Support for "else if" syntax

Ghidra/Features/Decompiler/certification.manifest

@@ -11,6 +11,7 @@ src/decompile/cpp/Doxyfile||GHIDRA|||Most of this file is autogenerated by doxyg
 src/decompile/cpp/Makefile||GHIDRA||||END|
 src/decompile/datatests/convert.xml||GHIDRA||||END|
 src/decompile/datatests/deadvolatile.xml||GHIDRA||||END|
+src/decompile/datatests/elseif.xml||GHIDRA||||END|
 src/decompile/datatests/floatprint.xml||GHIDRA||||END|
 src/decompile/datatests/forloop1.xml||GHIDRA||||END|
 src/decompile/datatests/forloop_loaditer.xml||GHIDRA||||END|

Ghidra/Features/Decompiler/src/decompile/cpp/prettyprint.cc

@@ -71,6 +71,7 @@ void EmitXml::endBlock(int4 id) {
 
 /// Tell the emitter that a new line is desired at the current indent level
 void EmitXml::tagLine(void) {
+  emitPending();
   *s << "<break " << highlight[(int4)no_color] << " indent=\"0x" << hex <<
     indentlevel << "\"/>";
 }
@@ -79,6 +80,7 @@ void EmitXml::tagLine(void) {
 /// is overridden only for the line, then it returns to its previous value.
 /// \param indent is the desired indent level for the new line
 void EmitXml::tagLine(int4 indent) {
+  emitPending();
   *s << "<break " << highlight[(int4)no_color] << " indent=\"0x" << hex <<
     indent << "\"/>";
 }
@@ -923,6 +925,7 @@ void EmitPrettyPrint::endBlock(int4 id)
 void EmitPrettyPrint::tagLine(void)
 
 {
+  emitPending();
   checkbreak();
   TokenSplit &tok( tokqueue.push() );
   tok.tagLine();
@@ -932,6 +935,7 @@ void EmitPrettyPrint::tagLine(void)
 void EmitPrettyPrint::tagLine(int4 indent)
 
 {
+  emitPending();
   checkbreak();
   TokenSplit &tok( tokqueue.push() );
   tok.tagLine(indent);

Ghidra/Features/Decompiler/src/decompile/cpp/prettyprint.hh

@@ -26,6 +26,7 @@ class PcodeOp;
 class FlowBlock;
 class Funcdata;
 class Symbol;
+class PendPrint;
 
 /// \brief Base class (and interface) for pretty printing and XML markup of tokens
 ///
@@ -80,9 +81,11 @@ protected:
   int4 indentlevel;			///< Current indent level (in fixed width characters)
   int4 parenlevel;			///< Current depth of parentheses
   int4 indentincrement;			///< Change in indentlevel per level of nesting
+  PendPrint *pendPrint;			///< Pending print callback
   void resetDefaultsInternal(void) { indentincrement = 2; }	///< Set options to default values for EmitXml
+  void emitPending(void);		///< Emit any pending print commands
 public:
-  EmitXml(void) { s = (ostream *)0; indentlevel=0; parenlevel=0; resetDefaultsInternal(); }	///< Constructor
+  EmitXml(void) { s = (ostream *)0; indentlevel=0; parenlevel=0; pendPrint=(PendPrint *)0; resetDefaultsInternal(); }	///< Constructor
 
