special_methods.html 18.9 KB
Newer Older
William Stein's avatar
William Stein committed
1 2 3 4 5
<!DOCTYPE doctype PUBLIC "-//w3c//dtd html 4.0 transitional//en">
<html><head>
         <meta http-equiv="Content-Type" content="text/html; charset=iso-8859-1">
         <meta name="GENERATOR" content="Mozilla/4.61 (Macintosh; I; PPC) [Netscape]"><title>Special Methods of Extenstion Types</title></head>
<body>
mathbunnyru's avatar
mathbunnyru committed
6
   <h1>  <hr width="100%">Special Methods of Extension Types
William Stein's avatar
William Stein committed
7
<hr width="100%"></h1>
8
  This page describes the special methods currently supported by Cython extension
mathbunnyru's avatar
mathbunnyru committed
9 10 11
 types. A complete list of all the special methods appears in the table at
the bottom. Some of these methods behave differently from their Python counterparts
or have no direct Python counterparts, and require special mention.
William Stein's avatar
William Stein committed
12 13 14 15 16 17 18 19 20 21 22
<p><span style="font-weight: bold;">Note:</span><i> Everything said on this page applies only to </i><span style="font-weight: bold;">extension</span><i style="font-weight: bold;">
</i><span style="font-weight: bold;">types</span><i>, defined with the </i><span style="font-weight: bold; font-family: monospace;">cdef class</span><i> statement. It doesn't apply&nbsp;
to classes defined with the Python </i><span style="font-family: monospace;">class</span><i><span style="font-family: monospace;"> </span>statement, where the normal
 Python rules apply.</i> </p>
 <h2><small>Declaration</small></h2>Special methods of extension types must be declared with <span style="font-family: monospace; font-weight: bold;">def</span>, <span style="font-style: italic;">not</span> <span style="font-family: monospace;">cdef</span>.<br>
<h2><font size="+1">Docstrings</font></h2>

  Currently, docstrings are not fully supported in special methods of extension
 types. You can place a docstring in the source to serve as a comment, but
 it won't show up in the corresponding <span style="font-family: monospace;">__doc__</span> attribute at run time. (This
 is a Python limitation -- there's nowhere in the PyTypeObject data structure
mathbunnyru's avatar
mathbunnyru committed
23
 to put such docstrings.)
William Stein's avatar
William Stein committed
24 25
<h2> <font size="+1">Initialisation methods: <tt>__new__</tt> and <tt>__init__</tt></font></h2>
  There are two methods concerned with initialising the object<tt>.</tt>
mathbunnyru's avatar
mathbunnyru committed
26 27 28 29
<p>The <b><tt>__new__</tt></b> method is where you should perform basic C-level
initialisation of the object, including allocation of any C data structures
that your object will own. You need to be careful what you do in the __new__
method, because the object may not yet be a valid Python object when it is
William Stein's avatar
William Stein committed
30 31 32 33
called. Therefore, you must not invoke any Python operations which might touch
the object; in particular, do not try to call any of its methods. </p>
 <p>Unlike the corresponding method in Python, your <tt>__new__</tt> method
 is <i>not</i> responsible for <i>creating</i> the object. By the time your
mathbunnyru's avatar
mathbunnyru committed
34 35
 <tt>__new__</tt> method is called, memory has been allocated for the object
and any C attributes it has have been initialised to 0 or null. (Any Python
William Stein's avatar
William Stein committed
36 37 38 39 40 41 42 43 44 45 46
attributes have also been initialised to <tt>None</tt>, but you probably shouldn't
rely on that.) Your <tt>__new__</tt> method is guaranteed to be called exactly
once.<br>
<br>
If your extension type has a base type, the <tt>__new__</tt> method of the
base type is automatically called <i>before</i> your <tt>__new__</tt> method
is called; you cannot explicitly call the inherited <tt>__new__</tt> method.
