Mercurial > genshi > genshi-test
annotate genshi/_speedups.c @ 784:67d324a62cc0 experimental-match-fastpaths
update to 0.5.x branch, up through r907
don't know how this fits in with SoC work, but I wanted to do due diligence and keep this branch working in case it someday gets considered for trunk
| author | aflett |
|---|---|
| date | Mon, 21 Jul 2008 23:17:52 +0000 |
| parents | 8f2c7023af94 |
| children |
| rev | line source |
|---|---|
| 541 | 1 /* |
|
724
8f2c7023af94
update to trunk to track r847, fixing python 2.4 compatibility issues in speedup (and fixing copyrights, apparently :))
aflett
parents:
718
diff
changeset
|
2 * Copyright (C) 2006-2008 Edgewall Software |
| 541 | 3 * All rights reserved. |
| 4 * | |
| 5 * This software is licensed as described in the file COPYING, which | |
| 6 * you should have received as part of this distribution. The terms | |
| 7 * are also available at http://genshi.edgewall.org/wiki/License. | |
| 8 * | |
| 9 * This software consists of voluntary contributions made by many | |
| 10 * individuals. For the exact contribution history, see the revision | |
| 11 * history and logs, available at http://genshi.edgewall.org/log/. | |
| 12 */ | |
| 13 | |
| 14 #include <Python.h> | |
| 15 #include <structmember.h> | |
| 16 | |
|
724
8f2c7023af94
update to trunk to track r847, fixing python 2.4 compatibility issues in speedup (and fixing copyrights, apparently :))
aflett
parents:
718
diff
changeset
|
17 #if PY_VERSION_HEX < 0x02050000 && !defined(PY_SSIZE_T_MIN) |
|
8f2c7023af94
update to trunk to track r847, fixing python 2.4 compatibility issues in speedup (and fixing copyrights, apparently :))
aflett
parents:
718
diff
changeset
|
18 typedef int Py_ssize_t; |
|
8f2c7023af94
update to trunk to track r847, fixing python 2.4 compatibility issues in speedup (and fixing copyrights, apparently :))
aflett
parents:
718
diff
changeset
|
19 #define PY_SSIZE_T_MAX INT_MAX |
|
8f2c7023af94
update to trunk to track r847, fixing python 2.4 compatibility issues in speedup (and fixing copyrights, apparently :))
aflett
parents:
718
diff
changeset
|
20 #define PY_SSIZE_T_MIN INT_MIN |
|
8f2c7023af94
update to trunk to track r847, fixing python 2.4 compatibility issues in speedup (and fixing copyrights, apparently :))
aflett
parents:
718
diff
changeset
|
21 #endif |
|
8f2c7023af94
update to trunk to track r847, fixing python 2.4 compatibility issues in speedup (and fixing copyrights, apparently :))
aflett
parents:
718
diff
changeset
|
22 |
| 541 | 23 static PyObject *amp1, *amp2, *lt1, *lt2, *gt1, *gt2, *qt1, *qt2; |
| 24 static PyObject *stripentities, *striptags; | |
| 25 | |
| 26 static void | |
| 27 init_constants(void) | |
| 28 { | |
| 29 PyObject *util = PyImport_ImportModule("genshi.util"); | |
| 30 stripentities = PyObject_GetAttrString(util, "stripentities"); | |
| 31 striptags = PyObject_GetAttrString(util, "striptags"); | |
| 32 Py_DECREF(util); | |
| 33 | |
| 34 amp1 = PyUnicode_DecodeASCII("&", 1, NULL); | |
| 35 amp2 = PyUnicode_DecodeASCII("&", 5, NULL); | |
| 36 lt1 = PyUnicode_DecodeASCII("<", 1, NULL); | |
| 37 lt2 = PyUnicode_DecodeASCII("<", 4, NULL); | |
| 38 gt1 = PyUnicode_DecodeASCII(">", 1, NULL); | |
| 39 gt2 = PyUnicode_DecodeASCII(">", 4, NULL); | |
| 40 qt1 = PyUnicode_DecodeASCII("\"", 1, NULL); | |
| 41 qt2 = PyUnicode_DecodeASCII(""", 5, NULL); | |
| 42 } | |
| 43 | |
| 44 /* Markup class */ | |
| 45 | |
| 46 PyAPI_DATA(PyTypeObject) MarkupType; | |
| 47 | |
| 48 PyDoc_STRVAR(Markup__doc__, | |
