DParser Manual


Contents
  1. Installation
  2. Getting Started
  3. Comments
  4. Productions
  5. Global Code
  6. Terminals
    1. Strings
    2. Unicode
    3. Regular Expressions
    4. Naming tokens
    5. External (C) scanners
    6. Tokenizers
    7. Longest Match
  7. Priorities and Associativity
    1. Token Priorities
    2. Operator Priorities
    3. Rule Priorities
  8. Actions
    1. Speculative Actions
    2. Final Actions
    3. Embedded
    4. Pass Actions
    5. Default Actions
  9. Attributes and Action Specifiers
    1. Global State
    2. Parse Nodes
    3. Misc
  10. Symbol Table
  11. Whitespace
  12. Ambiguities
  13. Error Recovery
  14. Parsing Options
  15. Grammar Grammar
1. Installation
   
To build: 'gmake'              (only available with source code package)

To test: 'gmake test'          (only available with source code package)

To install, 'gmake install'    (binary or source code packages)

2. Getting Started

2.1. Create your grammar, for example, in the file "my.g":
  
   E: E '+' E | "[abc]";
  
2.2. Convert grammar into parsing tables:

  % make_dparser my.g

2.3. Create a driver program, for example, in the file "my.c":  
          
#include <stdio.h>
#include <dparse.h>
extern D_ParserTables parser_tables_gram;
int
main(int argc, char *argv[]) {
  char s[256], *ss;
  D_Parser *p = new_D_Parser(&parser_tables_gram, 0);
  if (fgets(s,255,stdin) && dparse(p, s, strlen(s)) && !p->syntax_errors)
    printf("success\n");
  else
    printf("failure\n");
}

2.4. Compile:

  % cc -I/usr/local/include my.c my.g.d_parser.c -L/usr/local/lib -ldparse
 
2.5. Run:
  
  % a.out
  a=
  syntax error, '' line 1
  failure
  %
  
  % a.out
  a+b
  success
  %

 We'll come back to this example later.

3. Comments

  Grammars can include C/C++ style comments.

EXAMPLE

// My first grammar
   E: E '+' E | "[abc]";
/* is this right? */

4. Productions

  4.1. The first production is the root of your grammar (what you will be trying to parse).
  4.2. Productions start with the non-terminal being defined followed by a colon ':', a set of right hand sides separated by '|' (or) consisting of elements (non-terminals or terminals).
  4.3. Elements can be grouped with parens '(', and the normal regular expression symbols can be used ('+' '*' '?' '|').
  4.4. Elements can be repeated using '@', for example elem@3 or elem@1:3 for repeat 3 or between 1 and 3 times respectively.
EXAMPLE

program: statements+ |  comment* (function |  procedure)?;

  4.4. NOTE: Instead of using '[' ']' for optional elements we use the more familar and consistent '?' operator.  The square brackets are reserved for speculative actions (below).

5. Global Code

Global (or static) C code can be intermixed with productions by surrounding the code with brackets '{'.

EXAMPLE

{ void dr_s() { printf("Dr. S\n"); }
S: 'the' 'cat' 'and' 'the' 'hat' { dr_s(); } | T;
{ void twain() { printf("Mark Twain\n"); }
T: 'Huck' 'Finn' { twain(); };

6. Terminals

  6.1. Strings terminals are surrounded with single quotes.  For example:

block: '{' statements* '}';
whileblock: 'while' '(' expression ')' block;

  6.2. Unicode literals can appear in strings or as charaters with U+ or u+.  For example:

'φ' { printf("phi\n"); }
U+03c9 { printf("omega\n"); }

  6.3. Regular expressions are surrounded with double quotes.  For example:

hexint: "(0x|0X)[0-9a-fA-F]+[uUlL]?";

  NOTE: only the simple regular expression operators are currently supported (v1.3).  This include parens, square parens, ranges, and '*', '+', '?'.   If you need something more, request a feature or implement it yourself; the code is in scan.c.