If you need to pass a modified argument list to the base type, you will have
to do the relevant part of the initialisation in the <tt>__init__</tt> method
instead (where the normal rules for calling inherited methods apply).<br>
 </p>
mathbunnyru's avatar
mathbunnyru committed
47
 <p>Note that the first parameter of the <tt>__new__</tt> method is the object
William Stein's avatar
William Stein committed
48 49 50
to be initialised, not the class of the object as it is in Python. </p>
 <p>Any initialisation which cannot safely be done in the <tt>__new__</tt>
method should be done in the <b><tt>__init__</tt></b> method. By the time
mathbunnyru's avatar
mathbunnyru committed
51
 <tt>__init__</tt> is called, the object is a fully valid Python object and
William Stein's avatar
William Stein committed
52 53 54 55 56 57 58 59
all operations are safe. Under some circumstances it is possible for <tt>__init__</tt>
to be called more than once or not to be called at all, so your other methods
 should be designed to be robust in such situations. </p>
 <p>Keep in mind that any arguments passed to the constructor will be passed
 to the <tt>__new__</tt> method as well as the <tt>__init__</tt> method.
If you anticipate subclassing your extension type in Python, you may find
it useful to give the <tt>__new__</tt> method * and ** arguments so that
it can accept and ignore extra arguments. Otherwise, any Python subclass
mathbunnyru's avatar
mathbunnyru committed
60 61
which has an <tt>__init__</tt> with a different signature will have to override
<tt>__new__</tt> as well as <tt>__init__</tt>, which the writer of a Python
William Stein's avatar
William Stein committed
62 63 64 65 66
class wouldn't expect to have to do. </p>
 <h2> <font size="+1">Finalization method: <tt>__dealloc__</tt><tt></tt></font></h2>
  The counterpart to the <tt>__new__</tt> method is the <b><tt>__dealloc__</tt></b>
method, which should perform the inverse of the <tt>__new__</tt> method.
Any C data structures that you allocated in your <tt>__new__</tt> method
mathbunnyru's avatar
mathbunnyru committed
67 68 69 70 71 72
should be freed in your <tt>__dealloc__</tt> method.
<p>You need to be careful what you do in a <tt>__dealloc__</tt> method. By
the time your <tt>__dealloc__</tt> method is called, the object may already
have been partially destroyed and may not be in a valid state as far as Python
is concerned, so you should avoid invoking any Python operations which might
touch the object. In particular, don't call any other methods of the object
William Stein's avatar
William Stein committed
73 74
or do anything which might cause the object to be resurrected. It's best if
you stick to just deallocating C data. </p>
mathbunnyru's avatar
mathbunnyru committed
75
 <p>You don't need to worry about deallocating Python attributes of your object,
76
because that will be done for you by Cython after your <tt>__dealloc__</tt>
William Stein's avatar
William Stein committed
77 78
method returns.<br>
 <br>
mathbunnyru's avatar
mathbunnyru committed
79 80
 <b>Note:</b> There is no <tt>__del__</tt> method for extension types. (Earlier
versions of the Cython documentation stated that there was, but this turned
William Stein's avatar
William Stein committed
81 82 83 84 85 86 87
out to be incorrect.)<br>
  </p>
 <h2><font size="+1">Arithmetic methods</font></h2>
  Arithmetic operator methods, such as <tt>__add__</tt>, behave differently
 from their Python counterparts. There are no separate "reversed" versions
 of these methods (<tt>__radd__</tt>, etc.) Instead, if the first operand
cannot perform the operation, the <i>same</i> method of the second operand
mathbunnyru's avatar
mathbunnyru committed
88
is called, with the operands in the <i>same order</i>.
William Stein's avatar
William Stein committed
89 90 91 92 93 94 95 96 97 98
<p>This means that you can't rely on the first parameter of these methods
 being "self", and you should test the types of both operands before deciding
 what to do. If you can't handle the combination of types you've been given,
 you should return <tt>NotImplemented</tt>. </p>
 <p>This also applies to the in-place arithmetic method <tt>__ipow__</tt>.