| 49 "Marks a string as being safe for inclusion in HTML/XML output without\n\ | |
| 50 needing to be escaped."); | |
| 51 | |
| 52 static PyObject * | |
| 53 escape(PyObject *text, int quotes) | |
| 54 { | |
| 55 PyObject *args, *ret; | |
| 56 PyUnicodeObject *in, *out; | |
| 57 Py_UNICODE *inp, *outp; | |
| 58 int len, inn, outn; | |
| 59 | |
| 60 if (PyObject_TypeCheck(text, &MarkupType)) { | |
| 61 Py_INCREF(text); | |
| 62 return text; | |
| 63 } | |
| 784 | 64 if (PyObject_HasAttrString(text, "__html__")) { |
| 65 ret = PyObject_CallMethod(text, "__html__", NULL); | |
| 66 args = PyTuple_New(1); | |
| 67 if (args == NULL) { | |
| 68 Py_DECREF(ret); | |
| 69 return NULL; | |
| 70 } | |
| 71 PyTuple_SET_ITEM(args, 0, ret); | |
| 72 ret = MarkupType.tp_new(&MarkupType, args, NULL); | |
| 73 Py_DECREF(args); | |
| 74 return ret; | |
| 75 } | |
| 541 | 76 in = (PyUnicodeObject *) PyObject_Unicode(text); |
| 77 if (in == NULL) { | |
| 78 return NULL; | |
| 79 } | |
| 80 /* First we need to figure out how long the escaped string will be */ | |
| 81 len = inn = 0; | |
| 82 inp = in->str; | |
| 83 while (*(inp) || in->length > inp - in->str) { | |
| 84 switch (*inp++) { | |
| 85 case '&': len += 5; inn++; break; | |
| 86 case '"': len += quotes ? 5 : 1; inn += quotes ? 1 : 0; break; | |
| 87 case '<': | |
| 88 case '>': len += 4; inn++; break; | |
| 89 default: len++; | |
| 90 } | |
| 91 } | |
| 92 | |
| 93 /* Do we need to escape anything at all? */ | |
| 94 if (!inn) { | |
| 95 args = PyTuple_New(1); | |
| 96 if (args == NULL) { | |
| 97 Py_DECREF((PyObject *) in); | |
| 98 return NULL; | |
| 99 } | |
| 100 PyTuple_SET_ITEM(args, 0, (PyObject *) in); | |
| 101 ret = MarkupType.tp_new(&MarkupType, args, NULL); | |
| 102 Py_DECREF(args); | |
| 103 return ret; | |
| 104 } | |
| 105 | |
| 106 out = (PyUnicodeObject*) PyUnicode_FromUnicode(NULL, len); | |
| 107 if (out == NULL) { | |
|
665
281ab9fb4b04
Applied patch to fix a memory leak in the C implementation of the `Markup.escape()` function. Thanks to Christian Boos for reporting and figuring out the problem. Closes #166.
cmlenz
parents:
547
diff
changeset
|
108 Py_DECREF((PyObject *) in); |
| 541 | 109 return NULL; |
| 110 } | |
| 111 | |
| 112 outn = 0; | |
| 113 inp = in->str; | |
| 114 outp = out->str; | |
| 115 while (*(inp) || in->length > inp - in->str) { | |
| 116 if (outn == inn) { | |
| 117 /* copy rest of string if we have already replaced everything */ | |
| 118 Py_UNICODE_COPY(outp, inp, in->length - (inp - in->str)); | |
| 119 break; | |
| 120 } | |
| 121 switch (*inp) { | |
| 122 case '&': | |
| 123 Py_UNICODE_COPY(outp, ((PyUnicodeObject *) amp2)->str, 5); | |
| 124 outp += 5; | |
| 125 outn++; | |
| 126 break; | |
| 127 case '"': | |
| 128 if (quotes) { | |
| 129 Py_UNICODE_COPY(outp, ((PyUnicodeObject *) qt2)->str, 5); | |
| 130 outp += 5; | |
| 131 outn++; | |
| 132 } else { | |
| 133 *outp++ = *inp; | |
| 134 } | |
| 135 break; | |
| 136 case '<': | |
| 137 Py_UNICODE_COPY(outp, ((PyUnicodeObject *) lt2)->str, 4); | |
| 138 outp += 4; | |
| 139 outn++; | |
| 140 break; | |
| 141 case '>': | |
| 142 Py_UNICODE_COPY(outp, ((PyUnicodeObject *) gt2)->str, 4); | |
| 143 outp += 4; | |
| 144 outn++; | |
| 145 break; | |
| 146 default: | |
| 147 *outp++ = *inp; | |
| 148 } | |
| 149 inp++; | |
| 150 } | |
| 151 | |
|
665
281ab9fb4b04
Applied patch to fix a memory leak in the C implementation of the `Markup.escape()` function. Thanks to Christian Boos for reporting and figuring out the problem. Closes #166.
