The base class API for emitting a high-level language. More...

#include <printlanguage.hh>

Inheritance diagram for PrintLanguage:

Classes
struct	Atom
	A single non-operator token emitted by the decompiler. More...

struct	NodePending
	A pending data-flow node; waiting to be placed on the reverse polish notation stack. More...

struct	ReversePolish
	An entry on the reverse polish notation (RPN) stack. More...

Public Types
enum	modifiers { force_hex = 1, force_dec = 2, bestfit = 4, force_scinote = 8, force_pointer = 0x10, print_load_value = 0x20, print_store_value = 0x40, no_branch = 0x80, only_branch = 0x100, comma_separate = 0x200, flat = 0x400, falsebranch = 0x800, nofallthru = 0x1000, negatetoken = 0x2000, hide_thisparam = 0x4000 }
	Possible context sensitive modifiers to how tokens get emitted. More...

enum	tagtype { syntax, vartoken, functoken, optoken, typetoken, fieldtoken, blanktoken }
	Possible types of Atom. More...

Public Member Functions
	PrintLanguage (Architecture *g, const string &nm)
	Constructor. More...

virtual	~PrintLanguage (void)
	Destructor.

const string &	getName (void) const
	Get the language name.

CastStrategy *	getCastStrategy (void) const
	Get the casting strategy for the language.

ostream *	getOutputStream (void) const
	Get the output stream being emitted to.

void	setOutputStream (ostream *t)
	Set the output stream to emit to.

void	setScope (Scope *sc)
	Set the current Symbol scope.

void	setMaxLineSize (int4 mls)
	Set the maximum number of characters per line.

void	setIndentIncrement (int4 inc)
	Set the number of characters to indent per level of code nesting.

void	setLineCommentIndent (int4 val)
	Set the number of characters to indent comment lines. More...

void	setCommentDelimeter (const string &start, const string &stop, bool usecommentfill)
	Establish comment delimiters for the language. More...

uint4	getInstructionComment (void) const
	Get the type of comments suitable within the body of a function.

void	setInstructionComment (uint4 val)
	Set the type of comments suitable within the body of a function.

uint4	getHeaderComment (void) const
	Get the type of comments suitable for a function header.

void	setHeaderComment (uint4 val)
	Set the type of comments suitable for a function header.

bool	emitsXml (void) const
	Does the low-level emitter, emit XML markup.

void	setXML (bool val)
	Set whether the low-level emitter, emits XML markup. More...

void	setFlat (bool val)
	Set whether nesting code structure should be emitted. More...

virtual void	adjustTypeOperators (void)=0
	Set basic data-type information for p-code operators.

virtual void	clear (void)
	Clear the RPN stack and the low-level emitter.

virtual void	setIntegerFormat (const string &nm)
	Set the default integer format. More...

virtual void	setCommentStyle (const string &nm)=0
	Set the way comments are displayed in decompiler output. More...

virtual bool	isCharacterConstant (const uint1 *buf, int4 size, int4 charsize) const =0
	Decide is the given byte array looks like a character string. More...

virtual void	docTypeDefinitions (const TypeFactory *typegrp)=0
	Emit definitions of data-types. More...

virtual void	docAllGlobals (void)=0
	Emit declarations of global variables.

virtual void	docSingleGlobal (const Symbol *sym)=0
	Emit the declaration for a single (global) Symbol. More...

virtual void	docFunction (const Funcdata *fd)=0
	Emit the declaration (and body) of a function. More...

virtual void	emitBlockBasic (const BlockBasic *bb)=0
	Emit statements in a basic block.

virtual void	emitBlockGraph (const BlockGraph *bl)=0
	Emit (an unspecified) list of blocks.

virtual void	emitBlockCopy (const BlockCopy *bl)=0
	Emit a basic block (with any labels)

virtual void	emitBlockGoto (const BlockGoto *bl)=0
	Emit a block ending with a goto statement.

virtual void	emitBlockLs (const BlockList *bl)=0
	Emit a sequence of blocks.

virtual void	emitBlockCondition (const BlockCondition *bl)=0
	Emit a conditional statement.

virtual void	emitBlockIf (const BlockIf *bl)=0
	Emit an if/else style construct.

virtual void	emitBlockWhileDo (const BlockWhileDo *bl)=0
	Emit a loop structure, check at top.

virtual void	emitBlockDoWhile (const BlockDoWhile *bl)=0
	Emit a loop structure, check at bottom.

virtual void	emitBlockInfLoop (const BlockInfLoop *bl)=0
	Emit an infinite loop structure.

virtual void	emitBlockSwitch (const BlockSwitch *bl)=0
	Emit a switch structure.

virtual void	opCopy (const PcodeOp *op)=0
	Emit a COPY operator.

virtual void	opLoad (const PcodeOp *op)=0
	Emit a LOAD operator.

virtual void	opStore (const PcodeOp *op)=0
	Emit a STORE operator.

virtual void	opBranch (const PcodeOp *op)=0
	Emit a BRANCH operator.

virtual void	opCbranch (const PcodeOp *op)=0
	Emit a CBRANCH operator.

virtual void	opBranchind (const PcodeOp *op)=0
	Emit a BRANCHIND operator.

virtual void	opCall (const PcodeOp *op)=0
	Emit a CALL operator.

virtual void	opCallind (const PcodeOp *op)=0
	Emit a CALLIND operator.

virtual void	opCallother (const PcodeOp *op)=0
	Emit a CALLOTHER operator.

virtual void	opConstructor (const PcodeOp *op, bool withNew)=0
	Emit an operator constructing an object.

virtual void	opReturn (const PcodeOp *op)=0
	Emit a RETURN operator.

virtual void	opIntEqual (const PcodeOp *op)=0
	Emit a INT_EQUAL operator.

virtual void	opIntNotEqual (const PcodeOp *op)=0
	Emit a INT_NOTEQUAL operator.

virtual void	opIntSless (const PcodeOp *op)=0
	Emit a INT_SLESS operator.