 It doesn't apply to any of the <i>other</i> in-place methods (<tt>__iadd__</tt>,
 etc.) which always take self as the first argument. </p>
 <h2> <font size="+1">Rich comparisons</font></h2>
  There are no separate methods for the individual rich comparison operations
 (<tt>__eq__</tt>, <tt>__le__</tt>, etc.) Instead there is a single method
mathbunnyru's avatar
mathbunnyru committed
99 100
 <tt>__richcmp__</tt> which takes an integer indicating which operation is
to be performed, as follows:
William Stein's avatar
William Stein committed
101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142
<ul>
      <ul>
 &nbsp;         <table nosave="" border="0" cellpadding="5" cellspacing="0">
  <tbody>
         <tr nosave="">
  <td nosave="" bgcolor="#ffcc33" width="30">                     <div align="right">&lt;</div>
  </td>
   <td nosave="" bgcolor="#66ffff" width="30">0</td>
   <td><br>
           </td>
   <td nosave="" bgcolor="#ffcc33" width="30">                     <div align="right">==</div>
  </td>
   <td nosave="" bgcolor="#66ffff" width="30">2</td>
   <td><br>
           </td>
   <td nosave="" bgcolor="#ffcc33" width="30">                     <div align="right">&gt;</div>
  </td>
   <td nosave="" bgcolor="#66ffff" width="30">4</td>
  </tr>
   <tr nosave="">
  <td nosave="" bgcolor="#ffcc33">                     <div align="right">&lt;=</div>
  </td>
   <td nosave="" bgcolor="#66ffff">1</td>
   <td><br>
           </td>
   <td nosave="" bgcolor="#ffcc33">                     <div align="right">!=</div>
  </td>
   <td nosave="" bgcolor="#66ffff">3</td>
   <td><br>
           </td>
   <td nosave="" bgcolor="#ffcc33">                     <div align="right">&gt;=</div>
  </td>
   <td nosave="" bgcolor="#66ffff">5</td>
  </tr>
              </tbody>         </table>
      </ul>
  </ul>
   <h2> <font size="+1">The __next__ method</font></h2>
  Extension types wishing to implement the iterator interface should define
 a method called <b><tt>__next__</tt></b>, <i>not</i> <tt>next</tt>. The Python
 system will automatically supply a <tt>next</tt> method which calls your
<span style="font-family: monospace;">__next__</span>.  <b>Do NOT explicitly give your type a <tt>next</tt> method</b>,
mathbunnyru's avatar
mathbunnyru committed
143
or bad things could happen (see note 3).
William Stein's avatar
William Stein committed
144 145 146
<h2> <font size="+1">Special Method Table</font></h2>
  This table lists all of the special methods together with their parameter
 and return types. A parameter named <b>self</b> is of the type the method
mathbunnyru's avatar
mathbunnyru committed
147
 belongs to. Other untyped parameters are generic Python objects.
William Stein's avatar
William Stein committed
148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226
<p>You don't have to declare your method as taking these parameter types.
 If you declare different types, conversions will be performed as necessary.
 <br>
 &nbsp; <table nosave="" bgcolor="#ccffff" border="1" cellpadding="5" cellspacing="0">
  <tbody>
     <tr nosave="" bgcolor="#ffcc33">
  <td nosave=""><b>Name</b></td>
   <td><b>Parameters</b></td>
   <td><b>Return type</b></td>
   <td><b>Description</b></td>
  </tr>
   <tr nosave="" bgcolor="#66ffff">
  <td colspan="4" nosave=""><b>General</b></td>
  </tr>
   <tr>
  <td><tt>__new__</tt></td>
   <td>self, ...</td>
