parsermodule.c revision 52318d6215f9f9626d38a9b81b52d411dbbdb36a
1/* parsermodule.c 2 * 3 * Copyright 1995-1996 by Fred L. Drake, Jr. and Virginia Polytechnic 4 * Institute and State University, Blacksburg, Virginia, USA. 5 * Portions copyright 1991-1995 by Stichting Mathematisch Centrum, 6 * Amsterdam, The Netherlands. Copying is permitted under the terms 7 * associated with the main Python distribution, with the additional 8 * restriction that this additional notice be included and maintained 9 * on all distributed copies. 10 * 11 * This module serves to replace the original parser module written 12 * by Guido. The functionality is not matched precisely, but the 13 * original may be implemented on top of this. This is desirable 14 * since the source of the text to be parsed is now divorced from 15 * this interface. 16 * 17 * Unlike the prior interface, the ability to give a parse tree 18 * produced by Python code as a tuple to the compiler is enabled by 19 * this module. See the documentation for more details. 20 * 21 * I've added some annotations that help with the lint code-checking 22 * program, but they're not complete by a long shot. The real errors 23 * that lint detects are gone, but there are still warnings with 24 * Py_[X]DECREF() and Py_[X]INCREF() macros. The lint annotations 25 * look like "NOTE(...)". 26 */ 27 28#include "Python.h" /* general Python API */ 29#include "graminit.h" /* symbols defined in the grammar */ 30#include "node.h" /* internal parser structure */ 31#include "errcode.h" /* error codes for PyNode_*() */ 32#include "token.h" /* token definitions */ 33 /* ISTERMINAL() / ISNONTERMINAL() */ 34#include "compile.h" /* PyNode_Compile() */ 35 36#ifdef lint 37#include <note.h> 38#else 39#define NOTE(x) 40#endif 41 42/* String constants used to initialize module attributes. 43 * 44 */ 45static char parser_copyright_string[] = 46"Copyright 1995-1996 by Virginia Polytechnic Institute & State\n\ 47University, Blacksburg, Virginia, USA, and Fred L. Drake, Jr., Reston,\n\ 48Virginia, USA. Portions copyright 1991-1995 by Stichting Mathematisch\n\ 49Centrum, Amsterdam, The Netherlands."; 50 51 52PyDoc_STRVAR(parser_doc_string, 53"This is an interface to Python's internal parser."); 54 55static char parser_version_string[] = "0.5"; 56 57 58typedef PyObject* (*SeqMaker) (Py_ssize_t length); 59typedef int (*SeqInserter) (PyObject* sequence, 60 Py_ssize_t index, 61 PyObject* element); 62 63/* The function below is copyrighted by Stichting Mathematisch Centrum. The 64 * original copyright statement is included below, and continues to apply 65 * in full to the function immediately following. All other material is 66 * original, copyrighted by Fred L. Drake, Jr. and Virginia Polytechnic 67 * Institute and State University. Changes were made to comply with the 68 * new naming conventions. Added arguments to provide support for creating 69 * lists as well as tuples, and optionally including the line numbers. 