Merge pull request #202 from fktn-k/feature/200_allow_space_in_unquot…

…ed_strings #200 Allow a space in unquoted strings
fktn-k · Nov 8, 2023 · 6ed3e07 · 6ed3e07
2 parents 7df5b59 + d85d6e2
commit 6ed3e07
Show file tree

Hide file tree

Showing 3 changed files with 56 additions and 36 deletions.
diff --git a/docs/mkdocs/docs/tutorials/index.md b/docs/mkdocs/docs/tutorials/index.md
@@ -68,20 +68,20 @@ See [the CMake Integration section]() for the other ways and modify the implemen
 
     ```yaml
     novels:
-      - title: "Robinson Crusoe"
-        author: "Daniel Defoe"
+      - title: Robinson Crusoe
+        author: Daniel Defoe
         year: 1678
-      - title: "Frankenstein"
-        author: "Jane Austen"
+      - title: Frankenstein
+        author: Jane Austen
         year: 1818
-      - title: "Moby-Dick"
-        author: "Herman Melville"
+      - title: Moby-Dick
+        author: Herman Melville
         year: 1851
-      - title: "Brave New World"
-        author: "Aldous Huxley"
+      - title: Brave New World
+        author: Aldous Huxley
         year: 1932
-      - title: "Never Let Me Go"
-        author: "Kazuo Ishiguro"
+      - title: Never Let Me Go
+        author: Kazuo Ishiguro
         year: 2005
     ```
 === "tutorial.cpp"
@@ -133,24 +133,24 @@ If you run the tutorial executable file, you will see the output like:
 ```bash
 novels:
   -
-    title: "Robinson Crusoe"
-    author: "Daniel Defoe"
+    title: Robinson Crusoe
+    author: Daniel Defoe
     year: 1678
   -
-    title: "Frankenstein"
-    author: "Jane Austen"
+    title: Frankenstein
+    author: Jane Austen
     year: 1818
   -
-    title: "Moby-Dick"
-    author: "Herman Melville"
+    title: Moby-Dick
+    author: Herman Melville
     year: 1851
   -
-    title: "Brave New World"
-    author: "Aldous Huxley"
+    title: Brave New World
+    author: Aldous Huxley
     year: 1932
   -
-    title: "Never Let Me Go"
-    author: "Kazuo Ishiguro"
+    title: Never Let Me Go
+    author: Kazuo Ishiguro
     year: 2005
 ```
 
@@ -241,20 +241,20 @@ Rebuild and run the application, and you'll see the output like:
 ```bash
 recommends:
   -
-    title: "Robinson Crusoe"
-    author: "Daniel Defoe"
+    title: Robinson Crusoe
+    author: Daniel Defoe
   -
-    title: "Frankenstein"
-    author: "Jane Austen"
+    title: Frankenstein
+    author: Jane Austen
   -
-    title: "Moby-Dick"
-    author: "Herman Melville"
+    title: Moby-Dick
+    author: Herman Melville
   -
-    title: "Brave New World"
-    author: "Aldous Huxley"
+    title: Brave New World
+    author: Aldous Huxley
   -
-    title: "Never Let Me Go"
-    author: "Kazuo Ishiguro"
+    title: Never Let Me Go
+    author: Kazuo Ishiguro
 ```
 