virtual void	opIntSlessEqual (const PcodeOp *op)=0
	Emit a INT_SLESSEQUAL operator.

virtual void	opIntLess (const PcodeOp *op)=0
	Emit a INT_LESS operator.

virtual void	opIntLessEqual (const PcodeOp *op)=0
	Emit a INT_LESSEQUAL operator.

virtual void	opIntZext (const PcodeOp *op)=0
	Emit a INT_ZEXT operator.

virtual void	opIntSext (const PcodeOp *op)=0
	Emit a INT_SEXT operator.

virtual void	opIntAdd (const PcodeOp *op)=0
	Emit a INT_ADD operator.

virtual void	opIntSub (const PcodeOp *op)=0
	Emit a INT_SUB operator.

virtual void	opIntCarry (const PcodeOp *op)=0
	Emit a INT_CARRY operator.

virtual void	opIntScarry (const PcodeOp *op)=0
	Emit a INT_SCARRY operator.

virtual void	opIntSborrow (const PcodeOp *op)=0
	Emit a INT_SBORROW operator.

virtual void	opInt2Comp (const PcodeOp *op)=0
	Emit a INT_2COMP operator.

virtual void	opIntNegate (const PcodeOp *op)=0
	Emit a INT_NEGATE operator.

virtual void	opIntXor (const PcodeOp *op)=0
	Emit a INT_XOR operator.

virtual void	opIntAnd (const PcodeOp *op)=0
	Emit a INT_AND operator.

virtual void	opIntOr (const PcodeOp *op)=0
	Emit a INT_OR operator.

virtual void	opIntLeft (const PcodeOp *op)=0
	Emit a INT_LEFT operator.

virtual void	opIntRight (const PcodeOp *op)=0
	Emit a INT_RIGHT operator.

virtual void	opIntSright (const PcodeOp *op)=0
	Emit a INT_SRIGHT operator.

virtual void	opIntMult (const PcodeOp *op)=0
	Emit a INT_MULT operator.

virtual void	opIntDiv (const PcodeOp *op)=0
	Emit a INT_DIV operator.

virtual void	opIntSdiv (const PcodeOp *op)=0
	Emit a INT_SDIV operator.

virtual void	opIntRem (const PcodeOp *op)=0
	Emit a INT_REM operator.

virtual void	opIntSrem (const PcodeOp *op)=0
	Emit a INT_SREM operator.

virtual void	opBoolNegate (const PcodeOp *op)=0
	Emit a BOOL_NEGATE operator.

virtual void	opBoolXor (const PcodeOp *op)=0
	Emit a BOOL_XOR operator.

virtual void	opBoolAnd (const PcodeOp *op)=0
	Emit a BOOL_AND operator.

virtual void	opBoolOr (const PcodeOp *op)=0
	Emit a BOOL_OR operator.

virtual void	opFloatEqual (const PcodeOp *op)=0
	Emit a FLOAT_EQUAL operator.

virtual void	opFloatNotEqual (const PcodeOp *op)=0
	Emit a FLOAT_NOTEQUAL operator.

virtual void	opFloatLess (const PcodeOp *op)=0
	Emit a FLOAT_LESS operator.

virtual void	opFloatLessEqual (const PcodeOp *op)=0
	Emit a FLOAT_LESSEQUAL operator.

virtual void	opFloatNan (const PcodeOp *op)=0
	Emit a FLOAT_NAN operator.

virtual void	opFloatAdd (const PcodeOp *op)=0
	Emit a FLOAT_ADD operator.

virtual void	opFloatDiv (const PcodeOp *op)=0
	Emit a FLOAT_DIV operator.

virtual void	opFloatMult (const PcodeOp *op)=0
	Emit a FLOAT_MULT operator.

virtual void	opFloatSub (const PcodeOp *op)=0
	Emit a FLOAT_SUB operator.

virtual void	opFloatNeg (const PcodeOp *op)=0
	Emit a FLOAT_NEG operator.

virtual void	opFloatAbs (const PcodeOp *op)=0
	Emit a FLOAT_ABS operator.

virtual void	opFloatSqrt (const PcodeOp *op)=0
	Emit a FLOAT_SQRT operator.

virtual void	opFloatInt2Float (const PcodeOp *op)=0
	Emit a FLOAT_INT2FLOAT operator.

virtual void	opFloatFloat2Float (const PcodeOp *op)=0
	Emit a FLOAT_FLOAT2FLOAT operator.

virtual void	opFloatTrunc (const PcodeOp *op)=0
	Emit a FLOAT_TRUNC operator.

virtual void	opFloatCeil (const PcodeOp *op)=0
	Emit a FLOAT_CEIL operator.

virtual void	opFloatFloor (const PcodeOp *op)=0
	Emit a FLOAT_FLOOR operator.

virtual void	opFloatRound (const PcodeOp *op)=0
	Emit a FLOAT_ROUND operator.

virtual void	opMultiequal (const PcodeOp *op)=0
	Emit a MULTIEQUAL operator.

virtual void	opIndirect (const PcodeOp *op)=0
	Emit a INDIRECT operator.

virtual void	opPiece (const PcodeOp *op)=0
	Emit a PIECE operator.

virtual void	opSubpiece (const PcodeOp *op)=0
	Emit a SUBPIECE operator.

virtual void	opCast (const PcodeOp *op)=0
	Emit a CAST operator.

virtual void	opPtradd (const PcodeOp *op)=0
	Emit a PTRADD operator.

virtual void	opPtrsub (const PcodeOp *op)=0
	Emit a PTRSUB operator.

virtual void	opSegmentOp (const PcodeOp *op)=0
	Emit a SEGMENTOP operator.

virtual void	opCpoolRefOp (const PcodeOp *op)=0
	Emit a CPOOLREF operator.

virtual void	opNewOp (const PcodeOp *op)=0
	Emit a NEW operator.

Static Public Member Functions
static int4	mostNaturalBase (uintb val)
	Determine the most natural base for an integer. More...

static void	formatBinary (ostream &s, uintb val)
	Print a number in binary form. More...