70 */ 71 72 73static PyObject* 74node2tuple(node *n, /* node to convert */ 75 SeqMaker mkseq, /* create sequence */ 76 SeqInserter addelem, /* func. to add elem. in seq. */ 77 int lineno) /* include line numbers? */ 78{ 79 if (n == NULL) { 80 Py_INCREF(Py_None); 81 return (Py_None); 82 } 83 if (ISNONTERMINAL(TYPE(n))) { 84 int i; 85 PyObject *v; 86 PyObject *w; 87 88 v = mkseq(1 + NCH(n) + (TYPE(n) == encoding_decl)); 89 if (v == NULL) 90 return (v); 91 w = PyInt_FromLong(TYPE(n)); 92 if (w == NULL) { 93 Py_DECREF(v); 94 return ((PyObject*) NULL); 95 } 96 (void) addelem(v, 0, w); 97 for (i = 0; i < NCH(n); i++) { 98 w = node2tuple(CHILD(n, i), mkseq, addelem, lineno); 99 if (w == NULL) { 100 Py_DECREF(v); 101 return ((PyObject*) NULL); 102 } 103 (void) addelem(v, i+1, w); 104 } 105 106 if (TYPE(n) == encoding_decl) 107 (void) addelem(v, i+1, PyString_FromString(STR(n))); 108 return (v); 109 } 110 else if (ISTERMINAL(TYPE(n))) { 111 PyObject *result = mkseq(2 + lineno); 112 if (result != NULL) { 113 (void) addelem(result, 0, PyInt_FromLong(TYPE(n))); 114 (void) addelem(result, 1, PyString_FromString(STR(n))); 115 if (lineno == 1) 116 (void) addelem(result, 2, PyInt_FromLong(n->n_lineno)); 117 } 118 return (result); 119 } 120 else { 121 PyErr_SetString(PyExc_SystemError, 122 "unrecognized parse tree node type"); 123 return ((PyObject*) NULL); 124 } 125} 126/* 127 * End of material copyrighted by Stichting Mathematisch Centrum. 128 */ 129 130 131 132/* There are two types of intermediate objects we're interested in: 133 * 'eval' and 'exec' types. These constants can be used in the st_type 134 * field of the object type to identify which any given object represents. 135 * These should probably go in an external header to allow other extensions 136 * to use them, but then, we really should be using C++ too. ;-) 137 */ 138 139#define PyST_EXPR 1 140#define PyST_SUITE 2 141 142 143/* These are the internal objects and definitions required to implement the 144 * ST type. Most of the internal names are more reminiscent of the 'old' 145 * naming style, but the code uses the new naming convention. 146 */ 147 148static PyObject* 149parser_error = 0; 150 151 152typedef struct { 153 PyObject_HEAD /* standard object header */ 154 node* st_node; /* the node* returned by the parser */ 155 int st_type; /* EXPR or SUITE ? */ 156} PyST_Object; 157 158 159static void parser_free(PyST_Object *st); 160static int parser_compare(PyST_Object *left, PyST_Object *right); 161static PyObject *parser_getattr(PyObject *self, char *name); 162 163 164static 165PyTypeObject PyST_Type = { 166 PyObject_HEAD_INIT(NULL) 167 0, 168 "parser.st", /* tp_name */ 169 (int) sizeof(PyST_Object), /* tp_basicsize */ 170 0, /* tp_itemsize */ 171 (destructor)parser_free, /* tp_dealloc */ 172 0, /* tp_print */ 173 parser_getattr, /* tp_getattr */ 174 0, /* tp_setattr */ 175 (cmpfunc)parser_compare, /* tp_compare */ 176 0, /* tp_repr */ 177 0, /* tp_as_number */ 178 0, /* tp_as_sequence */ 179 0, /* tp_as_mapping */ 180 0, /* tp_hash */ 181 0, /* tp_call */ 182 0, /* tp_str */ 183 0, /* tp_getattro */ 184 0, /* tp_setattro */ 185 186 /* Functions to access object as input/output buffer */ 187 0, /* tp_as_buffer */ 188 189 Py_TPFLAGS_DEFAULT, /* tp_flags */ 190 191 /* __doc__ */ 192 "Intermediate representation of a Python parse tree." 193}; /* PyST_Type */ 194 195 196static int 197parser_compare_nodes(node *left, node *right) 198{ 199 int j; 200 201 if (TYPE(left) < TYPE(right)) 202 return (-1); 203 204 if (TYPE(right) < TYPE(left)) 205 return (1); 206 207 if (ISTERMINAL(TYPE(left))) 208 return (strcmp(STR(left), STR(right))); 209 210 if (NCH(left) < NCH(right)) 211 return (-1); 212 213 if (NCH(right) < NCH(left)) 214 return (1); 215 216 for (j = 0; j < NCH(left); ++j) { 217 int v = parser_compare_nodes(CHILD(left, j), CHILD(right, j)); 218 219 if (v != 0) 220 return (v); 221 } 222 return (0); 223} 224 225 226/* int parser_compare(PyST_Object* left, PyST_Object* right) 227 * 228 * Comparison function used by the Python operators ==, !