2 Unicode character type helpers.
4 Written by Marc-Andre Lemburg (mal@lemburg.com).
5 Modified for Python 2.0 by Fredrik Lundh (fredrik@pythonware.com)
7 Copyright (c) Corporation for National Research Initiatives.
12 #include "unicodeobject.h"
14 #define ALPHA_MASK 0x01
15 #define DECIMAL_MASK 0x02
16 #define DIGIT_MASK 0x04
17 #define LOWER_MASK 0x08
18 #define LINEBREAK_MASK 0x10
19 #define SPACE_MASK 0x20
20 #define TITLE_MASK 0x40
21 #define UPPER_MASK 0x80
22 #define NODELTA_MASK 0x100
23 #define NUMERIC_MASK 0x200
26 const Py_UNICODE upper;
27 const Py_UNICODE lower;
28 const Py_UNICODE title;
29 const unsigned char decimal;
30 const unsigned char digit;
31 const unsigned short flags;
32 } _PyUnicode_TypeRecord;
34 #include "unicodetype_db.h"
36 static const _PyUnicode_TypeRecord *
37 gettyperecord(Py_UNICODE code)
41 #ifdef Py_UNICODE_WIDE
47 index = index1[(code>>SHIFT)];
48 index = index2[(index<<SHIFT)+(code&((1<<SHIFT)-1))];
51 return &_PyUnicode_TypeRecords[index];
54 /* Returns the titlecase Unicode characters corresponding to ch or just
55 ch if no titlecase mapping is known. */
57 Py_UNICODE _PyUnicode_ToTitlecase(register Py_UNICODE ch)
59 const _PyUnicode_TypeRecord *ctype = gettyperecord(ch);
60 int delta = ctype->title;
62 if (ctype->flags & NODELTA_MASK)
71 /* Returns 1 for Unicode characters having the category 'Lt', 0
74 int _PyUnicode_IsTitlecase(Py_UNICODE ch)
76 const _PyUnicode_TypeRecord *ctype = gettyperecord(ch);
78 return (ctype->flags & TITLE_MASK) != 0;
81 /* Returns the integer decimal (0-9) for Unicode characters having
82 this property, -1 otherwise. */
84 int _PyUnicode_ToDecimalDigit(Py_UNICODE ch)
86 const _PyUnicode_TypeRecord *ctype = gettyperecord(ch);
88 return (ctype->flags & DECIMAL_MASK) ? ctype->decimal : -1;
91 int _PyUnicode_IsDecimalDigit(Py_UNICODE ch)
93 if (_PyUnicode_ToDecimalDigit(ch) < 0)
98 /* Returns the integer digit (0-9) for Unicode characters having
99 this property, -1 otherwise. */
101 int _PyUnicode_ToDigit(Py_UNICODE ch)
103 const _PyUnicode_TypeRecord *ctype = gettyperecord(ch);
105 return (ctype->flags & DIGIT_MASK) ? ctype->digit : -1;
108 int _PyUnicode_IsDigit(Py_UNICODE ch)
110 if (_PyUnicode_ToDigit(ch) < 0)
115 /* Returns the numeric value as double for Unicode characters having
116 this property, -1.0 otherwise. */
118 int _PyUnicode_IsNumeric(Py_UNICODE ch)
120 const _PyUnicode_TypeRecord *ctype = gettyperecord(ch);
122 return (ctype->flags & NUMERIC_MASK) != 0;
125 #ifndef WANT_WCTYPE_FUNCTIONS
127 /* Returns 1 for Unicode characters having the category 'Ll', 0
130 int _PyUnicode_IsLowercase(Py_UNICODE ch)
132 const _PyUnicode_TypeRecord *ctype = gettyperecord(ch);
134 return (ctype->flags & LOWER_MASK) != 0;
137 /* Returns 1 for Unicode characters having the category 'Lu', 0
140 int _PyUnicode_IsUppercase(Py_UNICODE ch)
142 const _PyUnicode_TypeRecord *ctype = gettyperecord(ch);
144 return (ctype->flags & UPPER_MASK) != 0;
147 /* Returns the uppercase Unicode characters corresponding to ch or just
148 ch if no uppercase mapping is known. */
150 Py_UNICODE _PyUnicode_ToUppercase(Py_UNICODE ch)
152 const _PyUnicode_TypeRecord *ctype = gettyperecord(ch);
153 int delta = ctype->upper;
154 if (ctype->flags & NODELTA_MASK)
161 /* Returns the lowercase Unicode characters corresponding to ch or just
162 ch if no lowercase mapping is known. */
164 Py_UNICODE _PyUnicode_ToLowercase(Py_UNICODE ch)
166 const _PyUnicode_TypeRecord *ctype = gettyperecord(ch);
167 int delta = ctype->lower;
168 if (ctype->flags & NODELTA_MASK)
175 /* Returns 1 for Unicode characters having the category 'Ll', 'Lu', 'Lt',
176 'Lo' or 'Lm', 0 otherwise. */
178 int _PyUnicode_IsAlpha(Py_UNICODE ch)
180 const _PyUnicode_TypeRecord *ctype = gettyperecord(ch);
182 return (ctype->flags & ALPHA_MASK) != 0;
187 /* Export the interfaces using the wchar_t type for portability
190 int _PyUnicode_IsLowercase(Py_UNICODE ch)
195 int _PyUnicode_IsUppercase(Py_UNICODE ch)
200 Py_UNICODE _PyUnicode_ToLowercase(Py_UNICODE ch)
205 Py_UNICODE _PyUnicode_ToUppercase(Py_UNICODE ch)
210 int _PyUnicode_IsAlpha(Py_UNICODE ch)