my language to parse contains statements like
public var a, b = 42, c;
I.e. the .g file looks something like:
statements
: (introduction | expression ';'! | ... )+
;
introduction
: head single+ -> ^(head single)+
;
single
: Name ('='^ expression)?
;
head
: modifiers* v='var' -> ^(VARIABLE[$v] modifiers*)
;
Generating a tree like that would be easy, but mostly useless (for me):
----------statements----------
/ | \
variable variable variable
/ \ / \ / \
'public' 'a' 'public' '=' 'public' 'c'
/ \
'b' expr
I would like to have the the '=' on top of the middle node:
----------statements----------
/ | \
variable '=' variable
/ \ / \ / \
'public' 'a' variable expr 'public' 'c'
/ \
'public' 'b'
but I can't find the rewrite rule to do that.
That is not easly done with the way you've set up your rules.
Here's a way it is possible:
grammar T;
options {
output=AST;
ASTLabelType=CommonTree;
}
tokens {
STATEMENTS;
VARIABLE;
DEFAULT_MODIFIER;
}
declaration
: modifier 'var' name[$modifier.tree] (',' name[$modifier.tree])* ';' -> ^(STATEMENTS name+)
;
modifier
: 'public'
| 'private'
| /* nothing */ -> DEFAULT_MODIFIER
;
name [CommonTree mod]
: ID '=' expression -> ^('=' ^(VARIABLE {new CommonTree(mod)} ID) expression)
| ID -> ^(VARIABLE {new CommonTree(mod)} ID)
;
// other parser & lexer rules
which produces the following AST:
for the input:
public var a, b = 42, c;
And produces:
for the input:
var a, b = 42, c;
Related
I wrote a PHP5 parser in ANTLR 3.4, which is almost ready, but I can not handle one of the tricky feature of PHP. My problem is with the precedence of assignment operator. As the PHP manual says the precedence of assignment is almost at the end of the list. Only and, xor, or and , are after it in the list.
But there is a note on this the manual page which says:
Although = has a lower precedence than most other operators, PHP will
still allow expressions similar to the following: if (!$a = foo()), in
which case the return value of foo() is put into $a.
The small example in the note isn't a problem for my parser, I can handle this as a special case in the assigment rule.
But there are more complex codes eg:
if ($a && $b = func()) {}
My parser fails here, because it recognizes first $a && $b and can not deal with the rest of the conditioin. This is because the && has higher precedence, than =.
If I put brackets around the right side of &&:
if ($a && ($b = func())) {}
In this way the parser recognizes the structure well.
The operators are built in the way that the ANTLR book recommends: there are the base exressions at the first step and each level of operators are coming after each other.
Is there any way to handle this precedence jumping?
Don't look at it as an assignment, but let's name it an assignment expression. Put this assignment expression "below" the unary expressions (so they have a higher precedence than the unary ones):
grammar T;
options {
output=AST;
}
tokens {
BLOCK;
FUNC_CALL;
EXPR_LIST;
}
parse
: stat* EOF!
;
stat
: assignment ';'!
| if_stat
;
assignment
: Var '='^ expr
;
if_stat
: If '(' expr ')' block -> ^(If expr block)
;
block
: '{' stat* '}' -> ^(BLOCK stat*)
;
expr
: or_expr
;
or_expr
: and_expr ('||'^ and_expr)*
;
and_expr
: unary_expr ('&&'^ unary_expr)*
;
unary_expr
: '!'^ assign_expr
| '-'^ assign_expr
| assign_expr
;
assign_expr
: Var ('='^ atom)*
| atom
;
atom
: Num
| func_call
;
func_call
: Id '(' expr_list ')' -> ^(FUNC_CALL Id expr_list)
;
expr_list
: (expr (',' expr)*)? -> ^(EXPR_LIST expr*)
;
If : 'if';
Num : '0'..'9'+;
Var : '$' Id;
Id : ('a'..'z')+;
Space : (' ' | '\t' | '\r' | '\n')+ {skip();};
If you'd now parse the source:
if (!$a = foo()) { $a = 1 && 2; }
if ($a && $b = func()) { $b = 2 && 3; }
if ($a = baz() && $b) { $c = 3 && 4; }
the following AST would get constructed:
I'm trying to parse a language using ANTLR which can contain the following syntax:
someVariable, somVariable.someMember, functionCall(param).someMember, foo.bar.baz(bjork).buffalo().xyzzy
This is the ANTLR grammar which i've come up with so far, and the access_operation throws the error
The following sets of rules are mutually left-recursive [access_operation, expression]:
grammar Test;
options {
output=AST;
ASTLabelType=CommonTree;
}
tokens {
LHS;
RHS;
CALL;
PARAMS;
}
start
: body? EOF
;
body
: expression (',' expression)*
;
expression
: function -> ^(CALL)
| access_operation
| atom
;
access_operation
: (expression -> ^(LHS)) '.'! (expression -> ^(RHS))
;
function
: (IDENT '(' body? ')') -> ^(IDENT PARAMS?)