Protected Member Functions
bool	isSet (uint4 m) const
	Is the given printing modification active.

void	pushScope (Scope *sc)
	Push a new symbol scope.

void	popScope (void)
	Pop to the previous symbol scope.

void	pushMod (void)
	Push current printing modifications to the stack.

void	popMod (void)
	Pop to the previous printing modifications.

void	setMod (uint4 m)
	Activate the given printing modification.

void	unsetMod (uint4 m)
	Deactivate the given printing modification.

void	pushOp (const OpToken tok, const PcodeOp op)
	Push an operator token onto the RPN stack. More...

void	pushAtom (const Atom &atom)
	Push a variable token onto the RPN stack. More...

void	pushVnImplied (const Varnode vn, const PcodeOp op, uint4 m)
	Push an implied variable onto the RPN stack. More...

void	pushVnExplicit (const Varnode vn, const PcodeOp op)
	Push an explicit variable onto the RPN stack. More...

void	pushVnLHS (const Varnode vn, const PcodeOp op)
	Push a variable as the left-hand side of an expression. More...

void	emitOp (const ReversePolish &entry)
	Send an operator token from the RPN to the emitter. More...

void	emitAtom (const Atom &atom)
	Send an variable token from the RPN to the emitter. More...

bool	escapeCharacterData (ostream &s, const uint1 *buf, int4 count, int4 charsize, bool bigend) const
	Emit a byte buffer to the stream as unicode characters. More...

void	recurse (void)
	Emit from the RPN stack as much as possible. More...

void	opBinary (const OpToken tok, const PcodeOp op)
	Push a binary operator onto the RPN stack. More...

void	opUnary (const OpToken tok, const PcodeOp op)
	Push a unary operator onto the RPN stack. More...

int4	getPending (void) const
	Get the number of pending nodes yet to be put on the RPN stack.

virtual void	printUnicode (ostream &s, int4 onechar) const =0
	Print a single unicode character as a character constant for the high-level language. More...

virtual void	pushType (const Datatype *ct)=0
	Push a data-type name onto the RPN expression stack. More...

virtual void	pushConstant (uintb val, const Datatype ct, const Varnode vn, const PcodeOp *op)=0
	Push a constant onto the RPN stack. More...

virtual bool	pushEquate (uintb val, int4 sz, const EquateSymbol sym, const Varnode vn, const PcodeOp *op)=0
	Push a constant marked up by and EquateSymbol onto the RPN stack. More...

virtual void	pushAnnotation (const Varnode vn, const PcodeOp op)=0
	Push an address which is not in the normal data-flow. More...

virtual void	pushSymbol (const Symbol sym, const Varnode vn, const PcodeOp *op)=0
	Push a specific Symbol onto the RPN stack. More...

virtual void	pushUnnamedLocation (const Address &addr, const Varnode vn, const PcodeOp op)=0
	Push an address as a substitute for a Symbol onto the RPN stack. More...

virtual void	pushPartialSymbol (const Symbol sym, int4 off, int4 sz, const Varnode vn, const PcodeOp op, Datatype outtype)=0
	Push a variable that represents only part of a symbol onto the RPN stack. More...

virtual void	pushMismatchSymbol (const Symbol sym, int4 off, int4 sz, const Varnode vn, const PcodeOp *op)=0
	Push an identifier for a variable that mismatches with its Symbol. More...

virtual void	emitLineComment (int4 indent, const Comment *comm)
	Emit a comment line. More...

virtual void	emitVarDecl (const Symbol *sym)=0
	Emit a variable declaration. More...

virtual void	emitVarDeclStatement (const Symbol *sym)=0
	Emit a variable declaration statement. More...

virtual bool	emitScopeVarDecls (const Scope *scope, int4 cat)=0
	Emit all the variable declarations for a given scope. More...

virtual void	emitExpression (const PcodeOp *op)=0
	Emit a full expression. More...

virtual void	emitFunctionDeclaration (const Funcdata *fd)=0
	Emit a function declaration. More...

virtual bool	checkPrintNegation (const Varnode *vn)=0
	Check whether a given boolean Varnode can be printed in negated form. More...

Static Protected Member Functions
static bool	unicodeNeedsEscape (int4 codepoint)
	Determine if the given codepoint needs to be escaped. More...

static void	writeUtf8 (ostream &s, int4 codepoint)
	Write unicode character to stream in UTF8 encoding. More...

static int4	readUtf16 (const uint1 *buf, bool bigend)
	Read a 2-byte UTF16 element from a byte array. More...

static int4	getCodepoint (const uint1 *buf, int4 charsize, bool bigend, int4 &skip)
	Extract the next unicode codepoint from an array of character data. More...

Protected Attributes
Architecture *	glb
	The Architecture owning the language emitter.

CastStrategy *	castStrategy
	The strategy for emitting explicit case operations.

EmitXml *	emit
	The low-level token emitter.

uint4	mods
	Currently active printing modifications.

uint4	instr_comment_type
	Type of instruction comments to display.

uint4	head_comment_type
	Type of header comments to display.

Private Attributes
string	name
	The name of the high-level language.

vector< uint4 >	modstack
	Printing modification stack.

vector< Scope * >	scopestack
	The symbol scope stack.

Scope *	curscope
	The current symbol scope.

vector< ReversePolish >	revpol
	The Reverse Polish Notation (RPN) token stack.

vector< NodePending >	nodepend
	Data-flow nodes waiting to be pushed onto the RPN stack.

int4	pending
	Number of data-flow nodes waiting to be pushed.

int4	line_commentindent
	Number of characters a comment line should be indented.

string	commentstart
	Delimiter characters for the start of a comment.

string	commentend
	Delimiter characters (if any) for the end of a comment.

Detailed Description

The base class API for emitting a high-level language.