cmlenz
parents:
547
diff
changeset
|
152 Py_DECREF((PyObject *) in); |
|
281ab9fb4b04
Applied patch to fix a memory leak in the C implementation of the `Markup.escape()` function. Thanks to Christian Boos for reporting and figuring out the problem. Closes #166.
cmlenz
parents:
547
diff
changeset
|
153 |
| 541 | 154 args = PyTuple_New(1); |
| 155 if (args == NULL) { | |
| 156 Py_DECREF((PyObject *) out); | |
| 157 return NULL; | |
| 158 } | |
| 159 PyTuple_SET_ITEM(args, 0, (PyObject *) out); | |
| 160 ret = MarkupType.tp_new(&MarkupType, args, NULL); | |
| 161 Py_DECREF(args); | |
| 162 return ret; | |
| 163 } | |
| 164 | |
| 165 PyDoc_STRVAR(escape__doc__, | |
| 166 "Create a Markup instance from a string and escape special characters\n\ | |
| 167 it may contain (<, >, & and \").\n\ | |
| 168 \n\ | |
| 547 | 169 >>> escape('\"1 < 2\"')\n\ |
| 170 <Markup u'"1 < 2"'>\n\ | |
| 171 \n\ | |
| 541 | 172 If the `quotes` parameter is set to `False`, the \" character is left\n\ |
| 173 as is. Escaping quotes is generally only required for strings that are\n\ | |
| 547 | 174 to be used in attribute values.\n\ |
| 175 \n\ | |
| 176 >>> escape('\"1 < 2\"', quotes=False)\n\ | |
| 177 <Markup u'\"1 < 2\"'>\n\ | |
| 178 \n\ | |
| 179 :param text: the text to escape\n\ | |
| 180 :param quotes: if ``True``, double quote characters are escaped in\n\ | |
| 181 addition to the other special characters\n\ | |
| 182 :return: the escaped `Markup` string\n\ | |
| 183 :rtype: `Markup`\n\ | |
| 184 "); | |
| 541 | 185 |
| 186 static PyObject * | |
| 187 Markup_escape(PyTypeObject* type, PyObject *args, PyObject *kwds) | |
| 188 { | |
| 189 static char *kwlist[] = {"text", "quotes", 0}; | |
| 190 PyObject *text = NULL; | |
| 191 char quotes = 1; | |
| 192 | |
| 193 if (!PyArg_ParseTupleAndKeywords(args, kwds, "O|b", kwlist, &text, "es)) { | |
| 194 return NULL; | |
| 195 } | |
| 196 if (PyObject_Not(text)) { | |
| 197 return type->tp_new(type, args, NULL); | |
| 198 } | |
| 199 if (PyObject_TypeCheck(text, type)) { | |
| 200 Py_INCREF(text); | |
| 201 return text; | |
| 202 } | |
| 203 return escape(text, quotes); | |
| 204 } | |
| 205 | |
| 784 | 206 static PyObject * |
| 207 Markup_html(PyObject *self) | |
| 208 { | |
| 209 Py_INCREF(self); | |
| 210 return self; | |
| 211 } | |
| 212 | |
| 547 | 213 PyDoc_STRVAR(join__doc__, |
| 214 "Return a `Markup` object which is the concatenation of the strings\n\ | |
| 215 in the given sequence, where this `Markup` object is the separator\n\ | |
| 216 between the joined elements.\n\ | |
| 217 \n\ | |
| 218 Any element in the sequence that is not a `Markup` instance is\n\ | |
| 219 automatically escaped.\n\ | |
| 220 \n\ | |
| 221 :param seq: the sequence of strings to join\n\ | |
| 222 :param escape_quotes: whether double quote characters in the elements\n\ | |
| 223 should be escaped\n\ | |
| 224 :return: the joined `Markup` object\n\ | |
| 225 :rtype: `Markup`\n\ | |
| 226 :see: `escape`\n\ | |
| 227 "); | |
| 228 | |
| 541 | 229 static PyObject * |
| 230 Markup_join(PyObject *self, PyObject *args, PyObject *kwds) | |
| 231 { | |
| 232 static char *kwlist[] = {"seq", "escape_quotes", 0}; | |
|
665
281ab9fb4b04
Applied patch to fix a memory leak in the C implementation of the `Markup.escape()` function. Thanks to Christian Boos for reporting and figuring out the problem. Closes #166.