;
atom
: IDENT
| NUMBER
;
fragment LETTER : ('a'..'z' | 'A'..'Z');
fragment DIGIT : '0'..'9';
IDENT : (LETTER)+ ;
NUMBER : (DIGIT)+ ;
SPACE : (' ' | '\t' | '\r' | '\n') { $channel=HIDDEN; };
What i could manage so far was to refactor the access_operation rule to '.' expression which generates an AST where the access_operation node only contains the right side of the operation.
What i'm looking for instead is something like this:
How can the left-recursion problem solved in this case?
By "wrong AST" I'll make a semi educated guess that, for input like "foo.bar.baz", you get an AST where foo is the root with bar as a child who in its turn has baz as a child, which is a leaf in the AST. You may want to have this reversed. But I'd not go for such an AST if I were you: I'd keep the AST as flat as possible:
foo
/ | \
/ | \
bar baz ...
That way, evaluating is far easier: you simply look up foo, and then walk from left to right through its children.
A quick demo:
grammar Test;
options {
output=AST;
ASTLabelType=CommonTree;
}
tokens {
BODY;
ACCESS;
CALL;
PARAMS;
}
start
: body EOF -> body
;
body
: expression (',' expression)* -> ^(BODY expression+)
;
expression
: atom
;
atom
: NUMBER
| (IDENT -> IDENT) ( tail -> ^(IDENT tail)
| call tail? -> ^(CALL IDENT call tail?)
)?
;
tail
: (access)+
;
access
: ('.' IDENT -> ^(ACCESS IDENT)) (call -> ^(CALL IDENT call))?
;
call
: '(' (expression (',' expression)*)? ')' -> ^(PARAMS expression*)
;
IDENT : LETTER+;
NUMBER : DIGIT+;
SPACE : (' ' | '\t' | '\r' | '\n') {$channel=HIDDEN;};
fragment LETTER : ('a'..'z' | 'A'..'Z');
fragment DIGIT : '0'..'9';
which can be tested with:
import org.antlr.runtime.*;
import org.antlr.runtime.tree.*;
import org.antlr.stringtemplate.*;
public class Main {
public static void main(String[] args) throws Exception {
String src = "someVariable, somVariable.someMember, functionCall(param).someMember, " +
"foo.bar.baz(bjork).buffalo().xyzzy";
TestLexer lexer = new TestLexer(new ANTLRStringStream(src));
TestParser parser = new TestParser(new CommonTokenStream(lexer));
CommonTree tree = (CommonTree)parser.start().getTree();
DOTTreeGenerator gen = new DOTTreeGenerator();
StringTemplate st = gen.toDOT(tree);
System.out.println(st);
}
}
The output of Main corresponds to the following AST:
EDIT
And since you indicated your ultimate goal is not evaluating the input, but that you rather need to conform the structure of the AST to some 3rd party API, here's a grammar that will create an AST like you indicated in your edited question:
grammar Test;
options {
output=AST;
ASTLabelType=CommonTree;
}
tokens {
BODY;
ACCESS_OP;
CALL;
PARAMS;
LHS;
RHS;
}
start
: body EOF -> body
;
body
: expression (',' expression)* -> ^(BODY expression+)
;
expression
: atom
;
atom
: NUMBER
| (ID -> ID) ( ('(' params ')' -> ^(CALL ID params))
('.' expression -> ^(ACCESS_OP ^(LHS ^(CALL ID params)) ^(RHS expression)))?
| '.' expression -> ^(ACCESS_OP ^(LHS ID) ^(RHS expression))
)?
