Cutlery/calibre
1
0
Fork 0
mirror of https://github.com/kovidgoyal/calibre.git synced 2025-07-09 03:04:10 -04:00
Code Issues Packages Projects Releases Wiki Activity

Replace use of deprecated ICU unorm.h API

Browse Source
This commit is contained in:
Kovid Goyal 2018-05-01 09:48:44 +05:30
parent ff952ad851
commit 88e9494e6b
No known key found for this signature in database
GPG Key ID: 06BC317B515ACE7C
3 changed files with 92 additions and 67 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

149
src/calibre/utils/icu.c
View File

@ -50,7 +50,7 @@ icu_Collator_new(PyTypeObject *type, PyObject *args, PyObject *kwds)
if (!PyArg_ParseTuple(args, "s", &loc)) return NULL; if (!PyArg_ParseTuple(args, "s", &loc)) return NULL;
collator = ucol_open(loc, &status); collator = ucol_open(loc, &status);
if (collator == NULL || U_FAILURE(status)) { if (collator == NULL || U_FAILURE(status)) {
PyErr_SetString(PyExc_Exception, "Failed to create collator."); PyErr_SetString(PyExc_Exception, "Failed to create collator.");
return NULL; return NULL;
} }
@ -144,7 +144,7 @@ icu_Collator_sort_key(icu_Collator *self, PyObject *input) {
UChar *buf = NULL; UChar *buf = NULL;
uint8_t *buf2 = NULL; uint8_t *buf2 = NULL;
PyObject *ans = NULL; PyObject *ans = NULL;
buf = python_to_icu(input, &sz, 1); buf = python_to_icu(input, &sz, 1);
if (buf == NULL) return NULL; if (buf == NULL) return NULL;
@ -173,7 +173,7 @@ icu_Collator_strcmp(icu_Collator *self, PyObject *args) {
int32_t asz = 0, bsz = 0; int32_t asz = 0, bsz = 0;
UChar *a = NULL, *b = NULL; UChar *a = NULL, *b = NULL;
UCollationResult res = UCOL_EQUAL; UCollationResult res = UCOL_EQUAL;
if (!PyArg_ParseTuple(args, "OO", &a_, &b_)) return NULL; if (!PyArg_ParseTuple(args, "OO", &a_, &b_)) return NULL;
a = python_to_icu(a_, &asz, 1); a = python_to_icu(a_, &asz, 1);
@ -182,7 +182,7 @@ icu_Collator_strcmp(icu_Collator *self, PyObject *args) {
if (b == NULL) goto end; if (b == NULL) goto end;
res = ucol_strcoll(self->collator, a, asz, b, bsz); res = ucol_strcoll(self->collator, a, asz, b, bsz);
end: end:
if (a != NULL) free(a); if (a != NULL) free(a);
if (b != NULL) free(b); if (b != NULL) free(b);
return (PyErr_Occurred()) ? NULL : Py_BuildValue("i", res); return (PyErr_Occurred()) ? NULL : Py_BuildValue("i", res);
@ -191,7 +191,7 @@ end:
// Collator.find {{{ // Collator.find {{{
static PyObject * static PyObject *
icu_Collator_find(icu_Collator *self, PyObject *args) { icu_Collator_find(icu_Collator *self, PyObject *args) {
#if PY_VERSION_HEX >= 0x03030000 #if PY_VERSION_HEX >= 0x03030000
#error Not implemented for python >= 3.3 #error Not implemented for python >= 3.3
#endif #endif
PyObject *a_ = NULL, *b_ = NULL; PyObject *a_ = NULL, *b_ = NULL;
@ -199,7 +199,7 @@ icu_Collator_find(icu_Collator *self, PyObject *args) {
int32_t asz = 0, bsz = 0, pos = -1, length = -1; int32_t asz = 0, bsz = 0, pos = -1, length = -1;
UErrorCode status = U_ZERO_ERROR; UErrorCode status = U_ZERO_ERROR;
UStringSearch *search = NULL; UStringSearch *search = NULL;