cmlenz
parents:
547
diff
changeset
|
233 PyObject *seq = NULL, *seq2, *tmp, *tmp2; |
| 541 | 234 char quotes = 1; |
| 235 int n, i; | |
| 236 | |
| 237 if (!PyArg_ParseTupleAndKeywords(args, kwds, "O|b", kwlist, &seq, "es)) { | |
| 238 return NULL; | |
| 239 } | |
| 240 if (!PySequence_Check(seq)) { | |
| 241 return NULL; | |
| 242 } | |
| 243 n = PySequence_Size(seq); | |
| 244 if (n < 0) { | |
| 245 return NULL; | |
| 246 } | |
| 247 seq2 = PyTuple_New(n); | |
| 248 if (seq2 == NULL) { | |
| 249 return NULL; | |
| 250 } | |
| 251 for (i = 0; i < n; i++) { | |
| 252 tmp = PySequence_GetItem(seq, i); | |
| 253 if (tmp == NULL) { | |
| 254 Py_DECREF(seq2); | |
| 255 return NULL; | |
| 256 } | |
|
665
281ab9fb4b04
Applied patch to fix a memory leak in the C implementation of the `Markup.escape()` function. Thanks to Christian Boos for reporting and figuring out the problem. Closes #166.
cmlenz
parents:
547
diff
changeset
|
257 tmp2 = escape(tmp, quotes); |
|
281ab9fb4b04
Applied patch to fix a memory leak in the C implementation of the `Markup.escape()` function. Thanks to Christian Boos for reporting and figuring out the problem. Closes #166.
cmlenz
parents:
547
diff
changeset
|
258 if (tmp2 == NULL) { |
| 541 | 259 Py_DECREF(seq2); |
| 260 return NULL; | |
| 261 } | |
|
665
281ab9fb4b04
Applied patch to fix a memory leak in the C implementation of the `Markup.escape()` function. Thanks to Christian Boos for reporting and figuring out the problem. Closes #166.
cmlenz
parents:
547
diff
changeset
|
262 PyTuple_SET_ITEM(seq2, i, tmp2); |
|
281ab9fb4b04
Applied patch to fix a memory leak in the C implementation of the `Markup.escape()` function. Thanks to Christian Boos for reporting and figuring out the problem. Closes #166.
cmlenz
parents:
547
diff
changeset
|
263 Py_DECREF(tmp); |
| 541 | 264 } |
| 265 tmp = PyUnicode_Join(self, seq2); | |
| 266 Py_DECREF(seq2); | |
| 267 if (tmp == NULL) | |
| 268 return NULL; | |
| 269 args = PyTuple_New(1); | |
| 270 if (args == NULL) { | |
| 271 Py_DECREF(tmp); | |
| 272 return NULL; | |
| 273 } | |
| 274 PyTuple_SET_ITEM(args, 0, tmp); | |
| 275 tmp = MarkupType.tp_new(&MarkupType, args, NULL); | |
| 276 Py_DECREF(args); | |
| 277 return tmp; | |
| 278 } | |
| 279 | |
| 280 static PyObject * | |
| 281 Markup_add(PyObject *self, PyObject *other) | |
| 282 { | |
| 283 PyObject *tmp, *tmp2, *args, *ret; | |
| 284 if (PyObject_TypeCheck(self, &MarkupType)) { | |
| 285 tmp = escape(other, 1); | |
| 286 if (tmp == NULL) | |
| 287 return NULL; | |
| 288 tmp2 = PyUnicode_Concat(self, tmp); | |
| 289 } else { // __radd__ | |
| 290 tmp = escape(self, 1); | |
| 291 if (tmp == NULL) | |
| 292 return NULL; | |
| 293 tmp2 = PyUnicode_Concat(tmp, other); | |
| 294 } | |
|
665
281ab9fb4b04
Applied patch to fix a memory leak in the C implementation of the `Markup.escape()` function. Thanks to Christian Boos for reporting and figuring out the problem. Closes #166.
cmlenz
parents:
547
diff
changeset
|
295 Py_DECREF(tmp); |
|
281ab9fb4b04
Applied patch to fix a memory leak in the C implementation of the `Markup.escape()` function. Thanks to Christian Boos for reporting and figuring out the problem. Closes #166.
