Fix deserialization of empty strings

.github/workflows/infra.yml

@@ -23,7 +23,7 @@ jobs:
     runs-on: ubuntu-24.04
     steps:
     - name: log github event
-      run: echo "${{toJSON(github.event)}}"
+      run: echo "${{toJSON(github.event)}}" || echo >/dev/null
   check_workflows:
     if: always()
     continue-on-error: false

.github/workflows/infra.ys

@@ -4,12 +4,12 @@
 :: workflow-setup()
 
 jobs:
-  log-github-event:
-    name: log github event
-    runs-on: ubuntu-24.04
-    steps:
-      - name: log github event
-        run: echo "${{toJSON(github.event)}}"
+  #log-github-event:
+  #  name: log github event
+  #  runs-on: ubuntu-24.04
+  #  steps:
+  #    - name: log github event
+  #      run: echo "${{toJSON(github.event)}}" || echo >/dev/null
 
   check_workflows:
     :: setup-job('infra' 'check_workflows')

changelog/current.md

@@ -1,7 +1,18 @@
 ## Fixes
 
+- Fix [#480](https://github.com/biojppm/rapidyaml/issues/480) ([PR#489](https://github.com/biojppm/rapidyaml/pull/489)):
+  - Deserializing an empty quoted string *will not* cause an error.
+  - Deserializing an empty string *will* cause an error.
+  - Ensure keys are deserialized using all the rules applying to vals.
+  - Added `KEYNIL` and `VALNIL` to `NodeType_e`, used by the parser to mark the key or val as empty. This changed the values of the `NodeType_e` enumeration.
+  - Added `NodeType::key_is_null()` and `NodeType::val_is_null()`.
 - [PR#488](https://github.com/biojppm/rapidyaml/pull/488):
   - add workarounds for problems with codegen of gcc 11,12,13.
   - improve CI coverage of gcc and clang optimization levels.
-- [BREAKING] Fix [#477](https://github.com/biojppm/rapidyaml/issues/477): changed `read<std::map>()` to overwrite existing entries. The provided implementations had an inconsistency between `std::map` (which wasn't overwriting) and `std::vector` (which *was* overwriting).
-- Fix [#475](https://github.com/biojppm/rapidyaml/issues/475): parse error on trailing whitespace in block containers.
+- [BREAKING] Fix [#477](https://github.com/biojppm/rapidyaml/issues/477) ([PR#479](https://github.com/biojppm/rapidyaml/pull/479)): changed `read<std::map>()` to overwrite existing entries. The provided implementations had an inconsistency between `std::map` (which wasn't overwriting) and `std::vector` (which *was* overwriting).
+
+
+## Thanks
+
+- @Delian0
+- @perlpunk

src/c4/yml/emit.def.hpp

@@ -147,18 +147,19 @@ void Emitter<Writer>::_emit_yaml(id_type id)
 template<class Writer>
 void Emitter<Writer>::_write_doc(id_type id)
 {
-    RYML_ASSERT(m_tree->is_doc(id));
-    RYML_ASSERT(!m_tree->has_key(id));
+    const NodeType ty = m_tree->type(id);
+    RYML_ASSERT(ty.is_doc());
+    RYML_ASSERT(!ty.has_key());
     if(!m_tree->is_root(id))
     {
         RYML_ASSERT(m_tree->is_stream(m_tree->parent(id)));
         this->Writer::_do_write("---");
     }
     //
-    if(!m_tree->has_val(id)) // this is more frequent
+    if(!ty.has_val()) // this is more frequent
     {
-        const bool tag = m_tree->has_val_tag(id);
-        const bool anchor = m_tree->has_val_anchor(id);
+        const bool tag = ty.has_val_tag();
+        const bool anchor = ty.has_val_anchor();
         if(!tag && !anchor)
         {
             ;
@@ -199,20 +200,20 @@ void Emitter<Writer>::_write_doc(id_type id)
     }
     else // docval
     {
-        _RYML_CB_ASSERT(m_tree->callbacks(), m_tree->has_val(id));
+        _RYML_CB_ASSERT(m_tree->callbacks(), ty.has_val());
         // some plain scalars such as '...' and '---' must not
         // appear at 0-indentation
         const csubstr val = m_tree->val(id);
         const bool preceded_by_3_dashes = !m_tree->is_root(id);
-        const type_bits style_marks = m_tree->type(id) & (KEY_STYLE|VAL_STYLE);
-        const bool is_plain = m_tree->type(id).is_val_plain();
+        const type_bits style_marks = ty & VAL_STYLE;
+        const bool is_plain = ty.is_val_plain();
         const bool is_ambiguous = (is_plain || !style_marks)
             && ((val.begins_with("...") || val.begins_with("---"))
                 ||
                 (val.find('\n') != npos));
         if(preceded_by_3_dashes)
         {
-            if(val.len == 0 && !m_tree->has_val_anchor(id) && !m_tree->has_val_tag(id))
+            if(is_plain && val.len == 0 && !ty.has_val_anchor() && !ty.has_val_tag())
             {
                 this->Writer::_do_write('\n');
                 return;
@@ -607,7 +608,7 @@ void Emitter<Writer>::_write(NodeScalar const& C4_RESTRICT sc, NodeType flags, i
     }
 