if (!PyArg_ParseTuple(args, "OO", &a_, &b_)) return NULL; if (!PyArg_ParseTuple(args, "OO", &a_, &b_)) return NULL;
a = python_to_icu(a_, &asz, 1); a = python_to_icu(a_, &asz, 1);
@ -238,7 +238,7 @@ icu_Collator_contains(icu_Collator *self, PyObject *args) {
uint8_t found = 0; uint8_t found = 0;
UErrorCode status = U_ZERO_ERROR; UErrorCode status = U_ZERO_ERROR;
UStringSearch *search = NULL; UStringSearch *search = NULL;
if (!PyArg_ParseTuple(args, "OO", &a_, &b_)) return NULL; if (!PyArg_ParseTuple(args, "OO", &a_, &b_)) return NULL;
a = python_to_icu(a_, &asz, 1); a = python_to_icu(a_, &asz, 1);
@ -276,7 +276,7 @@ icu_Collator_contractions(icu_Collator *self, PyObject *args) {
if (self->contractions == NULL) return PyErr_NoMemory(); if (self->contractions == NULL) return PyErr_NoMemory();
self->contractions = ucol_getTailoredSet(self->collator, &status); self->contractions = ucol_getTailoredSet(self->collator, &status);
} }
status = U_ZERO_ERROR; status = U_ZERO_ERROR;
count = uset_getItemCount(self->contractions); count = uset_getItemCount(self->contractions);
str = (UChar*)calloc(100, sizeof(UChar)); str = (UChar*)calloc(100, sizeof(UChar));
@ -299,7 +299,7 @@ icu_Collator_contractions(icu_Collator *self, PyObject *args) {
} }
end: end:
if (str != NULL) free(str); if (str != NULL) free(str);
return ans; return ans;
} // }}} } // }}}
@ -310,7 +310,7 @@ icu_Collator_startswith(icu_Collator *self, PyObject *args) {
int32_t asz = 0, bsz = 0; int32_t asz = 0, bsz = 0;
UChar *a = NULL, *b = NULL; UChar *a = NULL, *b = NULL;
uint8_t ans = 0; uint8_t ans = 0;
if (!PyArg_ParseTuple(args, "OO", &a_, &b_)) return NULL; if (!PyArg_ParseTuple(args, "OO", &a_, &b_)) return NULL;
a = python_to_icu(a_, &asz, 1); a = python_to_icu(a_, &asz, 1);
@ -320,7 +320,7 @@ icu_Collator_startswith(icu_Collator *self, PyObject *args) {
if (asz < bsz) goto end; if (asz < bsz) goto end;
if (bsz == 0) { ans = 1; goto end; } if (bsz == 0) { ans = 1; goto end; }
ans = ucol_equal(self->collator, a, bsz, b, bsz); ans = ucol_equal(self->collator, a, bsz, b, bsz);
end: end:
@ -340,7 +340,7 @@ icu_Collator_collation_order(icu_Collator *self, PyObject *a_) {
UErrorCode status = U_ZERO_ERROR; UErrorCode status = U_ZERO_ERROR;
UCollationElements *iter = NULL; UCollationElements *iter = NULL;
int order = 0, len = -1; int order = 0, len = -1;
a = python_to_icu(a_, &asz, 1); a = python_to_icu(a_, &asz, 1);
if (a == NULL) goto end; if (a == NULL) goto end;
@ -420,17 +420,17 @@ static PyMethodDef icu_Collator_methods[] = {
}; };
static PyGetSetDef icu_Collator_getsetters[] = { static PyGetSetDef icu_Collator_getsetters[] = {
{(char *)"actual_locale", {(char *)"actual_locale",
(getter)icu_Collator_actual_locale, NULL, (getter)icu_Collator_actual_locale, NULL,
(char *)"Actual locale used by this collator.", (char *)"Actual locale used by this collator.",
NULL}, NULL},
{(char *)"capsule", {(char *)"capsule",
(getter)icu_Collator_capsule, NULL, (getter)icu_Collator_capsule, NULL,
(char *)"A capsule enclosing the pointer to the ICU collator struct", (char *)"A capsule enclosing the pointer to the ICU collator struct",
