|
What this is
Other links
The source code/* * CTokenMarker.java - C token marker * Copyright (C) 1998, 1999 Slava Pestov * Copyright (C) 2001-2003 Romain Guy * * This program is free software; you can redistribute it and/or * modify it under the terms of the GNU General Public License * as published by the Free Software Foundation; either version 2 * of the License, or any later version. * * This program is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU General Public License for more details. * * You should have received a copy of the GNU General Public License * along with this program; if not, write to the Free Software * Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */ package org.gjt.sp.jedit.syntax; import javax.swing.text.Segment; /** * C token marker. * * @author Slava Pestov * @version $Id: CTokenMarker.java,v 1.9 2003/06/30 17:31:09 blaisorblade Exp $ */ public class CTokenMarker extends TokenMarker { public static final String METHOD_DELIMITERS = " \t~!%^*()-+=|\\#/{}[]:;\"'<>,.?@"; public CTokenMarker() { this(true, false, getKeywords()); } public CTokenMarker(boolean cpp, boolean javadoc, KeywordMap keywords) { this.cpp = cpp; this.javadoc = javadoc; this.keywords = keywords; } public byte markTokensImpl(byte token, Segment line, int lineIndex) { char[] array = line.array; int offset = line.offset; lastOffset = offset; lastKeyword = offset; lastWhitespace = offset - 1; int length = line.count + offset; boolean backslash = false; loop: for (int i = offset; i < length; i++) { int i1 = (i + 1); char c = array[i]; if (c == '\\') { backslash = !backslash; continue; } switch(token) { case Token.NULL: switch(c) { case '(': if (backslash) { doKeyword(line, i, c); backslash = false; } else { if (doKeyword(line, i, c)) break; addToken(lastWhitespace - lastOffset + 1, token); addToken(i - lastWhitespace - 1, Token.METHOD); addToken(1, Token.NULL); token = Token.NULL; lastOffset = lastKeyword = i1; lastWhitespace = i; } break; case '#': if (backslash) backslash = false; else if (cpp) { if (doKeyword(line, i, c)) break; addToken(i - lastOffset, token); addToken(length - i, Token.KEYWORD2); lastOffset = lastKeyword = length; break loop; } break; case '"': doKeyword(line, i, c); if(backslash) backslash = false; else { addToken(i - lastOffset, token); token = Token.LITERAL1; lastOffset = lastKeyword = i; } break; case '\'': doKeyword(line, i, c); if (backslash) backslash = false; else { addToken(i - lastOffset, token); token = Token.LITERAL2; lastOffset = lastKeyword = i; } break; case ':': if (lastKeyword == offset) { if (doKeyword(line, i, c)) break; else if (i1 < array.length && array[i1] == ':') addToken(i1 - lastOffset, Token.NULL); else addToken(i1 - lastOffset, Token.LABEL); lastOffset = lastKeyword = i1; lastWhitespace = i1; backslash = false; } else if (doKeyword(line, i, c)) break; break; case '/': backslash = false; doKeyword(line, i, c); if(length - i > 1) { switch(array[i1]) { case '*': addToken(i - lastOffset, token); lastOffset = lastKeyword = i; if(javadoc && length - i > 2 && array[i+2] == '*') token = Token.COMMENT2; else token = Token.COMMENT1; break; case '/': addToken(i - lastOffset, token); addToken(length - i, Token.COMMENT1); lastOffset = lastKeyword = length; break loop; } } break; default: backslash = false; if (!Character.isLetterOrDigit(c) && c != '_') doKeyword(line, i, c); if (METHOD_DELIMITERS.indexOf(c) != -1) { lastWhitespace = i; } break; } break; case Token.COMMENT1: case Token.COMMENT2: backslash = false; if (c == '*' && length - i > 1) { if (array[i1] == '/') { i++; addToken((i + 1) - lastOffset, token); token = Token.NULL; lastOffset = lastKeyword = i + 1; lastWhitespace = i; } } break; case Token.LITERAL1: if (backslash) backslash = false; else if (c == '"') { addToken(i1 - lastOffset, token); token = Token.NULL; lastOffset = lastKeyword = i1; lastWhitespace = i; } break; case Token.LITERAL2: if (backslash) backslash = false; else if (c == '\'') { addToken(i1 - lastOffset, token); token = Token.NULL; lastOffset = lastKeyword = i1; lastWhitespace = i; } break; default: throw new InternalError("Invalid state: " + token); } } if (token == Token.NULL) doKeyword(line, length, '\0'); switch(token) { case Token.LITERAL1: case Token.LITERAL2: addToken(length - lastOffset, Token.INVALID); token = Token.NULL; break; case Token.KEYWORD2: addToken(length - lastOffset, token); if (!backslash) token = Token.NULL; default: addToken(length - lastOffset, token); break; } return token; } public static KeywordMap getKeywords() { if (cKeywords == null) { cKeywords = new KeywordMap(false); cKeywords.add("char", Token.KEYWORD3); cKeywords.add("double", Token.KEYWORD3); cKeywords.add("enum", Token.KEYWORD3); cKeywords.add("float", Token.KEYWORD3); cKeywords.add("int", Token.KEYWORD3); cKeywords.add("long", Token.KEYWORD3); cKeywords.add("short", Token.KEYWORD3); cKeywords.add("signed", Token.KEYWORD3); cKeywords.add("struct", Token.KEYWORD3); cKeywords.add("typedef", Token.KEYWORD3); cKeywords.add("union", Token.KEYWORD3); cKeywords.add("unsigned", Token.KEYWORD3); cKeywords.add("void", Token.KEYWORD3); cKeywords.add("auto", Token.KEYWORD1); cKeywords.add("const", Token.KEYWORD1); cKeywords.add("extern", Token.KEYWORD1); cKeywords.add("register", Token.KEYWORD1); cKeywords.add("static", Token.KEYWORD1); cKeywords.add("volatile", Token.KEYWORD1); cKeywords.add("break", Token.KEYWORD1); cKeywords.add("case", Token.KEYWORD1); cKeywords.add("continue", Token.KEYWORD1); cKeywords.add("default", Token.KEYWORD1); cKeywords.add("do", Token.KEYWORD1); cKeywords.add("else", Token.KEYWORD1); cKeywords.add("for", Token.KEYWORD1); cKeywords.add("goto", Token.KEYWORD1); cKeywords.add("if", Token.KEYWORD1); cKeywords.add("return", Token.KEYWORD1); cKeywords.add("sizeof", Token.KEYWORD1); cKeywords.add("switch", Token.KEYWORD1); cKeywords.add("while", Token.KEYWORD1); cKeywords.add("asm", Token.KEYWORD2); cKeywords.add("asmlinkage", Token.KEYWORD2); cKeywords.add("far", Token.KEYWORD2); cKeywords.add("huge", Token.KEYWORD2); cKeywords.add("inline", Token.KEYWORD2); cKeywords.add("near", Token.KEYWORD2); cKeywords.add("pascal", Token.KEYWORD2); cKeywords.add("true", Token.LITERAL2); cKeywords.add("false", Token.LITERAL2); cKeywords.add("NULL", Token.LITERAL2); } return cKeywords; } // private members private static KeywordMap cKeywords; protected boolean cpp; protected boolean javadoc; protected KeywordMap keywords; protected int lastOffset; protected int lastKeyword; protected int lastWhitespace; protected boolean doKeyword(Segment line, int i, char c) { int i1 = i+1; int len = i - lastKeyword; byte id = keywords.lookup(line, lastKeyword, len); if (id != Token.NULL) { if (lastKeyword != lastOffset) addToken(lastKeyword - lastOffset, Token.NULL); addToken(len, id); lastOffset = i; lastKeyword = i1; lastWhitespace = i; return true; } lastKeyword = i1; return false; } } /* * ChangeLog: * $Log: CTokenMarker.java,v $ * Revision 1.9 2003/06/30 17:31:09 blaisorblade * Fix for line-ends. * * Revision 1.8 2003/06/29 13:37:27 gfx * Support of JDK 1.4.2 * * Revision 1.7 2003/03/13 22:52:48 gfx * Improved focus gain * * Revision 1.6 2002/03/22 21:01:00 gfx * Jext 3.1pre2 |
... this post is sponsored by my books ... | |
#1 New Release! |
FP Best Seller |
Copyright 1998-2024 Alvin Alexander, alvinalexander.com
All Rights Reserved.
A percentage of advertising revenue from
pages under the /java/jwarehouse
URI on this website is
paid back to open source projects.