* fix enum classes entries without initializer
* docs
* update grammar to support scoped identifiers in case labels
* add support for enum class constants in case statements
* add ScopableName to ConstDeclaration; formatting fixes; tests
* update docs
* further optimize shortcircuits
combine jumps if possible
dont emit logicalconvert instructions if the next instruction is another shortcircuit jump
* docs
* first working short curcuit for && and or
creates jumps after each expressions to skip the following
* fixed valuestack when short circuit jmp does not jump
* use specialized instructions for short circuit && and ||
1: lhs
2: logical jump if false/true goto 4 <- if jmp do logical convert
3: rhs
4: logical convert
5: rest
logical convert is needed since "normal" && || operations convert isTrue
to BLong
added BObjectRef BObjectImp set specialization to remove noise
* fixed converted objimp when jump on false values
* first version of short circuit warning
should be moved to analyzer
added whitelist of module functions which have no sideeffect to reduce
the number of warnings
* moved warning visitor to analyzer and added it as extra compile step
fixed sourceline print and cache the content
* missing include, unused member
* included the correct header
* ecompile.cfg to activate and warn
ecompile cmdline arg to activate it
run all tests also with it active
fixed that only the most right side statement was checked
* ecompile cmdline
* increase ecompile version
cleanup
* compilation error
* missing header
* allow -S- to deactivate shortcircuit like the other params do
* extended whitelist functions
* revert fileformat version increase
fixed escript test cmake
* use the correct arg
* escript testoutput can now be different if shortcircuit is active
* docs
* additional test
* addressed comments
* compile time optimization:
int with doubles and strings
doubles with ints and strings
strings with ints and doubles
* bool with other types, more unary ops, float branch optimizer
* more tests
fixed bool to dbl compare
* output cleanup
more tests
* optimize string values in if statements, optimize ternary operator
* optimize elvis, addes missing files, code cleanup
* use array to keep unoptimized if branch in funcexpr tests
* missing include
* better readable testdata
* removed file
* optimize while and dowhile loops if predicate is a compile time known
value
* cleaner variant of loop optimization?
* added ConstantPredicateLoop Node used by the optimizer for constant loop
predicates
optimize repeat until loop
change tests to run the loops more then once to be sure they work
correctly
* test break/continue with label for constant-loop
* docs
* Make ClassDeclaration only track class methods
* Keep methods, constructors in compiler workspace
- Fix UserFunctionBuilder to force reference to methods, constructors
- Fix Optimizer to include methods, constructors in referenced functions
* WIP for ClassDeclarationRegistrar
- Stub implementation of ClassDeclaration and its Registration
- Create "function table entry" via debug statements
* Use function links in methods
* Finish up ClassDescriptor
- Introduce MethodDescriptor for encapsulation
- ClassDeclarationRegistrar can register and take descriptors
* Handle read and write class table in eprog
* Pass Report to code generator for debug statements
* Fix implementation
- It _is_ possible for the constructor to be empty, ie. static classes
* Update DebugStoreSerializer
Print class table to dbg txt file
* Add missing method_names insertion
* Address GitHub CI Annotations
- unused variable 'uf'
- '=': conversion from 'size_t' to 'unsigned int'
* Remove DefaultConstructorFunction
* Modify class parameters; update semantic analysis
- Use ClassDeclarationParameter for class decls vs Identifiers
- Update ScopableName to have equality for constructors and unscoped
function calls
- Update SemanticAnalyzer (self referencing base class, no ctor)
- Update, add tests
* Semantic updates; debug messaging updates
- Ensure base class is constructable
- Various fmt implementations for debugging
* Address self-review comments
* Add Report.debug
* Implement function call scoping
* Implement variable scoping
- Rename `Identifier::scope` to `calling_scope`, and concatenate the identifier scope into `name` if necessary.
- Track current scope in SemanticAnalyzer, similar to UserFunctionVisitor.
- Use current scope in SemanticAnalyzer when visiting identifiers.