NULL}, NULL},
{(char *)"display_name", {(char *)"display_name",
(getter)icu_Collator_display_name, NULL, (getter)icu_Collator_display_name, NULL,
(char *)"Display name of this collator in English. The name reflects the actual data source used.", (char *)"Display name of this collator in English. The name reflects the actual data source used.",
NULL}, NULL},
@ -557,7 +557,7 @@ icu_BreakIterator_new(PyTypeObject *type, PyObject *args, PyObject *kwds)
if (!PyArg_ParseTuple(args, "is", &break_iterator_type, &locale)) return NULL; if (!PyArg_ParseTuple(args, "is", &break_iterator_type, &locale)) return NULL;
break_iterator = ubrk_open(break_iterator_type, locale, NULL, 0, &status); break_iterator = ubrk_open(break_iterator_type, locale, NULL, 0, &status);
if (break_iterator == NULL || U_FAILURE(status)) { if (break_iterator == NULL || U_FAILURE(status)) {
PyErr_SetString(PyExc_ValueError, u_errorName(status)); PyErr_SetString(PyExc_ValueError, u_errorName(status));
return NULL; return NULL;
} }
@ -577,7 +577,7 @@ icu_BreakIterator_set_text(icu_BreakIterator *self, PyObject *input) {
int32_t sz = 0; int32_t sz = 0;
UChar *buf = NULL; UChar *buf = NULL;
UErrorCode status = U_ZERO_ERROR; UErrorCode status = U_ZERO_ERROR;
buf = python_to_icu(input, &sz, 1); buf = python_to_icu(input, &sz, 1);
if (buf == NULL) return NULL; if (buf == NULL) return NULL;
ubrk_setText(self->break_iterator, buf, sz, &status); ubrk_setText(self->break_iterator, buf, sz, &status);
@ -595,13 +595,13 @@ icu_BreakIterator_set_text(icu_BreakIterator *self, PyObject *input) {
// BreakIterator.index {{{ // BreakIterator.index {{{
static PyObject * static PyObject *
icu_BreakIterator_index(icu_BreakIterator *self, PyObject *token) { icu_BreakIterator_index(icu_BreakIterator *self, PyObject *token) {
#if PY_VERSION_HEX >= 0x03030000 #if PY_VERSION_HEX >= 0x03030000
#error Not implemented for python >= 3.3 #error Not implemented for python >= 3.3
#endif #endif
UChar *buf = NULL, *needle = NULL; UChar *buf = NULL, *needle = NULL;
int32_t word_start = 0, p = 0, sz = 0, ans = -1, leading_hyphen = 0, trailing_hyphen = 0; int32_t word_start = 0, p = 0, sz = 0, ans = -1, leading_hyphen = 0, trailing_hyphen = 0;
buf = python_to_icu(token, &sz, 1); buf = python_to_icu(token, &sz, 1);
if (buf == NULL) return NULL; if (buf == NULL) return NULL;
if (sz < 1) goto end; if (sz < 1) goto end;
@ -613,7 +613,7 @@ icu_BreakIterator_index(icu_BreakIterator *self, PyObject *token) {
p = ubrk_first(self->break_iterator); p = ubrk_first(self->break_iterator);
while (p != UBRK_DONE) { while (p != UBRK_DONE) {
word_start = p; p = ubrk_next(self->break_iterator); word_start = p; p = ubrk_next(self->break_iterator);
if (self->type == UBRK_WORD && ubrk_getRuleStatus(self->break_iterator) == UBRK_WORD_NONE) if (self->type == UBRK_WORD && ubrk_getRuleStatus(self->break_iterator) == UBRK_WORD_NONE)
continue; // We are not at the start of a word continue; // We are not at the start of a word
if (self->text_len >= word_start + sz && memcmp(self->text + word_start, needle, sz * sizeof(UChar)) == 0) { if (self->text_len >= word_start + sz && memcmp(self->text + word_start, needle, sz * sizeof(UChar)) == 0) {
