Main Page | Namespace List | Class Hierarchy | Alphabetical List | Class List | Directories | File List | Namespace Members | Class Members | File Members | Related Pages

Ogre::Compiler2Pass Class Reference

Compiler2Pass is a generic 2 pass compiler/assembler. More...

#include <OgreCompiler2Pass.h>

Inheritance diagram for Ogre::Compiler2Pass:

Ogre::CompositorScriptCompiler Ogre::MaterialScriptCompiler List of all members.

Public Member Functions

 Compiler2Pass ()
 constructor
virtual ~Compiler2Pass ()
bool compile (const String &source, const String &sourceName)
 compile the source - performs 2 passes.
virtual const StringgetClientBNFGrammer (void)=0
 gets BNF Grammer.
virtual const StringgetClientGrammerName (void) const =0
 get the name of the BNF grammer.

Protected Types

typedef std::vector< TokenRuleTokenRuleContainer
typedef TokenRuleContainer::iterator TokenRuleIterator
typedef std::vector< LexemeTokenDefLexemeTokenDefContainer
typedef LexemeTokenDefContainer::iterator LexemeTokenDefIterator
typedef std::map< std::string,
size_t > 
LexemeTokenMap
typedef LexemeTokenMap::iterator TokenKeyIterator
typedef std::vector< TokenInstTokenInstContainer
typedef TokenInstContainer::iterator TokenInstIterator
enum  OperationType {
  otUNKNOWN, otRULE, otAND, otOR,
  otOPTIONAL, otREPEAT, otDATA, otNOT_TEST,
  otINSERT_TOKEN, otEND
}
enum  SystemRuleToken { _no_token_ = SystemTokenBase, _character_, _value_, _no_space_skip_ }
enum  BNF_ID {
  BNF_UNKOWN = 0, BNF_SYNTAX, BNF_RULE, BNF_IDENTIFIER,
  BNF_IDENTIFIER_RIGHT, BNF_IDENTIFIER_CHARACTERS, BNF_ID_BEGIN, BNF_ID_END,
  BNF_CONSTANT_BEGIN, BNF_SET_RULE, BNF_EXPRESSION, BNF_AND_TERM,
  BNF_OR_TERM, BNF_TERM, BNF_TERM_ID, BNF_CONSTANT,
  BNF_OR, BNF_TERMINAL_SYMBOL, BNF_TERMINAL_START, BNF_REPEAT_EXPRESSION,
  BNF_REPEAT_BEGIN, BNF_REPEAT_END, BNF_SET, BNF_SET_BEGIN,
  BNF_SET_END, BNF_NOT_TEST, BNF_NOT_TEST_BEGIN, BNF_CONDITIONAL_TOKEN_INSERT,
  BNF_OPTIONAL_EXPRESSION, BNF_NOT_EXPRESSION, BNF_NOT_CHK, BNF_OPTIONAL_BEGIN,
  BNF_OPTIONAL_END, BNF_NO_TOKEN_START, BNF_SINGLEQUOTE, BNF_SINGLE_QUOTE_EXC,
  BNF_SET_END_EXC, BNF_ANY_CHARACTER, BNF_SPECIAL_CHARACTERS1, BNF_SPECIAL_CHARACTERS2,
  BNF_WHITE_SPACE_CHK, BNF_LETTER, BNF_LETTER_DIGIT, BNF_DIGIT,
  BNF_WHITE_SPACE, BNF_ALPHA_SET, BNF_NUMBER_SET, BNF_SPECIAL_CHARACTER_SET1,
  BNF_SPECIAL_CHARACTER_SET2, BNF_SPECIAL_CHARACTER_SET3, BNF_NOT_CHARS
}

Protected Member Functions

bool doPass1 ()
 perform pass 1 of compile process scans source for lexemes that can be tokenized and then performs general semantic and context verification on each lexeme before it is tokenized.
bool doPass2 ()
 performs Pass 2 of compile process which is execution of the tokens
Remarks:
Pass 2 takes the token instructions generated in Pass 1 and builds the application specific instructions along with verifying symantic and context rules that could not be checked in Pass 1.

