Did you know that -0xF_ep-0_2
is a valid floating point literal in Swift? Well you probably didn't
(it's equal to -63.5), and as a macro author you shouldn't even have to care! Among many things, Macro
Toolkit shields you from edge cases so that users can use your macros in whatever weird (but correct)
manners they may desire.
You don't need in-depth knowledge of Swift's syntax to make a robust macro, you just need an idea.
If you find Swift Macro Toolkit useful, please consider supporting me by becoming a sponsor. I spend most of my spare time working on open-source projects, and each sponsorship helps me focus more time on making high quality tools and libraries for the community.
See for yourself;
Does -0xF_ep-0_2
look like the type of floating point literal you want to implement parsing for?
Nope; but you don't have to.
return literal.value
Without Macro Toolkit (worth a look)
let string = _syntax.floatingDigits.text
let isHexadecimal: Bool
let stringWithoutPrefix: String
switch string.prefix(2) {
case "0x":
isHexadecimal = true
stringWithoutPrefix = String(string.dropFirst(2))
default:
isHexadecimal = false
stringWithoutPrefix = string
}
let exponentSeparator: Character = isHexadecimal ? "p" : "e"
let parts = stringWithoutPrefix.lowercased().split(separator: exponentSeparator)
guard parts.count <= 2 else {
fatalError("Float literal cannot contain more than one exponent separator")
}
let exponentValue: Int
if parts.count == 2 {
// The exponent part is always decimal
let exponentPart = parts[1]
let exponentPartWithoutUnderscores = exponentPart.replacingOccurrences(of: "_", with: "")
guard
exponentPart.first != "_",
!exponentPart.starts(with: "-_"),
let exponent = Int(exponentPartWithoutUnderscores)
else {
fatalError("Float literal has invalid exponent part: \(string)")
}
exponentValue = exponent
} else {
exponentValue = 0
}
let partsBeforeExponent = parts[0].split(separator: ".")
guard partsBeforeExponent.count <= 2 else {
fatalError("Float literal cannot contain more than one decimal point: \(string)")
}
// The integer part can contain underscores anywhere except for the first character (which must be a digit).
let radix = isHexadecimal ? 16 : 10
let integerPart = partsBeforeExponent[0]
let integerPartWithoutUnderscores = integerPart.replacingOccurrences(of: "_", with: "")
guard
integerPart.first != "_",
let integerPartValue = Int(integerPartWithoutUnderscores, radix: radix).map(Double.init)
else {
fatalError("Float literal has invalid integer part: \(string)")
}
let fractionalPartValue: Double
if partsBeforeExponent.count == 2 {
// The fractional part can contain underscores anywhere except for the first character (which must be a digit).
let fractionalPart = partsBeforeExponent[1]
let fractionalPartWithoutUnderscores = fractionalPart.replacingOccurrences(of: "_", with: "")
guard
fractionalPart.first != "_",
let fractionalPartDigitsValue = Int(fractionalPartWithoutUnderscores, radix: radix)
else {
fatalError("Float literal has invalid fractional part: \(string)")
}
fractionalPartValue = Double(fractionalPartDigitsValue) / pow(Double(radix), Double(fractionalPart.count - 1))
} else {
fractionalPartValue = 0
}
let base: Double = isHexadecimal ? 2 : 10
let multiplier = pow(base, Double(exponentValue))
let sign: Double = _negationSyntax == nil ? 1 : -1
return (integerPartValue + fractionalPartValue) * multiplier * sign
guard case let .simple("Result", (successType, failureType))? = destructure(returnType) else {
throw MacroError("Invalid return type")
}
guard
let simpleReturnType = returnType.as(SimpleTypeIdentifierSyntax.self),
simpleReturnType.name.description == "Result",
let genericArguments = (simpleReturnType.genericArgumentClause?.arguments).map(Array.init),
genericArguments.count == 2
else {
throw MacroError("Invalid return type")
}
let successType = genericArguments[0]
let failureType = genericArguments[1]
Swift has many different ways to express a single type. To name a few such cases; () == Void
,
Int? == Optional<Int>
, and [Int] == Array<Int>
. Swift Macro Toolkit strives to hide these details
from you, so you don't have to handle all the edge cases.