@ -655,7 +655,7 @@ end:
// BreakIterator.split2 {{{ // BreakIterator.split2 {{{
static PyObject * static PyObject *
icu_BreakIterator_split2(icu_BreakIterator *self, PyObject *args) { icu_BreakIterator_split2(icu_BreakIterator *self, PyObject *args) {
#if PY_VERSION_HEX >= 0x03030000 #if PY_VERSION_HEX >= 0x03030000
#error Not implemented for python >= 3.3 #error Not implemented for python >= 3.3
#endif #endif
@ -663,14 +663,14 @@ icu_BreakIterator_split2(icu_BreakIterator *self, PyObject *args) {
int is_hyphen_sep = 0, leading_hyphen = 0, trailing_hyphen = 0; int is_hyphen_sep = 0, leading_hyphen = 0, trailing_hyphen = 0;
UChar sep = 0; UChar sep = 0;
PyObject *ans = NULL, *temp = NULL, *t = NULL; PyObject *ans = NULL, *temp = NULL, *t = NULL;
ans = PyList_New(0); ans = PyList_New(0);
if (ans == NULL) return PyErr_NoMemory(); if (ans == NULL) return PyErr_NoMemory();
p = ubrk_first(self->break_iterator); p = ubrk_first(self->break_iterator);
while (p != UBRK_DONE) { while (p != UBRK_DONE) {
word_start = p; p = ubrk_next(self->break_iterator); word_start = p; p = ubrk_next(self->break_iterator);
if (self->type == UBRK_WORD && ubrk_getRuleStatus(self->break_iterator) == UBRK_WORD_NONE) if (self->type == UBRK_WORD && ubrk_getRuleStatus(self->break_iterator) == UBRK_WORD_NONE)
continue; // We are not at the start of a word continue; // We are not at the start of a word
sz = (p == UBRK_DONE) ? self->text_len - word_start : p - word_start; sz = (p == UBRK_DONE) ? self->text_len - word_start : p - word_start;
if (sz > 0) { if (sz > 0) {
@ -703,12 +703,12 @@ icu_BreakIterator_split2(icu_BreakIterator *self, PyObject *args) {
} else { } else {
sz += leading_hyphen + trailing_hyphen; sz += leading_hyphen + trailing_hyphen;
last_sz = sz; last_sz = sz;
temp = Py_BuildValue("ll", (long)(word_start - leading_hyphen), (long)sz); temp = Py_BuildValue("ll", (long)(word_start - leading_hyphen), (long)sz);
if (temp == NULL) { if (temp == NULL) {
Py_DECREF(ans); ans = NULL; break; Py_DECREF(ans); ans = NULL; break;
} }
if (PyList_Append(ans, temp) != 0) { if (PyList_Append(ans, temp) != 0) {
Py_DECREF(temp); Py_DECREF(ans); ans = NULL; break; Py_DECREF(temp); Py_DECREF(ans); ans = NULL; break;
} }
Py_DECREF(temp); Py_DECREF(temp);
} }
@ -912,18 +912,18 @@ icu_get_available_transliterators(PyObject *self, PyObject *args) {
// character_name {{{ // character_name {{{
static PyObject * static PyObject *
icu_character_name(PyObject *self, PyObject *args) { icu_character_name(PyObject *self, PyObject *args) {
char name[512] = {0}; char name[512] = {0};
int32_t sz = 0, alias = 0; int32_t sz = 0, alias = 0;
UChar *buf; UChar *buf;
UErrorCode status = U_ZERO_ERROR; UErrorCode status = U_ZERO_ERROR;
PyObject *palias = NULL, *result = NULL, *input = NULL; PyObject *palias = NULL, *result = NULL, *input = NULL;
UChar32 code = 0; UChar32 code = 0;
if (!PyArg_ParseTuple(args, "O|O", &input, &palias)) return NULL; if (!PyArg_ParseTuple(args, "O|O", &input, &palias)) return NULL;