virtual void executeTokenAction (const size_t tokenID)=0
 execute the action associated with the token pointed to by the Pass 2 token instruction position.
virtual void setupTokenDefinitions (void)=0
 setup client token definitions.
const TokenInstgetNextToken (const size_t expectedTokenID=0) const
 Gets the next token from the instruction que.
const TokenInstgetCurrentToken (const size_t expectedTokenID=0) const
 Gets the current token from the instruction que.
bool testNextTokenID (const size_t expectedTokenID) const
 If a next token instruction exist then test if its token ID matches.
bool testCurrentTokenID (const size_t expectedTokenID) const
 If a current token instruction exist then test if its token ID matches.
void skipToken (void) const
 skip to the next token in the pass2 queue.
void replaceToken (void)
 go back to the previous token in the pass2 queue.
float getNextTokenValue (void) const
 Gets the next token's associated floating point value in the instruction que that was parsed from the text source.
float getCurrentTokenValue (void) const
 Gets the current token's associated floating point value in the instruction que that was parsed from the text source.
const StringgetNextTokenLabel (void) const
 Gets the next token's associated text label in the instruction que that was parsed from the text source.
const StringgetCurrentTokenLabel (void) const
 Gets the next token's associated text label in the instruction que that was parsed from the text source.
size_t getNextTokenID (void) const
 Get the next token's ID value.
size_t getCurrentTokenID (void) const
 Get the current token's ID value.
const StringgetNextTokenLexeme (void) const
 Get the next token's lexeme string.
const StringgetCurrentTokenLexeme (void) const
 Get the current token's lexeme string.
size_t getPass2TokenQueCount (void) const
 Gets the number of tokens waiting in the instruction que that need to be processed by an token action in pass 2.
size_t getRemainingTokensForAction (void) const
 Get the number of tokens not processed by action token.
void setPass2TokenQuePosition (size_t pos, const bool activateAction=false)
 Manualy set the Pass2 Token que position.
size_t getPass2TokenQuePosition (void) const
 Get the current position in the Pass2 Token Que.
bool setNextActionQuePosition (size_t pos, const bool search=false)
 Set the position of the next token action in the Pass2 Token Que.
void addLexemeToken (const String &lexeme, const size_t token, const bool hasAction=false, const bool caseSensitive=false)
 Add a lexeme token association.
void setClientBNFGrammer (void)
 Sets up the parser rules for the client based on the BNF Grammer text passed in.
void findEOL ()
 find the eol charater
bool isFloatValue (float &fvalue, size_t &charsize) const
 check to see if the text at the present position in the source is a numerical constant
bool isCharacterLabel (const size_t rulepathIDX)
 Check if source at current position is supposed to be a user defined character label.
bool isLexemeMatch (const String &lexeme, const bool caseSensitive) const
 check to see if the text is in the lexeme text library
bool positionToNextLexeme ()
 position to the next possible valid sysmbol
bool processRulePath (size_t rulepathIDX)
 process input source text using rulepath to determine allowed tokens
void setActiveContexts (const uint contexts)
 setup ActiveContexts - should be called by subclass to setup initial language contexts
void skipComments ()
 comment specifiers are hard coded
void skipEOL ()
 find end of line marker and move past it
void skipWhiteSpace ()
 skip all the white space which includes spaces and tabs
bool ValidateToken (const size_t rulepathIDX, const size_t activeRuleID)
 check if current position in source has the lexeme text equivalent to the TokenID
void verifyTokenRuleLinks (const String &grammerName)
 scan through all the rules and initialize token definition with index to rules for non-terminal tokens.
void checkTokenActionTrigger (void)
 Checks the last token instruction and if it has an action then it triggers the action of the previously found token having an action.
String getBNFGrammerTextFromRulePath (size_t ruleID, const size_t level=0)
 Get the text representation of the rule path.

Protected Attributes

TokenStatemClientTokenState
TokenStatemActiveTokenState
 Active token que, definitions, rules currntly being used by parser.
size_t mPass2TokenQuePosition
 the location within the token instruction container where pass 2 is
size_t mPreviousActionQuePosition
 the que position of the previous token that had an action.
size_t mNextActionQuePosition
 the que position for the next token that has an action.
const StringmSource
 pointer to the source to be compiled
String mSourceName
 name of the source to be compiled
size_t mEndOfSource
size_t mCurrentLine
size_t mCharPos
 current line number in source being tokenized
size_t mErrorCharPos
 position in current line in source being tokenized
std::map< size_t, float > mConstants
 storage container for constants defined in source container uses Token index as a key associated with a float constant
std::map< size_t, StringmLabels
 storage container for string labels defined in source container uses Token index as a key associated with a label
bool mLabelIsActive
 flag indicates when a label is being parsed.
size_t mActiveLabelKey
 the key of the active label being built during pass 1.
bool mNoSpaceSkip
 flag being true indicates that spaces are not to be skipped automatically gets set to false when mLabelIsActive goes to false
bool mNoTerminalToken
 if flag is true then next terminal token is not added to token que if found but does effect rule path flow
size_t mInsertTokenID
 TokenID to insert if next rule finds a terminal token if zero then no token inserted.
uint mActiveContexts
 Active Contexts pattern used in pass 1 to determine which tokens are valid for a certain context.

Static Protected Attributes

static const size_t SystemTokenBase = 1000

Private Types

typedef std::map< String,
TokenState
TokenStateContainer

Private Member Functions

void activatePreviousTokenAction (void)
 if a previous token action was setup then activate it now
void initBNFCompiler (void)
 initialize token definitions and rule paths
void buildClientBNFRulePaths (void)
 Convert BNF grammer token que created in pass 1 into a BNF rule path.
void modifyLastRule (const OperationType pendingRuleOp, const size_t tokenID)
 modify the last rule in the container. An end operation is added to the rule path.
size_t getClientLexemeTokenID (const String &lexeme, const bool isCaseSensitive=false)
 get the token ID for a lexeme in the client state.
void extractNonTerminal (const OperationType pendingRuleOp)
 Extract a Non Terminal identifier from the token que.
void extractTerminal (const OperationType pendingRuleOp, const bool notoken=false)
 Extract a Terminal lexeme from the token que and add to current rule expression.
void extractSet (const OperationType pendingRuleOp)
 Extract a set from the token que and add to current rule expression.
void extractNumericConstant (const OperationType pendingRuleOp)
 Extract a numeric constant from the token que and add it to the current rule expression.
void setConditionalTokenInsert (void)
 changes previous terminal token rule into a conditional terminal token insert rule
String getLexemeText (size_t &ruleID, const size_t level=0)
 get the lexem text of a rule.

