1259698Sdim//===- llvm/Support/Unicode.h - Unicode character properties  -*- C++ -*-=====//
2259698Sdim//
3259698Sdim//                     The LLVM Compiler Infrastructure
4259698Sdim//
5259698Sdim// This file is distributed under the University of Illinois Open Source
6259698Sdim// License. See LICENSE.TXT for details.
7259698Sdim//
8259698Sdim//===----------------------------------------------------------------------===//
9259698Sdim//
10259698Sdim// This file defines functions that allow querying certain properties of Unicode
11259698Sdim// characters.
12259698Sdim//
13259698Sdim//===----------------------------------------------------------------------===//
14259698Sdim
15259698Sdim#include "llvm/ADT/StringRef.h"
16259698Sdim
17259698Sdimnamespace llvm {
18259698Sdimnamespace sys {
19259698Sdimnamespace unicode {
20259698Sdim
21259698Sdimenum ColumnWidthErrors {
22259698Sdim  ErrorInvalidUTF8 = -2,
23259698Sdim  ErrorNonPrintableCharacter = -1
24259698Sdim};
25259698Sdim
26259698Sdim/// Determines if a character is likely to be displayed correctly on the
27259698Sdim/// terminal. Exact implementation would have to depend on the specific
28259698Sdim/// terminal, so we define the semantic that should be suitable for generic case
29259698Sdim/// of a terminal capable to output Unicode characters.
30259698Sdim///
31259698Sdim/// All characters from the Unicode code point range are considered printable
32259698Sdim/// except for:
33259698Sdim///   * C0 and C1 control character ranges;
34259698Sdim///   * default ignorable code points as per 5.21 of
35259698Sdim///     http://www.unicode.org/versions/Unicode6.2.0/UnicodeStandard-6.2.pdf
36259698Sdim///     except for U+00AD SOFT HYPHEN, as it's actually displayed on most
37259698Sdim///     terminals;
38259698Sdim///   * format characters (category = Cf);
39259698Sdim///   * surrogates (category = Cs);
40259698Sdim///   * unassigned characters (category = Cn).
41259698Sdim/// \return true if the character is considered printable.
42259698Sdimbool isPrintable(int UCS);
43259698Sdim
44259698Sdim/// Gets the number of positions the UTF8-encoded \p Text is likely to occupy
45259698Sdim/// when output on a terminal ("character width"). This depends on the
46259698Sdim/// implementation of the terminal, and there's no standard definition of
47259698Sdim/// character width.
48259698Sdim///
49259698Sdim/// The implementation defines it in a way that is expected to be compatible
50259698Sdim/// with a generic Unicode-capable terminal.
51259698Sdim///
52259698Sdim/// \return Character width:
53259698Sdim///   * ErrorNonPrintableCharacter (-1) if \p Text contains non-printable
54259698Sdim///     characters (as identified by isPrintable);
55259698Sdim///   * 0 for each non-spacing and enclosing combining mark;
56259698Sdim///   * 2 for each CJK character excluding halfwidth forms;
57259698Sdim///   * 1 for each of the remaining characters.
58259698Sdimint columnWidthUTF8(StringRef Text);
59259698Sdim
60259698Sdim} // namespace unicode
61259698Sdim} // namespace sys
62259698Sdim} // namespace llvm
63