org.antlr.codegen
Class CodeGenerator

java.lang.Object
  extended by org.antlr.codegen.CodeGenerator

public class CodeGenerator
extends java.lang.Object

ANTLR's code generator. Generate recognizers derived from grammars. Language independence achieved through the use of StringTemplateGroup objects. All output strings are completely encapsulated in the group files such as Java.stg. Some computations are done that are unused by a particular language. This generator just computes and sets the values into the templates; the templates are free to use or not use the information. To make a new code generation target, define X.stg for language X by copying from existing Y.stg most closely releated to your language; e.g., to do CSharp.stg copy Java.stg. The template group file has a bunch of templates that are needed by the code generator. You can add a new target w/o even recompiling ANTLR itself. The language=X option in a grammar file dictates which templates get loaded/used. Some language like C need both parser files and header files. Java needs to have a separate file for the cyclic DFA as ANTLR generates bytecodes directly (which cannot be in the generated parser Java file). To facilitate this, cyclic can be in same file, but header, output must be searpate. recognizer is in outptufile.


Field Summary
 ACyclicDFACodeGenerator acyclicDFAGenerator
          I have factored out the generation of acyclic DFAs to separate class
protected  StringTemplateGroup baseTemplates
          The basic output templates without AST or templates stuff; this will be the templates loaded for the language such as Java.stg *and* the Dbg stuff if turned on.
 java.lang.String classpathTemplateRootDirectoryName
           
protected  boolean debug
          Generate debugging event method calls
static boolean EMIT_TEMPLATE_DELIMITERS
           
 boolean GENERATE_SWITCHES_WHEN_POSSIBLE
           
 Grammar grammar
          Which grammar are we generating code for? Each generator is attached to a specific grammar.
protected  StringTemplate headerFileST
           
protected  java.lang.String language
          What language are we generating?
protected  int lineWidth
           
static int MAX_ACYCLIC_DFA_STATES_INLINE
           
 int MAX_SWITCH_CASE_LABELS
          When generating SWITCH statements, some targets might need to limit the size (based upon the number of case labels).
 int MIN_SWITCH_ALTS
           
protected  StringTemplate outputFileST
           
protected  boolean profile
          Track runtime parsing information about decisions etc...
protected  StringTemplate recognizerST
           
 Target target
          The target specifies how to write out files and do other language specific actions.
protected  StringTemplateGroup templates
          Where are the templates this generator should use to generate code?
protected  Tool tool
          A reference to the ANTLR tool so we can learn about output directories and such.
protected  boolean trace
          Create a Tracer object and make the recognizer invoke this.
protected  int uniqueLabelNumber
          Used to create unique labels
static java.lang.String VOCAB_FILE_EXTENSION
          I have factored out the generation of cyclic DFAs to separate class
protected static java.lang.String vocabFilePattern
           
 
Constructor Summary
CodeGenerator(Tool tool, Grammar grammar, java.lang.String language)
           
 
Method Summary
protected  boolean canGenerateSwitch(DFAState s)
          You can generate a switch rather than if-then-else for a DFA state if there are no semantic predicates and the number of edge label values is small enough; e.g., don't generate a switch for a state containing an edge label such as 20..52330 (the resulting byte codes would overflow the method 65k limit probably).
 java.lang.String createUniqueLabel(java.lang.String name)
          Create a label to track a token / rule reference's result.
 void generateLocalFOLLOW(GrammarAST referencedElementNode, java.lang.String referencedElementName, java.lang.String enclosingRuleName, int elementIndex)
          Error recovery in ANTLR recognizers.
 StringTemplate generateSpecialState(DFAState s)
          A special state is huge (too big for state tables) or has a predicated edge.
protected  StringTemplate genLabelExpr(StringTemplateGroup templates, Transition edge, int k)
          Generate an expression for traversing an edge.
 StringTemplate genLookaheadDecision(StringTemplate recognizerST, DFA dfa)
          Generate code that computes the predicted alt given a DFA.
 StringTemplate genRecognizer()
          Given the grammar to which we are attached, walk the AST associated with that grammar to create NFAs.
protected  StringTemplate genSemanticPredicateExpr(StringTemplateGroup templates, Transition edge)
           
