicit export and import in others
• minor type inferencer fixes for records and tuples

v0.2.4

• fixes for type aliases and unions involving them: https://github.com/danabr
• early record support with row polymorphism
• fixes for unification of the value type portion of maps

v0.2.3

• comment syntax fixes: https://github.com/danabr
• formatting and cleanup: https://github.com/tuncer
• polymorphic ADT unification fix: https://github.com/danabr
• renamed call_erlang to beam for clarity: https://github.com/tuncer
• built-in parametric types (e.g. list, map) can be parameterized by ADTs: https://github.com/danabr
• fix for polymorphic process spawning: https://github.com/j14159

v0.2.2

• Haskell-style comments, from https://github.com/tuncer
• nested comments are errors, from https://github.com/yurrriq
• file renaming (cleanup) from https://github.com/tuncer (scanner -> mlfe_scanner)
• new compiler entry points from https://github.com/tuncer

v0.2.1

• unary minus fix (parser corrections) from https://github.com/danabr. Makes symbol-1 work as symbol - 1 does.

v0.2.0

• dialyzer and xref fixes, Travis CI integration, courtesy of https://github.com/NobbZ
• module_info synthesized by the compiler, courtesy of https://github.com/erszcz
• typos and clarifications from https://github.com/tjweir, https://github.com/saem, https://github.com/ypaq, https://github.com/omarkj, https://github.com/yurrriq
• binaries, thanks to https://github.com/yurrriq and https://github.com/talentdeficit for feedback and ideas
• basic map support
• change to `spawn` syntax, it now takes a (syntactic) function application instead of function and arg list, thanks to https://github.com/ypaq for feedback
• a “language tour” document, thanks to https://github.com/ypaq and https://github.com/saem for initial feedback and corrections
• a simple unit test form/hook for creating basic EUnit-compatible tests
• UTF-8 strings, compiled as binaries
• character lists for Erlang string support
• lots of little type inferencer/checker fixes