Home | History | Annotate | Download | only in Support
      1 //===- llvm/Support/Unicode.h - Unicode character properties  -*- C++ -*-=====//
      2 //
      3 //                     The LLVM Compiler Infrastructure
      4 //
      5 // This file is distributed under the University of Illinois Open Source
      6 // License. See LICENSE.TXT for details.
      7 //
      8 //===----------------------------------------------------------------------===//
      9 //
     10 // This file defines functions that allow querying certain properties of Unicode
     11 // characters.
     12 //
     13 //===----------------------------------------------------------------------===//
     14 
     15 #ifndef LLVM_SUPPORT_UNICODE_H
     16 #define LLVM_SUPPORT_UNICODE_H
     17 
     18 namespace llvm {
     19 class StringRef;
     20 
     21 namespace sys {
     22 namespace unicode {
     23 
     24 enum ColumnWidthErrors {
     25   ErrorInvalidUTF8 = -2,
     26   ErrorNonPrintableCharacter = -1
     27 };
     28 
     29 /// Determines if a character is likely to be displayed correctly on the
     30 /// terminal. Exact implementation would have to depend on the specific
     31 /// terminal, so we define the semantic that should be suitable for generic case
     32 /// of a terminal capable to output Unicode characters.
     33 ///
     34 /// All characters from the Unicode code point range are considered printable
     35 /// except for:
     36 ///   * C0 and C1 control character ranges;
     37 ///   * default ignorable code points as per 5.21 of
     38 ///     http://www.unicode.org/versions/Unicode6.2.0/UnicodeStandard-6.2.pdf
     39 ///     except for U+00AD SOFT HYPHEN, as it's actually displayed on most
     40 ///     terminals;
     41 ///   * format characters (category = Cf);
     42 ///   * surrogates (category = Cs);
     43 ///   * unassigned characters (category = Cn).
     44 /// \return true if the character is considered printable.
     45 bool isPrintable(int UCS);
     46 
     47 /// Gets the number of positions the UTF8-encoded \p Text is likely to occupy
     48 /// when output on a terminal ("character width"). This depends on the
     49 /// implementation of the terminal, and there's no standard definition of
     50 /// character width.
     51 ///
     52 /// The implementation defines it in a way that is expected to be compatible
     53 /// with a generic Unicode-capable terminal.
     54 ///
     55 /// \return Character width:
     56 ///   * ErrorNonPrintableCharacter (-1) if \p Text contains non-printable
     57 ///     characters (as identified by isPrintable);
     58 ///   * 0 for each non-spacing and enclosing combining mark;
     59 ///   * 2 for each CJK character excluding halfwidth forms;
     60 ///   * 1 for each of the remaining characters.
     61 int columnWidthUTF8(StringRef Text);
     62 
     63 } // namespace unicode
     64 } // namespace sys
     65 } // namespace llvm
     66 
     67 #endif
     68