Instances of this object are responsible for converting a function's (transformed) data-flow graph into the final stream of tokens of a high-level source code language. There a few main entry points including:

The system is responsible for printing:

Control-flow structures
Expressions
Type declarations
Function prototypes
Comments

As part of all this printing, the system is also responsible for

Emitting integers, floats, and character constants
Placing parentheses within expressions to properly represent data-flow
Deciding whether cast operations need an explicit cast token
Indenting and line wrapping

To accomplish this, the API is broken up into three sections. The first section are the main entry point 'doc' methods. The second section are 'emit' methods, which are responsible for printing a representation of a particular high-level code construct. The third section are 'push' and 'op' methods, which are responsible for walking expression trees. The order in which tokens are emitted for an expression is determined by a Reverse Polish Notation (RPN) stack, that the 'push' methods manipulate. Operators and variables are pushed onto this stack and are ultimately emitted in the correct order.

The base class provides a generic printing modifications stack and a symbol scope stack to provide a printing context mechanism for derived classes.

Member Enumeration Documentation

◆ modifiers

enum PrintLanguage::modifiers

Possible context sensitive modifiers to how tokens get emitted.

Enumerator
force_hex	Force printing of hex.
force_dec	Force printing of dec.
bestfit	Decide on most aesthetic form.
force_scinote	Force scientific notation for floats.
force_pointer	Force '*' notation for pointers.
print_load_value	Hide pointer deref for load with other ops.
print_store_value	Hide pointer deref for store with other ops.
no_branch	Do not print branch instruction.
only_branch	Print only the branch instruction.
comma_separate	Statements within condition.
flat	Do not print block structure.
falsebranch	Print the false branch (for flat)
nofallthru	Fall-thru no longer exists.
negatetoken	Print the token representing the negation of current token.
hide_thisparam	Do not print the 'this' parameter in argument lists.

◆ tagtype

enum PrintLanguage::tagtype

Possible types of Atom.

Enumerator
syntax	Emit atom as syntax.
vartoken	Emit atom as variable.
functoken	Emit atom as function name.
optoken	Emit atom as operator.
typetoken	Emit atom as operator.
fieldtoken	Emit atom as structure field.
blanktoken	For anonymous types.

Constructor & Destructor Documentation

◆ PrintLanguage()

PrintLanguage::PrintLanguage	(	Architecture *	g,
		const string &	nm
	)

Constructor.

Parameters

g	is the Architecture that owns and will use this PrintLanguage
nm	is the formal name of the language

References castStrategy, curscope, emit, glb, head_comment_type, Comment::header, instr_comment_type, line_commentindent, mods, name, pending, Comment::user2, Comment::warning, and Comment::warningheader.

Member Function Documentation

◆ checkPrintNegation()

virtual bool PrintLanguage::checkPrintNegation ( const Varnode * vn )

protectedpure virtual

Check whether a given boolean Varnode can be printed in negated form.

In many situations a boolean value can be inverted by flipping the operator token producing it to a complementary token.

Parameters

vn	is the given boolean Varnode

Returns: true if the value can be easily inverted

Implemented in PrintC.

◆ docFunction()

virtual void PrintLanguage::docFunction ( const Funcdata * fd )

pure virtual

Emit the declaration (and body) of a function.

Parameters

fd	is the function to emit

Implemented in PrintC.

Referenced by DecompileAt::rawAction().

◆ docSingleGlobal()

virtual void PrintLanguage::docSingleGlobal ( const Symbol * sym )

pure virtual

Emit the declaration for a single (global) Symbol.

Parameters

sym	is the Symbol to declare

Implemented in PrintC.

◆ docTypeDefinitions()

virtual void PrintLanguage::docTypeDefinitions ( const TypeFactory * typegrp )

pure virtual

Emit definitions of data-types.

Parameters

typegrp is the container for the data-types that should be defined

Implemented in PrintC.

◆ emitAtom()

void PrintLanguage::emitAtom ( const Atom & atom )

protected

Send an variable token from the RPN to the emitter.

Send the given Atom to the low-level emitter, marking it up according to its type

Parameters

atom	is the given Atom to emit

References blanktoken, PrintLanguage::Atom::ct, emit, PrintLanguage::Atom::fd, fieldtoken, functoken, PrintLanguage::Atom::highlight, PrintLanguage::Atom::name, PrintLanguage::Atom::offset, PrintLanguage::Atom::op, optoken, EmitXml::print(), PrintLanguage::Atom::ptr_second, syntax, EmitXml::tagField(), EmitXml::tagFuncName(), EmitXml::tagOp(), EmitXml::tagType(), EmitXml::tagVariable(), PrintLanguage::Atom::type, typetoken, vartoken, and PrintLanguage::Atom::vn.

Referenced by pushAtom().

◆ emitExpression()

virtual void PrintLanguage::emitExpression ( const PcodeOp * op )

protectedpure virtual

Emit a full expression.

This can be an assignment statement, if the given PcodeOp has an output Varnode, or it can be a statement with no left-hand side.

Parameters

op	is the given PcodeOp performing the final operation of the expression

Implemented in PrintC.

◆ emitFunctionDeclaration()

virtual void PrintLanguage::emitFunctionDeclaration ( const Funcdata * fd )

protectedpure virtual

Emit a function declaration.

This prints the formal defining prototype for a function.

Parameters

fd	is the Funcdata object representing the function to be emitted

Implemented in PrintC.

◆ emitLineComment()

void PrintLanguage::emitLineComment	(	int4	indent,
		const Comment *	comm
	)

protectedvirtual

Emit a comment line.

The comment will get emitted as a single line using the high-level language's delimiters with the given indent level

Parameters

indent	is the number of characters to indent
comm	is the Comment object containing the character data and associated markup info

References EmitXml::comment_color, commentend, commentstart, emit, Comment::getAddr(), Address::getOffset(), Address::getSpace(), Comment::getText(), line_commentindent, EmitXml::spaces(), EmitXml::startComment(), EmitXml::stopComment(), EmitXml::tagComment(), and EmitXml::tagLine().

Referenced by PrintC::emitCommentFuncHeader(), and PrintC::emitCommentGroup().

◆ emitOp()

void PrintLanguage::emitOp ( const ReversePolish & entry )

protected

Send an operator token from the RPN to the emitter.