 StringTemplate genSetExpr(StringTemplateGroup templates, IntSet set, int k, boolean partOfDFA)
          For intervals such as [3..3, 30..35], generate an expression that tests the lookahead similar to LA(1)==3 || (LA(1)>=30&&LA(1)<=35)
protected  void genTokenTypeConstants(StringTemplate code)
          Set attributes tokens and literals attributes in the incoming code template.
protected  void genTokenTypeNames(StringTemplate code)
          Generate a token names table that maps token type to a printable name: either the label like INT or the literal like "begin".
protected  StringTemplate genTokenVocabOutput()
          Generate a token vocab file with all the token names/types.
 StringTemplateGroup getBaseTemplates()
           
static java.util.List<java.lang.String> getListOfArgumentsFromAction(java.lang.String actionText, int separatorChar)
           
static int getListOfArgumentsFromAction(java.lang.String actionText, int start, int targetChar, int separatorChar, java.util.List<java.lang.String> args)
          Given an arg action like [x, (*a).foo(21,33), 3.2+1, '\n', "a,oo\nick", {bl, "fdkj"eck}, ["cat\n,", x, 43]] convert to a list of arguments.
 java.lang.String getRecognizerFileName(java.lang.String name, int type)
          Generate TParser.java and TLexer.java from T.g if combined, else just use T.java as output regardless of type.
 StringTemplate getRecognizerST()
           
 StringTemplateGroup getTemplates()
           
 java.lang.String getTokenTypeAsTargetLabel(int ttype)
          Get a meaningful name for a token type useful during code generation.
 java.lang.String getVocabFileName()
          What is the name of the vocab file generated for this grammar? Returns null if no .tokens file should be generated.
 void issueInvalidAttributeError(java.lang.String x, Rule enclosingRule, antlr.Token actionToken, int outerAltNum)
           
 void issueInvalidAttributeError(java.lang.String x, java.lang.String y, Rule enclosingRule, antlr.Token actionToken, int outerAltNum)
           
 void issueInvalidScopeError(java.lang.String x, java.lang.String y, Rule enclosingRule, antlr.Token actionToken, int outerAltNum)
           
protected  void loadLanguageTarget(java.lang.String language)
           
 void loadTemplates(java.lang.String language)
          load the main language.stg template group file
 void setDebug(boolean debug)
           
 void setProfile(boolean profile)
           
 void setTrace(boolean trace)
           
 java.util.List translateAction(java.lang.String ruleName, GrammarAST actionTree)
           
protected  void translateActionAttributeReferences(java.util.Map actions)
          Actions may reference $x::y attributes, call translateAction on each action and replace that action in the Map.
protected  void translateActionAttributeReferencesForSingleScope(Rule r, java.util.Map scopeActions)
          Use for translating rule @init{...} actions that have no scope
 java.util.List<StringTemplate> translateArgAction(java.lang.String ruleName, GrammarAST actionTree)
          Translate an action like [3,"foo",a[3]] and return a List of the translated actions.
 StringTemplate translateTemplateConstructor(java.lang.String ruleName, int outerAltNum, antlr.Token actionToken, java.lang.String templateActionText)
          Given a template constructor action like %foo(a={...}) in an action, translate it to the appropriate template constructor from the templateLib.
protected  void verifyActionScopesOkForTarget(java.util.Map actions)
          Some targets will have some extra scopes like C++ may have '@headerfile:name {action}' or something.
 void write(StringTemplate code, java.lang.String fileName)
           
 
Methods inherited from class java.lang.Object
clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait
 

Field Detail

MAX_SWITCH_CASE_LABELS

public int MAX_SWITCH_CASE_LABELS
When generating SWITCH statements, some targets might need to limit the size (based upon the number of case labels). Generally, this limit will be hit only for lexers where wildcard in a UNICODE vocabulary environment would generate a SWITCH with 65000 labels.


MIN_SWITCH_ALTS

public int MIN_SWITCH_ALTS

GENERATE_SWITCHES_WHEN_POSSIBLE

public boolean GENERATE_SWITCHES_WHEN_POSSIBLE

EMIT_TEMPLATE_DELIMITERS

public static boolean EMIT_TEMPLATE_DELIMITERS

MAX_ACYCLIC_DFA_STATES_INLINE

public static int MAX_ACYCLIC_DFA_STATES_INLINE

classpathTemplateRootDirectoryName

public java.lang.String classpathTemplateRootDirectoryName

grammar

public Grammar grammar
Which grammar are we generating code for? Each generator is attached to a specific grammar.


language

protected java.lang.String language
What language are we generating?


target

public Target target
The target specifies how to write out files and do other language specific actions.


templates

protected StringTemplateGroup templates
Where are the templates this generator should use to generate code?


baseTemplates

protected StringTemplateGroup baseTemplates
The basic output templates without AST or templates stuff; this will be the templates loaded for the language such as Java.stg *and* the Dbg stuff if turned on. This is used for generating syntactic predicates.


