Bug 579628 - Merging of consecutive repeated commands creates poorly-structured HTML

Most issues from this report were already implemented. Extended for cross reference lists with the possibility to "style" intermediate items as well.
(also some code structure, i.e. mnemonics instead of numbers).

2 files changed, 82 insertions, 32 deletions

diff --git a/src/htmldocvisitor.cpp b/src/htmldocvisitor.cpp
index 6a9c142..a9439c9 100644
--- a/src/htmldocvisitor.cpp
+++ b/src/htmldocvisitor.cpp
@@ -38,6 +38,31 @@
 
 static const int NUM_HTML_LIST_TYPES = 4;
 static const char types[][NUM_HTML_LIST_TYPES] = {"1", "a", "i", "A"};
+enum contexts_t
+{
+    NONE,      // 0
+    STARTLI,   // 1
+    STARTDD,   // 2
+    ENDLI,     // 3
+    ENDDD,     // 4
+    STARTTD,   // 5
+    ENDTD,     // 6
+    INTERLI,   // 7
+    INTERDD,   // 8
+    INTERTD    // 9
+};
+static const char *contexts[10] = 
+{ "",          // 0
+  "startli",   // 1
+  "startdd",   // 2
+  "endli",     // 3
+  "enddd",     // 4
+  "starttd",   // 5
+  "endtd",     // 6
+  "interli",   // 7
+  "interdd",   // 8
+  "intertd"    // 9
+};
 