An OpToken directly from the RPN is sent to the low-level emitter, resolving any final spacing or parentheses.

Parameters

entry is the RPN entry to be emitted

References OpToken::binary, OpToken::bump, EmitXml::closeParen(), emit, PrintLanguage::ReversePolish::id2, EmitXml::no_color, PrintLanguage::ReversePolish::op, EmitXml::openParen(), OpToken::postsurround, OpToken::presurround, OpToken::print, OpToken::space, EmitXml::spaces(), OpToken::spacing, EmitXml::tagOp(), PrintLanguage::ReversePolish::tok, OpToken::type, OpToken::unary_prefix, and PrintLanguage::ReversePolish::visited.

Referenced by PrintC::emitBlockCondition(), pushAtom(), and pushOp().

◆ emitScopeVarDecls()

virtual bool PrintLanguage::emitScopeVarDecls	(	const Scope *	scope,
		int4	cat
	)

protectedpure virtual

Emit all the variable declarations for a given scope.

A subset of all variables can be declared by specifying a category, 0 for parameters, -1 for everything.

Parameters

scope	is the given Scope
cat	is the category of variable to declare

Implemented in PrintC.

◆ emitVarDecl()

virtual void PrintLanguage::emitVarDecl ( const Symbol * sym )

protectedpure virtual

Emit a variable declaration.

This can be part of a full a statement, or just the declaration of a function parameter

Parameters

sym	is the Symbol to be declared

Implemented in PrintC.

◆ emitVarDeclStatement()

virtual void PrintLanguage::emitVarDeclStatement ( const Symbol * sym )

protectedpure virtual

Emit a variable declaration statement.

Parameters

sym	is the Symbol to be declared

Implemented in PrintC.

◆ escapeCharacterData()

bool PrintLanguage::escapeCharacterData	(	ostream &	s,
		const uint1 *	buf,
		int4	count,
		int4	charsize,
		bool	bigend
	)		const

protected

Emit a byte buffer to the stream as unicode characters.

Characters are emitted until we reach a terminator character or count bytes is consumed.

Parameters

s	is the output stream
buf	is the byte buffer
count	is the maximum number of bytes to consume
charsize	is 1 for UTF8, 2 for UTF16, or 4 for UTF32
bigend	is true for a big endian encoding of UTF elements

Returns: true if we reach a terminator character

References getCodepoint(), and printUnicode().

Referenced by PrintJava::opCpoolRefOp(), PrintC::opCpoolRefOp(), and PrintC::printCharacterConstant().

◆ formatBinary()

void PrintLanguage::formatBinary	(	ostream &	s,
		uintb	val
	)

static

Print a number in binary form.

Print a string a '0' and '1' characters representing the given value

Parameters

s	is the output stream
val	is the given value

References mostsigbit_set().

Referenced by PrintC::push_integer().

◆ getCodepoint()

int4 PrintLanguage::getCodepoint	(	const uint1 *	buf,
		int4	charsize,
		bool	bigend,
		int4 &	skip
	)

staticprotected

Extract the next unicode codepoint from an array of character data.

One or more bytes is consumed from the array, and the number of bytes used is passed back.

Parameters

buf	is a pointer to the bytes in the character array
charsize	is 1 for UTF8, 2 for UTF16, or 4 for UTF32
bigend	is true for big endian encoding of the UTF element
skip	is a reference for passing back the number of bytes consumed

Returns: the codepoint or -1 if the encoding is invalid

References readUtf16().

Referenced by escapeCharacterData(), and PrintC::isCharacterConstant().

◆ isCharacterConstant()

virtual bool PrintLanguage::isCharacterConstant	(	const uint1 *	buf,
		int4	size,
		int4	charsize
	)		const

pure virtual

Decide is the given byte array looks like a character string.

This looks for encodings and/or a terminator that is appropriate for the high-level language

Parameters

buf	is a pointer to the byte array
size	is the number of bytes in the array
charsize	is the size in bytes of the encoding element (i.e. UTF8, UTF16, etc.) to assume

Implemented in PrintC.

Referenced by RulePtrsubCharConstant::applyOp().

◆ mostNaturalBase()

int4 PrintLanguage::mostNaturalBase ( uintb val )

static

Determine the most natural base for an integer.

Count '0' and '9' digits base 10. Count '0' and 'f' digits base 16. The highest count is the preferred base.

Parameters

val	is the given integer

Returns: 10 for decimal or 16 for hexidecimal

Referenced by PrintC::push_integer().

◆ opBinary()

void PrintLanguage::opBinary	(	const OpToken *	tok,
		const PcodeOp *	op
	)

protected

Push a binary operator onto the RPN stack.

Push an operator onto the stack that has a normal binary format. Both of its input expressions are also pushed.

Parameters

tok	is the operator token to push
op	is the associated PcodeOp

References PcodeOp::getIn(), isSet(), mods, OpToken::negate, negatetoken, pushOp(), pushVnImplied(), and unsetMod().

Referenced by PrintC::opBoolAnd(), PrintC::opBoolOr(), PrintC::opBoolXor(), PrintC::opFloatAdd(), PrintC::opFloatDiv(), PrintC::opFloatEqual(), PrintC::opFloatLess(), PrintC::opFloatLessEqual(), PrintC::opFloatMult(), PrintC::opFloatNotEqual(), PrintC::opFloatSub(), PrintC::opIntAdd(), PrintC::opIntAnd(), PrintC::opIntDiv(), PrintC::opIntEqual(), PrintC::opIntLeft(), PrintC::opIntLess(), PrintC::opIntLessEqual(), PrintC::opIntMult(), PrintC::opIntNotEqual(), PrintC::opIntOr(), PrintC::opIntRem(), PrintC::opIntRight(), PrintC::opIntSdiv(), PrintC::opIntSless(), PrintC::opIntSlessEqual(), PrintC::opIntSrem(), PrintC::opIntSright(), PrintC::opIntSub(), and PrintC::opIntXor().

