/*
* Sun Public License Notice
*
* The contents of this file are subject to the Sun Public License
* Version 1.0 (the "License"). You may not use this file except in
* compliance with the License. A copy of the License is available at
* http://www.sun.com/
*
* The Original Code is NetBeans. The Initial Developer of the Original
* Code is Sun Microsystems, Inc. Portions Copyright 1997-2003 Sun
* Microsystems, Inc. All Rights Reserved.
*/
package org.netbeans.editor;
/**
* Lexical analyzer that works on a given text buffer. It allows
* to sequentially parse a given character buffer by calling
* <tt>nextToken()</tt> that returns the token-ids.
*
* After the token is found by calling the <tt>nextToken</tt> method,
* the <tt>getTokenOffset()</tt> method can be used
* to get the starting offset of the current
* token in the buffer. The <tt>getTokenLength()</tt> gives the length
* of the current token.
*
* The heart of the analyzer is the <tt>parseToken()</tt> method which
* parses the text and returns the token-id of the last token found.
* The <tt>parseToken()</tt> method is called from the <tt>nextToken()</tt>.
* It operates with two important variables. The <tt>offset</tt>
* variable identifies the currently scanned character in the buffer.
* The <tt>tokenOffset</tt> is the begining of the current token.
* The <tt>state</tt> variable that identifies the current internal
* state of the analyzer is set accordingly when the characters are parsed.
* If the <tt>parseToken()</tt> recognizes a token, it returns its ID
* and the <tt>tokenOffset</tt> is its begining in the buffer and
* <tt>offset - tokenOffset</tt> is its length. When the token is processed
* the value of <tt>tokenOffset</tt> is set to be the same as current
* value of the <tt>offset</tt> and the parsing continues.
*
* Internal states are the integer constants used internally by analyzer.
* They are assigned to the <tt>state</tt> variable to express
* that the analyzer has moved from one state to another.
* They are usually numbered starting from zero but they don't
* have to. The only reserved value is -1 which is reserved
* for the INIT state - the initial internal state of the analyzer.
*
* There is also the support for defining the persistent info about
* the current state of the analyzer. This info can be later used
* to restore the parsing from some particular state instead of
* parsing from the begining of the buffer. This feature is very
* useful if there are the modifications performed in the document.
* The info is stored in the <tt>StateInfo</tt> interface
* with the <tt>BaseStateInfo</tt> as the basic implementation.
* It enables to get and set the two important values
* from the persistent point of view.
* The first one is the value of the <tt>state</tt> variable.
* The other one is the difference <tt>offset - tokenOffset</tt>
* which is called pre-scan. The particular analyzer can define
* additional values important for the persistent storage.
* The <tt>createStateInfo()</tt> can be overriden to create
* custom state-info and <tt>loadState()</tt> and <tt>storeState()</tt>
* can be overriden to get/set the additional values.
*
* The <tt>load()</tt> method sets the buffer to be parsed.
* There is a special parameter in the load() method called position
* that allows a relation of the character buffer passed to the load()
* method and the position of the buffer's data in the document.
* For this extended functionality the document must be passed
* to the constructor of the lexical analyzer at some level.
*
*
* @author Miloslav Metelka
* @version 1.00
*/
public class Syntax {
/** Is the state of analyzer equal to a given state info? */
public static final int EQUAL_STATE = 0;
/** Is the state of analyzer different from given state info? */
public static final int DIFFERENT_STATE = 1;
/** Initial internal state of the analyzer */
public static final int INIT = -1;
/** Internal state of the lexical analyzer. At the begining
* it's set to INIT value but it is changed by <tt>parseToken()</tt>
* as the characters are processed one by one.
*/
protected int state = INIT;
/** Text buffer to scan */
protected char buffer[];
/** Current offset in the buffer */
protected int offset;
/** Offset holding the begining of the current token */
protected int tokenOffset;
/** This variable is the length of the token that was found */
protected int tokenLength;
/** Path from which the found token-id comes from.
* The <tt>TokenContext.getContextPath()</tt> can be used
* to get the path. If the lexical analyzer doesn't use
* any children token-contexts it can assign
* the path in the constructor.
*/
protected TokenContextPath tokenContextPath;
/** Setting this flag to true means that there are currently no more
* buffers available so that analyzer should return all the tokens
* including those whose successful scanning would be otherwise
* left for later when the next buffer will be available. Setting
* this flag to true ensures that all the characters in the current
* buffer will be processed.
* The lexical analyzer should on one hand process all the characters
* but on the other hand it should "save" its context. For example
* if the scanner finds the unclosed comment at the end of the buffer
* it should return the comment token but
* stay in the "being in comment" internal state.
*/
protected boolean lastBuffer;
/** On which offset in the buffer scanning should stop. */
protected int stopOffset;
/** The position in the document that logically corresponds
* to the stopOffset value. If there's no relation
* to the document, it's -1. The reason why the relation
* to the document's data is expressed through
* the stopOffset to stopPosition relation is because
* the stopOffset is the only offset that doesn't change
* rapidly in the operation of the lexical analyzer.
