From 27eb53b48cf027345e7ab6d6dc18f900a14fd418 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Rainer=20Schl=C3=B6nvoigt?= <rainer.schlonvoigt@philips.com>
Date: Thu, 22 Jun 2023 16:24:29 +0200
Subject: [PATCH] Exposing a rawLiteral on DocString for situation where
 sanitization is too eager

---
 .../CucumberSwift/Gherkin/Lexer/Lexer.swift   | 27 ++++++++++++++-----
 .../Gherkin/Parser/DocString.swift            |  1 +
 .../Gherkin/DocstringTests.swift              | 26 ++++++++++++++++++
 3 files changed, 48 insertions(+), 6 deletions(-)

diff --git a/Sources/CucumberSwift/Gherkin/Lexer/Lexer.swift b/Sources/CucumberSwift/Gherkin/Lexer/Lexer.swift
index 3c7a31e6..104bd011 100644
--- a/Sources/CucumberSwift/Gherkin/Lexer/Lexer.swift
+++ b/Sources/CucumberSwift/Gherkin/Lexer/Lexer.swift
@@ -75,13 +75,18 @@ public class Lexer: StringReader {
         }
     }
 
-    @discardableResult internal func readDocString(_ evaluation: ((Character) -> Bool)) -> String {
+    @discardableResult internal func readDocString(
+        _ evaluation: ((Character) -> Bool)
+    ) -> (docString: String, rawDocString: String) {
         var str = ""
+        var rawStr = ""
         while let char = currentChar {
             if char.isEscapeCharacter,
                let next = nextChar,
                next.isDocStringLiteral {
                 str.append(next)
+                rawStr.append(char)
+                rawStr.append(next)
                 advanceIndex()
                 advanceIndex()
                 continue
@@ -90,9 +95,10 @@ public class Lexer: StringReader {
                 break
             }
             str.append(char)
+            rawStr.append(char)
             advanceIndex()
         }
-        return str
+        return (str, rawStr)
     }
 
     @discardableResult internal func stripSpaceIfNecessary() -> Bool {
@@ -159,13 +165,16 @@ public class Lexer: StringReader {
         let open = lookAheadAtLineUntil { !$0.isDocStringLiteral }
         if open.isDocStringLiteral() {
             readLineUntil { !$0.isDocStringLiteral }
-            let docStringValues = readDocString {
+
+            let (docString, rawDocString) = readDocString {
                 if $0.isDocStringLiteral {
                     let close = lookAheadAtLineUntil { !$0.isDocStringLiteral }
                     if close == open { return true }
                 }
                 return false
-            }.components(separatedBy: "\n")
+            }
+
+            let docStringValues = docString.components(separatedBy: "\n")
             .enumerated()
             .reduce(into: (whitespaceCount: 0, trimmedLines: [String]())) { res, e in
                 let (offset, line) = e
@@ -181,8 +190,14 @@ public class Lexer: StringReader {
             .dropLast { $0.trimmingCharacters(in: .whitespacesAndNewlines).isEmpty }
 
             readLineUntil { !$0.isDocStringLiteral }
-            return advance(.docString(position, DocString(literal: docStringValues.dropFirst().joined(separator: "\n"),
-                                                          contentType: docStringValues.first?.trimmingCharacters(in: .whitespacesAndNewlines))))
+            return advance(.docString(
+                position,
+                DocString(
+                    rawLiteral: rawDocString,
+                    literal: docStringValues.dropFirst().joined(separator: "\n"),
+                    contentType: docStringValues.first?.trimmingCharacters(in: .whitespacesAndNewlines)
+                )
+            ))
         } else if char == .quote {
             return advance(.match(position, "\(Character.quote)"))
         }
diff --git a/Sources/CucumberSwift/Gherkin/Parser/DocString.swift b/Sources/CucumberSwift/Gherkin/Parser/DocString.swift
index 8e4e4ab6..75c7a1a3 100644
--- a/Sources/CucumberSwift/Gherkin/Parser/DocString.swift
+++ b/Sources/CucumberSwift/Gherkin/Parser/DocString.swift
@@ -8,6 +8,7 @@
 
 import Foundation
 public struct DocString: Hashable {
+    public let rawLiteral: String
     public var literal: String
     public var contentType: String?
 }
diff --git a/Tests/CucumberSwiftTests/Gherkin/DocstringTests.swift b/Tests/CucumberSwiftTests/Gherkin/DocstringTests.swift
index 10b4c8df..87262558 100644
--- a/Tests/CucumberSwiftTests/Gherkin/DocstringTests.swift
+++ b/Tests/CucumberSwiftTests/Gherkin/DocstringTests.swift
@@ -162,4 +162,30 @@ class DocstringTests: XCTestCase {
         third line
         """)
     }
+
+    func testDocStringPreservesEscapedQuotesWithinJSONString() throws {
+        let cucumber = Cucumber(withString:
+            #"""
+            Feature: DocString variations
+
+            Scenario: minimalistic
+                Given a DocString with JSON with escaped quotes in a string
+                """
+                {
+                    "stringWithEscapedQuote":"String with a \"quote\" or two"
+                }
+                """
+            """#
+        )
+        let firstStep = cucumber.features.first?.scenarios.first?.steps.first
+        let jsonString = try XCTUnwrap(firstStep?.docString?.rawLiteral)
+        let jsonData = try XCTUnwrap(jsonString.data(using: .utf8))
+
+        struct MyJsonType: Decodable {
+            let stringWithEscapedQuote: String
+        }
+
+        let jsonObject = try JSONDecoder().decode(MyJsonType.self, from: jsonData)
+        XCTAssertEqual(jsonObject.stringWithEscapedQuote, "String with a \"quote\" or two")
+    }
 }