xref: /OpenGrok/opengrok-indexer/src/main/jflex/analysis/plain/PlainFullTokenizer.lex (revision 731916a808817deb91f063ac93b2a723abef5ef0)
1d219b4ceSAdam Hornacek /*
2d219b4ceSAdam Hornacek  * CDDL HEADER START
3d219b4ceSAdam Hornacek  *
4d219b4ceSAdam Hornacek  * The contents of this file are subject to the terms of the
5d219b4ceSAdam Hornacek  * Common Development and Distribution License (the "License").
6d219b4ceSAdam Hornacek  * You may not use this file except in compliance with the License.
7d219b4ceSAdam Hornacek  *
8d219b4ceSAdam Hornacek  * See LICENSE.txt included in this distribution for the specific
9d219b4ceSAdam Hornacek  * language governing permissions and limitations under the License.
10d219b4ceSAdam Hornacek  *
11d219b4ceSAdam Hornacek  * When distributing Covered Code, include this CDDL HEADER in each
12d219b4ceSAdam Hornacek  * file and include the License file at LICENSE.txt.
13d219b4ceSAdam Hornacek  * If applicable, add the following below this CDDL HEADER, with the
14d219b4ceSAdam Hornacek  * fields enclosed by brackets "[]" replaced with your own identifying
15d219b4ceSAdam Hornacek  * information: Portions Copyright [yyyy] [name of copyright owner]
16d219b4ceSAdam Hornacek  *
17d219b4ceSAdam Hornacek  * CDDL HEADER END
18d219b4ceSAdam Hornacek  */
19d219b4ceSAdam Hornacek 
20d219b4ceSAdam Hornacek /*
21d219b4ceSAdam Hornacek  * Copyright (c) 2005, 2021, Oracle and/or its affiliates. All rights reserved.
22d219b4ceSAdam Hornacek  * Portions Copyright (c) 2017-2018, Chris Fraire <cfraire@me.com>.
23d219b4ceSAdam Hornacek  */
24d219b4ceSAdam Hornacek 
25d219b4ceSAdam Hornacek package org.opengrok.indexer.analysis.plain;
26d219b4ceSAdam Hornacek 
27d219b4ceSAdam Hornacek import java.util.Locale;
28d219b4ceSAdam Hornacek import org.opengrok.indexer.analysis.JFlexSymbolMatcher;
29d219b4ceSAdam Hornacek %%
30d219b4ceSAdam Hornacek %public
31d219b4ceSAdam Hornacek %class PlainFullTokenizer
32d219b4ceSAdam Hornacek %extends JFlexSymbolMatcher
33d219b4ceSAdam Hornacek %unicode
34d219b4ceSAdam Hornacek %buffer 32766
35d219b4ceSAdam Hornacek %int
36d219b4ceSAdam Hornacek %include ../CommonLexer.lexh
37d219b4ceSAdam Hornacek %caseless
38d219b4ceSAdam Hornacek %char
39d219b4ceSAdam Hornacek 
40d219b4ceSAdam Hornacek //WhiteSpace     = [ \t\f\r]+|\n
41d219b4ceSAdam Hornacek Identifier = [a-zA-Z\p{Letter}_] [a-zA-Z\p{Letter}0-9\p{Number}_]*
42d219b4ceSAdam Hornacek Number = [0-9]+|[0-9]+\.[0-9]+| "0[xX]" [0-9a-fA-F]+
43d219b4ceSAdam Hornacek // No letters in the following, so no toLowerCase(Locale.ROOT) needed.
44*731916a8SAdam Hornacek Printable = [\@\$\%\^\&\-+=\?\.\:!\[\]\{\}*\/\|#\<\>\(\),;~]
45d219b4ceSAdam Hornacek 
46d219b4ceSAdam Hornacek %%
47d219b4ceSAdam Hornacek {Identifier}|{Number}|{Printable} {
48d219b4ceSAdam Hornacek     onSymbolMatched(yytext().toLowerCase(Locale.ROOT), yychar);
49d219b4ceSAdam Hornacek     return yystate();
50d219b4ceSAdam Hornacek }
51d219b4ceSAdam Hornacek [^]    {}
52