*/
protected int stopPosition;
/** This variable can be populated by the parseToken() method
* in case the user types an errorneous construction but
* it's clear what correct token he meant to write.
* For example if the user writes a single '0x' it's an errorneous
* construct but it's clear that the user wants to enter
* the hexa-number. In this situation the parseToken()
* should report error, but it should also set the supposedTokenID
* to the hexa-number token.
* This information is used while drawing the text. If the caret
* stand inside or around such token, it calls the getSupposedTokenID()
* after calling the nextToken() and if it's non-null it uses it
* instead of the original token.
*/
protected TokenID supposedTokenID;
/** Function that should be called externally to scan the text.
* It manages the call to parseToken() and cares about the proper
* setting of the offsets.
* It can be extended to support any custom debugging required.
*/
public TokenID nextToken() {
// Return immediately when at the end of buffer
if (tokenOffset >= stopOffset) {
tokenLength = 0;
return null; // signal no token found
}
// Divide non-debug and debug sections
supposedTokenID = null;
TokenID tokenID = parseToken();
if (tokenID != null) { // regular token found
tokenLength = offset - tokenOffset;
tokenOffset = offset; // move to the next token
if (tokenLength == 0) { // test for empty token
return nextToken(); // repeat until non-empty token is found
}
} else { // EOT reached
tokenLength = 0;
}
return tokenID;
}
/** This is core function of analyzer and it returns either the token-id
* or null to indicate that the end of buffer was found.
* The function scans the active character and does one or more
* of the following actions:
* 1. change internal analyzer state
* 2. set the token-context-path and return token-id
* 3. adjust current position to signal different end of token;
* the character that offset points to is not included in the token
*/
protected TokenID parseToken() {
return null;
}
/** Load the state from syntax mark into analyzer. This method is used when
* @param stateInfo info about the state of the lexical analyzer to load.
* It can be null to indicate there's no previous state so the analyzer
* starts from its initial state.
* @param buffer buffer that will be scanned
* @param offset offset of the first character that will be scanned
* @param len length of the area to be scanned
* @param lastBuffer whether this is the last buffer in the document. All the tokens
* will be returned including the last possibly incomplete one. If the data
* come from the document, the simple rule for this parameter
* is (doc.getLength() == stop-position) where stop-position
* is the position corresponding to the (offset + len) in the buffer
* that comes from the document data.
* @param stopPosition position in the document that corresponds to (offset + len) offset
* in the provided buffer. It has only sense if the data in the buffer come from the document.
* It helps in writing the advanced analyzers that need to interact with some other data
* in the document than only those provided in the character buffer.
* If there is no relation to the document data, the stopPosition parameter
* must be filled with -1 which means an invalid value.
* The stop-position is passed (instead of start-position) because it doesn't
* change through the analyzer operation. It corresponds to the <tt>stopOffset</tt>
* that also doesn't change through the analyzer operation so any
* buffer-offset can be transferred to position by computing
* <tt>stopPosition + buffer-offset - stopOffset</tt>
* where stopOffset is the instance variable that is assigned
* to <tt>offset + len</tt> in the body of relocate().
*/
public void load(StateInfo stateInfo, char buffer[], int offset, int len,
boolean lastBuffer, int stopPosition) {
this.buffer = buffer;
this.offset = offset;
this.tokenOffset = offset;
this.stopOffset = offset + len;
this.lastBuffer = lastBuffer;
this.stopPosition = stopPosition;
if (stateInfo != null) {
loadState(stateInfo);
} else {
loadInitState();
}
}
/** Relocate scanning to another buffer.
* This is used to continue scanning after previously
* reported EOT. Relocation delta between current offset and the requested offset
* is computed and all the offsets are relocated. If there's a non-zero preScan
* in the analyzer, it is a caller's responsibility to provide all the preScan
* characters in the relocation buffer.
* @param buffer next buffer where the scan will continue.
* @param offset offset where the scan will continue.
* It's not decremented by the current preScan.
* @param len length of the area to be scanned.
* It's not extended by the current preScan.
* @param lastBuffer whether this is the last buffer in the document. All the tokens
* will be returned including the last possibly incomplete one. If the data
* come from the document, the simple rule for this parameter
* is (doc.getLength() == stop-position) where stop-position
* is the position corresponding to the (offset + len) in the buffer
* that comes from the document data.
* @param stopPosition position in the document that corresponds to (offset + len) offset
* in the provided buffer. It has only sense if the data in the buffer come from the document.
* It helps in writing the advanced analyzers that need to interact with some other data
* in the document than only those provided in the character buffer.
* If there is no relation to the document data, the stopPosition parameter
* must be filled with -1 which means an invalid value.