  6.4 Terminal modifiers

  Terminals can contain embbed escape codes.  Including the standard C escape codes, the codes \x and \d permit inserting hex and decimal ASCII characters directly.

  Tokens can be given a name by appending the $name option.  This is useful when you have several
  tokens which which represent the same string (e.g. ',').   For example,
  
      function_call: function '(' parameter ( ',' $name 'parameter_comma' parameter) ')';

  It is now possible to use $0.symbol == ${string parameter_comma} to differentiate ParseNode ($0) between
a parameter comma node and say an initialization comma.

  Terminals ending in '/i' are case insensitive.    For example 'hi'/i matches 'HI', 'Hi' and "hI' in addition to 'hi'.

  6.3. External (C) Scanners

  There are two types of external scanners, those which read a single terminal, and those which are global (called for every terminal).  Here is an example of a scanner for a single terminal.  Notice how it can be mixed with regular string terminals.

{
extern char *ops;
extern void *ops_cache;
int ops_scan(char *ops, void *ops_cache, char **as,
    int *col, int *line, unsigned short *op_assoc, int *op_priority);
}

X: '1' (${scan ops_scan(ops, ops_cache)} '2')*;

  The user provides the 'ops_scan' function.  This example is from tests/g4.test.g in the source distribution.

  The second type of scanner is a global scanner:

{
#include "g7.test.g.d_parser.h"
int myscanner(char **s, int *col, int *line, unsigned short *symbol,
      int *term_priority, unsigned short *op_assoc, int *op_priority)
{
  if (**s == 'a') {
    (*s)++;
    *symbol = A;
    return 1;
  } else if (**s == 'b') {
    (*s)++;
    *symbol = BB;
    return 1;
  } else if (**s == 'c') {
    (*s)++;
    *symbol = CCC;
    return 1;
  } else if (**s == 'd') {
    (*s)++;
    *symbol = DDDD;
    return 1;
  } else
    return 0;
}
${scanner myscanner}
${token A BB CCC DDDD}

S: A (BB CCC)+ SS;
SS: DDDD;

  Notice how the you need to include the header file generated by make_dparser which contains the token definitions.

6.4. Tokenizers

  Tokenizers are non-context sensitive global scanners which produce only one token for any given input string.  Some programming languages (for example C) are easier to specify using a tokenizer because (for example) reserved words can be handled simply by lowering the terminal priority for identifiers.

EXAMPLE:

S : 'if' '(' S ')' S ';' | 'do' S 'while' '(' S ')' ';' | ident;
ident: "[a-z]+" $term -1;

  The sentence: if ( while ) a; is legal because while cannot appear at the start of S and so it doesn't conflict with the parsing of while as an ident in that position.  However, if a tokenizer is specified, all tokens will be possible at each position and the sentense will produce a syntax error.

  DParser provides two ways to specify tokenizers: globally as an option (-T) to make_dparser and locally with a ${declare tokenize ...} specifier (see the ANSI C grammar for an example).  The ${declare tokenize ...} declartion allows a tokenizer to be specified over a subset of the parsing states so that (for example) ANSI C could be a subgrammar of another larger grammar.  Currently the parse states are not split so that the productions for the substates must be disjoint.

6.5 Longest Match

  Longest match lexical ambiguity resolution is a technique used by separate phase lexers to help decide (along with lexical priorities) which single token to select for a given input string.  It is used in the definition of ANSI-C, but not in C++ because of a snafu in the definition of templates whereby templates of templates (List<List <Int>>) can end with the right shift token ('>>").  Since DParser does not have a separate lexical phase, it does not require longest match disambiguation, but provides it as an option.

  There are two ways to specify longest match disabiguation: globally as an option (-l) to make_dparser or locally with with a ${declare longest_match ...}.  If global longest match disambiguation is ON, it can be locally disabled with {$declare all_matches ...} .  As with Tokenizers above, local declarations operate on disjoint subsets of parsing states.

