1

/*

     2

 *  matiec - a compiler for the programming languages defined in IEC 61131-3

     3

 *

     4

 *  Copyright (C) 2009-2011  Mario de Sousa (msousa@fe.up.pt)

     5

 *  Copyright (C) 2007-2011  Laurent Bessard and Edouard Tisserant

     6

 *

     7

 *  This program is free software: you can redistribute it and/or modify

     8

 *  it under the terms of the GNU General Public License as published by

     9

 *  the Free Software Foundation, either version 3 of the License, or

    10

 *  (at your option) any later version.

    11

 *

    12

 *  This program is distributed in the hope that it will be useful,

    13

 *  but WITHOUT ANY WARRANTY; without even the implied warranty of

    14

 *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the

    15

 *  GNU General Public License for more details.

    16

 *

    17

 *  You should have received a copy of the GNU General Public License

    18

 *  along with this program.  If not, see <http://www.gnu.org/licenses/>.

    19

 *

    20

 *

    21

 * This code is made available on the understanding that it will not be

    22

 * used in safety-critical situations without a full and competent review.

    23

 */

    24

    25

/*

    26

 * An IEC 61131-3 compiler.

    27

 *

    28

 * Based on the

    29

 * FINAL DRAFT - IEC 61131-3, 2nd Ed. (2001-12-10)

    30

 *

    31

 */

    32

    33

    34

/*

    35

 * This is the main stage 3a file.

    36

 *

    37

 * In stage 3a some helpful symbol tables are instanciated and populated.

    38

 * These symbol tables wll then be used by stage3b and atage4 code generators.

    39

 */

    40

    41

    42

    43

    44

    45

    46

// #include <stdio.h>  /* required for NULL */

    47

#include <string>

    48

#include <iostream>

    49

#include <sstream>

    50

#include <typeinfo>

    51

#include <list>

    52

#include <strings.h>

    53

#include <string.h>  /* required for strlen() */

    54

#include <stdlib.h>  /* required for atoi() */

    55

#include <errno.h>   /* required for errno */

    56

    57

#include "../util/symtable.hh"

    58

#include "../util/dsymtable.hh"

    59

#include "../absyntax/visitor.hh"

    60

    61

    62

    63

//#define DEBUG

    64

#ifdef DEBUG

    65

#define TRACE(classname) printf("\n____%s____\n",classname);

    66

#else

    67

#define TRACE(classname)

    68

#endif

    69

    70

#define ERROR error_exit(__FILE__,__LINE__)

    71

/* function defined in main.cc */

    72

extern void error_exit(const char *file_name, int line_no);

    73

    74

    75

/***********************************************************************/

    76

/***********************************************************************/

    77

/***********************************************************************/

    78

/***********************************************************************/

    79

    80

    81

/* returns 0 if the names are equal!! */

    82

/* NOTE: it must ignore case!! */

    83

int compare_identifiers(symbol_c *ident1, symbol_c *ident2) {

    84

    85

  token_c *name1 = dynamic_cast<token_c *>(ident1);

    86

  token_c *name2 = dynamic_cast<token_c *>(ident2);

    87

    88

  if ((name1 == NULL) || (name2 == NULL))

    89

    /* invalid identifiers... */

    90

    return -1;

    91

    92

  if (strcasecmp(name1->value, name2->value) == 0)

    93

    return 0;

    94

    95

  /* identifiers do not match! */

    96

  return 1;

    97

}

    98

    99

   100

   101

  /* To allow the compiler to be portable, we cannot assume that int64_t is mapped onto long long int,

   102

   * so we cannot call strtoll() and strtoull() in extract_int64() and extract_uint64().

   103

   *

   104

   * So, we create our own strtouint64() and strtoint64() functions.

   105

   * (We actually call them matiec_strtoint64() so they will not clash with any function

   106

   *  that may be added to the standard library in the future).

   107

   * We actually create several of each, and let the compiler choose which is the correct one,

   108

   * by having it resolve the call to the overloaded function. For the C++ compiler to be able

   109

   * to resolve this ambiguity, we need to add a dummy parameter to each function!

   110

   *

   111

   * TODO: support platforms in which int64_t is mapped onto int !! Is this really needed?