     // ensure the style flags only have one of KEY or VAL
-    _RYML_CB_ASSERT(m_tree->callbacks(), ((flags & SCALAR_STYLE) == 0) || (((flags&KEY_STYLE) == 0) != ((flags&VAL_STYLE) == 0)));
+    _RYML_CB_ASSERT(m_tree->callbacks(), ((flags & SCALAR_STYLE) == 0) || (((flags & KEY_STYLE) == 0) != ((flags & VAL_STYLE) == 0)));
     type_bits style_marks = flags & SCALAR_STYLE;
     if(!style_marks)
         style_marks = scalar_style_choose(sc.scalar);

src/c4/yml/emit.hpp

@@ -472,7 +472,6 @@ inline OStream& operator<< (OStream& s, as_yaml const& y)
  * @param id the node where to start emitting.
  * @param opts emit options.
  * @param buf the output buffer.
- * @param opts emit options.
  * @param error_on_excess Raise an error if the space in the buffer is insufficient.
  * @return a substr trimmed to the result in the output buffer. If the buffer is
  * insufficient (when error_on_excess is false), the string pointer of the
@@ -493,7 +492,6 @@ inline substr emit_yaml(Tree const& t, id_type id, substr buf, bool error_on_exc
  * @param id the node where to start emitting.
  * @param opts emit options.
  * @param buf the output buffer.
- * @param opts emit options.
  * @param error_on_excess Raise an error if the space in the buffer is insufficient.
  * @return a substr trimmed to the result in the output buffer. If the buffer is
  * insufficient (when error_on_excess is false), the string pointer of the
@@ -515,6 +513,7 @@ inline substr emit_json(Tree const& t, id_type id, substr buf, bool error_on_exc
 
 /** (1) emit YAML to the given buffer. Return a substr trimmed to the emitted YAML.
  * @param t the tree; will be emitted from the root node.
+ * @param opts emit options.
  * @param buf the output buffer.
  * @param error_on_excess Raise an error if the space in the buffer is insufficient.
  * @return a substr trimmed to the result in the output buffer. If the buffer is
@@ -533,6 +532,7 @@ inline substr emit_yaml(Tree const& t, substr buf, bool error_on_excess=true)
 }
 /** (1) emit JSON to the given buffer. Return a substr trimmed to the emitted JSON.
  * @param t the tree; will be emitted from the root node.
+ * @param opts emit options.
  * @param buf the output buffer.
  * @param error_on_excess Raise an error if the space in the buffer is insufficient.
  * @return a substr trimmed to the result in the output buffer. If the buffer is

src/c4/yml/event_handler_tree.hpp

@@ -360,6 +360,19 @@ struct EventHandlerTree : public EventHandlerStack<EventHandlerTree, EventHandle
     /** @{ */
 
 
+    C4_ALWAYS_INLINE void set_key_scalar_plain_empty() noexcept
+    {
+        _c4dbgpf("node[{}]: set key scalar plain as empty", m_curr->node_id);
+        m_curr->tr_data->m_key.scalar = {};
+        _enable_(KEY|KEY_PLAIN|KEYNIL);
+    }
+    C4_ALWAYS_INLINE void set_val_scalar_plain_empty() noexcept
+    {
+        _c4dbgpf("node[{}]: set val scalar plain as empty", m_curr->node_id);
+        m_curr->tr_data->m_val.scalar = {};
+        _enable_(VAL|VAL_PLAIN|VALNIL);
+    }
+
     C4_ALWAYS_INLINE void set_key_scalar_plain(csubstr scalar) noexcept
     {
         _c4dbgpf("node[{}]: set key scalar plain: [{}]~~~{}~~~ ({})", m_curr->node_id, scalar.len, scalar, reinterpret_cast<void const*>(scalar.str));