Static Private Attributes

static TokenState mBNFTokenState
static TokenStateContainer mClientTokenStates

Classes

struct  LexemeTokenDef
 structure used to build lexeme Type library More...
struct  TokenInst
 structure for Token instructions that are constructed during first pass More...
struct  TokenRule
 structure used to build rule paths More...
struct  TokenState

Detailed Description

Compiler2Pass is a generic 2 pass compiler/assembler.

Remarks:
provides a tokenizer in pass 1 and relies on the subclass to provide the virtual method for pass 2
PASS 1 - tokenize source: this is a simple brute force lexical scanner/analyzer that also parses the formed token for proper semantics and context in one pass it uses top down (recursive descent) ruling based on Backus - Naur Form (BNF) notation for semantic checking.

During Pass1, if a terminal token is identified as having an action then that action gets triggered when the next terminal token is uncountered that has an action.

PASS 2 - generate application specific instructions ie native instructions based on the tokens in the instruction container.

this class must be subclassed with the subclass providing some implementation details for Pass 2. The subclass is responsible for setting up the token libraries along with defining the language syntax and handling token actions during the second pass.
The sub class normally supplies a simplified BNF text description in its constructor prior to doing any parsing/tokenizing of source. The simplified BNF text description defines the language syntax and rule structure. The meta-symbols used in the BNF text description are:
::= meaning "is defined as". "::=" starts the definition of a rule. The left side of ::= must contain an <identifier>
<> angle brackets are used to surround syntax rule names. A syntax rule name is also called a non-terminal in that it does not generate a terminal token in the instruction container for pass 2 processing.
| meaning "or". if the item on the left of the | fails then the item on the right is tested. Example: <true_false> ::= 'true' | 'false'; whitespace is used to imply AND operation between left and right items. Example: <terrain_shadaws> ::= 'terrain_shadows' <true_false> the 'terrain_shadows' terminal token must be found and <true_false> rule must pass in order for <terrain_shadows> rule to pass.
[] optional rule identifier is enclosed in meta symbols [ and ]. Note that only one identifier or terminal token can take [] modifier.
{} repetitive identifier (zero or more times) is enclosed in meta symbols { and } Note that only one identifier or terminal token can take {} modifier.
'' terminal tokens are surrounded by single quotes. A terminal token is always one or more characters. For example: 'Colour' defines a character sequence that must be matched in whole. Note that matching is case sensitive.
@ turn on single character scanning and don't skip white space. Mainly used for label processing that allow white space. Example: '@ ' prevents the white space between the quotes from being skipped
-'' no terminal token is generated when a - precedes the first single quote but the text in between the quotes is still tested against the characters in the source being parsed.
(?! ) negative lookahead (not test) inspired by Perl 5. Scans ahead for a non-terminal or terminal expression that should fail in order to make the rule production pass. Does not generate a token or advance the cursur. If the lookahead result fails ie token is found, then the current rule fails and rollback occurs. Mainly used to solve multiple contexts of a token. An Example of where not test is used to solve multiple contexts:
<rule> ::= <identifier> "::=" <expression>
<expression> ::= <and_term> { <or_term> }
<or_term> ::= "|" <and_term>
<and_term> ::= <term> { <term> }
<term> ::= <identifier_right> | <terminal_symbol> | <repeat_expression> | <optional_expression>
<identifier_right> ::= <identifier> (?!"::=")

<identiefier> appears on both sides of the ::= so (?!"::=") test to make sure that ::= is not on the right which would indicate that a new rule was being formed.

Works on both terminals and non-terminals. Note: lookahead failure causes the whole rule to fail and rollback to occur

<#name> # indicates that a numerical value is to be parsed to form a terminal token. Name is optional and is just a descriptor to help with understanding what the value will be used for. Example: <colour> ::= <#red> <#green> <#blue>
() parentheses enclose a set of characters that can be used to generate a user identifier. for example: (0123456789) matches a single character found in that set. An example of a user identifier:
<label> ::= <character> {<character>}
<character> ::= (abcdefghijklmnopqrstuvwxyz)
This will generate a rule that accepts one or more lowercase letters to make up the Label. The User identifier stops collecting the characters into a string when a match cannot be found in the rule.

(! ) if the first character in the set is a ! then any input character not found in the set will be accepted. An example:
<label> ::= <anycharacter_nolinebreak> {<anycharacter_nolinebreak>}
<anycharacter_nolinebreak> ::= (!
)

any character but
or is accepted in the input.

