blob: 89d776bbb0cdbe4973c3bd1a26b70215d6cd02be [file] [log] [blame]
/*===-- Lexer.l - Scanner for llvm assembly files ----------------*- C++ -*--=//
//
// This file implements the flex scanner for LLVM assembly languages files.
//
//===------------------------------------------------------------------------=*/
%option prefix="llvmAsm"
%option yylineno
%option nostdinit
%option never-interactive
%option batch
%option noyywrap
%option nodefault
%option 8bit
%option outfile="Lexer.cpp"
%option ecs
%option noreject
%option noyymore
%{
#include "ParserInternals.h"
#include "llvm/BasicBlock.h"
#include "llvm/Method.h"
#include "llvm/Module.h"
#include <list>
#include "llvmAsmParser.h"
#define RET_TOK(type, Enum, sym) \
llvmAsmlval.type = Instruction::Enum; return sym
// TODO: All of the static identifiers are figured out by the lexer,
// these should be hashed.
// atoull - Convert an ascii string of decimal digits into the unsigned long
// long representation... this does not have to do input error checking,
// because we know that the input will be matched by a suitable regex...
//
uint64_t atoull(const char *Buffer) {
uint64_t Result = 0;
for (; *Buffer; Buffer++) {
uint64_t OldRes = Result;
Result *= 10;
Result += *Buffer-'0';
if (Result < OldRes) { // Uh, oh, overflow detected!!!
ThrowException("constant bigger than 64 bits detected!");
}
}
return Result;
}
#define YY_NEVER_INTERACTIVE 1
%}
/* Comments start with a ; and go till end of line */
Comment ;.*
/* Variable(Def) identifiers start with a % sign */
VarID %[a-zA-Z$._][a-zA-Z$._0-9]*
/* Label identifiers end with a colon */
Label [a-zA-Z$._0-9]+:
/* Quoted names can contain any character except " and \ */
StringConstant \"[^\"]+\"
/* [PN]Integer: match positive and negative literal integer values that
* are preceeded by a '%' character. These represent unnamed variable slots.
*/
EPInteger %[0-9]+
ENInteger %-[0-9]+
/* E[PN]Integer: match positive and negative literal integer values */
PInteger [0-9]+
NInteger -[0-9]+
%%
{Comment} { /* Ignore comments for now */ }
begin { return BEGINTOK; }
end { return END; }
true { return TRUE; }
false { return FALSE; }
declare { return DECLARE; }
implementation { return IMPLEMENTATION; }
- { cerr << "deprecated argument '-' used!\n"; return '-'; }
bb { cerr << "deprecated type 'bb' used!\n"; llvmAsmlval.TypeVal = Type::LabelTy; return LABEL;}
void { llvmAsmlval.TypeVal = Type::VoidTy ; return VOID; }
bool { llvmAsmlval.TypeVal = Type::BoolTy ; return BOOL; }
sbyte { llvmAsmlval.TypeVal = Type::SByteTy ; return SBYTE; }
ubyte { llvmAsmlval.TypeVal = Type::UByteTy ; return UBYTE; }
short { llvmAsmlval.TypeVal = Type::ShortTy ; return SHORT; }
ushort { llvmAsmlval.TypeVal = Type::UShortTy; return USHORT; }
int { llvmAsmlval.TypeVal = Type::IntTy ; return INT; }
uint { llvmAsmlval.TypeVal = Type::UIntTy ; return UINT; }
long { llvmAsmlval.TypeVal = Type::LongTy ; return LONG; }
ulong { llvmAsmlval.TypeVal = Type::ULongTy ; return ULONG; }
float { llvmAsmlval.TypeVal = Type::FloatTy ; return FLOAT; }
double { llvmAsmlval.TypeVal = Type::DoubleTy; return DOUBLE; }
type { llvmAsmlval.TypeVal = Type::TypeTy ; return TYPE; }
label { llvmAsmlval.TypeVal = Type::LabelTy ; return LABEL; }
neg { RET_TOK(UnaryOpVal, Neg, NEG); }
not { RET_TOK(UnaryOpVal, Not, NOT); }
phi { return PHI; }
call { return CALL; }
add { RET_TOK(BinaryOpVal, Add, ADD); }
sub { RET_TOK(BinaryOpVal, Sub, SUB); }
mul { RET_TOK(BinaryOpVal, Mul, MUL); }
div { RET_TOK(BinaryOpVal, Div, DIV); }
rem { RET_TOK(BinaryOpVal, Rem, REM); }
setne { RET_TOK(BinaryOpVal, SetNE, SETNE); }
seteq { RET_TOK(BinaryOpVal, SetEQ, SETEQ); }
setlt { RET_TOK(BinaryOpVal, SetLT, SETLT); }
setgt { RET_TOK(BinaryOpVal, SetGT, SETGT); }
setle { RET_TOK(BinaryOpVal, SetLE, SETLE); }
setge { RET_TOK(BinaryOpVal, SetGE, SETGE); }
ret { RET_TOK(TermOpVal, Ret, RET); }
br { RET_TOK(TermOpVal, Br, BR); }
switch { RET_TOK(TermOpVal, Switch, SWITCH); }
malloc { RET_TOK(MemOpVal, Malloc, MALLOC); }
alloca { RET_TOK(MemOpVal, Alloca, ALLOCA); }
free { RET_TOK(MemOpVal, Free, FREE); }
load { RET_TOK(MemOpVal, Load, LOAD); }
store { RET_TOK(MemOpVal, Store, STORE); }
getfield { RET_TOK(MemOpVal, GetField, GETFIELD); }
putfield { RET_TOK(MemOpVal, PutField, PUTFIELD); }
{VarID} { llvmAsmlval.StrVal = strdup(yytext+1); return VAR_ID; }
{Label} {
yytext[strlen(yytext)-1] = 0; // nuke colon
llvmAsmlval.StrVal = strdup(yytext);
return LABELSTR;
}
{StringConstant} {
yytext[strlen(yytext)-1] = 0; // nuke end quote
llvmAsmlval.StrVal = strdup(yytext+1); // Nuke start quote
return STRINGCONSTANT;
}
{PInteger} { llvmAsmlval.UInt64Val = atoull(yytext); return EUINT64VAL; }
{NInteger} {
uint64_t Val = atoull(yytext+1);
// +1: we have bigger negative range
if (Val > (uint64_t)INT64_MAX+1)
ThrowException("Constant too large for signed 64 bits!");
llvmAsmlval.SInt64Val = -Val;
return ESINT64VAL;
}
{EPInteger} { llvmAsmlval.UIntVal = atoull(yytext+1); return UINTVAL; }
{ENInteger} {
uint64_t Val = atoull(yytext+2);
// +1: we have bigger negative range
if (Val > (uint64_t)INT32_MAX+1)
ThrowException("Constant too large for signed 32 bits!");
llvmAsmlval.SIntVal = -Val;
return SINTVAL;
}
[ \t\n] { /* Ignore whitespace */ }
. { /*printf("'%s'", yytext);*/ return yytext[0]; }
%%