◆ opUnary()

void PrintLanguage::opUnary	(	const OpToken *	tok,
		const PcodeOp *	op
	)

protected

Push a unary operator onto the RPN stack.

Push an operator onto the stack that has a normal unary format. Its input expression is also pushed.

Parameters

tok	is the operator token to push
op	is the associated PcodeOp

References PcodeOp::getIn(), mods, pushOp(), and pushVnImplied().

Referenced by PrintC::opFloatNeg(), PrintC::opInt2Comp(), and PrintC::opIntNegate().

◆ printUnicode()

virtual void PrintLanguage::printUnicode	(	ostream &	s,
		int4	onechar
	)		const

protectedpure virtual

Print a single unicode character as a character constant for the high-level language.

For most languages, this prints the character surrounded by single quotes.

Parameters

s	is the output stream
onechar	is the unicode code point of the character to print

Implemented in PrintC, and PrintJava.

Referenced by escapeCharacterData().

◆ pushAnnotation()

virtual void PrintLanguage::pushAnnotation	(	const Varnode *	vn,
		const PcodeOp *	op
	)

protectedpure virtual

Push an address which is not in the normal data-flow.

The given Varnode is treated as an address, which may or may not have a symbol name.

Parameters

vn	is the annotation Varnode
op	is the PcodeOp which takes the annotation as input

Implemented in PrintC.

Referenced by pushVnExplicit().

◆ pushAtom()

void PrintLanguage::pushAtom ( const Atom & atom )

protected

Push a variable token onto the RPN stack.

Push a single token (an Atom) onto the RPN stack. This may trigger some amount of the RPN stack to get emitted, depending on what was pushed previously. The 'emit' routines are called, popping off as much as possible.

Parameters

atom	is the token to be pushed

References EmitXml::closeGroup(), EmitXml::closeParen(), emit, emitAtom(), emitOp(), nodepend, pending, recurse(), and revpol.

Referenced by PrintC::emitPrototypeInputs(), PrintC::emitStructDefinition(), PrintC::opCall(), PrintJava::opCallind(), PrintC::opCallind(), PrintC::opCallother(), PrintC::opConstructor(), PrintJava::opCpoolRefOp(), PrintC::opCpoolRefOp(), PrintC::opFunc(), PrintC::opNewOp(), PrintC::opPtrsub(), PrintC::opReturn(), PrintC::push_float(), PrintC::push_integer(), PrintC::pushAnnotation(), PrintC::pushBoolConstant(), PrintC::pushCharConstant(), PrintC::pushConstant(), PrintC::pushEnumConstant(), PrintC::pushMismatchSymbol(), PrintC::pushPartialSymbol(), PrintC::pushPrototypeInputs(), PrintC::pushPtrCharConstant(), PrintC::pushPtrCodeConstant(), PrintC::pushSymbol(), PrintC::pushType(), PrintC::pushTypeEnd(), PrintJava::pushTypeStart(), PrintC::pushTypeStart(), and PrintC::pushUnnamedLocation().

◆ pushConstant()

virtual void PrintLanguage::pushConstant	(	uintb	val,
		const Datatype *	ct,
		const Varnode *	vn,
		const PcodeOp *	op
	)

protectedpure virtual

Push a constant onto the RPN stack.

The value is ultimately emitted based on its data-type and other associated mark-up

Parameters

val	is the value of the constant
ct	is the data-type of the constant
vn	is the Varnode holding the constant (optional)
op	is the PcodeOp using the constant (optional)

Implemented in PrintC.

Referenced by pushVnExplicit().

◆ pushEquate()

virtual bool PrintLanguage::pushEquate	(	uintb	val,
		int4	sz,
		const EquateSymbol *	sym,
		const Varnode *	vn,
		const PcodeOp *	op
	)

protectedpure virtual

Push a constant marked up by and EquateSymbol onto the RPN stack.

The equate may substitute a name or force a conversion for the constant

Parameters

val	is the value of the constant
sz	is the number of bytes to use for the encoding
sym	is the EquateSymbol that marks up the constant
vn	is the Varnode holding the constant (optional)
op	is the PcodeOp using the constant (optional)

Implemented in PrintC.

◆ pushMismatchSymbol()

virtual void PrintLanguage::pushMismatchSymbol	(	const Symbol *	sym,
		int4	off,
		int4	sz,
		const Varnode *	vn,
		const PcodeOp *	op
	)

protectedpure virtual

Push an identifier for a variable that mismatches with its Symbol.

This happens when a Varnode overlaps, but is not contained by a Symbol. This most commonly happens when the size of a Symbol is unknown

Parameters

sym	is the overlapped symbol
off	is the byte offset of the variable relative to the symbol
sz	is the size of the variable in bytes
vn	is the Varnode representing the variable
op	is a PcodeOp associated with the Varnode

Implemented in PrintC.

Referenced by pushVnExplicit(), and pushVnLHS().

◆ pushOp()

void PrintLanguage::pushOp	(	const OpToken *	tok,
		const PcodeOp *	op
	)

protected

Push an operator token onto the RPN stack.

This generally will recursively push an entire expression onto the RPN stack, up to Varnode objects marked as explicit, and will decide token order and parenthesis placement. As the ordering gets resolved, some amount of the expression may get emitted.

Parameters

tok	is the operator token to push
op	is the PcodeOp associated with the token

References emit, emitOp(), nodepend, EmitXml::openGroup(), EmitXml::openParen(), pending, recurse(), and revpol.

Referenced by PrintC::emitExpression(), PrintC::emitInplaceOp(), opBinary(), PrintC::opBoolNegate(), PrintC::opCall(), PrintJava::opCallind(), PrintC::opCallind(), PrintC::opCallother(), PrintC::opCbranch(), PrintC::opConstructor(), PrintJava::opCpoolRefOp(), PrintC::opCpoolRefOp(), PrintC::opFunc(), PrintJava::opLoad(), PrintC::opLoad(), PrintC::opNewOp(), PrintC::opPtradd(), PrintC::opPtrsub(), PrintC::opReturn(), PrintJava::opStore(), PrintC::opStore(), PrintC::opTypeCast(), opUnary(), PrintC::pushConstant(), PrintC::pushEnumConstant(), PrintC::pushEquate(), PrintC::pushPartialSymbol(), PrintC::pushPrototypeInputs(), PrintJava::pushTypeStart(), and PrintC::pushTypeStart().