* The stop-position is passed (instead of start-position) because it doesn't
* change through the analyzer operation. It corresponds to the <tt>stopOffset</tt>
* that also doesn't change through the analyzer operation so any
* buffer-offset can be transferred to position by computing
* <tt>stopPosition + buffer-offset - stopOffset</tt>
* where stopOffset is the instance variable that is assigned
* to <tt>offset + len</tt> in the body of relocate().
*/
public void relocate(char buffer[], int offset, int len,
boolean lastBuffer, int stopPosition) {
this.buffer = buffer;
this.lastBuffer = lastBuffer;
int delta = offset - this.offset; // delta according to current offset
this.offset += delta;
this.tokenOffset += delta;
this.stopOffset = offset + len;
this.stopPosition = stopPosition;
}
/** Get the current buffer */
public char[] getBuffer() {
return buffer;
}
/** Get the current scanning offset */
public int getOffset() {
return offset;
}
/** Get start of token in scanned buffer. */
public int getTokenOffset() {
return offset - tokenLength;
}
/** Get length of token in scanned buffer. */
public int getTokenLength() {
return tokenLength;
}
/** Get the token-context-path of the returned token. */
public TokenContextPath getTokenContextPath() {
return tokenContextPath;
}
public TokenID getSupposedTokenID() {
return supposedTokenID;
}
/** Get the pre-scan which is a number
* of characters between offset and tokenOffset.
* If there's no more characters in the current buffer,
* the analyzer returns EOT, but it can be in a state when
* there are already some characters parsed at the end of
* the current buffer but the token
* is still incomplete and it cannot be returned yet.
* The pre-scan value helps to determine how many characters
* from the end of the current buffer should be present
* at the begining of the next buffer so that the current
* incomplete token can be returned as the first token
* when parsing the next buffer.
*/
public int getPreScan() {
return offset - tokenOffset;
}
/** Initialize the analyzer when scanning from the begining
* of the document or when the state stored in syntax mark
* is null for some reason or to explicitly reset the analyzer
* to the initial state. The offsets must not be touched by this method.
*/
public void loadInitState() {
state = INIT;
}
public void reset() {
tokenLength = stopOffset = tokenOffset = offset = 0;
loadInitState();
}
/** Load valid mark state into the analyzer. Offsets
* are already initialized when this method is called. This method
* must get the state from the mark and set it to the analyzer. Then
* it must decrease tokenOffset by the preScan stored in the mark state.
* @param markState mark state to be loaded into syntax. It must be non-null value.
*/
public void loadState(StateInfo stateInfo) {
state = stateInfo.getState();
tokenOffset -= stateInfo.getPreScan();
}
/** Store state of this analyzer into given mark state. */
public void storeState(StateInfo stateInfo) {
stateInfo.setState(state);
stateInfo.setPreScan(getPreScan());
}
/** Compare state of this analyzer to given state info */
public int compareState(StateInfo stateInfo) {
if (stateInfo != null) {
return ((stateInfo.getState() == state) && stateInfo.getPreScan() == getPreScan())
? EQUAL_STATE : DIFFERENT_STATE;
} else {
return DIFFERENT_STATE;
}
}
/** Create state info appropriate for particular analyzer */
public StateInfo createStateInfo() {
return new BaseStateInfo();
}
/** Get state name as string. It can be used for debugging purposes
* by developer of new syntax analyzer. The states that this function
* recognizes can include all constants used in analyzer so that it can
* be used everywhere in analyzer to convert numbers to more practical strings.
*/
public String getStateName(int stateNumber) {
switch(stateNumber) {
case INIT:
return "INIT"; // NOI18N
default:
return "Unknown state " + stateNumber; // NOI18N
}
}
/** Syntax information as String */
public String toString() {
return "tokenOffset=" + tokenOffset // NOI18N
+ ", offset=" + offset // NOI18N
+ ", state=" + getStateName(state) // NOI18N
+ ", stopOffset=" + stopOffset // NOI18N
+ ", lastBuffer=" + lastBuffer; // NOI18N
}
/** Interface that stores two basic pieces of information about
* the state of the whole lexical analyzer - its internal state and preScan.
*/
public interface StateInfo {
/** Get the internal state */
public int getState();
/** Store the internal state */
public void setState(int state);
/** Get the preScan value */
public int getPreScan();
/** Store the preScan value */
public void setPreScan(int preScan);
}
/** Base implementation of the StateInfo interface */
public static class BaseStateInfo implements StateInfo {
/** analyzer state */
private int state;
/** Pre-scan length */
private int preScan;
public int getState() {
return state;
}
public void setState(int state) {
this.state = state;
}
public int getPreScan() {
return preScan;
}
public void setPreScan(int preScan) {
this.preScan = preScan;
}
public String toString(Syntax syntax) {
return "state="
+ ((syntax != null)
? syntax.getStateName(getState())
: Integer.toString(getState()))
+ ", preScan=" + getPreScan(); // NOI18N
}
public String toString() {
return toString(null);
}
}
}