   /// \brief Possible types of syntax highlighting
   enum syntax_highlight {
@@ -136,7 +139,7 @@ public:
   /// Inform the emitter that a printing group is ending.
   /// \param id is the id associated with the group (as returned by openGroup)
   virtual void closeGroup(int4 id) {}
-  virtual void clear(void) { parenlevel = 0; indentlevel=0; }	///< Reset the emitter to its initial state
+  virtual void clear(void) { parenlevel = 0; indentlevel=0; pendPrint=(PendPrint *)0; }	///< Reset the emitter to its initial state
   virtual void setOutputStream(ostream *t) { s = t; }		///< Set the output stream for the emitter
   virtual ostream *getOutputStream(void) const { return s; }	///< Get the current output stream
   virtual void spaces(int4 num,int4 bump=0);
@@ -214,6 +217,24 @@ public:
   ///
   /// \param val is the desired number of characters to indent
   void setIndentIncrement(int4 val) { indentincrement = val; }
+
+  /// \brief Set a pending print callback
+  ///
+  /// The callback will be issued prior to the the next call to tagLine() unless
+  /// a the method cancelPendingPrint() is called first.
+  /// \param pend is the callback to be issued
+  void setPendingPrint(PendPrint *pend) { pendPrint = pend; }
+
+  /// \brief Cancel any pending print callback
+  ///
+  /// If there is any print callback pending, cancel it
+  void cancelPendingPrint(void) { pendPrint = (PendPrint *)0; }
+
+  /// \brief Check if the given print callback is still pending
+  ///
+  /// \param pend is the given print callback to check
+  /// \return \b true if the specific print callback is pending
+  bool hasPendingPrint(PendPrint *pend) const { return (pendPrint == pend); }
 };
 
 /// \brief A trivial emitter that outputs syntax straight to the stream
@@ -779,4 +800,25 @@ public:
   void setXML(bool val);	///< Toggle whether the low-level emitter emits XML markup or not
 };
 
+/// \brief Helper class for sending cancelable print commands to an ExitXml
+///
+/// The PendPrint is issued as a placeholder for commands to the emitter using its
+/// setPendingPrint() method.  The callback() method is overridden to tailor the exact
+/// sequence of print commands.  The print commands will be executed prior to the next
+/// tagLine() call to the emitter, unless the PendPrint is cancelled.
+class PendPrint {
+public:
+  virtual ~PendPrint(void) {}			///< Destructor
+  virtual void callback(EmitXml *emit)=0;	///< Callback that executes the actual print commands
+};
+
+inline void EmitXml::emitPending(void)
+
+{
+  if (pendPrint != (PendPrint *)0) {
+    pendPrint->callback(this);
+    pendPrint = (PendPrint *)0;
+  }
+}
+
 #endif

Ghidra/Features/Decompiler/src/decompile/cpp/printc.cc

@@ -2596,17 +2596,28 @@ void PrintC::emitBlockCondition(const BlockCondition *bl)
   }
 }
 
+void PendingBrace::callback(EmitXml *emit)
+
+{
+  emit->print("{");
+  indentId = emit->startIndent();
+}
+
 void PrintC::emitBlockIf(const BlockIf *bl)
 
 {
   const PcodeOp *op;
+  PendingBrace pendingBrace;
+
+  if (isSet(pending_brace))
+    emit->setPendingPrint(&pendingBrace);
 
 				// if block never prints final branch
 				// so no_branch and only_branch don't matter
 				// and shouldn't be passed automatically to
 				// the subblocks
   pushMod();
-  unsetMod(no_branch|only_branch);
+  unsetMod(no_branch|only_branch|pending_brace);
 
   pushMod();
   setMod(no_branch);
@@ -2614,7 +2625,11 @@ void PrintC::emitBlockIf(const BlockIf *bl)
   condBlock->emit(this);
   popMod();
   emitCommentBlockTree(condBlock);
-  emit->tagLine();
+  if (emit->hasPendingPrint(&pendingBrace))	// If we issued a brace but it did not emit
+    emit->cancelPendingPrint();			// Cancel the brace in order to have "else if" syntax
+  else
+    emit->tagLine();				// Otherwise start the "if" on a new line
+
   op = condBlock->lastOp();
   emit->tagOp("if",EmitXml::keyword_color,op);
   emit->spaces(1);
@@ -2625,10 +2640,8 @@ void PrintC::emitBlockIf(const BlockIf *bl)
   if (bl->getGotoTarget() != (FlowBlock *)0) {
     emit->spaces(1);
     emitGotoStatement(condBlock,bl->getGotoTarget(),bl->getGotoType());
-    popMod();
-    return;
   }
-  
+  else {
     setMod(no_branch);
     emit->spaces(1);
     int4 id = emit->startIndent();
@@ -2639,20 +2652,36 @@ void PrintC::emitBlockIf(const BlockIf *bl)
     emit->stopIndent(id);
     emit->tagLine();
     emit->print("}");
-  if (bl->getSize()==3) {
+    if (bl->getSize() == 3) {
       emit->tagLine();
       emit->print("else",EmitXml::keyword_color);
       emit->spaces(1);
+      FlowBlock *elseBlock = bl->getBlock(2);
+      if (elseBlock->getType() == FlowBlock::t_if) {
+	// Attempt to merge the "else" and "if" syntax
+	setMod(pending_brace);
+	int4 id2 = emit->beginBlock(elseBlock);
+	elseBlock->emit(this);
+	emit->endBlock(id2);
+      }
+      else {
 	int4 id = emit->startIndent();
 	emit->print("{");
-    int4 id2 = emit->beginBlock(bl->getBlock(2));
-    bl->getBlock(2)->emit(this);
+	int4 id2 = emit->beginBlock(elseBlock);
+	elseBlock->emit(this);
 	emit->endBlock(id2);
 	emit->stopIndent(id);
 	emit->tagLine();
 	emit->print("}");
       }
+    }
+  }
   popMod();
+  if (pendingBrace.getIndentId() >= 0) {
+    emit->stopIndent(pendingBrace.getIndentId());
+    emit->tagLine();
+    emit->print("}");
+  }
 }
 