;
params
: (expression (',' expression)*)? -> ^(PARAMS expression*)
;
ID : LETTER+;
NUMBER : DIGIT+;
SPACE : (' ' | '\t' | '\r' | '\n') {$channel=HIDDEN;};
fragment LETTER : ('a'..'z' | 'A'..'Z');
fragment DIGIT : '0'..'9';
which creates the following AST if you run the Main class:
The atom rule may be a bit daunting, but you can't shorten it much since the left ID needs to be available to most of the alternatives. ANTLRWorks helps in visualizing the alternative paths this rule may take:
which means atom can be any of the 5 following alternatives (with their corresponding AST's):
+----------------------+--------------------------------------------------------+
| alternative | generated AST |
+----------------------+--------------------------------------------------------+
| NUMBER | NUMBER |
| ID | ID |
| ID params | ^(CALL ID params) |
| ID params expression | ^(ACCESS_OP ^(LHS ^(CALL ID params)) ^(RHS expression))|
| ID expression | ^(ACCESS_OP ^(LHS ID) ^(RHS expression) |
+----------------------+--------------------------------------------------------+
I wrote a grammar for a language and now I want to treat some syntactic sugar constructions, for that I was thinking of writing a template translator.
The problem is I want my template grammar to translate only some constructions of the language and leave the rest as it is.
For example:
I have this as input:
class Main {
int a[10];
}
and I want to translate that into something like:
class Main {
Array a = new Array(10);
}
Ideally I would like to do some think like this in ANTLR
grammer Translator
options { output=template;}
decl
: TYPE ID '[' INT ']' -> template(name = {$ID.text}, size ={$INT.text})
"Array <name> = new Array(<size>);
I would like it to leave the rest of the input that doesn't match rule decl as it is.
How can I achieve this in ANTLR without writing the full grammar for the language ?
I would simply handle such things in the parser grammar.
Assuming you're constructing an AST in your parser grammar, I guess you'll have a rule to parse input like Array a = new Array(10); similar to:
decl
: TYPE ID '=' expr ';' -> ^(DECL TYPE ID expr)
;
where expr eventually matches a term like this:
term
: NUMBER
| 'new' ID '(' (expr (',' expr)*)? ')' -> ^('new' ID expr*)
| ...
;
To account for your short-hand declaration int a[10];, all you have to do is expand decl like this:
decl
: TYPE ID '=' expr ';' -> ^(DECL TYPE ID expr)
| TYPE ID '[' expr ']' ';' -> ^(DECL 'Array' ID ^(NEW ARRAY expr))
;
which will rewrite the input int a[10]; into the following AST:
which is exactly the same as the AST created for input Array a = new Array(10);.
EDIT
Here's a small working demo:
grammar T;
options {
output=AST;
}
tokens {
ROOT;
DECL;
NEW='new';
INT='int';
ARRAY='Array';
}
parse
: decl+ EOF -> ^(ROOT decl+)
;
decl
: type ID '=' expr ';' -> ^(DECL type ID expr)
| type ID '[' expr ']' ';' -> ^(DECL ARRAY ID ^(NEW ARRAY expr))
;
expr
: Number
| NEW type '(' (expr (',' expr)*)? ')' -> ^(NEW ID expr*)
;
type
: INT
| ARRAY
| ID
;
ID : ('a'..'z' | 'A'..'Z')+;
Number : '0'..'9'+;
Space : (' ' | '\t' | '\r' | '\n') {skip();};
which can be tested with the class:
import org.antlr.runtime.*;
import org.antlr.runtime.tree.*;
import org.antlr.stringtemplate.*;
public class Main {
public static void main(String[] args) throws Exception {
String src = "Array a = new Array(10); int a[10];";
TLexer lexer = new TLexer(new ANTLRStringStream(src));
TParser parser = new TParser(new CommonTokenStream(lexer));
CommonTree tree = (CommonTree)parser.parse().getTree();
DOTTreeGenerator gen = new DOTTreeGenerator();
StringTemplate st = gen.toDOT(tree);
System.out.println(st);
}
}
I am trying to create a grammar which accepts any character or number or just about anything, provided its length is equal to 1.
Is there a function to check the length?
EDIT
Let me make my question more clear with an example.
I wrote the following code:
grammar first;
tokens {
SET = 'set';
VAL = 'val';
UND = 'und';
CON = 'con';
ON = 'on';
OFF = 'off';
}
#parser::members {
private boolean inbounds(Token t, int min, int max) {
int n = Integer.parseInt(t.getText());
return n >= min && n <= max;
}
}
parse : SET expr;
expr : VAL('u'('e')?)? String |
UND('e'('r'('l'('i'('n'('e')?)?)?)?)?)? (ON | OFF) |
CON('n'('e'('c'('t')?)?)?)? oneChar
;
CHAR : 'a'..'z';
DIGIT : '0'..'9';
String : (CHAR | DIGIT)+;
dot : .;
oneChar : dot { $dot.text.length() == 1;} ;
Space : (' ' | '\t' | '\r' | '\n') {$channel=HIDDEN;};
I want my grammar to do the following things:
Accept commands like: 'set value abc' , 'set underli on' , 'set conn #'. The grammar should be intelligent enough to accept incomplete words like 'underl' instead of 'underline. etc etc.