* Update grammar for `::identifier` global scoping
* Introduce ScopableName for function resolution
* Address review comments
- Check for display_debugs missing
* Add ScopableName for identifier resolution
* Some cleanup
- Remove `Identifier::calling_scope`, as it is tracked in the semantic
analyzer
- Refactor `ScopeName::exists` to `global` for better clarification of
its use
- Removed `X::maybe_scoped_string/name`
* Address self-review comments
* Rename `Function::module_name` to `scope`
* Update grammar
* WIP with tracking classes
- Refactor AvailableUserFunction to AvailableParseTree
- Track ClassDeclarationContexts similarly to user functions
* Skeleton AST + tracking
* Finish up AST building
* Implement AST building
- TODO decide better name for UserFunctionBuilder/Visitor,
since it does both user functions and classes
* Update Prettifier
* Crude, first-round semantic analysis + codegen
- Currently, no real difference from regular functions
* Move class var statements to top level statements
* Add semantic analysis for base class existence
* Introduce ClassInstance AST node for generating `this` parameter
* Skeleton create class instance instruction
* Default ctor; 'this' byref; Error if multiple same baseclass
* Update grammar for scoped funcrefs and scoped identifiers
* Update prettifier, builders for scoped funcref, identifiers
* Fix CI issues
- Styling
- Shadowing
* Add grammar tests for scope; Remove scope from switch label
* Final draft todos
- Rename `VarStatement::class_name` to `scope`
- Comment why identifiers dont have scopes in enum declarations
- Remove code comment
* initial poc of function expressions
- update grammar
- mock AST builder to return BBoolean(true) for a func expr
- update prettifier for skeleton implementation
* more skeleton work
create AST class FunctionExpression mimicking boolean value
* create ast UserFunction, add to workspace from functexpr
* can generate instructions
* reorg tests; add test for instructions
* can track FunctionDepth in Variable
* can get captures for funcexprs inside funcs.. tbd if this way of nesting works
* Implement create-functor instruction
- Move function depth from Variable to Variables
- Introduce stacking of `Variables` for function expresions via `FunctionVariableScope`
- Add `TOK_FUNCTOR` instruction for 'create-functor'
- Handle emitting a `FunctionExpression` AST node
- Update `emit.declare_variable` and `emit.access_variable` to account for function captures
- Remove `in_function` from instruction generator, as it is tracked via `UserFunction` stack
- Update tests
* Address Discord comments
- Swap pop param order
* Bubble up captured variables through nested functions
* Improve testing infrastructure; add some test cases
* Fix compilation error
* prepend captures to function parameters in funcref mth_call
* update tests
* move from function{} to @{}
* implementation fix
* some more tests
* update docs
* Some cleanup
* fix CI annotation warning
- 'argument': conversion from 'size_t' to 'VariableIndex', possible loss of data
* Address review comments
- Add `passert_always`
- Use better example in docs
* Some cleanup
- Remove unused functions
* Move more specialized expression types from optimizer to AST builder
VarStatement now handles its own value consumption
* Move assign-multisubscript-consume logic to the instruction generator
Renamed:
- GetMember to MemberAccess
- SetMember to MemberAssignment
- SetMemberByOperator to MemberAssignmentByOperator
- AssignVariableConsume to VariableAssignmentStatement
Adds:
- ast/ElementAccess: AST node for x[a] or x[a,b,...] (may be get or set)
- ast/ElementAssignment: AST node for x[a] := b;
- ast/ElementIndexes: AST node to hold all indexes for element lookup
- optimizer/AssignmentOptimizer: optimizes := in different circumstances
Adds:
- ast/AssignVariableConsume: AST node for assignment to a variable, while consuming the result
- optimizer/ValueConsumerOptimizer: Optimizes expressions where the resulting value will be consumed
1. Optimize if statements (parity with OG compiler):
- discard empty 'alternative' blocks
- optimize to the consequent if the predicate is a nonzero integer
- optimize to the alternative, or an empty block if there is none, if the predicate is a zero integer.
2. In comparison mode, emit module functions in the same order as the OG compiler.
This is because the OG compiler can emit declarations for module functions that it doesn't actually call, and in a different order than calls to actual module functions.
Adds:
- optimizer/BinaryOperatorOptimizer: optimizes expressions with a binary operator
- assignments will be handled by optimizer/AssignmentOptimizer
- optimizer/BinaryOperatorWithFloatOptimizer: optimizes binary operator expressions with a FloatValue left-hand side
- optimizer/BinaryOperatorWithIntegerOptimizer: optimizes binary operator expressions with an IntegerValue left-hand side
- optimizer/BinaryOperatorWithStringOptimizer: optimizes binary operator expressions with a StringValue left-hand side
Adds:
- analyzer/Constants: holds constant name -> value for lookup
- ast/ConstDeclaration: AST node for a const declaration
- optimizer/ConstValidator: validates that an optimized expression is valid to use as a constant.
the Optimizer converts Identifiers that refer to a constant to the optimized constant value.
Adds:
- UnaryOperator: AST node for unary operators -, ++, --, and so forth
- UnaryOperatorOptimizer: optimizes a unary operator with its operand
- now integer and float negation, and integer inversion
- later x[y]++ to a single instruction
Also:
- automatically include basic.em, which includes some parameter defaults like -1.
This allows the compiler to generate output for a hello, world script with the exact same .ecl output as the legacy compiler.
Adds:
- ast/Argument: AST node for an argument passed to a function.
- ast/FunctionCall: AST node for a function call.
- codegen/ModuleDeclarationRegistrar: The code generator registers module function declarations with this in order to determine module indexes and function indexes for instructions.
- model/FunctionLink: this is a reference either:
- from: a FunctionCall or a FunctionReference
- to: a ModuleFunctionDeclaration or a UserFunction
- optimizer/ReferencedFunctionGatherer: a visitor that determines which module functions and user functions are referenced, by looking at function calls and function references.
Also:
- CodeGenerator registers module functions
- InstructionGenerator generates code for function calls
- InstructionEmitter generates TOK_FUNC instructions
- StoredTokenDecoder decodes TOK_FUNC instructions
After all of this, the compiler can compile print("hello, world");