 ### :pill: Integrate with user-defined types

diff --git a/include/fkYAML/detail/input/lexical_analyzer.hpp b/include/fkYAML/detail/input/lexical_analyzer.hpp
@@ -222,7 +222,7 @@ class lexical_analyzer
         case 'F':
         case 'f': {
             // YAML specifies that only these words represent the boolean value `false`.
-            // See "10.3.2. Tag Resolution" section in https://yaml.org/spec/1.2.2/
+            // See "10.3.2 Tag Resolution" section in https://yaml.org/spec/1.2.2/
             char_int_type ret = m_input_handler.get_range(5, m_value_buffer);
             if (ret == end_of_input)
             {
@@ -247,7 +247,7 @@ class lexical_analyzer
         case 'n': {
             // YAML specifies that these words and a tilde represent a null value.
             // Tildes are already checked above, so no check is needed here.
-            // See "10.3.2. Tag Resolution" section in https://yaml.org/spec/1.2.2/
+            // See "10.3.2 Tag Resolution" section in https://yaml.org/spec/1.2.2/
             char_int_type ret = m_input_handler.get_range(4, m_value_buffer);
             if (ret == end_of_input)
             {
@@ -271,7 +271,7 @@ class lexical_analyzer
         case 'T':
         case 't': {
             // YAML specifies that only these words represent the boolean value `true`.
-            // See "10.3.2. Tag Resolution" section in https://yaml.org/spec/1.2.2/
+            // See "10.3.2 Tag Resolution" section in https://yaml.org/spec/1.2.2/
             char_int_type ret = m_input_handler.get_range(4, m_value_buffer);
             if (ret == end_of_input)
             {
@@ -636,7 +636,7 @@ class lexical_analyzer
         case 'o':
             // Do not store 'o' since std::strtoull does not support "0o" but "0" as the prefix for octal numbers.
             // YAML specifies octal values start with the prefix "0o".
-            // See "10.3.2. Tag Resolution" section in https://yaml.org/spec/1.2.2/
+            // See "10.3.2 Tag Resolution" section in https://yaml.org/spec/1.2.2/
             return scan_octal_number();
         case 'x':
             m_value_buffer.push_back(char_traits_type::to_char_type(next));
@@ -813,7 +813,26 @@ class lexical_analyzer
             {
                 if (!needs_last_double_quote && !needs_last_single_quote)
                 {
-                    return lexical_token_t::STRING_VALUE;
+                    // Allow a space in an unquoted string only if the space is surrounded by non-space characters.
+                    // See "7.3.3 Plain Style" section in https://yaml.org/spec/1.2.2/
+                    current = m_input_handler.get_next();
+                    switch (current)
+                    {
+                    case ' ':
+                    case '\r':
+                    case '\n':
+                    case '{':
+                    case '}':
+                    case '[':
+                    case ']':
+                    case ',':
+                    case ':':
+                    case '#':
+                    case '\\':
+                        return lexical_token_t::STRING_VALUE;
+                    }
+                    m_input_handler.unget();
+                    current = m_input_handler.get_current();
                 }
                 m_value_buffer.push_back(char_traits_type::to_char_type(current));
                 continue;
@@ -932,7 +951,7 @@ class lexical_analyzer
             }
 
             // Handle escaped characters.
-            // See "5.7. Escaped Characters" section in https://yaml.org/spec/1.2.2/
+            // See "5.7 Escaped Characters" section in https://yaml.org/spec/1.2.2/
             if (current == '\\')
             {
                 if (!needs_last_double_quote)

diff --git a/test/unit_test/test_lexical_analyzer_class.cpp b/test/unit_test/test_lexical_analyzer_class.cpp
@@ -398,6 +398,7 @@ TEST_CASE("LexicalAnalyzerClassTest_ScanStringTokenTest", "[LexicalAnalyzerClass
         value_pair_t(std::string(".on"), fkyaml::node::string_type(".on")),
         value_pair_t(std::string("foo]"), fkyaml::node::string_type("foo")),
         value_pair_t(std::string("foo:bar"), fkyaml::node::string_type("foo:bar")),
+        value_pair_t(std::string("foo bar"), fkyaml::node::string_type("foo bar")),
         value_pair_t(std::string("\"foo bar\""), fkyaml::node::string_type("foo bar")),
         value_pair_t(std::string("\"foo:bar\""), fkyaml::node::string_type("foo:bar")),
         value_pair_t(std::string("\"foo,bar\""), fkyaml::node::string_type("foo,bar")),