if (palias != NULL && PyObject_IsTrue(palias)) alias = 1; if (palias != NULL && PyObject_IsTrue(palias)) alias = 1;
buf = python_to_icu(input, &sz, 1); buf = python_to_icu(input, &sz, 1);
if (buf == NULL) goto end; if (buf == NULL) goto end;
U16_GET(buf, 0, 0, sz, code); U16_GET(buf, 0, 0, sz, code);
if (alias) { if (alias) {
sz = u_charName(code, U_CHAR_NAME_ALIAS, name, 511, &status); sz = u_charName(code, U_CHAR_NAME_ALIAS, name, 511, &status);
@ -941,16 +941,16 @@ end:
// character_name_from_code {{{ // character_name_from_code {{{
static PyObject * static PyObject *
icu_character_name_from_code(PyObject *self, PyObject *args) { icu_character_name_from_code(PyObject *self, PyObject *args) {
char name[512] = {0}; char name[512] = {0};
int32_t sz, alias = 0; int32_t sz, alias = 0;
UErrorCode status = U_ZERO_ERROR; UErrorCode status = U_ZERO_ERROR;
PyObject *palias = NULL, *result = NULL; PyObject *palias = NULL, *result = NULL;
UChar32 code = 0; UChar32 code = 0;
if (!PyArg_ParseTuple(args, "I|O", &code, &palias)) return NULL; if (!PyArg_ParseTuple(args, "I|O", &code, &palias)) return NULL;
if (palias != NULL && PyObject_IsTrue(palias)) alias = 1; if (palias != NULL && PyObject_IsTrue(palias)) alias = 1;
if (alias) { if (alias) {
sz = u_charName(code, U_CHAR_NAME_ALIAS, name, 511, &status); sz = u_charName(code, U_CHAR_NAME_ALIAS, name, 511, &status);
} else { } else {
@ -969,7 +969,7 @@ icu_chr(PyObject *self, PyObject *args) {
UChar32 code = 0; UChar32 code = 0;
UChar buf[5] = {0}; UChar buf[5] = {0};
int32_t sz = 0; int32_t sz = 0;
if (!PyArg_ParseTuple(args, "I", &code)) return NULL; if (!PyArg_ParseTuple(args, "I", &code)) return NULL;
u_strFromUTF32(buf, 4, &sz, &code, 1, &status); u_strFromUTF32(buf, 4, &sz, &code, 1, &status);
@ -996,26 +996,50 @@ icu_ord_string(PyObject *self, PyObject *input) {
end: end:
if (input_buf != NULL) free(input_buf); if (input_buf != NULL) free(input_buf);
return ans; return ans;
} // }}} } // }}}
// normalize {{{ // normalize {{{
typedef enum { NFC, NFKC, NFD, NFKD } NORM_MODES;
static PyObject * static PyObject *
icu_normalize(PyObject *self, PyObject *args) { icu_normalize(PyObject *self, PyObject *args) {
UErrorCode status = U_ZERO_ERROR; UErrorCode status = U_ZERO_ERROR;
int32_t sz = 0, mode = UNORM_DEFAULT, cap = 0, rsz = 0; int32_t sz = 0, cap = 0, rsz = 0;
NORM_MODES mode;
UChar *dest = NULL, *source = NULL; UChar *dest = NULL, *source = NULL;
PyObject *ret = NULL, *src = NULL; PyObject *ret = NULL, *src = NULL;
if (!PyArg_ParseTuple(args, "iO", &mode, &src)) return NULL; if (!PyArg_ParseTuple(args, "iO", &mode, &src)) return NULL;
const UNormalizer2 *n = NULL;
switch (mode) {
case NFC:
n = unorm2_getNFCInstance(&status);
break;
case NFKC:
n = unorm2_getNFKCInstance(&status);
break;
case NFD:
n = unorm2_getNFDInstance(&status);
break;
case NFKD:
n = unorm2_getNFKDInstance(&status);
break;
}
if (U_FAILURE(status)) {
PyErr_SetString(PyExc_ValueError, u_errorName(status));
goto end;
}
source = python_to_icu(src, &sz, 1); source = python_to_icu(src, &sz, 1);