 static QCString convertIndexWordToAnchor(const QString &word)
 {
@@ -887,24 +912,24 @@ static int getParagraphContext(DocPara *p,bool &isFirst,bool &isLast)
           }
           isFirst=isFirstChildNode((DocParBlock*)p->parent(),p);
           isLast =isLastChildNode ((DocParBlock*)p->parent(),p);
-          t=0;
+          t=NONE;
           if (isFirst)
           {
             if (kind==DocNode::Kind_HtmlListItem ||
                 kind==DocNode::Kind_SecRefItem)
             {
-              t=1;
+              t=STARTLI;
             }
             else if (kind==DocNode::Kind_HtmlDescData ||
                      kind==DocNode::Kind_XRefItem ||
                      kind==DocNode::Kind_SimpleSect)
             {
-              t=2;
+              t=STARTDD;
             }
             else if (kind==DocNode::Kind_HtmlCell ||
                      kind==DocNode::Kind_ParamList)
             {
-              t=5;
+              t=STARTTD;
             }
           }
           if (isLast)
@@ -912,18 +937,37 @@ static int getParagraphContext(DocPara *p,bool &isFirst,bool &isLast)
             if (kind==DocNode::Kind_HtmlListItem ||
                 kind==DocNode::Kind_SecRefItem)
             {
-              t=3;
+              t=ENDLI;
             }
             else if (kind==DocNode::Kind_HtmlDescData ||
                      kind==DocNode::Kind_XRefItem ||
                      kind==DocNode::Kind_SimpleSect)
             {
-              t=4;
+              t=ENDDD;
             }
             else if (kind==DocNode::Kind_HtmlCell ||
                      kind==DocNode::Kind_ParamList)
             {
-              t=6;
+              t=ENDTD;
+            }
+          }
+          if (!isFirst && !isLast)
+          {
+            if (kind==DocNode::Kind_HtmlListItem ||
+                kind==DocNode::Kind_SecRefItem)
+            {
+              t=INTERLI;
+            }
+            else if (kind==DocNode::Kind_HtmlDescData ||
+                     kind==DocNode::Kind_XRefItem ||
+                     kind==DocNode::Kind_SimpleSect)
+            {
+              t=INTERDD;
+            }
+            else if (kind==DocNode::Kind_HtmlCell ||
+                     kind==DocNode::Kind_ParamList)
+            {
+              t=INTERTD;
             }
           }
           break;
@@ -931,47 +975,51 @@ static int getParagraphContext(DocPara *p,bool &isFirst,bool &isLast)
       case DocNode::Kind_AutoListItem:
         isFirst=isFirstChildNode((DocAutoListItem*)p->parent(),p);
         isLast =isLastChildNode ((DocAutoListItem*)p->parent(),p);
-        t=1; // not used
+        t=STARTLI; // not used
         break;
       case DocNode::Kind_SimpleListItem:
         isFirst=TRUE;
         isLast =TRUE;
-        t=1; // not used
+        t=STARTLI; // not used
         break;
       case DocNode::Kind_ParamList:
         isFirst=TRUE;
         isLast =TRUE;
-        t=1; // not used
+        t=STARTLI; // not used
         break;
       case DocNode::Kind_HtmlListItem:
         isFirst=isFirstChildNode((DocHtmlListItem*)p->parent(),p);
         isLast =isLastChildNode ((DocHtmlListItem*)p->parent(),p);
-        if (isFirst) t=1;
-        if (isLast)  t=3;
+        if (isFirst) t=STARTLI;
+        if (isLast)  t=ENDLI;
+        if (!isFirst && !isLast) t = INTERLI;
         break;
       case DocNode::Kind_SecRefItem:
         isFirst=isFirstChildNode((DocSecRefItem*)p->parent(),p);
         isLast =isLastChildNode ((DocSecRefItem*)p->parent(),p);
-        if (isFirst) t=1;
-        if (isLast)  t=3;
+        if (isFirst) t=STARTLI;
+        if (isLast)  t=ENDLI;
+        if (!isFirst && !isLast) t = INTERLI;
         break;
       case DocNode::Kind_HtmlDescData:
         isFirst=isFirstChildNode((DocHtmlDescData*)p->parent(),p);
         isLast =isLastChildNode ((DocHtmlDescData*)p->parent(),p);
-        if (isFirst) t=2;
-        if (isLast)  t=4;
+        if (isFirst) t=STARTDD;
+        if (isLast)  t=ENDDD;
+        if (!isFirst && !isLast) t = INTERDD;
         break;
       case DocNode::Kind_XRefItem:
         isFirst=isFirstChildNode((DocXRefItem*)p->parent(),p);
         isLast =isLastChildNode ((DocXRefItem*)p->parent(),p);
-        if (isFirst) t=2;
-        if (isLast)  t=4;
+        if (isFirst) t=STARTDD;
+        if (isLast)  t=ENDDD;
+        if (!isFirst && !isLast) t = INTERDD;
         break;
       case DocNode::Kind_SimpleSect:
         isFirst=isFirstChildNode((DocSimpleSect*)p->parent(),p);
         isLast =isLastChildNode ((DocSimpleSect*)p->parent(),p);
-        if (isFirst) t=2;
-        if (isLast)  t=4;
+        if (isFirst) t=STARTDD;
+        if (isLast)  t=ENDDD;
         if (isSeparatedParagraph((DocSimpleSect*)p->parent(),p))
           // if the paragraph is enclosed with separators it will
           // be included in <dd>..</dd> so avoid addition paragraph
@@ -979,12 +1027,14 @@ static int getParagraphContext(DocPara *p,bool &isFirst,bool &isLast)
         {
           isFirst=isLast=TRUE;
         }
+        if (!isFirst && !isLast) t = INTERDD;
         break;
       case DocNode::Kind_HtmlCell:
         isFirst=isFirstChildNode((DocHtmlCell*)p->parent(),p);
         isLast =isLastChildNode ((DocHtmlCell*)p->parent(),p);
-        if (isFirst) t=5;
-        if (isLast)  t=6;
+        if (isFirst) t=STARTTD;
+        if (isLast)  t=ENDTD;
+        if (!isFirst && !isLast) t = INTERTD;
         break;
       default:
         break;
@@ -1052,19 +1102,10 @@ void HtmlDocVisitor::visitPre(DocPara *p)
     }
   }
 
-  // check if this paragraph is the first or last child of a <li> or <dd>.
+  // check if this paragraph is the first or last or intermediate child of a <li> or <dd>.
   // this allows us to mark the tag with a special class so we can
   // fix the otherwise ugly spacing.
   int t;
-  static const char *contexts[7] = 
-  { "",          // 0
-    "startli",   // 1
-    "startdd",   // 2
-    "endli",     // 3
-    "enddd",     // 4
-    "starttd",   // 5
-    "endtd"      // 6
-  };
   bool isFirst;
   bool isLast;
   t = getParagraphContext(p,isFirst,isLast);
diff --git a/templates/html/doxygen.css b/templates/html/doxygen.css
index 8a1235a..500a70e 100644
--- a/templates/html/doxygen.css
+++ b/templates/html/doxygen.css
@@ -80,6 +80,15 @@ p.endtd {
 	margin-bottom: 2px;
 }
 
+p.interli {
+}
+
+p.interdd {
+}
+
+p.intertd {
+}
+
 /* @end */
 
 caption {