   112

   */

   113

static  int64_t matiec_strtoint64 (         long      int *dummy, const char *nptr, char **endptr, int base) {return strtol  (nptr, endptr, base);}

   114

static  int64_t matiec_strtoint64 (         long long int *dummy, const char *nptr, char **endptr, int base) {return strtoll (nptr, endptr, base);}

   115

   116

static uint64_t matiec_strtouint64(unsigned long      int *dummy, const char *nptr, char **endptr, int base) {return strtoul (nptr, endptr, base);}

   117

static uint64_t matiec_strtouint64(unsigned long long int *dummy, const char *nptr, char **endptr, int base) {return strtoull(nptr, endptr, base);}

   118

   119

   120

/* extract the value of an integer from an integer_c object !! */

   121

/* NOTE: it must ignore underscores! */

   122

int64_t extract_int64_value(symbol_c *sym, bool *overflow) {

   123

  std::string str = "";

   124

  integer_c *integer;

   125

  neg_integer_c * neg_integer;

   126

  char *endptr;

   127

  int64_t ret;

   128

   129

  if ((integer = dynamic_cast<integer_c *>(sym)) == NULL) ERROR;

   130

  for(unsigned int i = 0; i < strlen(integer->value); i++)

   131

    if (integer->value[i] != '_')  str += integer->value[i];

   132

   133

  errno = 0; // since strtoXX() may legally return 0, we must set errno to 0 to detect errors correctly!

   134

  ret = matiec_strtoint64((int64_t *)NULL, str.c_str(), &endptr, 10);

   135

  if (overflow != NULL)

   136

    *overflow = (errno == ERANGE);

   137

  if ((errno != 0) && (errno != ERANGE))

   138

    ERROR;

   139

   140

  return ret;

   141

}

   142

   143

uint64_t extract_uint64_value(symbol_c *sym, bool *overflow) {

   144

  std::string str = "";

   145

  integer_c *integer;

   146

  neg_integer_c * neg_integer;

   147

  char *endptr;

   148

  uint64_t ret;

   149

   150

  if ((integer = dynamic_cast<integer_c *>(sym)) == NULL) ERROR;

   151

  for(unsigned int i = 0; i < strlen(integer->value); i++)

   152

    if (integer->value[i] != '_')  str += integer->value[i];

   153

   154

  errno = 0; // since strtoXX() may legally return 0, we must set errno to 0 to detect errors correctly!

   155

  ret = matiec_strtouint64((uint64_t *)NULL, str.c_str(), &endptr, 10);

   156

  if (overflow != NULL)

   157

    *overflow = (errno == ERANGE);

   158

  if ((errno != 0) && (errno != ERANGE))

   159

    ERROR;

   160

   161

  return ret;

   162

}

   163

   164

   165

/* extract the value of an hex integer from an hex_integer_c object !! */

   166

/* NOTE: it must ignore underscores! */

   167

uint64_t extract_hex_value(symbol_c *sym) {

   168

  std::string str = "";

   169

  char *endptr;

   170

  hex_integer_c * hex_integer;

   171

  uint64_t ret;

   172

   173

  if ((hex_integer = dynamic_cast<hex_integer_c *>(sym)) == NULL) ERROR;

   174

  for(unsigned int i = 3; i < strlen(hex_integer->value); i++)

   175

    if (hex_integer->value[i] != '_') str += hex_integer->value[i];

   176

   177

  errno = 0; // since strtoXX() may legally return 0, we must set errno to 0 to detect errors correctly!

   178

  ret = strtoull(str.c_str(), &endptr, 16);

   179

  if (errno != 0) ERROR;

   180

   181

  return ret;

   182

}

   183

   184

   185

/* extract the value of a real from an real_c object !! */

   186

/* NOTE: it must ignore underscores! */

   187

/* From iec_bison.yy

   188

 *  real:

   189

 *   real_token		{$$ = new real_c($1, locloc(@$));}

   190

 * | fixed_point_token	{$$ = new real_c($1, locloc(@$));}

   191

 *

   192

 * From iec_flex.ll

   193

 * {real}			{yylval.ID=strdup(yytext); return real_token;}

   194

 * {fixed_point}		{yylval.ID=strdup(yytext); return fixed_point_token;}

   195

 *

   196

 * real		{integer}\.{integer}{exponent}

   197

 * fixed_point		{integer}\.{integer}

   198

 * exponent        [Ee]([+-]?){integer}

   199

 * integer         {digit}((_?{digit})*)

