1

/*

     2

 * (c) 2003-2007 Mario de Sousa

     3

 *

     4

 * Offered to the public under the terms of the GNU General Public License

     5

 * as published by the Free Software Foundation; either version 2 of the

     6

 * License, or (at your option) any later version.

     7

 *

     8

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

     9

 * WITHOUT ANY WARRANTY; without even the implied warranty of

    10

 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General

    11

 * Public License for more details.

    12

 *

    13

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

    14

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

    15

 */

    16

    17

/*

    18

 * An IEC 61131-3 IL, ST and SFC compiler.

    19

 *

    20

 * Based on the

    21

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

    22

 *

    23

 */

    24

    25

/*

    26

 * Stage 2

    27

 * =======

    28

 *

    29

 *  This file contains the syntax definition of the textual

    30

 *  languages IL and ST, as well as the textual version of SFC.

    31

 *  The syntax parser, comprising the 2nd stage of the overall 

    32

 *  compiler, is generated by runing bison on this file.

    33

 */

    34

    35

    36

    37

    38

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

    39

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

    40

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

    41

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

    42

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

    43

/*******   The following syntax does not have any conflicts.    *******/

    44

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

    45

/*******   P L E A S E    K E E P    I T    T H A T    W A Y !  *******/

    46

/*******   ===================================================  *******/

    47

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

    48

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

    49

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

    50

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

    51

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

    52

    53

    54

    55

    56

%{

    57

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

    58

    59

    60

/* declare the token parser generated by flex... */

    61

int yylex(void);

    62

    63

/* declare the error handler defined at the end of this file */

    64

void yyerror (const char *error_msg);

    65

    66

/* produce a more verbose parsing error message */

    67

#define YYERROR_VERBOSE

    68

    69

/* Include debuging code.

    70

 * Printing of debug info must then be activated by setting

    71

 * the variable yydebug to 1.

    72

 */

    73

#define YYDEBUG 0

    74

    75

    76

/* file with declaration of absyntax classes... */

    77

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

    78

    79

/* file with declaration of token constants. Generated by bison! */

    80

// #include "iec.y.hh"

    81

    82

/* The interface through which bison and flex interact. */

    83

#include "stage1_2_priv.hh"

    84

    85

    86

/* an ugly hack!!

    87

 * We will probably not need it when we decide

    88

 *  to cut down the abstract syntax down to size.

    89

 *  We keep it as it is until we get to write

    90

 *  stages 3 and 4 of the compiler. Who knows,

    91

 *  we might just find out that we really do need

    92

 *  the abstract syntax tree to stay as it is

    93

 *  afterall!

    94

 */

    95

/* for each element <elem> in list_c * <list>

    96

 * execute the code <code>

    97

 */

    98

#define FOR_EACH_ELEMENT(elem, list, code) {		\

    99

  symbol_c *elem;					\

   100

  for(int i = 0; i < list->n; i++) {			\

   101

    elem = list->elements[i];				\

   102

    code;						\

   103

  }							\

   104

}

   105

   106

   107

   108

/* Macros used to pass the line and column locations when

   109

 * creating a new object for the abstract syntax tree.

   110

 */

   111

#define locloc(foo) foo.first_line, foo.first_column, foo.last_line, foo.last_column

   112

#define   locf(foo) foo.first_line, foo.first_column

   113

#define   locl(foo) foo.last_line,  foo.last_column

   114

   115

   116

   117

/* A macro for printing out internal parser errors... */

   118

#define ERROR error_exit(__FILE__,__LINE__)

   119

/* function defined in main.cc */

   120

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

   121

   122

   123

   124

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

   125

/* global variables...   */

   126

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

   127

/* NOTE: For some strange reason bison ver 2.3 is including these declarations

   128

 *       in the iec.y.hh file, which is in turn included by flex.

   129

 *       We cannot therefore define any variables over here, but merely declare 

   130

 *       their existance (otherwise we get errors when linking the code, since we

   131

 *       would get a new variable defined each time iec.y.hh is included!).

   132

 *       Even though the variables are declared 'extern' over here, they will in

   133

 *       fact be defined towards the end of this same file (i.e. in the prologue)

   134

 */

   135

   136

/* A global flag used to tell the parser if overloaded funtions should be allowed.