recognizerST

protected StringTemplate recognizerST

outputFileST

protected StringTemplate outputFileST

headerFileST

protected StringTemplate headerFileST

uniqueLabelNumber

protected int uniqueLabelNumber
Used to create unique labels


tool

protected Tool tool
A reference to the ANTLR tool so we can learn about output directories and such.


debug

protected boolean debug
Generate debugging event method calls


trace

protected boolean trace
Create a Tracer object and make the recognizer invoke this.


profile

protected boolean profile
Track runtime parsing information about decisions etc... This requires the debugging event mechanism to work.


lineWidth

protected int lineWidth

acyclicDFAGenerator

public ACyclicDFACodeGenerator acyclicDFAGenerator
I have factored out the generation of acyclic DFAs to separate class


VOCAB_FILE_EXTENSION

public static final java.lang.String VOCAB_FILE_EXTENSION
I have factored out the generation of cyclic DFAs to separate class

See Also:
Constant Field Values

vocabFilePattern

protected static final java.lang.String vocabFilePattern
See Also:
Constant Field Values
Constructor Detail

CodeGenerator

public CodeGenerator(Tool tool,
                     Grammar grammar,
                     java.lang.String language)
Method Detail

loadLanguageTarget

protected void loadLanguageTarget(java.lang.String language)

loadTemplates

public void loadTemplates(java.lang.String language)
load the main language.stg template group file


genRecognizer

public StringTemplate genRecognizer()
Given the grammar to which we are attached, walk the AST associated with that grammar to create NFAs. Then create the DFAs for all decision points in the grammar by converting the NFAs to DFAs. Finally, walk the AST again to generate code. Either 1 or 2 files are written: recognizer: the main parser/lexer/treewalker item header file: language like C/C++ need extern definitions The target, such as JavaTarget, dictates which files get written.


verifyActionScopesOkForTarget

protected void verifyActionScopesOkForTarget(java.util.Map actions)
Some targets will have some extra scopes like C++ may have '@headerfile:name {action}' or something. Make sure the target likes the scopes in action table.


translateActionAttributeReferences

protected void translateActionAttributeReferences(java.util.Map actions)
Actions may reference $x::y attributes, call translateAction on each action and replace that action in the Map.


translateActionAttributeReferencesForSingleScope

protected void translateActionAttributeReferencesForSingleScope(Rule r,
                                                                java.util.Map scopeActions)
Use for translating rule @init{...} actions that have no scope


generateLocalFOLLOW

public void generateLocalFOLLOW(GrammarAST referencedElementNode,
                                java.lang.String referencedElementName,
                                java.lang.String enclosingRuleName,
                                int elementIndex)
Error recovery in ANTLR recognizers. Based upon original ideas: Algorithms + Data Structures = Programs by Niklaus Wirth and A note on error recovery in recursive descent parsers: http://portal.acm.org/citation.cfm?id=947902.947905 Later, Josef Grosch had some good ideas: Efficient and Comfortable Error Recovery in Recursive Descent Parsers: ftp://www.cocolab.com/products/cocktail/doca4.ps/ell.ps.zip Like Grosch I implemented local FOLLOW sets that are combined at run-time upon error to avoid parsing overhead.


genLookaheadDecision

public StringTemplate genLookaheadDecision(StringTemplate recognizerST,
                                           DFA dfa)
Generate code that computes the predicted alt given a DFA. The recognizerST can be either the main generated recognizerTemplate for storage in the main parser file or a separate file. It's up to the code that ultimately invokes the codegen.g grammar rule. Regardless, the output file and header file get a copy of the DFAs.


generateSpecialState

public StringTemplate generateSpecialState(DFAState s)
A special state is huge (too big for state tables) or has a predicated edge. Generate a simple if-then-else. Cannot be an accept state as they have no emanating edges. Don't worry about switch vs if-then-else because if you get here, the state is super complicated and needs an if-then-else. This is used by the new DFA scheme created June 2006.


genLabelExpr

protected StringTemplate genLabelExpr(StringTemplateGroup templates,
                                      Transition edge,
                                      int k)
Generate an expression for traversing an edge.


genSemanticPredicateExpr

protected StringTemplate genSemanticPredicateExpr(StringTemplateGroup templates,
                                                  Transition edge)

genSetExpr

public StringTemplate genSetExpr(StringTemplateGroup templates,
                                 IntSet set,
                                 int k,
                                 boolean partOfDFA)
For intervals such as [3..3, 30..35], generate an expression that tests the lookahead similar to LA(1)==3 || (LA(1)>=30&&LA(1)<=35)


genTokenTypeConstants

protected void genTokenTypeConstants(StringTemplate code)
Set attributes tokens and literals attributes in the incoming code template. This is not the token vocab interchange file, but rather a list of token type ID needed by the recognizer.