   200

 */

   201

real64_t extract_real_value(symbol_c *sym, bool *overflow) {

   202

  std::string str = "";

   203

  real_c * real_sym;

   204

  real64_t ret;

   205

   206

  if ((real_sym = dynamic_cast<real_c *>(sym)) == NULL) ERROR;

   207

  for(unsigned int i = 0; i < strlen(real_sym->value); i++)

   208

    if (real_sym->value[i] != '_') str += real_sym->value[i];

   209

   210

  errno = 0; // since strtoXX() may legally return 0, we must set errno to 0 to detect errors correctly!

   211

  #if    (real64_t  == float)

   212

    ret = strtof(str.c_str(), NULL);

   213

  #elif  (real64_t  == double)

   214

    ret = strtod(str.c_str(), NULL);

   215

  #elif  (real64_t  == long_double)

   216

    ret = strtold(str.c_str(), NULL);

   217

  #else 

   218

    #error Could not determine which data type is being used for real64_t (defined in absyntax.hh). Aborting!

   219

  #endif

   220

  if (overflow != NULL)

   221

    *overflow = (errno == ERANGE);

   222

  if ((errno != 0) && (errno != ERANGE)) 

   223

    ERROR;

   224

   225

  return ret;

   226

}

   227

   228

   229

   230

/***********************************************************************/

   231

/***********************************************************************/

   232

/***********************************************************************/

   233

/***********************************************************************/

   234

   235

   236

   237

/* A symbol table with all globally declared functions... */

   238

function_declaration_c null_symbol1(NULL,NULL,NULL,NULL);

   239

dsymtable_c<function_declaration_c *, &null_symbol1> function_symtable;

   240

   241

/* A symbol table with all globally declared functions block types... */

   242

function_block_declaration_c null_symbol2(NULL,NULL,NULL);

   243

symtable_c<function_block_declaration_c *, &null_symbol2> function_block_type_symtable;

   244

   245

/* A symbol table with all globally declared program types... */

   246

program_declaration_c null_symbol3(NULL,NULL,NULL);

   247

symtable_c<program_declaration_c *, &null_symbol3> program_type_symtable;

   248

   249

/* A symbol table with all user declared type definitions... */

   250

/* Note that function block types and program types have their

   251

 * own symbol tables, so do not get placed in this symbol table!

   252

 *

   253

 * The symbol_c * associated to the value will point to the data type declaration.

   254

 */

   255

symbol_c null_symbol4;

   256

symtable_c<symbol_c *, &null_symbol4> type_symtable;

   257

   258

/* A symbol table with all values declared for enumerated type... */

   259

/* Notes:

   260

 * - if the value is defined multiple times the value

   261

 * is the null pointer.

   262

 *

   263

 * - The stored symbol_c * associated to the value points to the enumerated_type_name

   264

 * (i.e. the name of the enumerated data type) in which the the value/identifier

   265

 * is used/declared.

   266

 *

   267

 * - We could re-use the null_symbol4 object, but it is safer to use a distinct object

   268

 *   (i.e. it might make it easier to find strange bugs).

   269

 */

   270

symbol_c null_symbol5;

   271

symtable_c<symbol_c *, &null_symbol5> enumerated_value_symtable;

   272

   273

   274

/***********************************************************************/

   275

/***********************************************************************/

   276

/***********************************************************************/

   277

/***********************************************************************/

   278

   279

   280

class populate_symtables_c: public iterator_visitor_c {

   281

   282

  private:

   283

	symbol_c *current_enumerated_type;

   284

   285

  public:

   286

    populate_symtables_c(void) {

   287

    	current_enumerated_type = NULL;

   288

    };

   289

    virtual ~populate_symtables_c(void) {}

   290

   291

   292

  public:

   293

   294

  /*************************/

   295

  /* B.1 - Common elements */

   296

  /*************************/

   297

  /*******************************************/

   298

  /* B 1.1 - Letters, digits and identifiers */

   299

  /*******************************************/

   300

  /*********************/

   301

  /* B 1.2 - Constants */

   302

  /*********************/

   303

  /******************************/

   304

  /* B 1.2.1 - Numeric Literals */

   305

  /******************************/

   306

  /*******************************/

   307

  /* B.1.2.2   Character Strings */

   308

  /*******************************/

   309

  /***************************/

   310

  /* B 1.2.3 - Time Literals */

   311

  /***************************/

   312

  /************************/

   313

  /* B 1.2.3.1 - Duration */

   314

  /************************/

   315

  /************************************/

   316

  /* B 1.2.3.2 - Time of day and Date */

   317

  /************************************/

   318

  /**********************/

   319

  /* B.1.3 - Data types */

   320

  /**********************/

   321

  /***********************************/

   322

  /* B 1.3.1 - Elementary Data Types */

   323

  /***********************************/

   324

  /********************************/

   325

  /* B.1.3.2 - Generic data types */

   326

  /********************************/

   327

  /********************************/

   328

  /* B 1.3.3 - Derived data types */

   329

  /********************************/

   330

   331

  /*  subrange_type_name ':' subrange_spec_init */

   332

  void *visit(subrange_type_declaration_c *symbol) {

   333

    TRACE("subrange_type_declaration_c");  

   334

    type_symtable.insert(symbol->subrange_type_name, symbol->subrange_spec_init);

   335

    return NULL;

   336

  }

   337

   338

   339

  /*  enumerated_type_name ':' enumerated_spec_init */

   340

  void *visit(enumerated_type_declaration_c *symbol) {

   341

    TRACE("enumerated_type_declaration_c");

   342

    type_symtable.insert(symbol->enumerated_type_name, symbol->enumerated_spec_init);

   343

    current_enumerated_type = symbol->enumerated_type_name;

   344

    symbol->enumerated_spec_init->accept(*this);

   345

    current_enumerated_type = NULL;

   346

    return NULL;

   347

  }

   348

   349

  /* enumerated_specification ASSIGN enumerated_value */

   350

  void *visit(enumerated_spec_init_c *symbol) {

   351

    return symbol->enumerated_specification->accept(*this);

   352

  }

   353

   354

  /* [enumerated_type_name '#'] identifier */

   355

  void *visit(enumerated_value_c *symbol) {

   356

    if (current_enumerated_type != NULL) {

   357

      if (symbol->type != NULL) ERROR;

   358

   359

      symbol_c *value_type = enumerated_value_symtable.find_value(symbol->value);

   360

      /* NOTE: The following condition checks whether the same identifier is used more than once

   361

       *       when defining the enumerated values of the type declaration of the new enumerated type.

   362

       *       If this occurs, then the program beeing compiled contains a semantic error, which

   363

       *       must be caught and reported by the semantic analyser. However, since

   364

       *       this code is run before the semantic analyser, we must not yet raise the ERROR (internal

   365

       *       compiler error message).

   366

       *       For this reason, the follosing check is commented out.

   367

       */

   368

      /* if (value_type == current_enumerated_type) ERROR; */

   369

   370

      if (value_type == enumerated_value_symtable.end_value())

   371

	/* This identifier has not yet been used in any previous declaration of an enumeration data type.

   372

	 * so we add it to the symbol table.

   373

	 */

   374

        enumerated_value_symtable.insert(symbol->value, current_enumerated_type);

   375

      else if (value_type != NULL)

   376

	/* This identifier has already been used in a previous declaration of an enumeration data type.

   377

	 * so we set the symbol in symbol table pointing to NULL.

   378

	 */

   379

        enumerated_value_symtable.set(symbol->value, NULL);

   380

    }

   381

    return NULL;

   382

  }

   383

   384

  /*  identifier ':' array_spec_init */

   385

  void *visit(array_type_declaration_c *symbol) {

   386

    TRACE("array_type_declaration_c");

   387

    type_symtable.insert(symbol->identifier, symbol->array_spec_init);

   388

    return NULL;

   389

  }

   390

   391

   392

  /*  simple_type_name ':' simple_spec_init */

   393

  void *visit(simple_type_declaration_c *symbol) {

   394

    TRACE("simple_type_declaration_c");

   395

    type_symtable.insert(symbol->simple_type_name, symbol->simple_spec_init);

   396

    return NULL;