   137

 * The IEC 61131-3 standard allows overloaded funtions in the standard library,

   138

 * but disallows them in user code...

   139

 */

   140

extern bool allow_function_overloading;

   141

   142

/* A pointer to the root of the parsing tree that will be generated 

   143

 * by bison.

   144

 */

   145

extern symbol_c *tree_root;

   146

   147

   148

   149

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

   150

/* forward declarations */

   151

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

   152

/* The functions declared here are defined at the end of this file... */

   153

   154

/* Convert an il_operator_c into an identifier_c */

   155

symbol_c *il_operator_c_2_identifier_c(symbol_c *il_operator);

   156

   157

/* return if current token is a syntax element */

   158

bool is_current_syntax_token();

   159

   160

/* print an error message */

   161

void print_err_msg(int first_line,

   162

                   int first_column,

   163

                   int last_line,

   164

                   int last_column,

   165

                   const char *additional_error_msg);

   166

%}

   167

   168

   169

   170

   171

// %glr-parser

   172

// %expect-rr 1

   173

   174

   175

   176

   177

%union {

   178

    symbol_c 	*leaf;

   179

    list_c	*list;

   180

    char 	*ID;	/* token value */

   181

}

   182

   183

/*

   184

 TODO: DO we need to define a destructor do free

   185

       memory when recovering from errors, or do the

   186

       class destructors already handle this?

   187

       Following is example on how to define

   188

       detructors, using the syntax:

   189

       %destructor { CODE } SYMBOLS

   190

%union

   191

      {

   192

        char *string;

   193

      }

   194

      %token <string> STRING

   195

      %type  <string> string

   196

      %destructor { free ($$); } STRING string

   197

*/

   198

   199

   200

   201

   202

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

   203

/* Prelimenary constructs... */

   204

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

   205

/* A token used to identify the very end of the input file

   206

 * after all includes have already been processed.

   207

 *

   208

 * Flex automatically returns the token with value 0

   209

 * at the end of the file. We therefore specify here

   210

 * a token with that exact same value here, so we can use it

   211

 * to detect the very end of the input files.

   212

 */

   213

%token END_OF_INPUT 0

   214

   215

/* A bogus token that, in principle, flex MUST NEVER generate */

   216

/* USE 1:

   217

 * ======

   218

 * This token is currently also being used as the default

   219

 * initialisation value of the token_id member in

   220

 * the symbol_c base class.

   221

 *

   222

 * USE 2

   223

 * =====

   224

 * This token may also be used in the future to remove

   225

 * mysterious reduce/reduce conflicts due to the fact

   226

 * that our grammar may not be LALR(1) but merely LR(1).

   227

 * This means that bison cannot handle it without some

   228

 * caoxing from ourselves. We will then need this token

   229

 * to do the coaxing...

   230

 */

   231

%token BOGUS_TOKEN_ID

   232

   233

%type <leaf>	start

   234

   235

%type <leaf>	any_identifier

   236

   237

%token <ID>	prev_declared_variable_name_token

   238

%token <ID>	prev_declared_fb_name_token

   239

%type <leaf>	prev_declared_variable_name

   240

%type <leaf>	prev_declared_fb_name

   241

   242

%token  <ID>	prev_declared_simple_type_name_token

   243

%token  <ID>	prev_declared_subrange_type_name_token

   244

%token  <ID>	prev_declared_enumerated_type_name_token

   245

%token  <ID>	prev_declared_array_type_name_token

   246

%token  <ID>	prev_declared_structure_type_name_token

   247

%token  <ID>	prev_declared_string_type_name_token

   248

   249

%type  <leaf>	prev_declared_simple_type_name

   250

%type  <leaf>	prev_declared_subrange_type_name

   251

%type  <leaf>	prev_declared_enumerated_type_name

   252

%type  <leaf>	prev_declared_array_type_name

   253

%type  <leaf>	prev_declared_structure_type_name

   254

%type  <leaf>	prev_declared_string_type_name

   255

   256

%token <ID>	prev_declared_derived_function_name_token

   257

%token <ID>	prev_declared_derived_function_block_name_token

   258

%token <ID>	prev_declared_program_type_name_token

   259

%type  <leaf>	prev_declared_derived_function_name

   260

%type  <leaf>	prev_declared_derived_function_block_name

   261

%type  <leaf>	prev_declared_program_type_name

   262

   263

   264

   265

   266

/* The pragmas... */

   267

%token <ID>	pragma_token

   268

%type <leaf>	pragma

   269

   270

   271

/* Where do these tokens belong ?? */

   272

/* TODO: get the syntax parser to handle these tokens... */

   273

%token	EN

   274

%token	ENO

   277

   278

   279

   280

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

   281

/* B 0 - Programming Model */

   282

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

   283

%type <list>	library

   284

%type <leaf>	library_element_declaration

   285

   286

   287

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

   288

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

   289

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

   290

/* Done totally within flex...