◆ pushPartialSymbol()

virtual void PrintLanguage::pushPartialSymbol	(	const Symbol *	sym,
		int4	off,
		int4	sz,
		const Varnode *	vn,
		const PcodeOp *	op,
		Datatype *	outtype
	)

protectedpure virtual

Push a variable that represents only part of a symbol onto the RPN stack.

Generally member syntax specifying a field within a structure gets emitted.

Parameters

sym	is the root Symbol
off	is the byte offset, within the Symbol, of the partial variable
sz	is the number of bytes in the partial variable
vn	is the Varnode holding the partial value
op	is a PcodeOp associate with the Varnode
outtype	is the data-type expected by expression using the partial variable

Implemented in PrintC.

Referenced by pushVnExplicit(), and pushVnLHS().

◆ pushSymbol()

virtual void PrintLanguage::pushSymbol	(	const Symbol *	sym,
		const Varnode *	vn,
		const PcodeOp *	op
	)

protectedpure virtual

Push a specific Symbol onto the RPN stack.

Parameters

sym	is the given Symbol
vn	is the Varnode holding the Symbol value
op	is a PcodeOp associated with the Varnode

Implemented in PrintC.

Referenced by pushVnExplicit(), and pushVnLHS().

◆ pushType()

virtual void PrintLanguage::pushType ( const Datatype * ct )

protectedpure virtual

Push a data-type name onto the RPN expression stack.

The data-type is generally emitted as if for a cast.

Parameters

ct	is the data-type to push

Implemented in PrintC.

◆ pushUnnamedLocation()

virtual void PrintLanguage::pushUnnamedLocation	(	const Address &	addr,
		const Varnode *	vn,
		const PcodeOp *	op
	)

protectedpure virtual

Push an address as a substitute for a Symbol onto the RPN stack.

If there is no Symbol or other name source for an explicit variable, this method is used to print something to represent the variable based on its storage address.

Parameters

addr	is the storage address
vn	is the Varnode representing the variable (if present)
op	is a PcodeOp associated with the variable

Implemented in PrintC.

Referenced by pushVnExplicit(), and pushVnLHS().

◆ pushVnExplicit()

void PrintLanguage::pushVnExplicit	(	const Varnode *	vn,
		const PcodeOp *	op
	)

protected

Push an explicit variable onto the RPN stack.

This method pushes a given Varnode as a leaf of the current expression. It decides how the Varnode should get emitted, as a symbol, constant, etc., and then pushes the resulting leaf Atom onto the stack.

Parameters

vn	is the given explicit Varnode
op	is the PcodeOp incorporating the Varnode into the current expression

References Varnode::getAddr(), Varnode::getHigh(), HighVariable::getNameRepresentative(), Varnode::getOffset(), Datatype::getSize(), Varnode::getSize(), HighVariable::getSymbol(), HighVariable::getSymbolOffset(), HighVariable::getType(), Symbol::getType(), Varnode::isAnnotation(), Varnode::isConstant(), pushAnnotation(), pushConstant(), pushMismatchSymbol(), pushPartialSymbol(), pushSymbol(), and pushUnnamedLocation().

Referenced by PrintC::emitInplaceOp(), and recurse().

◆ pushVnImplied()

void PrintLanguage::pushVnImplied	(	const Varnode *	vn,
		const PcodeOp *	op,
		uint4	m
	)

protected

Push an implied variable onto the RPN stack.

For a given implied Varnode, the entire expression producing it is recursively pushed onto the RPN stack.

When calling this method multiple times to push Varnode inputs for a single p-code op, the inputs must be pushed in reverse order.

Parameters

vn	is the given implied Varnode
op	is PcodeOp taking the Varnode as input
m	is the set of printing modifications to apply for this sub-expression

References nodepend.

Referenced by PrintC::emitInplaceOp(), opBinary(), PrintC::opBoolNegate(), PrintC::opBranch(), PrintC::opBranchind(), PrintC::opCall(), PrintJava::opCallind(), PrintC::opCallind(), PrintC::opCallother(), PrintC::opCbranch(), PrintC::opConstructor(), PrintC::opCopy(), PrintJava::opCpoolRefOp(), PrintC::opCpoolRefOp(), PrintC::opFunc(), PrintJava::opLoad(), PrintC::opLoad(), PrintC::opNewOp(), PrintC::opPtradd(), PrintC::opPtrsub(), PrintC::opReturn(), PrintC::opSegmentOp(), PrintJava::opStore(), PrintC::opStore(), PrintC::opTypeCast(), and opUnary().

◆ pushVnLHS()

void PrintLanguage::pushVnLHS	(	const Varnode *	vn,
		const PcodeOp *	op
	)

protected

Push a variable as the left-hand side of an expression.

The given Varnode will ultimately be emitted as an explicit variable on the left-hand side of an assignment statement. As with pushVnExplicit(), this method decides how the Varnode will be emitted and pushes the resulting Atom onto the RPN stack.

Parameters

vn	is the given LSH Varnode
op	is the PcodeOp which produces the Varnode as an output

References Varnode::getAddr(), Varnode::getHigh(), HighVariable::getNameRepresentative(), Datatype::getSize(), Varnode::getSize(), HighVariable::getSymbol(), HighVariable::getSymbolOffset(), Symbol::getType(), pushMismatchSymbol(), pushPartialSymbol(), pushSymbol(), and pushUnnamedLocation().

Referenced by PrintC::emitExpression().

◆ readUtf16()

int4 PrintLanguage::readUtf16	(	const uint1 *	buf,
		bool	bigend
	)

inlinestaticprotected

Read a 2-byte UTF16 element from a byte array.