: Insert the terminal token on the left before the next terminal token on the right if the next terminal token on right parses. Usefull for when terminal tokens don't have a definate text state but only context state based on another terminal or character token. An example:
<last_resort> ::= 'external_command' : <special_label>
<special_label> ::= (!
)

In the example, <last_resort> gets processed when all other rules fail to parse. if <special_label> parses (reads in any character but
) then the terminal token 'external_command' is inserted prior to the Special_Label for pass 2 processing. 'external_command' does not have have an explicit text representation but based on the context of no other rules matching and <special_label> parsing, 'external_command' is considered parsed.

Definition at line 148 of file OgreCompiler2Pass.h.


Member Typedef Documentation

typedef std::vector<LexemeTokenDef> Ogre::Compiler2Pass::LexemeTokenDefContainer [protected]
 

Definition at line 221 of file OgreCompiler2Pass.h.

typedef LexemeTokenDefContainer::iterator Ogre::Compiler2Pass::LexemeTokenDefIterator [protected]
 

Definition at line 222 of file OgreCompiler2Pass.h.

typedef std::map<std::string, size_t> Ogre::Compiler2Pass::LexemeTokenMap [protected]
 

Definition at line 224 of file OgreCompiler2Pass.h.

typedef std::vector<TokenInst> Ogre::Compiler2Pass::TokenInstContainer [protected]
 

Definition at line 239 of file OgreCompiler2Pass.h.

typedef TokenInstContainer::iterator Ogre::Compiler2Pass::TokenInstIterator [protected]
 

Definition at line 240 of file OgreCompiler2Pass.h.

typedef LexemeTokenMap::iterator Ogre::Compiler2Pass::TokenKeyIterator [protected]
 

Definition at line 225 of file OgreCompiler2Pass.h.

typedef std::vector<TokenRule> Ogre::Compiler2Pass::TokenRuleContainer [protected]
 

Definition at line 170 of file OgreCompiler2Pass.h.

typedef TokenRuleContainer::iterator Ogre::Compiler2Pass::TokenRuleIterator [protected]
 

Definition at line 171 of file OgreCompiler2Pass.h.

typedef std::map<String, TokenState> Ogre::Compiler2Pass::TokenStateContainer [private]
 

Definition at line 586 of file OgreCompiler2Pass.h.


Member Enumeration Documentation

enum Ogre::Compiler2Pass::BNF_ID [protected]
 

Enumeration values:
BNF_UNKOWN 
BNF_SYNTAX 
BNF_RULE 
BNF_IDENTIFIER 
BNF_IDENTIFIER_RIGHT 
BNF_IDENTIFIER_CHARACTERS 
BNF_ID_BEGIN 
BNF_ID_END 
BNF_CONSTANT_BEGIN 
BNF_SET_RULE 
BNF_EXPRESSION 
BNF_AND_TERM 
BNF_OR_TERM 
BNF_TERM 
BNF_TERM_ID 
BNF_CONSTANT 
BNF_OR 
BNF_TERMINAL_SYMBOL 
BNF_TERMINAL_START 
BNF_REPEAT_EXPRESSION 
BNF_REPEAT_BEGIN 
BNF_REPEAT_END 
BNF_SET 
BNF_SET_BEGIN 
BNF_SET_END 
BNF_NOT_TEST 
BNF_NOT_TEST_BEGIN 
BNF_CONDITIONAL_TOKEN_INSERT 
BNF_OPTIONAL_EXPRESSION 
BNF_NOT_EXPRESSION 
BNF_NOT_CHK 
BNF_OPTIONAL_BEGIN 
BNF_OPTIONAL_END 
BNF_NO_TOKEN_START 
BNF_SINGLEQUOTE 
BNF_SINGLE_QUOTE_EXC 
BNF_SET_END_EXC 
BNF_ANY_CHARACTER 
BNF_SPECIAL_CHARACTERS1 
BNF_SPECIAL_CHARACTERS2 
BNF_WHITE_SPACE_CHK 
BNF_LETTER 
BNF_LETTER_DIGIT 
BNF_DIGIT 
BNF_WHITE_SPACE 
BNF_ALPHA_SET 
BNF_NUMBER_SET 
BNF_SPECIAL_CHARACTER_SET1 
BNF_SPECIAL_CHARACTER_SET2 
BNF_SPECIAL_CHARACTER_SET3 
BNF_NOT_CHARS 

Definition at line 181 of file OgreCompiler2Pass.h.

enum Ogre::Compiler2Pass::OperationType [protected]
 

Enumeration values:
otUNKNOWN 
otRULE 
otAND 
otOR 
otOPTIONAL 
otREPEAT 
otDATA 
otNOT_TEST 
otINSERT_TOKEN 
otEND 

Definition at line 154 of file OgreCompiler2Pass.h.

enum Ogre::Compiler2Pass::SystemRuleToken [protected]
 

Enumeration values:
_no_token_ 
_character_ 
_value_ 
_no_space_skip_ 

Definition at line 174 of file OgreCompiler2Pass.h.