if (source == NULL) goto end; if (source == NULL) goto end;
cap = 2 * sz; cap = 2 * sz;
dest = (UChar*) calloc(cap, sizeof(UChar)); dest = (UChar*) calloc(cap, sizeof(UChar));
if (dest == NULL) { PyErr_NoMemory(); goto end; } if (dest == NULL) { PyErr_NoMemory(); goto end; }
while (1) { while (1) {
rsz = unorm_normalize(source, sz, (UNormalizationMode)mode, 0, dest, cap, &status); rsz = unorm2_normalize(n, source, sz, dest, cap, &status);
if (status == U_BUFFER_OVERFLOW_ERROR) { if (status == U_BUFFER_OVERFLOW_ERROR) {
cap *= 2; cap *= 2;
dest = (UChar*) realloc(dest, cap*sizeof(UChar)); dest = (UChar*) realloc(dest, cap*sizeof(UChar));
@ -1029,7 +1053,7 @@ icu_normalize(PyObject *self, PyObject *args) {
PyErr_SetString(PyExc_ValueError, u_errorName(status)); PyErr_SetString(PyExc_ValueError, u_errorName(status));
goto end; goto end;
} }
ret = icu_to_python(dest, rsz); ret = icu_to_python(dest, rsz);
end: end:
@ -1044,7 +1068,7 @@ icu_roundtrip(PyObject *self, PyObject *src) {
int32_t sz = 0; int32_t sz = 0;
UChar *icu = NULL; UChar *icu = NULL;
PyObject *ret = NULL; PyObject *ret = NULL;
icu = python_to_icu(src, &sz, 1); icu = python_to_icu(src, &sz, 1);
if (icu != NULL) { if (icu != NULL) {
ret = icu_to_python(icu, sz); ret = icu_to_python(icu, sz);
@ -1071,7 +1095,7 @@ icu_break_iterator_locales(PyObject *self, PyObject *args) {
PyTuple_SET_ITEM(ret, i, t); PyTuple_SET_ITEM(ret, i, t);
} }
} }
return ret; return ret;
} // }}} } // }}}
@ -1080,7 +1104,7 @@ static PyObject *
icu_string_length(PyObject *self, PyObject *src) { icu_string_length(PyObject *self, PyObject *src) {
int32_t sz = 0; int32_t sz = 0;
UChar *icu = NULL; UChar *icu = NULL;
icu = python_to_icu(src, &sz, 1); icu = python_to_icu(src, &sz, 1);
if (icu == NULL) return NULL; if (icu == NULL) return NULL;
sz = u_countChar32(icu, sz); sz = u_countChar32(icu, sz);
@ -1091,7 +1115,7 @@ icu_string_length(PyObject *self, PyObject *src) {
// utf16_length {{{ // utf16_length {{{
static PyObject * static PyObject *
icu_utf16_length(PyObject *self, PyObject *src) { icu_utf16_length(PyObject *self, PyObject *src) {
#if PY_VERSION_HEX >= 0x03030000 #if PY_VERSION_HEX >= 0x03030000
#error Not implemented for python >= 3.3 #error Not implemented for python >= 3.3
#endif #endif
@ -1100,7 +1124,7 @@ icu_utf16_length(PyObject *self, PyObject *src) {
int32_t i = 0, t = 0; int32_t i = 0, t = 0;
Py_UNICODE *data = NULL; Py_UNICODE *data = NULL;
#endif #endif
if (!PyUnicode_Check(src)) { PyErr_SetString(PyExc_TypeError, "Must be a unicode object"); return NULL; } if (!PyUnicode_Check(src)) { PyErr_SetString(PyExc_TypeError, "Must be a unicode object"); return NULL; }
sz = (int32_t)PyUnicode_GET_SIZE(src); sz = (int32_t)PyUnicode_GET_SIZE(src);
#ifdef Py_UNICODE_WIDE #ifdef Py_UNICODE_WIDE
@ -1135,39 +1159,39 @@ static PyMethodDef icu_methods[] = {
"get_available_transliterators() -> Return list of available transliterators. This list is rather limited on OS X." "get_available_transliterators() -> Return list of available transliterators. This list is rather limited on OS X."
