shrimp/src/parser/shrimp.grammar

@external propSource highlighting from "./highlight"

@context trackScope from "./scopeTracker"

@skip { space }

@top Program { item* }

@external tokens operatorTokenizer from "./operatorTokenizer" { Star, Slash, Plus, Minus, And, Or, Eq, Neq, Lt, Lte, Gt, Gte }

@tokens {
  @precedence { Number Regex }

  StringFragment { !['\\$]+ }
  NamedArgPrefix { $[a-z]+ "=" }
  Number { ("-" | "+")? $[0-9]+ ('.' $[0-9]+)? }
  Boolean { "true" | "false" }
  newlineOrSemicolon { "\n" | ";" }
  eof { @eof }
  space { " " | "\t" }
  leftParen { "(" }
  rightParen { ")" }
  colon[closedBy="end", @name="colon"] { ":" }
  Underscore { "_" }
  Regex { "//" (![/\\\n[] | "\\" ![\n] | "[" (![\n\\\]] | "\\" ![\n])* "]")+ ("//" $[gimsuy]*)? } // Stolen from the lezer JavaScript grammar
  "|"[@name=operator]

}

@external tokens tokenizer from "./tokenizer" { Identifier, AssignableIdentifier, Word, IdentifierBeforeDot }

@precedence {
  pipe @left,
  multiplicative @left,
  additive @left,
  call
}

item { 
  consumeToTerminator newlineOrSemicolon |
  consumeToTerminator eof |
  newlineOrSemicolon // allow blank lines
}


consumeToTerminator {
  PipeExpr |
  ambiguousFunctionCall |
  DotGet |
  IfExpr |
  FunctionDef |
  Assign |
  BinOp |
  expressionWithoutIdentifier
}

PipeExpr {
  pipeOperand (!pipe "|" pipeOperand)+
}

pipeOperand {
  FunctionCall | FunctionCallOrIdentifier
}

FunctionCallOrIdentifier {
  Identifier
}

ambiguousFunctionCall {
  FunctionCall | FunctionCallOrIdentifier
}

FunctionCall {
  Identifier arg+
}

arg {
  PositionalArg | NamedArg
}


PositionalArg {
  expression | FunctionDef | Underscore
}

NamedArg {
  NamedArgPrefix (expression | FunctionDef | Underscore)
}

FunctionDef {
  singleLineFunctionDef | multilineFunctionDef
}

singleLineFunctionDef {
  @specialize[@name=keyword]<Identifier, "do"> Params colon consumeToTerminator @specialize[@name=keyword]<Identifier, "end">
}

multilineFunctionDef {
  @specialize[@name=keyword]<Identifier, "do"> Params colon newlineOrSemicolon block @specialize[@name=keyword]<Identifier, "end">
}

IfExpr {
  singleLineIf | multilineIf
}

singleLineIf {
  @specialize[@name=keyword]<Identifier, "if"> (ConditionalOp | expression) colon ThenBlock { consumeToTerminator }
}

multilineIf {
  @specialize[@name=keyword]<Identifier, "if"> (ConditionalOp | expression) colon newlineOrSemicolon ThenBlock ElseIfExpr* ElseExpr? @specialize[@name=keyword]<Identifier, "end">
}

ElseIfExpr {
  @specialize[@name=keyword]<Identifier, "elseif"> (ConditionalOp | expression) colon newlineOrSemicolon ThenBlock
}

ElseExpr {
  @specialize[@name=keyword]<Identifier, "else"> colon newlineOrSemicolon ThenBlock
}

ThenBlock {
  block
}

ConditionalOp {
  expression Eq expression |
  expression Neq expression |
  expression Lt expression |
  expression Lte expression |
  expression Gt expression |
  expression Gte expression |
  expression And (expression | ConditionalOp) |
  expression Or (expression | ConditionalOp)
}

Params {
  AssignableIdentifier*
}

Assign {
  AssignableIdentifier Eq consumeToTerminator
}

BinOp {
  (expression | BinOp) !multiplicative Star (expression | BinOp) |
  (expression | BinOp) !multiplicative Slash (expression | BinOp) |
  (expression | BinOp) !additive Plus (expression | BinOp) |
  (expression | BinOp) !additive Minus (expression | BinOp)
}

ParenExpr {
  leftParen (ambiguousFunctionCall | BinOp | expressionWithoutIdentifier | ConditionalOp | PipeExpr | FunctionDef) rightParen
}

expression {
  expressionWithoutIdentifier | DotGet | Identifier
}


@local tokens {
  dot { "." }
}

@skip {} {
  DotGet {
    IdentifierBeforeDot dot Identifier
  }

  String { "'" stringContent* "'" }

}

stringContent {
  StringFragment |
  Interpolation |
  EscapeSeq
}

Interpolation {
  "$" Identifier |
  "$" ParenExpr
}

EscapeSeq {
  "\\" ("$" | "n" | "t" | "r" | "\\" | "'")
}

// We need expressionWithoutIdentifier to avoid conflicts in consumeToTerminator.
// Without this, when parsing "my-var" at statement level, the parser can't decide:
//   - ambiguousFunctionCall → FunctionCallOrIdentifier → Identifier
//   - expression → Identifier  
// Both want the same Identifier token! So we use expressionWithoutIdentifier
// to remove Identifier from the second path, forcing standalone identifiers
// to go through ambiguousFunctionCall (which is what we want semantically).
// Yes, it is annoying and I gave up trying to use GLR to fix it.
expressionWithoutIdentifier {
  ParenExpr | Word | String | Number | Boolean | Regex | @specialize[@name=Null]<Identifier, "null">
}

block {
  (consumeToTerminator newlineOrSemicolon)*
}