The third syntax: 'set connect oneChar' should accept any character, but just one character. It can be a numeric digit or alphabet or any special character. I am getting a compiler error in the generated parser file because of this.
The first syntax: 'set value' should accept all the possible strings, even on and off. But when I give something like: 'set value offer', the grammar is failing. I think this is happening because I already have a token 'OFF'.
In my grammar all the three requirements I have listed above are not working fine. Don't know why.
There are some mistakes and/or bad practices in your grammar:
#1
The following is not a validating predicate:
{$dot.text.length() == 1;}
A proper validating predicate in ANTLR has a question mark at the end, and the inner code has no semi colon at the end. So it should be:
{$dot.text.length() == 1}?
instead.
#2
You should not be handling these alternative commands:
expr
: VAL('u'('e')?)? String
| UND('e'('r'('l'('i'('n'('e')?)?)?)?)?)? (ON | OFF)
| CON('n'('e'('c'('t')?)?)?)? oneChar
;
in a parser rule. You should let the lexer handle this instead. Something like this will do it:
expr
: VAL String
| UND (ON | OFF)
| CON oneChar
;
// ...
VAL : 'val' ('u' ('e')?)?;
UND : 'und' ( 'e' ( 'r' ( 'l' ( 'i' ( 'n' ( 'e' )?)?)?)?)?)?;
CON : 'con' ( 'n' ( 'e' ( 'c' ( 't' )?)?)?)?;
(also see #5!)
#3
Your lexer rules:
CHAR : 'a'..'z';
DIGIT : '0'..'9';
String : (CHAR | DIGIT)+;
are making things complicated for you. The lexer can produce three different kind of tokens because of this: CHAR, DIGIT or String. Ideally, you should only create String tokens since a String can already be a single CHAR or DIGIT. You can do that by adding the fragment keyword before these rules:
fragment CHAR : 'a'..'z' | 'A'..'Z';
fragment DIGIT : '0'..'9';
String : (CHAR | DIGIT)+;
There will now be no CHAR and DIGIT tokens in your token stream, only String tokens. In short: fragment rules are only used inside lexer rules, by other lexer rules. They will never be tokens of their own (and can therefor never appear in any parser rule!).
#4
The rule:
dot : .;
does not do what you think it does. It matches "any token", not "any character". Inside a lexer rule, the . matches any character but in parser rules, it matches any token. Realize that parser rules can only make use of the tokens created by the lexer.
The input source is first tokenized based on the lexer-rules. After that has been done, the parser (though its parser rules) can then operate on these tokens (not characters!!!). Make sure you understand this! (if not, ask for clarification or grab a book about ANTLR)
- an example -
Take the following grammar:
p : . ;
A : 'a' | 'A';
B : 'b' | 'B';
The parser rule p will now match any token that the lexer produces: which is only a A- or B-token. So, p can only match one of the characters 'a', 'A', 'b' or 'B', nothing else.
And in the following grammar:
prs : . ;
FOO : 'a';
BAR : . ;
the lexer rule BAR matches any single character in the range \u0000 .. \uFFFF, but it can never match the character 'a' since the lexer rule FOO is defined before the BAR rule and captures this 'a' already. And the parser rule prs again matches any token, which is either FOO or BAR.
#5
Putting single characters like 'u' inside your parser rules, will cause the lexer to tokenize an u as a separate token: you don't want that. Also, by putting them in parser rules, it is unclear which token has precedence over other tokens. You should keep all such literals outside your parser rules and make them explicit lexer rules instead. Only use lexer rules in your parser rules.
So, don't do:
pRule : 'u' ':' String
String : ...
but do:
pRule : U ':' String
U : 'u';
String : ...
You could make ':' a lexer rule, but that is of less importance. The 'u' however can also be a String so it must appear as a lexer rule before the String rule.