cmlenz
parents:
547
diff
changeset
|
296 if (tmp2 == NULL) |
| 541 | 297 return NULL; |
| 298 args = PyTuple_New(1); | |
| 299 if (args == NULL) { | |
| 300 Py_DECREF(tmp2); | |
| 301 return NULL; | |
| 302 } | |
| 303 PyTuple_SET_ITEM(args, 0, tmp2); | |
| 304 ret = MarkupType.tp_new(&MarkupType, args, NULL); | |
| 305 Py_DECREF(args); | |
| 306 return ret; | |
| 307 } | |
| 308 | |
| 309 static PyObject * | |
| 310 Markup_mod(PyObject *self, PyObject *args) | |
| 311 { | |
| 312 PyObject *tmp, *tmp2, *ret, *args2; | |
| 718 | 313 int i, nargs = 0; |
| 314 PyObject *kwds = NULL; | |
| 541 | 315 |
| 718 | 316 if (PyDict_Check(args)) { |
| 317 kwds = args; | |
| 318 } | |
| 319 if (kwds && PyDict_Size(kwds)) { | |
| 320 PyObject *kwcopy, *key, *value; | |
|
724
8f2c7023af94
update to trunk to track r847, fixing python 2.4 compatibility issues in speedup (and fixing copyrights, apparently :))
aflett
parents:
718
diff
changeset
|
321 #if PY_VERSION_HEX >= 0x02050000 |
| 718 | 322 Py_ssize_t pos = 0; |
|
724
8f2c7023af94
update to trunk to track r847, fixing python 2.4 compatibility issues in speedup (and fixing copyrights, apparently :))
aflett
parents:
718
diff
changeset
|
323 #else |
|
8f2c7023af94
update to trunk to track r847, fixing python 2.4 compatibility issues in speedup (and fixing copyrights, apparently :))
aflett
parents:
718
diff
changeset
|
324 int pos = 0; |
|
8f2c7023af94
update to trunk to track r847, fixing python 2.4 compatibility issues in speedup (and fixing copyrights, apparently :))
aflett
parents:
718
diff
changeset
|
325 #endif |
| 718 | 326 |
| 327 kwcopy = PyDict_Copy( kwds ); | |
| 328 if (kwcopy == NULL) { | |
| 329 return NULL; | |
| 330 } | |
| 331 while (PyDict_Next(kwcopy, &pos, &key, &value)) { | |
| 332 tmp = escape(value, 1); | |
| 333 if (tmp == NULL) { | |
| 334 Py_DECREF(kwcopy); | |
| 335 return NULL; | |
| 336 } | |
| 337 if (PyDict_SetItem(kwcopy, key, tmp) < 0) { | |
| 338 Py_DECREF(tmp); | |
| 339 Py_DECREF(kwcopy); | |
| 340 return NULL; | |
| 341 } | |
| 342 } | |
| 343 tmp = PyUnicode_Format(self, kwcopy); | |
| 344 Py_DECREF(kwcopy); | |
| 345 if (tmp == NULL) { | |
| 346 return NULL; | |
| 347 } | |
| 348 } else if (PyTuple_Check(args)) { | |
| 541 | 349 nargs = PyTuple_GET_SIZE(args); |
| 350 args2 = PyTuple_New(nargs); | |
| 351 if (args2 == NULL) { | |
| 352 return NULL; | |
| 353 } | |
| 354 for (i = 0; i < nargs; i++) { | |
| 355 tmp = escape(PyTuple_GET_ITEM(args, i), 1); | |
| 356 if (tmp == NULL) { | |
| 357 Py_DECREF(args2); | |
| 358 return NULL; | |
| 359 } | |
| 360 PyTuple_SET_ITEM(args2, i, tmp); | |
| 361 } | |
| 362 tmp = PyUnicode_Format(self, args2); | |
| 363 Py_DECREF(args2); | |
| 364 if (tmp == NULL) { | |
| 365 return NULL; | |
| 366 } | |
| 367 } else { | |
| 368 tmp2 = escape(args, 1); | |
| 369 if (tmp2 == NULL) { | |
| 370 return NULL; | |
| 371 } | |
| 372 tmp = PyUnicode_Format(self, tmp2); | |
| 373 Py_DECREF(tmp2); | |
| 374 if (tmp == NULL) { | |
| 375 return NULL; | |
| 376 } | |
| 377 } | |
| 378 args = PyTuple_New(1); | |
| 379 if (args == NULL) { | |
| 380 Py_DECREF(tmp); | |
| 381 return NULL; | |
| 382 } | |
| 383 PyTuple_SET_ITEM(args, 0, tmp); | |
| 384 ret = PyUnicode_Type.tp_new(&MarkupType, args, NULL); | |
| 385 Py_DECREF(args); | |
| 386 return ret; | |
| 387 } | |
| 388 | |
| 389 static PyObject * | |
| 390 Markup_mul(PyObject *self, PyObject *num) | |
| 391 { | |
| 392 PyObject *unicode, *result, *args; | |
| 393 | |
| 394 if (PyObject_TypeCheck(self, &MarkupType)) { | |
| 395 unicode = PyObject_Unicode(self); | |
| 396 if (unicode == NULL) return NULL; | |
| 397 result = PyNumber_Multiply(unicode, num); | |
| 398 } else { // __rmul__ | |
| 399 unicode = PyObject_Unicode(num); | |
| 400 if (unicode == NULL) return NULL; | |
| 401 result = PyNumber_Multiply(unicode, self); | |
| 402 } | |
|
665
281ab9fb4b04
Applied patch to fix a memory leak in the C implementation of the `Markup.escape()` function. Thanks to Christian Boos for reporting and figuring out the problem. Closes #166.