7. Priorities and Associativity

  Priorities can very from MININT to MAXINT and are specified as integers.  Associativity can take the values:

assoc : '$unary_op_right' | '$unary_op_left' | '$binary_op_right'
           | '$binary_op_left' | '$unary_right' | '$unary_left'
           | '$binary_right' | '$binary_left' | '$right' | '$left' ;

7.1. Token Prioritites

  Termininal priorities apply after the set of matching strings has been found and the terminal(s) with the highest priority is selected.
  Terminal priorities are introduced after a terminal by the specifier $term.  We saw an example of token priorities with the definition of ident.

EXAMPLE:

S : 'if' '(' S ')' S ';' | 'do' S 'while' '(' S ')' ';' | ident;
ident: "[a-z]+" $term -1;

7.2. Operator Priorities

  Operator priorities specify the priority of a operator symbol (either a terminal or a non-terminal).  This corresponds to the yacc or bison %left etc. declaration.  However, since DParser is doesn't require a global tokenizer, operator priorities and associativities are specified on the reduction which creates the token.  Moreover, the associativity includes the operator usage as well since it cannot be infered from rule context.  Possible operator associativies are:

operator_assoc : '$unary_op_right' | '$unary_op_left' | '$binary_op_right'
           | '$binary_op_left' | '$unary_right' | '$unary_left'
           | '$binary_right' | '$binary_left';

EXAMPLE:

E: ident op ident;
ident: '[a-z]+';
op: '*' $binary_op_left 2 |
      '+' $binary_op_left 1;

7.3. Rule Priorities

  Rule priorities specify the priority of the reduction itself and have the possible associativies:

rule_assoc: '$right' | '$left';

  Rule and operator priorities can be intermixed and are interpreted at run time (not when the tables are built).  This make it possible for user-defined scanners to return the associativities and priorities of tokens.

8. Actions

  Actions are the bits of code which run when a reduction occurs.

EXAMPLE

S: this | that;
this: 'this' { printf("got this\n"); };
that: 'that' { printf("got that\n"); };

8.1 Speculative Action

  Speculative actions occur when the reduction takes place during the speculative parsing process.  It is possible
that the reduction will not be part of the final parse or that it will occur a different number of times.  For example:

S: this | that;
this: hi 'mom';
that: ho 'dad';
ho: 'hello' [ printf("ho\n"); ];
hi: 'hello' [ printf("hi\n"); ];

Will print both 'hi' and 'ho' when given the input 'hello dad' because at the time hello is reduced, the following token is not known.

8.2 Final Actions

  Final actions occur only when the reduction must be part of any legal final parse (committed).  It is possible to do final actions during parsing or delay them till the entire parse tree is constructed (see Options).  Final actions are executed in order and in number according the the single final unambiguous parse.

S: A S 'b' | 'x';
A: [ printf("speculative e-reduce A\n"); ]
   { printf("final e-reduce A\n"); };

  On input:

xbbb

  Will produce:

speculative e-reduce A
final e-reduce A
final e-reduce A
final e-reduce A

8.3 Embedded Actions

  Actions can be embedded into rule. These actions are executed as if they were replaced with a synthetic production with a single null rule containing the actions.  For example:

S: A { printf("X"); } B;
A: 'a' { printf("a"); };
B: 'b' { printf("b"); };

  On input:

ab

  Will produce:

aXb

Note that in the above example, the print("X") is evaluated in a context null rule context while in:

S: A (A B { printf("X"); }) B;

The print is evalated in the context of the "A B" subrule because it appears at the end of the subrule and is therefor treated as a normal action for the subrule.

8.4 Pass Actions

  DParser supports multiple pass compilation.  The passes are declared at the top of the grammar, and the actions are associated with individual rules.