   <td>&nbsp;</td>
   <td>Basic initialisation (no direct Python equivalent)</td>
  </tr>
   <tr>
  <td><tt>__init__</tt></td>
   <td>self, ...</td>
   <td>&nbsp;</td>
   <td>Further initialisation</td>
  </tr>
   <tr>
  <td><tt>__dealloc__</tt></td>
   <td>self</td>
   <td>&nbsp;</td>
   <td>Basic deallocation (no direct Python equivalent)</td>
  </tr>
   <tr>
  <td><tt>__cmp__</tt></td>
   <td>x, y</td>
   <td>int</td>
   <td>3-way comparison</td>
  </tr>
   <tr>
  <td><tt>__richcmp__</tt></td>
   <td>x, y, int op</td>
   <td>object</td>
   <td>Rich comparison (no direct Python equivalent)</td>
  </tr>
   <tr>
  <td><tt>__str__</tt></td>
   <td>self</td>
   <td>object</td>
   <td>str(self)</td>
  </tr>
   <tr>
  <td><tt>__repr__</tt></td>
   <td>self</td>
   <td>object</td>
   <td>repr(self)</td>
  </tr>
   <tr nosave="">
  <td nosave=""><tt>__hash__</tt></td>
   <td>self</td>
   <td>int</td>
   <td>Hash function</td>
  </tr>
   <tr>
  <td><tt>__call__</tt></td>
   <td>self, ...</td>
   <td>object</td>
   <td>self(...)</td>
  </tr>
   <tr>
  <td><tt>__iter__</tt></td>
   <td>self</td>
   <td>object</td>
   <td>Return iterator for sequence</td>
  </tr>
   <tr>
  <td><tt>__getattr__</tt></td>
   <td>self, name</td>
   <td>object</td>
   <td>Get attribute</td>
227 228 229 230 231
  </tr>
  <td><tt>__getattribute__</tt></td>
   <td>self, name</td>
   <td>object</td>
   <td>Get attribute, unconditionally</td>
William Stein's avatar
William Stein committed
232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588
  </tr>
   <tr>
  <td><tt>__setattr__</tt></td>
   <td>self, name, val</td>
   <td>&nbsp;</td>
   <td>Set attribute</td>
  </tr>
   <tr>
  <td><tt>__delattr__</tt></td>
   <td>self, name</td>
   <td>&nbsp;</td>
   <td>Delete attribute</td>
  </tr>
   <tr nosave="" bgcolor="#66ffff">
  <td colspan="4" nosave=""><b>Arithmetic operators</b></td>
  </tr>
   <tr>
  <td><tt>__add__</tt></td>
   <td>x, y</td>
   <td>object</td>
   <td>binary + operator</td>
  </tr>
   <tr>
  <td><tt>__sub__</tt></td>
   <td>x, y</td>
   <td>object</td>
   <td>binary - operator</td>
  </tr>
   <tr>
  <td><tt>__mul__</tt></td>
   <td>x, y</td>
   <td>object</td>
   <td>* operator</td>
  </tr>
   <tr>
  <td><tt>__div__</tt></td>
   <td>x, y</td>
   <td>object</td>
   <td>/&nbsp; operator for old-style division</td>
  </tr>
   <tr>
  <td><tt>__floordiv__</tt></td>
   <td>x, y</td>
   <td>object</td>
   <td>//&nbsp; operator</td>
  </tr>
   <tr>
  <td><tt>__truediv__</tt></td>
   <td>x, y</td>
   <td>object</td>
   <td>/&nbsp; operator for new-style division</td>
  </tr>
   <tr>
  <td><tt>__mod__</tt></td>
   <td>x, y</td>
   <td>object</td>
   <td>% operator</td>
  </tr>
   <tr>
  <td><tt>__divmod__</tt></td>
   <td>x, y</td>
   <td>object</td>
   <td>combined div and mod</td>
  </tr>
   <tr>
  <td><tt>__pow__</tt></td>
   <td>x, y, z</td>
   <td>object</td>
   <td>** operator or pow(x, y, z)</td>
  </tr>
   <tr>
  <td><tt>__neg__</tt></td>
   <td>self</td>
   <td>object</td>
   <td>unary - operator</td>
  </tr>
   <tr>
  <td><tt>__pos__</tt></td>
   <td>self</td>
   <td>object</td>
   <td>unary + operator</td>
  </tr>
   <tr>
  <td><tt>__abs__</tt></td>
   <td>self</td>
   <td>object</td>
   <td>absolute value</td>
  </tr>
   <tr>
  <td><tt>__nonzero__</tt></td>
   <td>self</td>
   <td>int</td>
   <td>convert to boolean</td>