function.returnsVoid
func returnsVoid(_ function: FunctionDeclSyntax) -> Bool {
// Function can either have no return type annotation, `()`, `Void`, or a nested single
// element tuple with a Void-like inner type (e.g. `((((()))))` or `(((((Void)))))`)
func isVoid(_ type: TypeSyntax) -> Bool {
if type.description == "Void" || type.description == "()" {
return true
}
guard let tuple = type.as(TupleTypeSyntax.self) else {
return false
}
if let element = tuple.elements.first, tuple.elements.count == 1 {
let isUnlabeled = element.name == nil && element.secondName == nil
return isUnlabeled && isVoid(TypeSyntax(element.type))
}
return false
}
guard let returnType = function.output?.returnType else {
return false
}
return isVoid(returnType)
}
Getting the value of a string literal (without interpolations) can be incredibly
tedious if you want to do it the right way. You have to evaluate all escape sequences
yourself (unicode ones are particularly annoying e.g. \u{2020}
). And then if a user
wants to use a raw string literal (e.g. #"This isn't a newline \n"#
), things get even
more difficult to get right. Don't fear though, Swift Macro Toolkit has you covered.
return literal.value
Without Macro Toolkit
let segments = _syntax.segments.compactMap { (segment) -> String? in
guard case let .stringSegment(segment) = segment else {
return nil
}
return segment.content.text
}
guard segments.count == _syntax.segments.count else {
return nil
}
let map: [Character: Character] = [
"\\": "\\",
"n": "\n",
"r": "\r",
"t": "\t",
"0": "\0",
"\"": "\"",
"'": "'"
]
let hexadecimalCharacters = "0123456789abcdefABCDEF"
// The length of the `\###...` sequence that starts an escape sequence (zero hashes if not a raw string)
let escapeSequenceDelimiterLength = (_syntax.openDelimiter?.text.count ?? 0) + 1
// Evaluate backslash escape sequences within each segment before joining them together
let transformedSegments = segments.map { segment in
var characters: [Character] = []
var inEscapeSequence = false
var iterator = segment.makeIterator()
var escapeSequenceDelimiterPosition = 0 // Tracks the current position in the delimiter if parsing one
while let c = iterator.next() {
if inEscapeSequence {
if let replacement = map[c] {
characters.append(replacement)
} else if c == "u" {
var count = 0
var digits: [Character] = []
var iteratorCopy = iterator
guard iterator.next() == "{" else {
fatalError("Expected '{' in unicode scalar escape sequence")
}
var foundClosingBrace = false
while let c = iterator.next() {
if c == "}" {
foundClosingBrace = true
break
}
guard hexadecimalCharacters.contains(c) else {
iterator = iteratorCopy
break
}
iteratorCopy = iterator
digits.append(c)
count += 1
}
guard foundClosingBrace else {
fatalError("Expected '}' in unicode scalar escape sequence")
}
if !(1...8).contains(count) {
fatalError("Invalid unicode character escape sequence (must be 1 to 8 digits)")
}
guard
let value = UInt32(digits.map(String.init).joined(separator: ""), radix: 16),
let scalar = Unicode.Scalar(value)
else {
fatalError("Invalid unicode scalar hexadecimal value literal")
}
characters.append(Character(scalar))
}
inEscapeSequence = false
} else if c == "\\" && escapeSequenceDelimiterPosition == 0 {
escapeSequenceDelimiterPosition += 1
} else if !inEscapeSequence && c == "#" && escapeSequenceDelimiterPosition != 0 {
escapeSequenceDelimiterPosition += 1
} else {
if escapeSequenceDelimiterPosition != 0 {
characters.append("\\")
for _ in 0..<(escapeSequenceDelimiterPosition - 1) {
characters.append("#")
}
escapeSequenceDelimiterPosition = 0
}
characters.append(c)
}
if escapeSequenceDelimiterPosition == escapeSequenceDelimiterLength {
inEscapeSequence = true
escapeSequenceDelimiterPosition = 0
}
}
return characters.map(String.init).joined(separator: "")
}
return transformedSegments.joined(separator: "")
let diagnostic = DiagnosticBuilder(for: function._syntax.funcKeyword)
.message("can only add a completion-handler variant to an 'async' function")
.messageID(domain: "AddCompletionHandlerMacro", id: "MissingAsync")
.suggestReplacement(
"add 'async'",
old: function._syntax.signature,
new: newSignature
)
.build()
let diagnostic = Diagnostic(
node: Syntax(funcDecl.funcKeyword),
message: SimpleDiagnosticMessage(
message: "can only add a completion-handler variant to an 'async' function",
diagnosticID: MessageID(domain: "AddCompletionHandlerMacro", id: "MissingAsync"),
severity: .error
),
fixIts: [
FixIt(
message: SimpleDiagnosticMessage(
message: "add 'async'",
diagnosticID: MessageID(domain: "AddCompletionHandlerMacro", id: "MissingAsync"),
severity: .error
),
changes: [
FixIt.Change.replace(
oldNode: Syntax(funcDecl.signature),
newNode: Syntax(newSignature)
)
]
),
]
)