 /// Print the loop using the keyword \e for, followed by a semicolon separated

Ghidra/Features/Decompiler/src/decompile/cpp/printc.hh

@@ -301,4 +301,16 @@ public:
   virtual void opPopcountOp(const PcodeOp *op) { opFunc(op); }
 };
 
+/// \brief Set of print commands for displaying an open brace '{' and setting a new indent level
+///
+/// These are the print commands sent to the emitter prior to printing and \e else block.
+/// The open brace can be canceled if the block decides it wants to use "else if" syntax.
+class PendingBrace : public PendPrint {
+  int4 indentId;		///< Id associated with the new indent level
+public:
+  PendingBrace(void) { indentId = -1; }			///< Constructor
+  int4 getIndentId(void) const { return indentId; }	///< If commands have been issued, returns the new indent level id.
+  virtual void callback(EmitXml *emit);
+};
+
 #endif

Ghidra/Features/Decompiler/src/decompile/cpp/printlanguage.hh

@@ -149,7 +149,8 @@ public:
     falsebranch = 0x800,	///< Print the false branch (for flat)
     nofallthru = 0x1000,       	///< Fall-thru no longer exists
     negatetoken = 0x2000,	///< Print the token representing the negation of current token
-    hide_thisparam = 0x4000	///< Do not print the 'this' parameter in argument lists
+    hide_thisparam = 0x4000,	///< Do not print the 'this' parameter in argument lists
+    pending_brace = 0x8000	///< The current block may need to surround itself with additional braces
   };
   /// \brief Possible types of Atom
   enum tagtype {
@@ -282,7 +283,6 @@ protected:
   int4 getPending(void) const { return pending; }			///< Get the number of pending nodes yet to be put on the RPN stack
   void resetDefaultsInternal(void);					///< Reset options to default for PrintLanguage
 
-
   /// \brief Print a single unicode character as a \e character \e constant for the high-level language
   ///
   /// For most languages, this prints the character surrounded by single quotes.