Constructor & Destructor Documentation

Ogre::Compiler2Pass::Compiler2Pass  ) 
 

constructor

virtual Ogre::Compiler2Pass::~Compiler2Pass  )  [virtual]
 

Definition at line 618 of file OgreCompiler2Pass.h.


Member Function Documentation

void Ogre::Compiler2Pass::activatePreviousTokenAction void   )  [private]
 

if a previous token action was setup then activate it now

void Ogre::Compiler2Pass::addLexemeToken const String lexeme,
const size_t  token,
const bool  hasAction = false,
const bool  caseSensitive = false
[protected]
 

Add a lexeme token association.

Remarks:
The backend compiler uses the associations between lexeme and token when building the rule base from the BNF script so all associations must be done prior to compiling a source.
Parameters:
lexeme is the name of the token and use when parsing the source to determin a match for a token.
token is the ID associated with the lexeme
hasAction must be set true if the client wants an action triggered when this token is generated
caseSensitive should be set true if lexeme match should use case sensitivity

void Ogre::Compiler2Pass::buildClientBNFRulePaths void   )  [private]
 

Convert BNF grammer token que created in pass 1 into a BNF rule path.

void Ogre::Compiler2Pass::checkTokenActionTrigger void   )  [protected]
 

Checks the last token instruction and if it has an action then it triggers the action of the previously found token having an action.

bool Ogre::Compiler2Pass::compile const String source,
const String sourceName
 

compile the source - performs 2 passes.

First pass is to tokinize, check semantics and context. The second pass is performed by using tokens to look up function implementors and executing them which convert tokens to application specific instructions.

Remarks:
Pass 2 only gets executed if Pass 1 has built enough tokens to complete a rule path and found no errors
Parameters:
source a pointer to the source text to be compiled
Returns:
true if Pass 1 and Pass 2 are successfull false if any errors occur in Pass 1 or Pass 2

Referenced by Ogre::MaterialScriptCompiler::parseScript(), and Ogre::CompositorScriptCompiler::parseScript().

bool Ogre::Compiler2Pass::doPass1  )  [protected]
 

perform pass 1 of compile process scans source for lexemes that can be tokenized and then performs general semantic and context verification on each lexeme before it is tokenized.

A tokenized instruction list is built to be used by Pass 2. A rule path can trigger Pass 2 execution if enough tokens have been generated in Pass 1. Pass 1 will then pass control to pass 2 temporarily until the current tokens have been consumed.

bool Ogre::Compiler2Pass::doPass2  )  [protected]
 

performs Pass 2 of compile process which is execution of the tokens

Remarks:
Pass 2 takes the token instructions generated in Pass 1 and builds the application specific instructions along with verifying symantic and context rules that could not be checked in Pass 1.

Pass 2 execution consumes tokens and moves the Pass 2 token instruction position towards the end of the token container. Token execution can insert new tokens into the token container.

virtual void Ogre::Compiler2Pass::executeTokenAction const size_t  tokenID  )  [protected, pure virtual]
 

execute the action associated with the token pointed to by the Pass 2 token instruction position.

Remarks:
Its upto the child class to implement how it will associate a token key with and action. Actions should be placed at positions withing the BNF grammer (instruction que) that indicate enough tokens exist for pass 2 processing to take place.

Implemented in Ogre::CompositorScriptCompiler, and Ogre::MaterialScriptCompiler.

void Ogre::Compiler2Pass::extractNonTerminal const OperationType  pendingRuleOp  )  [private]
 

Extract a Non Terminal identifier from the token que.

void Ogre::Compiler2Pass::extractNumericConstant const OperationType  pendingRuleOp  )  [private]
 

Extract a numeric constant from the token que and add it to the current rule expression.

void Ogre::Compiler2Pass::extractSet const OperationType  pendingRuleOp  )  [private]
 

Extract a set from the token que and add to current rule expression.

void Ogre::Compiler2Pass::extractTerminal const OperationType  pendingRuleOp,
const bool  notoken = false
[private]
 

Extract a Terminal lexeme from the token que and add to current rule expression.

void Ogre::Compiler2Pass::findEOL  )  [protected]
 

find the eol charater

String Ogre::Compiler2Pass::getBNFGrammerTextFromRulePath size_t  ruleID,
const size_t  level = 0
[protected]
 

Get the text representation of the rule path.

This is a good way to way to visually verify that the BNF grammer did compile correctly.

Parameters:
ruleID is the index into the rule path.
level is the number of levels a non-terminal will expand to. Defaults to 0 if not set which will cause non-terminals to not expand.

virtual const String& Ogre::Compiler2Pass::getClientBNFGrammer void   )  [pure virtual]
 

gets BNF Grammer.

Gets called when BNF grammer has to be compiled for the first time.

Implemented in Ogre::CompositorScriptCompiler, and Ogre::MaterialScriptCompiler.

virtual const String& Ogre::Compiler2Pass::getClientGrammerName void   )  const [pure virtual]
 

get the name of the BNF grammer.

Implemented in Ogre::CompositorScriptCompiler, and Ogre::MaterialScriptCompiler.

size_t Ogre::Compiler2Pass::getClientLexemeTokenID const String lexeme,
const bool  isCaseSensitive = false
[private]
 

get the token ID for a lexeme in the client state.