}, },
{"character_name", icu_character_name, METH_VARARGS, {"character_name", icu_character_name, METH_VARARGS,
"character_name(char, alias=False) -> Return name for the first character in char, which must be a unicode string." "character_name(char, alias=False) -> Return name for the first character in char, which must be a unicode string."
}, },
{"character_name_from_code", icu_character_name_from_code, METH_VARARGS, {"character_name_from_code", icu_character_name_from_code, METH_VARARGS,
"character_name_from_code(code, alias=False) -> Return the name for the specified unicode code point" "character_name_from_code(code, alias=False) -> Return the name for the specified unicode code point"
}, },
{"chr", icu_chr, METH_VARARGS, {"chr", icu_chr, METH_VARARGS,
"chr(code) -> Return a python unicode string corresponding to the specified character code. The string can have length 1 or 2 (for non BMP codes on narrow python builds)." "chr(code) -> Return a python unicode string corresponding to the specified character code. The string can have length 1 or 2 (for non BMP codes on narrow python builds)."
}, },
{"ord_string", icu_ord_string, METH_O, {"ord_string", icu_ord_string, METH_O,
"ord_string(code) -> Convert a python unicode string to a tuple of unicode codepoints." "ord_string(code) -> Convert a python unicode string to a tuple of unicode codepoints."
}, },
{"normalize", icu_normalize, METH_VARARGS, {"normalize", icu_normalize, METH_VARARGS,
"normalize(mode, unicode_text) -> Return a python unicode string which is normalized in the specified mode." "normalize(mode, unicode_text) -> Return a python unicode string which is normalized in the specified mode."
}, },
{"roundtrip", icu_roundtrip, METH_O, {"roundtrip", icu_roundtrip, METH_O,
"roundtrip(string) -> Roundtrip a unicode object from python to ICU back to python (useful for testing)" "roundtrip(string) -> Roundtrip a unicode object from python to ICU back to python (useful for testing)"
}, },
{"available_locales_for_break_iterator", icu_break_iterator_locales, METH_NOARGS, {"available_locales_for_break_iterator", icu_break_iterator_locales, METH_NOARGS,
"available_locales_for_break_iterator() -> Return tuple of all available locales for the BreakIterator" "available_locales_for_break_iterator() -> Return tuple of all available locales for the BreakIterator"
}, },
{"string_length", icu_string_length, METH_O, {"string_length", icu_string_length, METH_O,
"string_length(string) -> Return the length of a string (number of unicode code points in the string). Useful on narrow python builds where len() returns an incorrect answer if the string contains surrogate pairs." "string_length(string) -> Return the length of a string (number of unicode code points in the string). Useful on narrow python builds where len() returns an incorrect answer if the string contains surrogate pairs."
}, },
{"utf16_length", icu_utf16_length, METH_O, {"utf16_length", icu_utf16_length, METH_O,
"utf16_length(string) -> Return the length of a string (number of UTF-16 code points in the string). Useful on wide python builds where len() returns an incorrect answer if the string contains surrogate pairs." "utf16_length(string) -> Return the length of a string (number of UTF-16 code points in the string). Useful on wide python builds where len() returns an incorrect answer if the string contains surrogate pairs."