cmlenz
parents:
547
diff
changeset
|
403 Py_DECREF(unicode); |
| 541 | 404 |
| 405 if (result == NULL) return NULL; | |
| 406 args = PyTuple_New(1); | |
| 407 if (args == NULL) { | |
| 408 Py_DECREF(result); | |
| 409 return NULL; | |
| 410 } | |
| 411 PyTuple_SET_ITEM(args, 0, result); | |
| 412 result = PyUnicode_Type.tp_new(&MarkupType, args, NULL); | |
| 413 Py_DECREF(args); | |
| 414 | |
| 415 return result; | |
| 416 } | |
| 417 | |
| 418 static PyObject * | |
| 419 Markup_repr(PyObject *self) | |
| 420 { | |
| 421 PyObject *format, *result, *args; | |
| 422 | |
| 423 format = PyString_FromString("<Markup %r>"); | |
| 424 if (format == NULL) return NULL; | |
| 425 result = PyObject_Unicode(self); | |
|
665
281ab9fb4b04
Applied patch to fix a memory leak in the C implementation of the `Markup.escape()` function. Thanks to Christian Boos for reporting and figuring out the problem. Closes #166.
cmlenz
parents:
547
diff
changeset
|
426 if (result == NULL) { |
|
281ab9fb4b04
Applied patch to fix a memory leak in the C implementation of the `Markup.escape()` function. Thanks to Christian Boos for reporting and figuring out the problem. Closes #166.
cmlenz
parents:
547
diff
changeset
|
427 Py_DECREF(format); |
|
281ab9fb4b04
Applied patch to fix a memory leak in the C implementation of the `Markup.escape()` function. Thanks to Christian Boos for reporting and figuring out the problem. Closes #166.
cmlenz
parents:
547
diff
changeset
|
428 return NULL; |
|
281ab9fb4b04
Applied patch to fix a memory leak in the C implementation of the `Markup.escape()` function. Thanks to Christian Boos for reporting and figuring out the problem. Closes #166.
cmlenz
parents:
547
diff
changeset
|
429 } |
| 541 | 430 args = PyTuple_New(1); |
| 431 if (args == NULL) { | |
|
665
281ab9fb4b04
Applied patch to fix a memory leak in the C implementation of the `Markup.escape()` function. Thanks to Christian Boos for reporting and figuring out the problem. Closes #166.
cmlenz
parents:
547
diff
changeset
|
432 Py_DECREF(format); |
| 541 | 433 Py_DECREF(result); |
| 434 return NULL; | |
| 435 } | |
| 436 PyTuple_SET_ITEM(args, 0, result); | |
| 437 result = PyString_Format(format, args); | |
|
681
ae5378f81398
Fix another memory leak in the C speedups code. Thanks to Erik Bray for finding this one and providing a patch. Closes #166 (again).
cmlenz
parents:
665
diff
changeset
|
438 Py_DECREF(format); |
| 541 | 439 Py_DECREF(args); |
| 440 return result; | |
| 441 } | |
| 442 | |
| 443 PyDoc_STRVAR(unescape__doc__, | |
| 547 | 444 "Reverse-escapes &, <, >, and \" and returns a `unicode` object.\n\ |
| 445 \n\ | |
| 446 >>> Markup('1 < 2').unescape()\n\ | |
| 447 u'1 < 2'\n\ | |
| 448 \n\ | |
| 449 :return: the unescaped string\n\ | |
| 450 :rtype: `unicode`\n\ | |
| 451 :see: `genshi.core.unescape`\n\ | |
| 452 "); | |
| 541 | 453 |
| 454 static PyObject * | |
| 455 Markup_unescape(PyObject* self) | |
| 456 { | |
| 457 PyObject *tmp, *tmp2; | |
| 458 | |
| 459 tmp = PyUnicode_Replace(self, qt2, qt1, -1); | |
| 460 if (tmp == NULL) return NULL; | |
| 461 tmp2 = PyUnicode_Replace(tmp, gt2, gt1, -1); | |
| 462 Py_DECREF(tmp); | |
| 463 if (tmp2 == NULL) return NULL; | |
| 464 tmp = PyUnicode_Replace(tmp2, lt2, lt1, -1); | |
| 465 Py_DECREF(tmp2); | |