Okay, those were the most obvious things that come to mind. Based on them, here's a proposed grammar:
grammar first;
parse
: (SET expr {System.out.println("expr = " + $expr.text);} )+ EOF
;
expr
: VAL String {System.out.print("A :: ");}
| UL (ON | OFF) {System.out.print("B :: ");}
| CON oneChar {System.out.print("C :: ");}
;
oneChar
: String {$String.text.length() == 1}?
;
SET : 'set';
VAL : 'val' ('u' ('e')?)?;
UL : 'und' ( 'e' ( 'r' ( 'l' ( 'i' ( 'n' ( 'e' )?)?)?)?)?)?;
CON : 'con' ( 'n' ( 'e' ( 'c' ( 't' )?)?)?)?;
ON : 'on';
OFF : 'off';
String : (CHAR | DIGIT)+;
fragment CHAR : 'a'..'z' | 'A'..'Z';
fragment DIGIT : '0'..'9';
Space : (' ' | '\t' | '\r' | '\n') {$channel=HIDDEN;};
that can be tested with the following class:
import org.antlr.runtime.*;
public class Main {
public static void main(String[] args) throws Exception {
String source =
"set value abc \n" +
"set underli on \n" +
"set conn x \n" +
"set conn xy ";
ANTLRStringStream in = new ANTLRStringStream(source);
firstLexer lexer = new firstLexer(in);
CommonTokenStream tokens = new CommonTokenStream(lexer);
firstParser parser = new firstParser(tokens);
System.out.println("parsing:\n======\n" + source + "\n======");
parser.parse();
}
}
which, after generating the lexer and parser:
java -cp antlr-3.2.jar org.antlr.Tool first.g
javac -cp antlr-3.2.jar *.java
java -cp .:antlr-3.2.jar Main
prints the following output:
parsing:
======
set value abc
set underli on
set conn x
set conn xy
======
A :: expr = value abc
B :: expr = underli on
C :: expr = conn x
line 0:-1 rule oneChar failed predicate: {$String.text.length() == 1}?
C :: expr = conn xy
As you can see, the last command, C :: expr = conn xy, produces an error, as expected.
I have created the following grammar: I would like some idea how to build an interpreter that returns a tree in java, which I can later use for printing in the screen, Im bit stack on how to start on it.
grammar myDSL;
options {
language = Java;
}
#header {
package DSL;
}
#lexer::header {
package DSL;
}
program
: IDENT '={' components* '}'
;
components
: IDENT '=('(shape)(shape|connectors)* ')'
;
shape
: 'Box' '(' (INTEGER ','?)* ')'
| 'Cylinder' '(' (INTEGER ','?)* ')'
| 'Sphere' '(' (INTEGER ','?)* ')'
;
connectors
: type '(' (INTEGER ','?)* ')'
;
type
: 'MG'
| 'EL'
;
IDENT: ('a'..'z' | 'A'..'Z')('a'..'z' | 'A'..'Z' | '0'..'0')*;
INTEGER: '0'..'9'+;
// This if for the empty spaces between tokens and avoids them in the parser
WS: (' ' | '\t' | '\n' | '\r' | '\f')+ {$channel=HIDDEN;};
COMMENT: '//' .* ('\n' | '\r') {$channel=HIDDEN;};
A couple of remarks:
There's no need to set the language for Java, which is the default target language. So you can remove this:
options {
language = Java;
}
Your IDENT contains an error:
IDENT: ('a'..'z' | 'A'..'Z')('a'..'z' | 'A'..'Z' | '0'..'0')*;
the '0'..'0') should most probably be '0'..'9').
The sub rule (INTEGER ','?)* also matches source like 1 2 3 4 (no comma's at all!). Perhaps you meant to do: (INTEGER (',' INTEGER)*)?
Now, as to your question: how to let ANTLR construct a proper AST? This can be done by adding output = AST; in your options block:
options {
//language = Java;
output = AST;
}
And then either adding the "tree operators" ^ and ! in your parser rules, or by using tree rewrite rules: rule: a b c -> ^(c b a).
The "tree operator" ^ is used to define the root of the (sub) tree and ! is used to exclude a token from the (sub) tree.
Rewrite rules have ^( /* tokens here */ ) where the first token (right after ^() is the root of the (sub) tree, and all following tokens are child nodes of the root.
An example might be in order. Let's take your first rule:
program
: IDENT '={' components* '}'
;
and you want to let IDENT be the root, components* the children and you want to exclude ={ and } from the tree. You can do that by doing:
program
: IDENT^ '={'! components* '}'!
;
or by doing:
program
: IDENT '={' components* '}' -> ^(IDENT components*)
;