  </tr>
   <tr>
  <td><tt>__invert__</tt></td>
   <td>self</td>
   <td>object</td>
   <td>~ operator</td>
  </tr>
   <tr>
  <td><tt>__lshift__</tt></td>
   <td>x, y</td>
   <td>object</td>
   <td>&lt;&lt; operator</td>
  </tr>
   <tr>
  <td><tt>__rshift__</tt></td>
   <td>x, y</td>
   <td>object</td>
   <td>&gt;&gt; operator</td>
  </tr>
   <tr>
  <td><tt>__and__</tt></td>
   <td>x, y</td>
   <td>object</td>
   <td>&amp; operator</td>
  </tr>
   <tr>
  <td><tt>__or__</tt></td>
   <td>x, y</td>
   <td>object</td>
   <td>| operator</td>
  </tr>
   <tr>
  <td><tt>__xor__</tt></td>
   <td>x, y</td>
   <td>object</td>
   <td>^ operator</td>
  </tr>
   <tr nosave="" bgcolor="#66ffff">
  <td colspan="4" nosave=""><b>Numeric conversions</b></td>
  </tr>
   <tr>
  <td><tt>__int__</tt></td>
   <td>self</td>
   <td>object</td>
   <td>Convert to integer</td>
  </tr>
   <tr>
  <td><tt>__long__</tt></td>
   <td>self</td>
   <td>object</td>
   <td>Convert to long integer</td>
  </tr>
   <tr>
  <td><tt>__float__</tt></td>
   <td>self</td>
   <td>object</td>
   <td>Convert to float</td>
  </tr>
   <tr>
  <td><tt>__oct__</tt></td>
   <td>self</td>
   <td>object</td>
   <td>Convert to octal</td>
  </tr>
   <tr>
  <td><tt>__hex__</tt></td>
   <td>self</td>
   <td>object</td>
   <td>Convert to hexadecimal</td>
  </tr>
   <tr nosave="" bgcolor="#66ffff">
  <td colspan="4" nosave=""><b>In-place arithmetic operators</b></td>
  </tr>
   <tr>
  <td><tt>__iadd__</tt></td>
   <td>self, x</td>
   <td>object</td>
   <td>+= operator</td>
  </tr>
   <tr>
  <td><tt>__isub__</tt></td>
   <td>self, x</td>
   <td>object</td>
   <td>-= operator</td>
  </tr>
   <tr>
  <td><tt>__imul__</tt></td>
   <td>self, x</td>
   <td>object</td>
   <td>*= operator</td>
  </tr>
   <tr>
  <td><tt>__idiv__</tt></td>
   <td>self, x</td>
   <td>object</td>
   <td>/= operator for old-style division</td>
  </tr>
   <tr>
  <td><tt>__ifloordiv__</tt></td>
   <td>self, x</td>
   <td>object</td>
   <td>//= operator</td>
  </tr>
   <tr>
  <td><tt>__itruediv__</tt></td>
   <td>self, x</td>
   <td>object</td>
   <td>/= operator for new-style division</td>
  </tr>
   <tr>
  <td><tt>__imod__</tt></td>
   <td>self, x</td>
   <td>object</td>
   <td>%= operator</td>
  </tr>
   <tr>
  <td><tt>__ipow__</tt></td>
   <td>x, y, z</td>
   <td>object</td>
   <td>**= operator</td>
  </tr>
   <tr>
  <td><tt>__ilshift__</tt></td>
   <td>self, x</td>
   <td>object</td>
   <td>&lt;&lt;= operator</td>
  </tr>
   <tr>
  <td><tt>__irshift__</tt></td>
   <td>self, x</td>
   <td>object</td>
   <td>&gt;&gt;= operator</td>
  </tr>
   <tr>
  <td><tt>__iand__</tt></td>
   <td>self, x</td>
   <td>object</td>
   <td>&amp;= operator</td>
  </tr>
   <tr>
  <td><tt>__ior__</tt></td>
   <td>self, x</td>
   <td>object</td>
   <td>|= operator</td>
  </tr>
   <tr>
  <td><tt>__ixor__</tt></td>
   <td>self, x</td>
   <td>object</td>
   <td>^= operator</td>
  </tr>
   <tr nosave="" bgcolor="#66ffff">
  <td colspan="4" nosave=""><b>Sequences and mappings</b></td>
  </tr>
   <tr>
  <td><tt>__len__</tt></td>
   <td>self</td>
   <td>int</td>
   <td>len(self)</td>
  </tr>
   <tr>
  <td><tt>__getitem__</tt></td>
   <td>self, x</td>
   <td>object</td>
   <td>self[x]</td>
  </tr>
   <tr>
  <td><tt>__setitem__</tt></td>
   <td>self, x, y</td>
   <td>&nbsp;</td>
   <td>self[x] = y</td>
  </tr>
   <tr>