| 466 if (tmp == NULL) return NULL; | |
| 467 tmp2 = PyUnicode_Replace(tmp, amp2, amp1, -1); | |
| 468 Py_DECREF(tmp); | |
| 469 return tmp2; | |
| 470 } | |
| 471 | |
| 472 PyDoc_STRVAR(stripentities__doc__, | |
| 473 "Return a copy of the text with any character or numeric entities\n\ | |
| 474 replaced by the equivalent UTF-8 characters.\n\ | |
| 475 \n\ | |
| 476 If the `keepxmlentities` parameter is provided and evaluates to `True`,\n\ | |
| 547 | 477 the core XML entities (``&``, ``'``, ``>``, ``<`` and\n\ |
| 478 ``"``) are not stripped.\n\ | |
| 479 \n\ | |
| 480 :return: a `Markup` instance with entities removed\n\ | |
| 481 :rtype: `Markup`\n\ | |
| 482 :see: `genshi.util.stripentities`\n\ | |
| 483 "); | |
| 541 | 484 |
| 485 static PyObject * | |
| 486 Markup_stripentities(PyObject* self, PyObject *args, PyObject *kwds) | |
| 487 { | |
| 488 static char *kwlist[] = {"keepxmlentities", 0}; | |
| 489 PyObject *result, *args2; | |
| 490 char keepxml = 0; | |
| 491 | |
| 492 if (!PyArg_ParseTupleAndKeywords(args, kwds, "|b", kwlist, &keepxml)) { | |
| 493 return NULL; | |
| 494 } | |
| 495 | |
| 496 if (stripentities == NULL) return NULL; | |
| 497 result = PyObject_CallFunction(stripentities, "Ob", self, keepxml); | |
| 498 if (result == NULL) return NULL; | |
| 499 args2 = PyTuple_New(1); | |
| 500 if (args2 == NULL) { | |
| 501 Py_DECREF(result); | |
| 502 return NULL; | |
| 503 } | |
| 504 PyTuple_SET_ITEM(args2, 0, result); | |
| 505 result = MarkupType.tp_new(&MarkupType, args2, NULL); | |
| 506 Py_DECREF(args2); | |
| 507 return result; | |
| 508 } | |
| 509 | |
| 510 PyDoc_STRVAR(striptags__doc__, | |
| 547 | 511 """Return a copy of the text with all XML/HTML tags removed.\n\ |
| 512 \n\ | |
| 513 :return: a `Markup` instance with all tags removed\n\ | |
| 514 :rtype: `Markup`\n\ | |
| 515 :see: `genshi.util.striptags`\n\ | |
| 516 "); | |
| 541 | 517 |
| 518 static PyObject * | |
| 519 Markup_striptags(PyObject* self) | |
| 520 { | |
| 521 PyObject *result, *args; | |
| 522 | |
| 523 if (striptags == NULL) return NULL; | |
| 524 result = PyObject_CallFunction(striptags, "O", self); | |
| 525 if (result == NULL) return NULL; | |
| 526 args = PyTuple_New(1); | |
| 527 if (args == NULL) { | |
| 528 Py_DECREF(result); | |
| 529 return NULL; | |
| 530 } | |
| 531 PyTuple_SET_ITEM(args, 0, result); | |
| 532 result = MarkupType.tp_new(&MarkupType, args, NULL); | |
| 533 Py_DECREF(args); | |
| 534 return result; | |
| 535 } | |
| 536 | |
| 537 typedef struct { | |
| 538 PyUnicodeObject HEAD; | |
| 539 } MarkupObject; | |
| 540 | |
| 541 static PyMethodDef Markup_methods[] = { | |
| 784 | 542 {"__html__", (PyCFunction) Markup_html, METH_NOARGS, NULL}, |
| 541 | 543 {"escape", (PyCFunction) Markup_escape, |
| 547 | 544 METH_VARARGS|METH_CLASS|METH_KEYWORDS, escape__doc__}, |
| 545 {"join", (PyCFunction)Markup_join, METH_VARARGS|METH_KEYWORDS, join__doc__}, | |
| 541 | 546 {"unescape", (PyCFunction)Markup_unescape, METH_NOARGS, unescape__doc__}, |
| 547 {"stripentities", (PyCFunction) Markup_stripentities, | |
| 548 METH_VARARGS|METH_KEYWORDS, stripentities__doc__}, | |
| 549 {"striptags", (PyCFunction) Markup_striptags, METH_NOARGS, | |
| 550 striptags__doc__}, | |
| 551 {NULL} /* Sentinel */ | |
