[fix] re #361: fix parse of scalars with : and starting on the next line

changelog/current.md

@@ -1,6 +1,16 @@
 
 ### Fixes
 
+- Fix [#361](https://github.com/biojppm/rapidyaml/pull/361) - parse error on map scalars containing `:` and starting on the next line:
+  ```yaml
+  ---
+  # failed to parse:
+  description:
+    foo:bar
+  ---
+  # but this was ok:
+  description: foo:bar
+  ```
 - [PR#368](https://github.com/biojppm/rapidyaml/pull/368) - fix pedantic compiler warnings.
 - Fix [#373](https://github.com/biojppm/rapidyaml/issues/373) - false parse error with empty quoted keys in block-style map ([PR#374](https://github.com/biojppm/rapidyaml/pull/374)).
 - Fix [#356](https://github.com/biojppm/rapidyaml/issues/356) - fix overzealous check in `emit_as()`. An id may be larger than the tree's size, eg when nodes were removed.

src/c4/yml/parse.cpp

@@ -2784,20 +2784,34 @@ bool Parser::_scan_scalar_unk(csubstr *C4_RESTRICT scalar, bool *C4_RESTRICT quo
     }
     size_t pos = s.find(" #");
     if(pos != npos)
+    {
+        _c4dbgpf("RUNK: found ' #' at {}", pos);
         s = s.left_of(pos);
+    }
     pos = s.find(": ");
     if(pos != npos)
+    {
+        _c4dbgpf("RUNK: found ': ' at {}", pos);
         s = s.left_of(pos);
+    }
     else if(s.ends_with(':'))
+    {
+        _c4dbgp("RUNK: ends with ':'");
         s = s.left_of(s.len-1);
+    }
     _RYML_WITH_TAB_TOKENS(
     else if((pos = s.find(":\t")) != npos) // TABS
+    {
+        _c4dbgp("RUNK: ends with ':\\t'");
         s = s.left_of(pos);
-    )
+    })
     else
+    {
+        _c4dbgp("RUNK: trimming left of ,");
         s = s.left_of(s.first_of(','));
+    }
     s = s.trim(" \t");
-    _c4dbgpf("RUNK: scalar='{}'", s);
+    _c4dbgpf("RUNK: scalar=[{}]~~~{}~~~", s.len, s);
 
     if(s.empty())
         return false;
@@ -2808,11 +2822,11 @@ bool Parser::_scan_scalar_unk(csubstr *C4_RESTRICT scalar, bool *C4_RESTRICT quo
 
     if(_at_line_end() && s != '~')
     {
-        _c4dbgpf("at line end. curr='{}'", s);
+        _c4dbgpf("at line end. curr=[{}]~~~{}~~", s.len, s);
         s = _extend_scanned_scalar(s);
     }
 
-    _c4dbgpf("scalar was '{}'", s);
+    _c4dbgpf("scalar was [{}]~~~{}~~~", s.len, s);
 
     *scalar = s;
     *quoted = false;
@@ -4046,7 +4060,7 @@ bool Parser::_handle_indentation()
         _RYML_CB_ASSERT(m_stack.m_callbacks, ind > m_state->indref);
         if(has_all(RMAP|RVAL))
         {
-            if(_is_scalar_next__rmap_val(remt) && remt.first_of(":?") == npos)
+            if(_is_scalar_next__rmap_val(remt) && (!remt.first_of_any(": ", "? ")) && (!remt.ends_with(":")))
             {
                 _c4dbgpf("actually it seems a value: '{}'", remt);
             }

test/test_simple_map.cpp

@@ -3,6 +3,66 @@
 namespace c4 {
 namespace yml {
 
+TEST(simple_map, issue361)
+{
+    {
+        const Tree t = parse_in_arena(R"(desc: foo bar)");
+        ASSERT_EQ(t["desc"].val(), "foo bar");
+    }
+    //
+    {
+        const Tree t = parse_in_arena(R"(desc:
+  foo bar)");
+        ASSERT_EQ(t["desc"].val(), "foo bar");
+    }
+    //
+    {
+        const Tree t = parse_in_arena(R"(desc: foo:bar)");
+        ASSERT_EQ(t["desc"].val(), "foo:bar");
+    }
+    //
+    {
+        const Tree t = parse_in_arena(R"(desc:
+  foo:bar)");
+        ASSERT_EQ(t["desc"].val(), "foo:bar");
+    }
+    //
+    {
+        const Tree t = parse_in_arena(R"(desc: Timecode in "HH:MM:SS:FF" format where the last ':' can be replaced by ',' or ';'.)");
+        ASSERT_EQ(t["desc"].val(), "Timecode in \"HH:MM:SS:FF\" format where the last ':' can be replaced by ',' or ';'.");
+    }
+    //
+    {
+        const Tree t = parse_in_arena(R"(desc:
+  Timecode in "HH:MM:SS:FF" format where the last ':' can be replaced by ',' or ';'.)");
+        ASSERT_EQ(t["desc"].val(), "Timecode in \"HH:MM:SS:FF\" format where the last ':' can be replaced by ',' or ';'.");
+    }
+    //
+    {
+        const Tree t = parse_in_arena(R"(value:
+  desc:
+    Timecode in "HH:MM:SS:FF" format where the last ':' can be replaced by ',' or ';'.)");
+        ASSERT_EQ(t["value"]["desc"].val(), "Timecode in \"HH:MM:SS:FF\" format where the last ':' can be replaced by ',' or ';'.");
+    }
+}
+
+TEST(simple_map, issue377)
+{
+    csubstr yaml = R"(
+const game::perception::PerceiveData:
+      PerceiveData: {default: nullptr}
+)";
+    test_check_emit_check(yaml, [](Tree const &t){
+        ASSERT_TRUE(t.rootref().has_child("const game::perception::PerceiveData"));
+        const ConstNodeRef cpd = t["const game::perception::PerceiveData"];
+        ASSERT_TRUE(cpd.is_map());
+        ASSERT_TRUE(cpd.has_child("PerceiveData"));
+        ASSERT_EQ(cpd["PerceiveData"].num_children(), 1u);
+        ASSERT_TRUE(cpd["PerceiveData"].has_child("default"));
+        ASSERT_EQ(cpd["PerceiveData"]["default"].val(), "nullptr");
+    });
+}
+
 TEST(simple_map, issue373)
 {
     {
@@ -833,8 +893,7 @@ h3 ,i3: val3 ,0003
 })",
     L{ // this is crazy...
         N(KEYVAL, "a0", /*"~"*/{}),
-        N("b0", "val0"),
-        N(KEYVAL, "0000 c0", /*"~"*/{}),
+        N("b0", "val0"), N(KEYVAL, "0000 c0", /*"~"*/{}),
         N("d0", "val0"), N(KEYVAL, "0000 e0", /*"~"*/{}),
         N("f0", "val0"), N(KEYVAL, "0000 h0", /*"~"*/{}),
         N("i0", "val0"), N(KEYVAL, "0000 a1", /*"~"*/{}),