   397

  }

   398

   399

   400

  /*  structure_type_name ':' structure_specification */

   401

  void *visit(structure_type_declaration_c *symbol) {

   402

    TRACE("structure_type_declaration_c");

   403

    type_symtable.insert(symbol->structure_type_name, symbol->structure_specification);

   404

    return NULL;

   405

  }

   406

   407

   408

  /*  string_type_name ':' elementary_string_type_name string_type_declaration_size string_type_declaration_init */

   409

  // SYM_REF4(string_type_declaration_c,	string_type_name,

   410

  //    					elementary_string_type_name,

   411

  //    					string_type_declaration_size,

   412

  //    					string_type_declaration_init) /* may be == NULL! */

   413

  void *visit(string_type_declaration_c *symbol)	{

   414

    TRACE("string_type_declaration_c");

   415

    type_symtable.insert(symbol->string_type_name, symbol);

   416

    return NULL;

   417

}

   418

   419

  /*********************/

   420

  /* B 1.4 - Variables */

   421

  /*********************/

   422

  /********************************************/

   423

  /* B.1.4.1   Directly Represented Variables */

   424

  /********************************************/

   425

  /*************************************/

   426

  /* B.1.4.2   Multi-element Variables */

   427

  /*************************************/

   428

  /******************************************/

   429

  /* B 1.4.3 - Declaration & Initialisation */

   430

  /******************************************/

   431

  /**************************************/

   432

  /* B.1.5 - Program organization units */

   433

  /**************************************/

   434

  /***********************/

   435

  /* B 1.5.1 - Functions */

   436

  /***********************/

   437

  public:

   438

  /*   FUNCTION derived_function_name ':' elementary_type_name io_OR_function_var_declarations_list function_body END_FUNCTION */

   439

  /* | FUNCTION derived_function_name ':' derived_type_name io_OR_function_var_declarations_list function_body END_FUNCTION */

   440

  void *visit(function_declaration_c *symbol) {

   441

    TRACE("function_declaration_c");

   442

    function_symtable.insert(symbol->derived_function_name, symbol);

   443

   444

    /* symbol->derived_function_name->accept(*this);  */ /* Function name */

   445

    /* symbol->type_name->accept(*this);              */ /* return data type */

   446

    /* symbol->var_declarations_list->accept(*this);  */ /* Function parameters and variables */

   447

    /* symbol->function_body->accept(*this);          */ /* Function body */

   448

    return NULL;

   449

  }

   450

   451

   452

  /*****************************/

   453

  /* B 1.5.2 - Function Blocks */

   454

  /*****************************/

   455

  public:

   456

  /*  FUNCTION_BLOCK derived_function_block_name io_OR_other_var_declarations function_block_body END_FUNCTION_BLOCK */

   457

  //SYM_REF4(function_block_declaration_c, fblock_name, var_declarations, fblock_body, unused)

   458

  void *visit(function_block_declaration_c *symbol) {

   459

    TRACE("function_block_declaration_c");

   460

    function_block_type_symtable.insert(symbol->fblock_name, symbol);

   461

  /*

   462

    symbol->fblock_name->accept(*this);

   463

    symbol->var_declarations->accept(*this);

   464

    symbol->fblock_body->accept(*this);

   465

  */

   466

    return NULL;

   467

  }

   468

   469

   470

  /**********************/

   471

  /* B 1.5.3 - Programs */

   472

  /**********************/

   473

  public:

   474

  /*  PROGRAM program_type_name program_var_declarations_list function_block_body END_PROGRAM */

   475

  //SYM_REF4(program_declaration_c, program_type_name, var_declarations, function_block_body, unused)

   476

  void *visit(program_declaration_c *symbol) {

   477

    TRACE("program_declaration_c");

   478

    program_type_symtable.insert(symbol->program_type_name, symbol);

   479

  /*

   480

    symbol->program_type_name->accept(*this);

   481

    symbol->var_declarations->accept(*this);

   482

    symbol->function_block_body->accept(*this);

   483

  */

   484

    return NULL;

   485

  }

   486

   487

}; /* populate_symtables_c */

   488

   489

   490

   491

   492

   493

void absyntax_utils_init(symbol_c *tree_root) {

   494

  populate_symtables_c populate_symbols;

   495

   496

  tree_root->accept(populate_symbols);

   497

}

   498