Pull the first two bytes from the byte array and combine them in the indicated endian order

Parameters

buf	is the byte array
bigend	is true to request big endian encoding

Returns: the decoded UTF16 element

Referenced by getCodepoint().

◆ recurse()

void PrintLanguage::recurse ( void )

protected

Emit from the RPN stack as much as possible.

Any complete sub-expressions that are still on the RPN will get emitted.

References Varnode::getDef(), Varnode::isImplied(), mods, nodepend, pending, PcodeOp::push(), and pushVnExplicit().

Referenced by PrintC::emitEnumDefinition(), PrintC::emitExpression(), PrintC::emitInplaceOp(), PrintC::emitPrototypeInputs(), PrintC::emitPrototypeOutput(), PrintC::emitSwitchCase(), PrintC::emitVarDecl(), PrintC::opBranchind(), PrintC::opCbranch(), pushAtom(), and pushOp().

◆ setCommentDelimeter()

void PrintLanguage::setCommentDelimeter	(	const string &	start,
		const string &	stop,
		bool	usecommentfill
	)

Establish comment delimiters for the language.

By default, comments are indicated in the high-level language by preceding them with a specific sequence of delimiter characters, and optionally by ending the comment with another set of delimiter characters.

Parameters

start	is the initial sequence of characters delimiting a comment
stop	if not empty is the sequence delimiting the end of the comment
usecommentfill	is true if the delimiter needs to be emitted after every line break

References commentend, commentstart, emit, and EmitXml::setCommentFill().

Referenced by PrintC::setCPlusPlusStyleComments(), and PrintC::setCStyleComments().

◆ setCommentStyle()

virtual void PrintLanguage::setCommentStyle ( const string & nm )

pure virtual

Set the way comments are displayed in decompiler output.

This method can either be provided a formal name or a sample of the initial delimiter, then it will choose from among the schemes it knows

Parameters

nm	is the configuration description

Implemented in PrintC.

Referenced by OptionCommentStyle::apply().

◆ setFlat()

void PrintLanguage::setFlat ( bool val )

Set whether nesting code structure should be emitted.

Emitting formal code structuring can be turned off, causing all control-flow to be represented as goto statements and labels.

Parameters

val	is true if no code structuring should be emitted

References flat, and mods.

◆ setIntegerFormat()

void PrintLanguage::setIntegerFormat ( const string & nm )

virtual

Set the default integer format.

This determines how integers are displayed by default. Possible values are "hex" and "dec" to force a given format, or "best" can be used to let the decompiler select what it thinks best for each individual integer.

Parameters

nm	is "hex", "dec", or "best"

References force_dec, force_hex, and mods.

Referenced by OptionIntegerFormat::apply().

◆ setLineCommentIndent()

void PrintLanguage::setLineCommentIndent ( int4 val )

Set the number of characters to indent comment lines.

Parameters

val	is the number of characters

References emit, EmitXml::getMaxLineSize(), and line_commentindent.

Referenced by OptionCommentIndent::apply().

◆ setXML()

void PrintLanguage::setXML ( bool val )

Set whether the low-level emitter, emits XML markup.

Tell the emitter whether to emit just the raw tokens or if output is in XML format with additional mark-up on the raw tokens.

Parameters

val	is true for XML mark-up

References emit.

Referenced by ArchitectureGhidra::ArchitectureGhidra(), and Architecture::setPrintLanguage().

◆ unicodeNeedsEscape()

bool PrintLanguage::unicodeNeedsEscape ( int4 codepoint )

staticprotected

Determine if the given codepoint needs to be escaped.

Separate unicode characters that can be clearly emitted in a source code string (letters, numbers, punctuation, symbols) from characters that are better represented in source code with an escape sequence (control characters, unusual spaces, separators, private use characters etc.

Parameters

codepoint is the given unicode codepoint to categorize.

Returns: true if the codepoint needs to be escaped

Referenced by PrintJava::printUnicode(), and PrintC::printUnicode().

◆ writeUtf8()

void PrintLanguage::writeUtf8	(	ostream &	s,
		int4	codepoint
	)

staticprotected

Write unicode character to stream in UTF8 encoding.

Encode the given unicode codepoint as UTF8 (1, 2, 3, or 4 bytes) and write the bytes to the stream.

Parameters

s	is the output stream
codepoint	is the unicode codepoint

References mostsigbit_set().

Referenced by PrintJava::printUnicode(), and PrintC::printUnicode().

The documentation for this class was generated from the following files:

printlanguage.hh
printlanguage.cc

Classes

Public Types

Public Member Functions

Static Public Member Functions

Protected Member Functions

Static Protected Member Functions

Protected Attributes

Private Attributes

Detailed Description

Member Enumeration Documentation

◆ modifiers

◆ tagtype

Constructor & Destructor Documentation

◆ PrintLanguage()

Member Function Documentation

◆ checkPrintNegation()

◆ docFunction()

◆ docSingleGlobal()

◆ docTypeDefinitions()

◆ emitAtom()

◆ emitExpression()

◆ emitFunctionDeclaration()

◆ emitLineComment()

◆ emitOp()

◆ emitScopeVarDecls()

◆ emitVarDecl()

◆ emitVarDeclStatement()

◆ escapeCharacterData()

◆ formatBinary()

◆ getCodepoint()

◆ isCharacterConstant()

◆ mostNaturalBase()

◆ opBinary()

◆ opUnary()

◆ printUnicode()

◆ pushAnnotation()

◆ pushAtom()

◆ pushConstant()

◆ pushEquate()

◆ pushMismatchSymbol()

◆ pushOp()

◆ pushPartialSymbol()

◆ pushSymbol()

◆ pushType()

◆ pushUnnamedLocation()

◆ pushVnExplicit()

◆ pushVnImplied()

◆ pushVnLHS()

◆ readUtf16()

◆ recurse()

◆ setCommentDelimeter()

◆ setCommentStyle()

◆ setFlat()

◆ setIntegerFormat()

◆ setLineCommentIndent()

◆ setXML()

◆ unicodeNeedsEscape()

◆ writeUtf8()