| 552 }; | |
| 553 | |
| 554 static PyNumberMethods Markup_as_number = { | |
| 547 | 555 Markup_add, /*nb_add*/ |
| 556 0, /*nb_subtract*/ | |
| 557 Markup_mul, /*nb_multiply*/ | |
| 558 0, /*nb_divide*/ | |
| 559 Markup_mod, /*nb_remainder*/ | |
| 541 | 560 }; |
| 561 | |
| 562 PyTypeObject MarkupType = { | |
| 563 PyObject_HEAD_INIT(NULL) | |
| 564 0, | |
| 565 "genshi._speedups.Markup", | |
| 566 sizeof(MarkupObject), | |
| 567 0, | |
| 568 0, /*tp_dealloc*/ | |
|
724
8f2c7023af94
update to trunk to track r847, fixing python 2.4 compatibility issues in speedup (and fixing copyrights, apparently :))
aflett
parents:
718
diff
changeset
|
569 0, /*tp_print*/ |
| 541 | 570 0, /*tp_getattr*/ |
| 571 0, /*tp_setattr*/ | |
| 572 0, /*tp_compare*/ | |
| 573 Markup_repr, /*tp_repr*/ | |
| 574 &Markup_as_number, /*tp_as_number*/ | |
| 575 0, /*tp_as_sequence*/ | |
| 576 0, /*tp_as_mapping*/ | |
| 577 0, /*tp_hash */ | |
| 578 | |
| 579 0, /*tp_call*/ | |
| 580 0, /*tp_str*/ | |
| 581 0, /*tp_getattro*/ | |
| 582 0, /*tp_setattro*/ | |
| 583 0, /*tp_as_buffer*/ | |
| 584 | |
| 585 Py_TPFLAGS_DEFAULT | Py_TPFLAGS_BASETYPE | Py_TPFLAGS_CHECKTYPES, /*tp_flags*/ | |
| 586 Markup__doc__,/*tp_doc*/ | |
|
724
8f2c7023af94
update to trunk to track r847, fixing python 2.4 compatibility issues in speedup (and fixing copyrights, apparently :))
aflett
parents:
718
diff
changeset
|
587 |
| 541 | 588 0, /*tp_traverse*/ |
| 589 0, /*tp_clear*/ | |
| 590 | |
| 591 0, /*tp_richcompare*/ | |
| 592 0, /*tp_weaklistoffset*/ | |
| 593 | |
| 594 0, /*tp_iter*/ | |
| 595 0, /*tp_iternext*/ | |
| 596 | |
| 597 /* Attribute descriptor and subclassing stuff */ | |
| 598 | |
| 599 Markup_methods,/*tp_methods*/ | |
| 600 0, /*tp_members*/ | |
| 601 0, /*tp_getset*/ | |
| 602 0, /*tp_base*/ | |
| 603 0, /*tp_dict*/ | |
|
724
8f2c7023af94
update to trunk to track r847, fixing python 2.4 compatibility issues in speedup (and fixing copyrights, apparently :))
aflett
parents:
718
diff
changeset
|
604 |
| 541 | 605 0, /*tp_descr_get*/ |
| 606 0, /*tp_descr_set*/ | |
| 607 0, /*tp_dictoffset*/ | |
|
724
8f2c7023af94
update to trunk to track r847, fixing python 2.4 compatibility issues in speedup (and fixing copyrights, apparently :))
aflett
parents:
718
diff
changeset
|
608 |
| 541 | 609 0, /*tp_init*/ |
| 610 0, /*tp_alloc will be set to PyType_GenericAlloc in module init*/ | |
| 718 | 611 0, /*tp_new*/ |
| 541 | 612 0, /*tp_free Low-level free-memory routine */ |
| 613 0, /*tp_is_gc For PyObject_IS_GC */ | |
| 614 0, /*tp_bases*/ | |
| 615 0, /*tp_mro method resolution order */ | |
| 616 0, /*tp_cache*/ | |
| 617 0, /*tp_subclasses*/ | |
| 618 0 /*tp_weaklist*/ | |
| 619 }; | |
| 620 | |
| 621 PyMODINIT_FUNC | |
| 622 init_speedups(void) | |
| 623 { | |
| 624 PyObject *module; | |
| 625 | |
| 626 /* Workaround for quirk in Visual Studio, see | |
| 627 <http://www.python.it/faq/faq-3.html#3.24> */ | |
| 628 MarkupType.tp_base = &PyUnicode_Type; | |
| 629 | |
| 630 if (PyType_Ready(&MarkupType) < 0) | |
| 631 return; | |
| 632 | |
| 633 init_constants(); | |
| 634 | |
| 635 module = Py_InitModule("_speedups", NULL); | |
| 636 Py_INCREF(&MarkupType); | |
| 637 PyModule_AddObject(module, "Markup", (PyObject *) &MarkupType); | |
| 638 } |