   291

  letter

   292

  digit

   293

  octal_digit

   294

  hex_digit

   295

*/

   296

%token <ID>	identifier_token

   297

%type  <leaf>	identifier

   298

%token INVALID_IDENTIFIER

   299

   300

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

   301

/* B 1.2 - Constants */

   302

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

   303

%type <leaf>	constant

   304

/* a helper symbol for expression */

   305

%type  <leaf> non_negative_constant

   306

   307

   308

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

   309

/* B 1.2.1 - Numeric Literals */

   310

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

   311

/* Done totally within flex...

   312

  bit

   313

*/

   314

%type  <leaf> numeric_literal

   315

/* helper symbol for non_negative_constant */

   316

%type  <leaf> non_negative_numeric_literal

   317

%type  <leaf> integer_literal

   318

%type  <leaf> signed_integer

   319

/* a helper symbol for non_negative_constant */

   320

%type  <leaf> non_negative_signed_integer

   321

%token <ID>   integer_token

   322

%type  <leaf> integer

   323

%token <ID>   binary_integer_token

   324

%type  <leaf> binary_integer

   325

%token <ID>   octal_integer_token

   326

%type  <leaf> octal_integer

   327

%token <ID>   hex_integer_token

   328

%type  <leaf> hex_integer

   329

%token <ID>   real_token

   330

%type  <leaf> real

   331

%type  <leaf> signed_real

   332

/* helper symbol for non_negative_real_literal */

   333

%type  <leaf> non_negative_signed_real

   334

%type  <leaf> real_literal

   335

/* helper symbol for non_negative_numeric_literal */

   336

%type  <leaf> non_negative_real_literal

   337

// %type  <leaf> exponent

   338

%type  <leaf> bit_string_literal

   339

%type  <leaf> boolean_literal

   340

   341

%token FALSE

   342

%token TRUE

   343

   344

   345

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

   346

/* B 1.2.2 - Character Strings */

   347

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

   348

%token <ID>   single_byte_character_string_token

   349

%token <ID>   double_byte_character_string_token

   350

   351

%type  <leaf> character_string

   352

%type  <leaf> single_byte_character_string

   353

%type  <leaf> double_byte_character_string

   354

   355

   356

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

   357

/* B 1.2.3 - Time Literals */

   358

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

   359

%type  <leaf> time_literal

   360

   361

   362

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

   363

/* B 1.2.3.1 - Duration */

   364

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

   365

%type  <leaf>	duration

   366

%type  <leaf>	interval

   367

%type  <leaf>	days

   368

%type  <leaf>	fixed_point

   369

%type  <leaf>	hours

   370

%type  <leaf>	minutes

   371

%type  <leaf>	seconds

   372

%type  <leaf>	milliseconds

   373

   374

%type  <leaf>	integer_d

   375

%type  <leaf>	integer_h

   376

%type  <leaf>	integer_m

   377

%type  <leaf>	integer_s

   378

%type  <leaf>	integer_ms

   379

%type  <leaf>	fixed_point_d

   380

%type  <leaf>	fixed_point_h

   381

%type  <leaf>	fixed_point_m

   382

%type  <leaf>	fixed_point_s

   383

%type  <leaf>	fixed_point_ms

   384

   385

%token <ID>	fixed_point_token

   386

%token <ID>	fixed_point_d_token

   387

%token <ID>	integer_d_token

   388

%token <ID>	fixed_point_h_token

   389

%token <ID>	integer_h_token

   390

%token <ID>	fixed_point_m_token

   391

%token <ID>	integer_m_token

   392

%token <ID>	fixed_point_s_token

   393

%token <ID>	integer_s_token

   394

%token <ID>	fixed_point_ms_token

   395

%token <ID>	integer_ms_token

   396

   397

%token TIME

   398

%token T_SHARP

   399

   400

   401

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

   402

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

   403

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

   404

%type  <leaf>	time_of_day

   405

%type  <leaf>	daytime

   406

%type  <leaf>	day_hour

   407

%type  <leaf>	day_minute

   408

%type  <leaf>	day_second

   409

%type  <leaf>	date

   410

