Mercurial > gnulib

annotate lib/unictype/identsyntaxmap.h @ 40057:b06060465f09

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
maint: Run 'make update-copyright'
author Paul Eggert <eggert@cs.ucla.edu>
date Tue, 01 Jan 2019 00:25:11 +0100
parents 10eb9086bea0
children
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
9471
6dc73c76eced Unicode character classification functions.
Bruno Haible <bruno@clisp.org>
parents:
diff changeset
1 /* Three-level bitmap lookup.
40057
b06060465f09 maint: Run 'make update-copyright'
Paul Eggert <eggert@cs.ucla.edu>
parents: 19484
diff changeset
2 Copyright (C) 2000-2002, 2005-2007, 2009-2019 Free Software Foundation, Inc.
9471
6dc73c76eced Unicode character classification functions.
Bruno Haible <bruno@clisp.org>
parents:
diff changeset
3 Written by Bruno Haible <bruno@clisp.org>, 2000-2002.
6dc73c76eced Unicode character classification functions.
Bruno Haible <bruno@clisp.org>
parents:
diff changeset
4
6dc73c76eced Unicode character classification functions.
Bruno Haible <bruno@clisp.org>
parents:
diff changeset
5 This program is free software: you can redistribute it and/or modify it
6dc73c76eced Unicode character classification functions.
Bruno Haible <bruno@clisp.org>
parents:
diff changeset
6 under the terms of the GNU Lesser General Public License as published
6dc73c76eced Unicode character classification functions.
Bruno Haible <bruno@clisp.org>
parents:
diff changeset
7 by the Free Software Foundation; either version 3 of the License, or
6dc73c76eced Unicode character classification functions.
Bruno Haible <bruno@clisp.org>
parents:
diff changeset
8 (at your option) any later version.
6dc73c76eced Unicode character classification functions.
Bruno Haible <bruno@clisp.org>
parents:
diff changeset
9
6dc73c76eced Unicode character classification functions.
Bruno Haible <bruno@clisp.org>
parents:
diff changeset
10 This program is distributed in the hope that it will be useful,
6dc73c76eced Unicode character classification functions.
Bruno Haible <bruno@clisp.org>
parents:
diff changeset
11 but WITHOUT ANY WARRANTY; without even the implied warranty of
6dc73c76eced Unicode character classification functions.
Bruno Haible <bruno@clisp.org>
parents:
diff changeset
12 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
6dc73c76eced Unicode character classification functions.
Bruno Haible <bruno@clisp.org>
parents:
diff changeset
13 Lesser General Public License for more details.
6dc73c76eced Unicode character classification functions.
Bruno Haible <bruno@clisp.org>
parents:
diff changeset
14
6dc73c76eced Unicode character classification functions.
Bruno Haible <bruno@clisp.org>
parents:
diff changeset
15 You should have received a copy of the GNU Lesser General Public License
19190
9759915b2aca all: prefer https: URLs
Paul Eggert <eggert@cs.ucla.edu>
parents: 18626
diff changeset
16 along with this program. If not, see <https://www.gnu.org/licenses/>. */
9471
6dc73c76eced Unicode character classification functions.
Bruno Haible <bruno@clisp.org>
parents:
diff changeset
17
6dc73c76eced Unicode character classification functions.
Bruno Haible <bruno@clisp.org>
parents:
diff changeset
18 static inline int identsyntax_lookup (ucs4_t uc);
6dc73c76eced Unicode character classification functions.
Bruno Haible <bruno@clisp.org>
parents:
diff changeset
19
6dc73c76eced Unicode character classification functions.
Bruno Haible <bruno@clisp.org>
parents:
diff changeset
20 static inline int
6dc73c76eced Unicode character classification functions.
Bruno Haible <bruno@clisp.org>
parents:
diff changeset
21 identsyntax_lookup (ucs4_t uc)
6dc73c76eced Unicode character classification functions.
Bruno Haible <bruno@clisp.org>
parents:
diff changeset
22 {
6dc73c76eced Unicode character classification functions.
Bruno Haible <bruno@clisp.org>
parents:
diff changeset
23 unsigned int index1 = uc >> identsyntax_header_0;
6dc73c76eced Unicode character classification functions.
Bruno Haible <bruno@clisp.org>
parents:
diff changeset
24 if (index1 < identsyntax_header_1)
6dc73c76eced Unicode character classification functions.
Bruno Haible <bruno@clisp.org>
parents:
diff changeset
25 {
6dc73c76eced Unicode character classification functions.
Bruno Haible <bruno@clisp.org>
parents:
diff changeset
26 int lookup1 = TABLE.level1[index1];
6dc73c76eced Unicode character classification functions.
Bruno Haible <bruno@clisp.org>
parents:
diff changeset
27 if (lookup1 >= 0)
12421
e8d2c6fc33ad Use spaces for indentation, not tabs.
Bruno Haible <bruno@clisp.org>
parents: 9471
diff changeset
28 {
e8d2c6fc33ad Use spaces for indentation, not tabs.
Bruno Haible <bruno@clisp.org>
parents: 9471
diff changeset
29 unsigned int index2 = (uc >> identsyntax_header_2) & identsyntax_header_3;
e8d2c6fc33ad Use spaces for indentation, not tabs.
Bruno Haible <bruno@clisp.org>
parents: 9471
diff changeset
30 int lookup2 = TABLE.level2[lookup1 + index2];
e8d2c6fc33ad Use spaces for indentation, not tabs.
Bruno Haible <bruno@clisp.org>
parents: 9471
diff changeset
31 if (lookup2 >= 0)
e8d2c6fc33ad Use spaces for indentation, not tabs.
Bruno Haible <bruno@clisp.org>
parents: 9471
diff changeset
32 {
e8d2c6fc33ad Use spaces for indentation, not tabs.
Bruno Haible <bruno@clisp.org>
parents: 9471
diff changeset
33 unsigned int index3 = (uc & identsyntax_header_4) + lookup2;
e8d2c6fc33ad Use spaces for indentation, not tabs.
Bruno Haible <bruno@clisp.org>
parents: 9471
diff changeset
34 /* level3 contains 2-bit values. */
e8d2c6fc33ad Use spaces for indentation, not tabs.
Bruno Haible <bruno@clisp.org>
parents: 9471
diff changeset
35 unsigned int lookup3 = TABLE.level3[index3 >> 3];
9471
6dc73c76eced Unicode character classification functions.
Bruno Haible <bruno@clisp.org>
parents:
diff changeset
36
12421
e8d2c6fc33ad Use spaces for indentation, not tabs.
Bruno Haible <bruno@clisp.org>
parents: 9471
diff changeset
37 return (lookup3 >> (2 * (index3 & 7))) & 3;
e8d2c6fc33ad Use spaces for indentation, not tabs.
Bruno Haible <bruno@clisp.org>
parents: 9471
diff changeset
38 }
e8d2c6fc33ad Use spaces for indentation, not tabs.
Bruno Haible <bruno@clisp.org>
parents: 9471
diff changeset
39 }
9471
6dc73c76eced Unicode character classification functions.
Bruno Haible <bruno@clisp.org>
parents:
diff changeset
40 }
6dc73c76eced Unicode character classification functions.
Bruno Haible <bruno@clisp.org>
parents:
diff changeset
41 return UC_IDENTIFIER_INVALID;
6dc73c76eced Unicode character classification functions.
Bruno Haible <bruno@clisp.org>
parents:
diff changeset
42 }