Ghidra/Features/Decompiler/src/decompile/datatests/elseif.xml

@@ -0,0 +1,84 @@
+<decompilertest>
+<binaryimage arch="x86:LE:64:default:gcc">
+<!--
+    Contrived example testing various forms of nested if/else structures,
+    where the some should collapse to "else if" syntax and others should not.
+-->
+<bytechunk space="ram" offset="0x1006ca" readonly="true">
+                    554889e54883
+ec10897dfc8975f8837dfc0a75188b45
+f889c6488d3d3a030000b800000000e8
+9cfeffffeb17837dfc017511488d3d2a
+030000b800000000e883feffff837df8
+01751b8b45fc89c6488d3d05030000b8
+00000000e867feffffe9bb000000837d
+f80b751e8b45fc83c06489c6488d3de1
+020000b800000000e843feffffe99700
+0000837df815751b8b45fc6bc06489c6
+488d3dbd020000b800000000e81ffeff
+ffeb76837df81f751d8b55fc8b45f801
+d089c6488d3d9a020000b800000000e8
+fcfdffffeb53837df8f9751b8b45fc2b
+45f889c6488d3d79020000b800000000
+e8dbfdffffeb32837df800751b8b45fc
+83e80189c6488d3d58020000b8000000
+00e8bafdffffeb11488d3d53020000b8
+00000000e8a7fdffff8b45fc3b45f875
+1b8b45fc83e80289c6488d3d24020000
+b800000000e886fdffffeb28837dfc25
+7511488d3d1f020000b800000000e86d
+fdffff488d3d18020000b800000000e8
+5cfdffff8b45fc3b45f87d1b8b45fc83
+e80389c6488d3dd9010000b800000000
+e83bfdffffeb1e8b55fc8b45f801d083
+f80a7511488d3ddd010000b800000000
+e81bfdffff817df8c8000000751b8b45
+fc83c06789c6488d3d97010000b80000
+0000e8f9fcffffeb23817df8c9000000
+751a817dfc2b0100007f11488d3d7b01
+0000b800000000e8d4fcffff837df81b
+751b8b45fc83c01b89c6488d3d530100
+00b800000000e8b5fcffffeb5d837df8
+1d7557eb2e488d3d64010000e88ffcff
+ff837dfc64751b8b45f883c06f89c648
+8d3d4e010000b800000000e880fcffff
+eb1890837dfc677511488d3d3c010000
+b800000000e866fcffff488d3d310100
+00e84afcffff8345fc018b45fc3b45f8
+7ca390c9c3
+</bytechunk>
+<bytechunk space="ram" offset="0x100a24" readonly="true">
+  426f64792025640a00456c73650046
+  696e616c004e6f7420717569746500
+  457874726100436f6d6d656e740054
+  6f7000426f6479202564004c616265
+  6c00426f74746f6d00
+</bytechunk>
+</binaryimage>
+<script>
+  <com>map fun r0x1006ca testElseIf</com>
+  <com>map fun r0x100590 printf</com>
+  <com>parse line extern void testElseIf(int4 a,int4 b);</com>
+  <com>parse line extern void printf(char *,...);</com>
+  <com>map label mylabel r0x100913</com>
+  <com>lo fu testElseIf</com>
+  <com>comment instr r0x10085a Comment</com>
+  <com>decompile</com>
+  <com>print C</com>
+  <com>quit</com>
+</script>
+<stringmatch name="Else-if #1" min="1" max="1">else if \(a == 1\)</stringmatch>
+<stringmatch name="Else-if #2" min="1" max="1">else if \(b == 0xb\)</stringmatch>
+<stringmatch name="Else-if #3" min="1" max="1">else if \(b == 0x15\)</stringmatch>
+<stringmatch name="Else-if #4" min="1" max="1">else if \(b == 0x1f\)</stringmatch>
+<stringmatch name="Else-if #5" min="1" max="1">else if \(b == -7\)</stringmatch>
+<stringmatch name="Else-if #6" min="1" max="1">else if \(b == 0\)</stringmatch>
+<stringmatch name="Else-if #7" min="0" max="0">else if \(a == 0x25\)</stringmatch>
+<stringmatch name="Else-if #8" min="1" max="1">if \(a == 0x25\)</stringmatch>
+<stringmatch name="Else-if #9" min="0" max="0">else if \(b \+ a == 10\)</stringmatch>
+<stringmatch name="Else-if #10" min="1" max="1">if \(b \+ a == 10\)</stringmatch>
+<stringmatch name="Else-if #11" min="1" max="1">else if \(b == 0x1d\) goto</stringmatch>
+<stringmatch name="Else-if #12" min="0" max="0">else if \(.* == 0x67\)</stringmatch>
+<stringmatch name="Else-if #13" min="1" max="1">if \(.* == 0x67\)</stringmatch>
+<stringmatch name="Else-if #14" min="1" max="1">else if \(\(b == 0xc9\) &amp;&amp; \(a &lt; 300\)\)</stringmatch>
+</decompilertest>