An experimental tool to check regular expressions at compile time

.buildkite/check_regex.sh

@@ -0,0 +1,9 @@
+#!/bin/bash -eu
+
+cd Tools
+
+echo '--- :swift: Build'
+swift build --product RegexChecker
+
+echo '--- :face_with_monocle: Check Regular Expressions'
+swift run --skip-build RegexChecker

.buildkite/pipeline.yml

@@ -39,6 +39,15 @@ steps:
     env: *common_env
     plugins: *common_plugins
 
+  #################
+  # Check Regex
+  #################
+  - label: ":face_with_monocle: Check Regular Expressions"
+    key: "check-regex"
+    command: ".buildkite/check_regex.sh"
+    env: *common_env
+    plugins: *common_plugins
+
   #################
   # Publish the Podspec (if we're building a tag)
   #################
@@ -52,4 +61,5 @@ steps:
       - "test"
       - "validate"
       - "lint"
+      - "check-regex"
     if: build.tag != null

.gitignore

@@ -33,3 +33,5 @@ vendor/
 fastlane/test_output
 fastlane/README.md
 fastlane/report.xml
+
+.swiftpm

Tools/Package.swift

@@ -0,0 +1,25 @@
+// swift-tools-version: 5.7
+
+import PackageDescription
+
+let package = Package(
+    name: "Tools",
+    platforms: [.macOS(.v12)],
+    products: [
+        .executable(
+            name: "RegexChecker",
+            targets: ["RegexChecker"]),
+    ],
+    dependencies: [
+        .package(url: "https://github.com/apple/swift-syntax.git", exact: "0.50700.1"),
+    ],
+    targets: [
+        .executableTarget(
+            name: "RegexChecker",
+            dependencies: [
+                .product(name: "SwiftSyntax", package: "swift-syntax"),
+                .product(name: "SwiftSyntaxParser", package: "swift-syntax"),
+            ]
+        )
+    ]
+)

Tools/Sources/RegexChecker/main.swift

@@ -0,0 +1,95 @@
+import Foundation
+import SwiftSyntax
+import SwiftSyntaxParser
+
+class RegexCheckerVisitor: SyntaxVisitor {
+    var declaration: VariableDeclSyntax?
+
+    var regexes = [String]()
+    var shouldFail: Bool = false
+
+    override func visit(_ node: VariableDeclSyntax) -> SyntaxVisitorContinueKind {
+        for piece in (node.leadingTrivia ?? []) {
+            if case let .lineComment(comment) = piece,
+               comment.hasPrefix("// CHECK-REGEX") {
+                self.declaration = node
+                break
+            }
+        }
+
+        return .visitChildren
+    }
+
+    override func visitPost(_ node: VariableDeclSyntax) {
+        self.declaration = nil
+    }
+
+    override func visit(_ literal: StringLiteralExprSyntax) -> SyntaxVisitorContinueKind {
+        guard declaration != nil else {
+            return .visitChildren
+        }
+
+
+        for line in declaration!.description.split(separator: "\n") {
+            print("> \(line.trimmingCharacters(in: .whitespaces))")
+        }
+
+        guard literal.firstToken?.tokenKind == .rawStringDelimiter("#"),
+              literal.lastToken?.tokenKind == .rawStringDelimiter("#") else {
+            print("❌ Regex declaration must use raw string literal (i.e. #\"regex\"#).")
+            shouldFail = true
+            return .skipChildren
+        }
+
+        var regex = literal.description
+        regex.removeFirst()
+        regex.removeLast()
+
+        do {
+            let _ = try NSRegularExpression(pattern: regex)
+            print("✅ Valid regular expression.")
+        } catch {
+            print("❌ Not a regular expression.")
+            shouldFail = true
+        }
+
+        return .skipChildren
+    }
+
+    private func reportFailure(_ message: String) {
+        print("^ \(message)")
+
+        shouldFail = true
+    }
+}
+
+func checkRegexLiterals(directory: URL) throws -> Bool {
+    guard let enumerator = FileManager.default.enumerator(at: directory, includingPropertiesForKeys: nil) else {
+        fatalError("Can't enumerate directory at \(directory)")
+    }
+
+    var hasInvalidRegex = false
+
+    for case let file as URL in enumerator {
+        guard file.lastPathComponent.hasSuffix(".swift") else { continue }
+
+        let visitor = RegexCheckerVisitor()
+        let syntax = try SyntaxParser.parse(file)
+        visitor.walk(syntax)
+
+        if visitor.shouldFail {
+            hasInvalidRegex = true
+        }
+    }
+
+    return !hasInvalidRegex
+}
+
+func main() throws {
+    let thisFile = URL(fileURLWithPath: #filePath)
+    let sourcesDirectory = URL(string: "../../../WordPressShared", relativeTo: thisFile)!
+    let success = try checkRegexLiterals(directory: sourcesDirectory)
+    exit(success ? 0 : 1)
+}
+
+try main()

WordPressShared/Core/Utility/String+StripGutenbergContentForExcerpt.swift

@@ -14,7 +14,8 @@ extension String {
     /// Strips Gutenberg galleries from strings.
     ///
     private func strippingGutenbergGalleries() -> String {
-        let pattern = "(?s)<!--\\swp:gallery?(.*?)wp:gallery\\s-->"
+        // CHECK-REGEX
+        let pattern = #"(?s)<!--\swp:gallery?(.*?)wp:gallery\s-->"#
 
         return removingMatches(pattern: pattern, options: .caseInsensitive)
     }

WordPressShared/Core/Utility/String+StripShortcodes.swift

@@ -5,7 +5,8 @@ extension String {
     /// Creates a new string by stripping all shortcodes from this string.
     ///
     func strippingShortcodes() -> String {
-        let pattern = "\\[[^\\]]+\\]"
+        // CHECK-REGEX
+        let pattern = #"\[[^\]]+\]"#
 
         return removingMatches(pattern: pattern, options: .caseInsensitive)
     }