If the lexeme is not found then it is added to the map and definition container and a new tokenID created.

Returns:
the ID of the token.

const TokenInst& Ogre::Compiler2Pass::getCurrentToken const size_t  expectedTokenID = 0  )  const [protected]
 

Gets the current token from the instruction que.

Remarks:
If an unkown token is found then an exception is raised. The subclass should catch the exception, provide an error message, and attempt recovery.
Parameters:
expectedTokenID if greater than 0 then an exception is raised if tokenID does not match.

size_t Ogre::Compiler2Pass::getCurrentTokenID void   )  const [protected]
 

Get the current token's ID value.

Definition at line 413 of file OgreCompiler2Pass.h.

const String& Ogre::Compiler2Pass::getCurrentTokenLabel void   )  const [protected]
 

Gets the next token's associated text label in the instruction que that was parsed from the text source.

If an unkown token is found or no associated label was found then an exception is raised. The subclass should catch the exception, provide an error message, and attempt recovery.

const String& Ogre::Compiler2Pass::getCurrentTokenLexeme void   )  const [protected]
 

Get the current token's lexeme string.

Handy when you don't want the ID but want the string representation.

float Ogre::Compiler2Pass::getCurrentTokenValue void   )  const [protected]
 

Gets the current token's associated floating point value in the instruction que that was parsed from the text source.

Remarks:
If an unkown token is found or no associated value was found then an exception is raised. The subclass should catch the exception, provide an error message, and attempt recovery.

String Ogre::Compiler2Pass::getLexemeText size_t &  ruleID,
const size_t  level = 0
[private]
 

get the lexem text of a rule.

const TokenInst& Ogre::Compiler2Pass::getNextToken const size_t  expectedTokenID = 0  )  const [protected]
 

Gets the next token from the instruction que.

Remarks:
If an unkown token is found then an exception is raised but the instruction pointer is still moved passed the unknown token. The subclass should catch the exception, provide an error message, and attempt recovery.
Parameters:
expectedTokenID if greater than 0 then an exception is raised if tokenID does not match.

Definition at line 339 of file OgreCompiler2Pass.h.

size_t Ogre::Compiler2Pass::getNextTokenID void   )  const [protected]
 

Get the next token's ID value.

Definition at line 410 of file OgreCompiler2Pass.h.

const String& Ogre::Compiler2Pass::getNextTokenLabel void   )  const [protected]
 

Gets the next token's associated text label in the instruction que that was parsed from the text source.

Remarks:
If an unkown token is found or no associated label was found then an exception is raised but the instruction pointer is still moved passed the unknown token. The subclass should catch the exception, provide an error message, and attempt recovery.

Definition at line 398 of file OgreCompiler2Pass.h.

const String& Ogre::Compiler2Pass::getNextTokenLexeme void   )  const [protected]
 

Get the next token's lexeme string.

Handy when you don't want the ID but want the string representation.

Definition at line 417 of file OgreCompiler2Pass.h.

float Ogre::Compiler2Pass::getNextTokenValue void   )  const [protected]
 

Gets the next token's associated floating point value in the instruction que that was parsed from the text source.

If an unkown token is found or no associated value was found then an exception is raised but the instruction pointer is still moved passed the unknown token. The subclass should catch the exception, provide an error message, and attempt recovery.

Definition at line 379 of file OgreCompiler2Pass.h.

size_t Ogre::Compiler2Pass::getPass2TokenQueCount void   )  const [protected]
 

Gets the number of tokens waiting in the instruction que that need to be processed by an token action in pass 2.

size_t Ogre::Compiler2Pass::getPass2TokenQuePosition void   )  const [protected]
 

Get the current position in the Pass2 Token Que.

Definition at line 447 of file OgreCompiler2Pass.h.

size_t Ogre::Compiler2Pass::getRemainingTokensForAction void   )  const [protected]
 

Get the number of tokens not processed by action token.

Client Actions should use this method to retreive the number of parameters(tokens) remaining to be processed in the action.

void Ogre::Compiler2Pass::initBNFCompiler void   )  [private]
 

initialize token definitions and rule paths

bool Ogre::Compiler2Pass::isCharacterLabel const size_t  rulepathIDX  )  [protected]
 

Check if source at current position is supposed to be a user defined character label.

A new label is processed when previous operation was not _character_ otherwise the processed character (if match was found) is added to the current label. This allows _character_ operations to be chained together to form a crude regular expression to build a label.