}, },
@ -1177,7 +1201,7 @@ static PyMethodDef icu_methods[] = {
#define ADDUCONST(x) PyModule_AddIntConstant(m, #x, x) #define ADDUCONST(x) PyModule_AddIntConstant(m, #x, x)
CALIBRE_MODINIT_FUNC CALIBRE_MODINIT_FUNC
initicu(void) initicu(void)
{ {
PyObject* m; PyObject* m;
UVersionInfo ver, uver; UVersionInfo ver, uver;
@ -1232,13 +1256,10 @@ initicu(void)
ADDUCONST(UCOL_LOWER_FIRST); ADDUCONST(UCOL_LOWER_FIRST);
ADDUCONST(UCOL_UPPER_FIRST); ADDUCONST(UCOL_UPPER_FIRST);
ADDUCONST(UNORM_NONE); ADDUCONST(NFD);
ADDUCONST(UNORM_NFD); ADDUCONST(NFKD);
ADDUCONST(UNORM_NFKD); ADDUCONST(NFC);
ADDUCONST(UNORM_NFC); ADDUCONST(NFKC);
ADDUCONST(UNORM_DEFAULT);
ADDUCONST(UNORM_NFKC);
ADDUCONST(UNORM_FCD);
ADDUCONST(UPPER_CASE); ADDUCONST(UPPER_CASE);
ADDUCONST(LOWER_CASE); ADDUCONST(LOWER_CASE);

8
src/calibre/utils/icu.py
View File

@ -28,7 +28,7 @@ if _icu is None:
raise RuntimeError('Failed to load icu with error: %s' % err) raise RuntimeError('Failed to load icu with error: %s' % err)
del err del err
icu_unicode_version = getattr(_icu, 'unicode_version', None) icu_unicode_version = getattr(_icu, 'unicode_version', None)
_nmodes = {m:getattr(_icu, 'UNORM_'+m, None) for m in ('NFC', 'NFD', 'NFKC', 'NFKD', 'NONE', 'DEFAULT', 'FCD')} _nmodes = {m:getattr(_icu, m) for m in ('NFC', 'NFD', 'NFKC', 'NFKD')}
# Ensure that the python internal filesystem and default encodings are not ASCII # Ensure that the python internal filesystem and default encodings are not ASCII
@ -38,6 +38,8 @@ def is_ascii(name):
return codecs.lookup(name).name == b'ascii' return codecs.lookup(name).name == b'ascii'
except (TypeError, LookupError): except (TypeError, LookupError):
return True return True
try: try:
if is_ascii(sys.getdefaultencoding()): if is_ascii(sys.getdefaultencoding()):
_icu.set_default_encoding(b'utf-8') _icu.set_default_encoding(b'utf-8')
@ -119,6 +121,7 @@ def case_sensitive_collator():
# function implementations based on different collators, to allow lazy loading # function implementations based on different collators, to allow lazy loading
# of collators, with maximum runtime performance # of collators, with maximum runtime performance
_sort_key_template = ''' _sort_key_template = '''
def {name}(obj): def {name}(obj):
try: try:
@ -222,6 +225,7 @@ def capitalize(x):
except (IndexError, TypeError, AttributeError): except (IndexError, TypeError, AttributeError):
return x return x
try: try:
swapcase = _icu.swap_case swapcase = _icu.swap_case
except AttributeError: # For people running from source except AttributeError: # For people running from source
@ -300,6 +304,7 @@ def partition_by_first_letter(items, reverse=False, key=lambda x:x):
ans[last_c] = [item] ans[last_c] = [item]
return ans return ans
# Return the number of unicode codepoints in a string # Return the number of unicode codepoints in a string
string_length = _icu.string_length if is_narrow_build else len string_length = _icu.string_length if is_narrow_build else len
@ -311,4 +316,3 @@ utf16_length = len if is_narrow_build else _icu.utf16_length
if __name__ == '__main__': if __name__ == '__main__':
from calibre.utils.icu_test import run from calibre.utils.icu_test import run
run(verbosity=4) run(verbosity=4)

2
src/calibre/utils/icu_calibre_utils.h
View File

@ -19,7 +19,7 @@
#include <unicode/ustring.h> #include <unicode/ustring.h>
#include <unicode/usearch.h> #include <unicode/usearch.h>
#include <unicode/utrans.h> #include <unicode/utrans.h>
#include <unicode/unorm.h> #include <unicode/unorm2.h>
#include <unicode/ubrk.h> #include <unicode/ubrk.h>
#if PY_VERSION_HEX >= 0x03030000 #if PY_VERSION_HEX >= 0x03030000