genTokenTypeNames

protected void genTokenTypeNames(StringTemplate code)
Generate a token names table that maps token type to a printable name: either the label like INT or the literal like "begin".


getTokenTypeAsTargetLabel

public java.lang.String getTokenTypeAsTargetLabel(int ttype)
Get a meaningful name for a token type useful during code generation. Literals without associated names are converted to the string equivalent of their integer values. Used to generate x==ID and x==34 type comparisons etc... Essentially we are looking for the most obvious way to refer to a token type in the generated code. If in the lexer, return the char literal translated to the target language. For example, ttype=10 will yield '\n' from the getTokenDisplayName method. That must be converted to the target languages literals. For most C-derived languages no translation is needed.


genTokenVocabOutput

protected StringTemplate genTokenVocabOutput()
Generate a token vocab file with all the token names/types. For example: ID=7 FOR=8 'for'=8 This is independent of the target language; used by antlr internally


translateAction

public java.util.List translateAction(java.lang.String ruleName,
                                      GrammarAST actionTree)

translateArgAction

public java.util.List<StringTemplate> translateArgAction(java.lang.String ruleName,
                                                         GrammarAST actionTree)
Translate an action like [3,"foo",a[3]] and return a List of the translated actions. Because actions are themselves translated to a list of chunks, must cat together into a StringTemplate>. Don't translate to strings early as we need to eval templates in context.


getListOfArgumentsFromAction

public static java.util.List<java.lang.String> getListOfArgumentsFromAction(java.lang.String actionText,
                                                                            int separatorChar)

getListOfArgumentsFromAction

public static int getListOfArgumentsFromAction(java.lang.String actionText,
                                               int start,
                                               int targetChar,
                                               int separatorChar,
                                               java.util.List<java.lang.String> args)
Given an arg action like [x, (*a).foo(21,33), 3.2+1, '\n', "a,oo\nick", {bl, "fdkj"eck}, ["cat\n,", x, 43]] convert to a list of arguments. Allow nested square brackets etc... Set separatorChar to ';' or ',' or whatever you want.


translateTemplateConstructor

public StringTemplate translateTemplateConstructor(java.lang.String ruleName,
                                                   int outerAltNum,
                                                   antlr.Token actionToken,
                                                   java.lang.String templateActionText)
Given a template constructor action like %foo(a={...}) in an action, translate it to the appropriate template constructor from the templateLib. This translates a *piece* of the action.


issueInvalidScopeError

public void issueInvalidScopeError(java.lang.String x,
                                   java.lang.String y,
                                   Rule enclosingRule,
                                   antlr.Token actionToken,
                                   int outerAltNum)

issueInvalidAttributeError

public void issueInvalidAttributeError(java.lang.String x,
                                       java.lang.String y,
                                       Rule enclosingRule,
                                       antlr.Token actionToken,
                                       int outerAltNum)

issueInvalidAttributeError

public void issueInvalidAttributeError(java.lang.String x,
                                       Rule enclosingRule,
                                       antlr.Token actionToken,
                                       int outerAltNum)

getTemplates

public StringTemplateGroup getTemplates()

getBaseTemplates

public StringTemplateGroup getBaseTemplates()

setDebug

public void setDebug(boolean debug)

setTrace

public void setTrace(boolean trace)

setProfile

public void setProfile(boolean profile)

getRecognizerST

public StringTemplate getRecognizerST()

getRecognizerFileName

public java.lang.String getRecognizerFileName(java.lang.String name,
                                              int type)
Generate TParser.java and TLexer.java from T.g if combined, else just use T.java as output regardless of type.


getVocabFileName

public java.lang.String getVocabFileName()
What is the name of the vocab file generated for this grammar? Returns null if no .tokens file should be generated.


write

public void write(StringTemplate code,
                  java.lang.String fileName)
           throws java.io.IOException
Throws:
java.io.IOException

canGenerateSwitch

protected boolean canGenerateSwitch(DFAState s)
You can generate a switch rather than if-then-else for a DFA state if there are no semantic predicates and the number of edge label values is small enough; e.g., don't generate a switch for a state containing an edge label such as 20..52330 (the resulting byte codes would overflow the method 65k limit probably).


createUniqueLabel

public java.lang.String createUniqueLabel(java.lang.String name)
Create a label to track a token / rule reference's result. Technically, this is a place where I break model-view separation as I am creating a variable name that could be invalid in a target language, however, label ::= is probably ok in all languages we care about.