Parameters:
rulepathIDX index into rule path database of token to validate.
Returns:
true if token was found for character label.

bool Ogre::Compiler2Pass::isFloatValue float &  fvalue,
size_t &  charsize
const [protected]
 

check to see if the text at the present position in the source is a numerical constant

Parameters:
fvalue is a reference that will receive the float value that is in the source
charsize reference to receive number of characters that make of the value in the source
Returns:
true if characters form a valid float representation false if a number value could not be extracted

bool Ogre::Compiler2Pass::isLexemeMatch const String lexeme,
const bool  caseSensitive
const [protected]
 

check to see if the text is in the lexeme text library

Parameters:
lexeme points to begining of text where a lexem token might exist
caseSensitive set to true if match should be case sensitive
Returns:
true if a matching token could be found in the token type library false if could not be tokenized

void Ogre::Compiler2Pass::modifyLastRule const OperationType  pendingRuleOp,
const size_t  tokenID
[private]
 

modify the last rule in the container. An end operation is added to the rule path.

bool Ogre::Compiler2Pass::positionToNextLexeme  )  [protected]
 

position to the next possible valid sysmbol

bool Ogre::Compiler2Pass::processRulePath size_t  rulepathIDX  )  [protected]
 

process input source text using rulepath to determine allowed tokens

Remarks:
the method is reentrant and recursive if a non-terminal token is encountered in the current rule path then the method is called using the new rule path referenced by the non-terminal token Tokens can have the following operation states which effects the flow path of the rule RULE: defines a rule path for the non-terminal token AND: the token is required for the rule to pass OR: if the previous tokens failed then try these ones OPTIONAL: the token is optional and does not cause the rule to fail if the token is not found REPEAT: the token is required but there can be more than one in a sequence DATA: Used by a previous token ie for character sets NOTTEST: performs negative lookahead ie make sure the next token is not of a certain type END: end of the rule path - the method returns the succuss of the rule
Parameters:
rulepathIDX index into an array of Token Rules that define a rule path to be processed
Returns:
true if rule passed - all required tokens found false if one or more tokens required to complete the rule were not found

void Ogre::Compiler2Pass::replaceToken void   )  [protected]
 

go back to the previous token in the pass2 queue.

void Ogre::Compiler2Pass::setActiveContexts const uint  contexts  )  [protected]
 

setup ActiveContexts - should be called by subclass to setup initial language contexts

Definition at line 540 of file OgreCompiler2Pass.h.

void Ogre::Compiler2Pass::setClientBNFGrammer void   )  [protected]
 

Sets up the parser rules for the client based on the BNF Grammer text passed in.

Remarks:
Raises an exception if the grammer did not compile successfully. This method gets called when a call to compile occurs and no compiled BNF grammer exists, otherwise nothing will happen since the compiler has no rules to work with. The grammer only needs to be set once during the lifetime of the compiler unless the grammer changes.
Note:
BNF Grammer rules are cached once the BNF grammer source is compiled. The client should never have to call this method directly.

void Ogre::Compiler2Pass::setConditionalTokenInsert void   )  [private]
 

changes previous terminal token rule into a conditional terminal token insert rule

bool Ogre::Compiler2Pass::setNextActionQuePosition size_t  pos,
const bool  search = false
[protected]
 

Set the position of the next token action in the Pass2 Token Que.

Remarks:
If the position is not within the que or there is no action associated with the token at the position in the que then NextActionQuePosition is not set.
Parameters:
pos is the position in the Pass2 Token Que where the next action is.
search if true then the que is searched from pos until an action is found. If the end of the que is reached and no action has been found then NextActionQuePosition is set to the end of the que and false is returned.

void Ogre::Compiler2Pass::setPass2TokenQuePosition size_t  pos,
const bool  activateAction = false
[protected]
 

Manualy set the Pass2 Token que position.

Remarks:
This method will also set the position of the next token in the pass2 que that has an action ensuring that getRemainingTokensForAction works currectly. This method is useful for when the token que must be reprocessed after pass1 and the position in the que must be changed so that an action will be triggered.
Parameters:
pos is the new position within the Pass2 que
activateAction if set true and the token at the new position has an action then the action is activated.

virtual void Ogre::Compiler2Pass::setupTokenDefinitions void   )  [protected, pure virtual]
 

setup client token definitions.

Gets called when BNF grammer is being setup.

Implemented in Ogre::CompositorScriptCompiler, and Ogre::MaterialScriptCompiler.

void Ogre::Compiler2Pass::skipComments  )  [protected]
 

comment specifiers are hard coded

void Ogre::Compiler2Pass::skipEOL  )  [protected]
 

find end of line marker and move past it

void Ogre::Compiler2Pass::skipToken void   )  const [protected]
 

skip to the next token in the pass2 queue.

void Ogre::Compiler2Pass::skipWhiteSpace  )  [protected]
 

skip all the white space which includes spaces and tabs

bool Ogre::Compiler2Pass::testCurrentTokenID const size_t  expectedTokenID  )  const [protected]
 

If a current token instruction exist then test if its token ID matches.

Parameters:
expectedTokenID is the ID of the token to match.

Definition at line 364 of file OgreCompiler2Pass.h.

bool Ogre::Compiler2Pass::testNextTokenID const size_t  expectedTokenID  )  const [protected]
 

If a next token instruction exist then test if its token ID matches.