%type  <leaf>	date_literal

   411

%type  <leaf>	year

   412

%type  <leaf>	month

   413

%type  <leaf>	day

   414

%type  <leaf>	date_and_time

   415

   416

%token TIME_OF_DAY

   417

%token DATE

   418

%token D_SHARP

   419

%token DATE_AND_TIME

   420

   421

   422

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

   423

/* B 1.3 - Data Types */

   424

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

   425

/* Strangely, the following symbol does seem to be required! */

   426

// %type  <leaf> data_type_name

   427

%type  <leaf> non_generic_type_name

   428

   429

   430

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

   431

/* B 1.3.1 - Elementary Data Types */

   432

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

   433

/* NOTES:

   434

 *

   435

 *    - To make the definition of bit_string_literal more

   436

 *      concise, it is useful to use an extra non-terminal

   437

 *      symbol (i.e. a grouping or construct) that groups the

   438

 *      following elements (BYTE, WORD, DWORD, LWORD).

   439

 *      Note that the definition of bit_string_type_name

   440

 *      (according to the spec) includes the above elements

   441

 *      and an extra BOOL.

   442

 *      We could use an extra construct with the first four

   443

 *      elements to be used solely in the definition of

   444

 *      bit_string_literal, but with the objective of not

   445

 *      having to replicate the actions (if we ever need

   446

 *      to change them, they would need to be changed in both

   447

 *      bit_string_type_name and the extra grouping), we

   448

 *      have re-defined bit_string_type_name as only including

   449

 *      the first four elements.

   450

 *      In order to have our parser implement the specification

   451

 *      correctly we have augmented every occurence of

   452

 *      bit_string_type_name in other rules with the BOOL

   453

 *      token. Since bit_string_type_name only appears in

   454

 *      the rule for elementary_type_name, this does not

   455

 *      seem to be a big concession to make!

   456

 *

   457

 *    - We have added a helper symbol to concentrate the

   458

 *      instantiation of STRING and WSTRING into a single

   459

 *      location (elementary_string_type_name).

   460

 *      These two elements show up in several other rules,

   461

 *      but we want to create the equivalent abstract syntax

   462

 *      in a single location of this file, in order to make

   463

 *      possible future changes easier to edit...

   464

 */

   465

%type  <leaf>	elementary_type_name

   466

%type  <leaf>	numeric_type_name

   467

%type  <leaf>	integer_type_name

   468

%type  <leaf>	signed_integer_type_name

   469

%type  <leaf>	unsigned_integer_type_name

   470

%type  <leaf>	real_type_name

   471

%type  <leaf>	date_type_name

   472

%type  <leaf>	bit_string_type_name

   473

/* helper symbol to concentrate the instantiation

   474

 * of STRING and WSTRING into a single location

   475

 */

   476

%type  <leaf>	elementary_string_type_name

   477

   478

%token BYTE

   479

%token WORD

   480

%token DWORD

   481

%token LWORD

   482

   483

%token LREAL

   484

%token REAL

   485

   486

%token SINT

   487

%token INT

   488

%token DINT

   489

%token LINT

   490

   491

%token USINT

   492

%token UINT

   493

%token UDINT

   494

%token ULINT

   495

   496

%token WSTRING

   497

%token STRING

   498

%token BOOL

   499

   500

// %token TIME

   501

// %token DATE

   502

// %token DATE_AND_TIME

   503

%token DT

   504

// %token TIME_OF_DAY

   505

%token TOD

   506

   507

   508

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

   509

/* B 1.3.2 - Generic data types */

   510

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

   511

/* Strangely, the following symbol does seem to be required! */

   512

// %type  <leaf>	generic_type_name

   513

   514

/* The following tokens do not seem to be used either

   515

 * but we declare them so they become reserved words...

   516

 */

   517

%token ANY

   518

%token ANY_DERIVED

   519

%token ANY_ELEMENTARY