From 326f4aba1dfdb5c62fa2146dc7668b08a924d622 Mon Sep 17 00:00:00 2001
From: Simon Gene Gottlieb <simon@gottliebtfreitag.de>
Date: Thu, 22 Aug 2024 10:29:29 +0200
Subject: [PATCH] fix: parse files with '\r' symbols as line ending correctly

---
 src/stream.cpp                      | 17 ++++++++++++++++-
 src/stream.h                        |  1 +
 test/integration/load_node_test.cpp | 16 ++++++++++++++++
 3 files changed, 33 insertions(+), 1 deletion(-)

diff --git a/src/stream.cpp b/src/stream.cpp
index b1aa092f6..b70940ed7 100644
--- a/src/stream.cpp
+++ b/src/stream.cpp
@@ -262,7 +262,22 @@ char Stream::get() {
   AdvanceCurrent();
   m_mark.column++;
 
-  if (ch == '\n') {
+  // if line ending symbol is unknown, set it to the first
+  // encountered '\n' or '\r' symbol
+  if (!m_lineEndingSymbol) {
+    if (ch == '\n') { // line ending is '\n'
+      m_lineEndingSymbol = '\n';
+    } else if (ch == '\r') {
+      auto ch2 = peek();
+      if (ch2 == '\n') { // line ending is '\r\n'
+        m_lineEndingSymbol = '\n';
+      } else { // line ending is '\r'
+        m_lineEndingSymbol = '\r';
+      }
+    }
+  }
+
+  if (ch == m_lineEndingSymbol) {
     m_mark.column = 0;
     m_mark.line++;
   }
diff --git a/src/stream.h b/src/stream.h
index 2bc7a1521..214104ade 100644
--- a/src/stream.h
+++ b/src/stream.h
@@ -53,6 +53,7 @@ class Stream {
   Mark m_mark;
 
   CharacterSet m_charSet;
+  char m_lineEndingSymbol{}; // 0 means it is not determined yet, must be '\n' or '\r'
   mutable std::deque<char> m_readahead;
   unsigned char* const m_pPrefetched;
   mutable size_t m_nPrefetchedAvailable;
diff --git a/test/integration/load_node_test.cpp b/test/integration/load_node_test.cpp
index 9d0c790fd..1cc84a45a 100644
--- a/test/integration/load_node_test.cpp
+++ b/test/integration/load_node_test.cpp
@@ -360,5 +360,21 @@ TEST(LoadNodeTest, BlockCRNLEncoded) {
   EXPECT_EQ(1, node["followup"].as<int>());
 }
 
+TEST(LoadNodeTest, BlockCREncoded) {
+  Node node = Load(
+      "blockText: |\r"
+      "  some arbitrary text \r"
+      "  spanning some \r"
+      "  lines, that are split \r"
+      "  by CR and NL\r"
+      "followup: 1");
+  EXPECT_EQ(
+      "some arbitrary text \nspanning some \nlines, that are split \nby CR and "
+      "NL\n",
+      node["blockText"].as<std::string>());
+  EXPECT_EQ(1, node["followup"].as<int>());
+}
+
+
 }  // namespace
 }  // namespace YAML