EXAMPLE

${pass sym for_all postorder}
${pass gen for_all postorder}

translation_unit: statement*;

statement
  : expression ';' {
    d_pass(${parser}, &$n, ${pass sym});
    d_pass(${parser}, &$n, ${pass gen});
  }
  ;

expression :  integer
  gen: { printf("gen integer\n"); }
  sym: { printf("sym integer\n"); }
  | expression '+' expression $right 2
  sym: { printf("sym +\n"); }
  ;

  A pass name then a colon indicate that the following action is associated with a particular pass. Passes can be either for_all or for_undefined (which means that the automatic traversal only applies to rules without actions defined for this pass).   Furthermore, passes can be postorder, preorder, and manual (you have to call d_pass yourself).  Passes can be initiated in the final action of any rule.

8.5 Default Actions

  The special production "_" can be defined with a single rule whose actions become the default when no other action is specified.  Default actions can be specified for speculative, final and pass actions and apply to each separately.

EXAMPLE

_: { printf("final action"); }
    gen: { printf("default gen action"); }
    sym: { printf("default sym action"); }
    ;

9. Attributes and Action Specifiers

9.1.  Global State ($g)

  Global state is declared by define'ing D_ParseNode_Globals (see the ANSI C grammar for a similar declaration for symbols). Global state can be accessed in any action with $g.  Because DParser handles ambiguous parsing global state can be accessed on different speculative parses.  In the future automatic splitting of global state may be implemented (if there is demand). Currently, the global state can be copied and assigned to $g to ensure that the changes made only effect subsequent speculative parses derived from the particular parse.

EXAMPLE

  [ $g = copy_globals($g);
    $g->my_variable = 1;
  ]

  The symbol table (Section 10) can be used to manage state information safely for different speculative parses.

9.2. Parse Node State

  Each parse node includes a set of system state variables and can have a set of user-defined state variables.  User defined parse node state is declared by define'ing D_ParseNodeUser. The size of the parse node state must be passed into new_D_Parser() to ensure that the appropriate amount of space is allocated for parse nodes.   Parse node state is accessed with:

 $# - number of child nodes
 $$ - user parse node state for parent node (non-terminal defined by the production)
 $X (where X is a number) - the user parse node state of element X of the production
 $n - the system parse node state of the rule node
 $nX - the system parse node state of element X of the production

  The system parse node state is defined in dparse.h which is installed with DParser.  It contains such information as the symbol, the location of the parsed string, and pointers to the start and end of the parsed string.

9.3. Misc

  ${scope} - the current symbol table scope
  ${reject} - in speculative actions permits the current parse to be rejected

10. Symbol Table

  The symbol table can be updated down different speculative paths while sharing the bulk of the data.  It defines the following functions in the file (dsymtab.h):
 
struct D_Scope *new_D_Scope(struct D_Scope *st);
struct D_Scope *enter_D_Scope(struct D_Scope *current, struct D_Scope *scope);
D_Sym *NEW_D_SYM(struct D_Scope *st, char *name, char *end);
D_Sym *find_D_Sym(struct D_Scope *st, char *name, char *end);
D_Sym *UPDATE_D_SYM(struct D_Scope *st, D_Sym *sym);
D_Sym *current_D_Sym(struct D_Scope *st, D_Sym *sym);
D_Sym *find_D_Sym_in_Scope(struct D_Scope *st, char *name, char *end);

'new_D_Scope' creates a new scope below 'st' or NULL for a 'top level' scope.  'enter_D_Scope' returns to a previous scoping level.  NOTE: do not simply assign ${scope} to a previous scope as any updated symbol information will be lost.  'commit_D_Scope' can be used in final actions to compress the update list for the top level scope and improve efficiency.

'find_D_Sym' finds the most current version of a symbol in a given scope.  'UPDATE_D_SYM' updates the value of symbol (creates a difference record on the current speculative parse path).  'current_D_Sym' is used to retrive the current version of a symbol, the pointer to which may have been stored in some other attribute or variable.  Symbols with the same name should not be created in the same scope.  The function 'find_D_Sym_in_Scope' is provided to detect this case.
 