Remarks:
This method is usefull for peeking ahead during pass 2 to see if a certain token exists.
Parameters:
expectedTokenID is the ID of the token to match.

bool Ogre::Compiler2Pass::ValidateToken const size_t  rulepathIDX,
const size_t  activeRuleID
[protected]
 

check if current position in source has the lexeme text equivalent to the TokenID

Parameters:
rulepathIDX index into rule path database of token to validate
activeRuleID index of non-terminal rule that generated the token
Returns:
true if token was found false if token lexeme text does not match the source text if token is non-terminal then processRulePath is called

void Ogre::Compiler2Pass::verifyTokenRuleLinks const String grammerName  )  [protected]
 

scan through all the rules and initialize token definition with index to rules for non-terminal tokens.

Gets called when internal grammer is being verified or after client grammer has been parsed.

Parameters:
grammerName is the name of the grammer the token rules are for


Member Data Documentation

uint Ogre::Compiler2Pass::mActiveContexts [protected]
 

Active Contexts pattern used in pass 1 to determine which tokens are valid for a certain context.

Definition at line 298 of file OgreCompiler2Pass.h.

size_t Ogre::Compiler2Pass::mActiveLabelKey [protected]
 

the key of the active label being built during pass 1.

a new key is calculated when mLabelIsActive switches from false to true

Definition at line 286 of file OgreCompiler2Pass.h.

TokenState* Ogre::Compiler2Pass::mActiveTokenState [protected]
 

Active token que, definitions, rules currntly being used by parser.

Definition at line 254 of file OgreCompiler2Pass.h.

TokenState Ogre::Compiler2Pass::mBNFTokenState [static, private]
 

Definition at line 584 of file OgreCompiler2Pass.h.

size_t Ogre::Compiler2Pass::mCharPos [protected]
 

current line number in source being tokenized

Definition at line 272 of file OgreCompiler2Pass.h.

TokenState* Ogre::Compiler2Pass::mClientTokenState [protected]
 

Definition at line 251 of file OgreCompiler2Pass.h.

TokenStateContainer Ogre::Compiler2Pass::mClientTokenStates [static, private]
 

Definition at line 587 of file OgreCompiler2Pass.h.

std::map<size_t, float> Ogre::Compiler2Pass::mConstants [protected]
 

storage container for constants defined in source container uses Token index as a key associated with a float constant

Definition at line 277 of file OgreCompiler2Pass.h.

size_t Ogre::Compiler2Pass::mCurrentLine [protected]
 

Definition at line 271 of file OgreCompiler2Pass.h.

size_t Ogre::Compiler2Pass::mEndOfSource [protected]
 

Definition at line 269 of file OgreCompiler2Pass.h.

size_t Ogre::Compiler2Pass::mErrorCharPos [protected]
 

position in current line in source being tokenized

Definition at line 273 of file OgreCompiler2Pass.h.

size_t Ogre::Compiler2Pass::mInsertTokenID [protected]
 

TokenID to insert if next rule finds a terminal token if zero then no token inserted.

Definition at line 295 of file OgreCompiler2Pass.h.

bool Ogre::Compiler2Pass::mLabelIsActive [protected]
 

flag indicates when a label is being parsed.

It gets set false when a terminal token not of _character_ is encountered

Definition at line 283 of file OgreCompiler2Pass.h.

std::map<size_t, String> Ogre::Compiler2Pass::mLabels [protected]
 

storage container for string labels defined in source container uses Token index as a key associated with a label

Definition at line 280 of file OgreCompiler2Pass.h.

size_t Ogre::Compiler2Pass::mNextActionQuePosition [protected]
 

the que position for the next token that has an action.

Definition at line 263 of file OgreCompiler2Pass.h.

bool Ogre::Compiler2Pass::mNoSpaceSkip [protected]
 

flag being true indicates that spaces are not to be skipped automatically gets set to false when mLabelIsActive goes to false

Definition at line 289 of file OgreCompiler2Pass.h.

bool Ogre::Compiler2Pass::mNoTerminalToken [protected]
 

if flag is true then next terminal token is not added to token que if found but does effect rule path flow

Definition at line 292 of file OgreCompiler2Pass.h.

size_t Ogre::Compiler2Pass::mPass2TokenQuePosition [mutable, protected]
 

the location within the token instruction container where pass 2 is

Definition at line 256 of file OgreCompiler2Pass.h.

size_t Ogre::Compiler2Pass::mPreviousActionQuePosition [protected]
 

the que position of the previous token that had an action.

A token's action is fired on the next token having an action.

Definition at line 260 of file OgreCompiler2Pass.h.

const String* Ogre::Compiler2Pass::mSource [protected]
 

pointer to the source to be compiled

Definition at line 266 of file OgreCompiler2Pass.h.

String Ogre::Compiler2Pass::mSourceName [protected]
 

name of the source to be compiled

Definition at line 268 of file OgreCompiler2Pass.h.

const size_t Ogre::Compiler2Pass::SystemTokenBase = 1000 [static, protected]
 

Definition at line 173 of file OgreCompiler2Pass.h.


The documentation for this class was generated from the following file:

Copyright © 2000-2005 by The OGRE Team
Creative Commons License
This work is licensed under a Creative Commons Attribution-ShareAlike 2.5 License.
Last modified Sun Jul 23 10:05:58 2006