  <td><tt>__delitem__</tt></td>
   <td>self, x</td>
   <td>&nbsp;</td>
   <td>del self[x]</td>
  </tr>
   <tr>
  <td><tt>__getslice__</tt></td>
   <td>self, int i, int j</td>
   <td>object</td>
   <td>self[i:j]</td>
  </tr>
   <tr>
  <td><tt>__setslice__</tt></td>
   <td>self, int i, int j, x</td>
   <td>&nbsp;</td>
   <td>self[i:j] = x</td>
  </tr>
   <tr>
  <td><tt>__delslice__</tt></td>
   <td>self, int i, int j</td>
   <td>&nbsp;</td>
   <td>del self[i:j]</td>
  </tr>
   <tr>
  <td><tt>__contains__</tt></td>
   <td>self, x</td>
   <td>int</td>
   <td>x in self</td>
  </tr>
   <tr nosave="" bgcolor="#66ffff">
  <td colspan="4" nosave=""><b>Iterators</b></td>
  </tr>
   <tr>
  <td><tt>__next__</tt></td>
   <td>self</td>
   <td>object</td>
   <td>Get next item (called <tt>next</tt> in Python)</td>
  </tr>
   <tr nosave="" bgcolor="#66ffff">
  <td colspan="4" nosave=""><b>Buffer interface</b>&nbsp; (no Python equivalents
 - see note 1)</td>
  </tr>
   <tr>
  <td><tt>__getreadbuffer__</tt></td>
   <td>self, int i, void **p</td>
   <td>&nbsp;</td>
   <td>&nbsp;</td>
  </tr>
   <tr>
  <td><tt>__getwritebuffer__</tt></td>
   <td>self, int i, void **p</td>
   <td>&nbsp;</td>
   <td>&nbsp;</td>
  </tr>
   <tr>
  <td><tt>__getsegcount__</tt></td>
   <td>self, int *p</td>
   <td>&nbsp;</td>
   <td>&nbsp;</td>
  </tr>
   <tr>
  <td><tt>__getcharbuffer__</tt></td>
   <td>self, int i, char **p</td>
   <td>&nbsp;</td>
   <td>&nbsp;</td>
  </tr>
   <tr nosave="" bgcolor="#66ffff">
  <td colspan="4" nosave=""><b>Descriptor objects</b>&nbsp; (no Python equivalents
 - see note 2)</td>
  </tr>
   <tr>
  <td><tt>__get__</tt></td>
   <td>self, instance, class</td>
   <td>object</td>
   <td>Get value of attribute</td>
  </tr>
   <tr>
  <td><tt>__set__</tt></td>
   <td>self, instance, value</td>
   <td>&nbsp;</td>
   <td>Set value of attribute</td>
  </tr>
   <tr>
  <td style="font-family: monospace;">__delete__</td>
   <td>self, instance</td>
   <td>&nbsp;</td>
   <td>Delete attribute</td>
  </tr>
      </tbody> </table>
   </p>
 <p>Note 1: The buffer interface is intended for use by C code and is not
mathbunnyru's avatar
mathbunnyru committed
589
directly accessible from Python. It is described in the <a href="http://www.python.org/doc/current/api/api.html">Python/C API Reference
William Stein's avatar
William Stein committed
590 591 592 593 594
Manual</a> under sections <a href="http://www.python.org/doc/current/api/abstract-buffer.html">6.6</a>
and <a href="http://www.python.org/doc/current/api/buffer-structs.html">10.6</a>.
 </p>
 <p>Note 2: Descriptor objects are part of the support mechanism for new-style
 Python classes. See the <a href="http://www.python.org/doc/2.2.1/whatsnew/sect-rellinks.html#SECTION000320000000000000000">discussion
mathbunnyru's avatar
mathbunnyru committed
595 596
 of descriptors in the Python documentation</a>. See also <a href="http://www.python.org/peps/pep-0252.html">PEP 252, "Making Types Look
More Like Classes"</a>, and <a href="http://www.python.org/peps/pep-0253.html">PEP 253, "Subtyping Built-In
William Stein's avatar
William Stein committed
597 598 599 600 601 602 603
Types"</a>. </p>
 <p>Note 3: If your type defines a <tt>__new__</tt> method, any method called
 <tt>new</tt> that you define will be overwritten with the system-supplied
 <tt>new</tt> at module import time. </p>
 <br>
 <br>
</body></html>