  User data can be attached to symbols by define'ing D_UserSym.  See the ANSI C grammar for an example.

Here is a full example of scope usage (from tests/g29.test.g):

#include <stdio.h>

typedef struct My_Sym {
  int value;
} My_Sym;
#define D_UserSym My_Sym
typedef struct My_ParseNode {
  int value;
  struct D_Scope *scope;
} My_ParseNode;
#define D_ParseNode_User My_ParseNode
}

translation_unit: statement*;
 
statement
  : expression ';'
  { printf("%d\n", $0.value); }
  | '{' new_scope statement* '}'
  [ ${scope} = enter_D_Scope(${scope}, $n0.scope); ]
  { ${scope} = commit_D_Scope(${scope}); }
  ;

new_scope: [ ${scope} = new_D_Scope(${scope}); ];

expression
  : identifier ':' expression
  [
    D_Sym *s;
    if (find_D_Sym_in_Scope(${scope}, $n0.start_loc.s, $n0.end))
      printf("duplicate identifier line %d\n", $n0.start_loc.line);
    s = NEW_D_SYM(${scope}, $n0.start_loc.s, $n0.end);
    s->user.value = $2.value;
    $$.value = s->user.value;
  ]
  | identifier '=' expression
  [ D_Sym *s = find_D_Sym(${scope}, $n0.start_loc.s, $n0.end);
    s = UPDATE_D_SYM(${scope}, s);
    s->user.value = $2.value;
    $$.value = s->user.value;
  ]
  | integer
  [ $$.value = atoi($n0.start_loc.s); ]
  | identifier
  [ D_Sym *s = find_D_Sym(${scope}, $n0.start_loc.s, $n0.end);
    if (s)
      $$.value = s->user.value;
  ]
  | expression '+' expression
  [ $$.value = $0.value + $1.value; ]
  ;

integer: "-?([0-9]|0(x|X))[0-9]*(u|U|b|B|w|W|L|l)*" $term -1;
identifier: "[a-zA-Z_][a-zA-Z_0-9]*";
11. Whitespace

  Whitespace can be specified two ways: C function which can be user-defined, or as a subgrammar.  The default whitespace parser is compatible with C/C++ #line directives and comments.  It can be replaced with any user specified function as a parsing option (see Options).

  Additionally, if the (optionally) reserved production whitespace is defined, the subgrammar it defines will be used to consume whitespace for the main grammar.   This subgrammar can include normal actions.

EXAMPLE

S: 'a' 'b' 'c';
whitespace: "[ \t\n]*";

  Whitespace can be accessed on a per parse node basis using the functions: d_ws_before and d_ws_after, which return the start of the whitespace before start_loc.s and after end respectively.

12. Ambiguities
  Ambiguities are resolved automatically based on priorities and associativities.  In addition, when the other resolution techniques fail, user defined ambiguity resolution is possible.   The default ambiguity handler produces a fatal error on an unresolved ambiguity.  This behavior can be replaced with a user defined resolvers the signature of which is provided in dparse.h.

  If the verbose_level flag is set, the default ambiguity handler will print out parenthesized versions of the ambiguous parse trees.   This may be of some assistence in disambiguating a grammar.

13. Error Recovery

  DParser implements an error recovery scheme appropriate to scannerless parsers.  I haven't had time to investigate all the prior work in this area, so I am not sure if it is novel.  Suffice for now that it is optional and works well with C/C++ like grammars.

14. Parsing Options

  Parser are instantiated with the function new_D_Parser.  The resulting data structure contains a number of user configurable options (see dparser.h).  These are provided reasonable default values and include:
In addtion, there are the following user configurables:
An the following result values:
This final value should be checked to see if parse was successful.

15. Grammar Grammar

  DParser is fully self-hosted (would you trust a parser generator which wasn't?).   The grammar grammar is here (Grammar Grammar).