=, <, >, <=, >= 229 * This really just wraps a call to parser_compare_nodes() with some easy 230 * checks and protection code. 231 * 232 */ 233static int 234parser_compare(PyST_Object *left, PyST_Object *right) 235{ 236 if (left == right) 237 return (0); 238 239 if ((left == 0) || (right == 0)) 240 return (-1); 241 242 return (parser_compare_nodes(left->st_node, right->st_node)); 243} 244 245 246/* parser_newstobject(node* st) 247 * 248 * Allocates a new Python object representing an ST. This is simply the 249 * 'wrapper' object that holds a node* and allows it to be passed around in 250 * Python code. 251 * 252 */ 253static PyObject* 254parser_newstobject(node *st, int type) 255{ 256 PyST_Object* o = PyObject_New(PyST_Object, &PyST_Type); 257 258 if (o != 0) { 259 o->st_node = st; 260 o->st_type = type; 261 } 262 else { 263 PyNode_Free(st); 264 } 265 return ((PyObject*)o); 266} 267 268 269/* void parser_free(PyST_Object* st) 270 * 271 * This is called by a del statement that reduces the reference count to 0. 272 * 273 */ 274static void 275parser_free(PyST_Object *st) 276{ 277 PyNode_Free(st->st_node); 278 PyObject_Del(st); 279} 280 281 282/* parser_st2tuple(PyObject* self, PyObject* args, PyObject* kw) 283 * 284 * This provides conversion from a node* to a tuple object that can be 285 * returned to the Python-level caller. The ST object is not modified. 286 * 287 */ 288static PyObject* 289parser_st2tuple(PyST_Object *self, PyObject *args, PyObject *kw) 290{ 291 PyObject *line_option = 0; 292 PyObject *res = 0; 293 int ok; 294 295 static char *keywords[] = {"ast", "line_info", NULL}; 296 297 if (self == NULL) { 298 ok = PyArg_ParseTupleAndKeywords(args, kw, "O!|O:st2tuple", keywords, 299 &PyST_Type, &self, &line_option); 300 } 301 else 302 ok = PyArg_ParseTupleAndKeywords(args, kw, "|O:totuple", &keywords[1], 303 &line_option); 304 if (ok != 0) { 305 int lineno = 0; 306 if (line_option != NULL) { 307 lineno = (PyObject_IsTrue(line_option) != 0) ? 1 : 0; 308 } 309 /* 310 * Convert ST into a tuple representation. Use Guido's function, 311 * since it's known to work already. 312 */ 313 res = node2tuple(((PyST_Object*)self)->st_node, 314 PyTuple_New, PyTuple_SetItem, lineno); 315 } 316 return (res); 317} 318 319 320/* parser_st2list(PyObject* self, PyObject* args, PyObject* kw) 321 * 322 * This provides conversion from a node* to a list object that can be 323 * returned to the Python-level caller. The ST object is not modified. 324 * 325 */ 326static PyObject* 327parser_st2list(PyST_Object *self, PyObject *args, PyObject *kw) 328{ 329 PyObject *line_option = 0; 330 PyObject *res = 0; 331 int ok; 332 333 static char *keywords[] = {"ast", "line_info", NULL}; 334 335 if (self == NULL) 336 ok = PyArg_ParseTupleAndKeywords(args, kw, "O!|O:st2list", keywords, 337 &PyST_Type, &self, &line_option); 338 else 339 ok = PyArg_ParseTupleAndKeywords(args, kw, "|O:tolist", &keywords[1], 340 &line_option); 341 if (ok) { 342 int lineno = 0; 343 if (line_option != 0) { 344 lineno = PyObject_IsTrue(line_option) ? 1 : 0; 345 } 346 /* 347 * Convert ST into a tuple representation. Use Guido's function, 348 * since it's known to work already. 349 */ 350 res = node2tuple(self->st_node, 351 PyList_New, PyList_SetItem, lineno); 352 } 353 return (res); 354} 355 356 357/* parser_compilest(PyObject* self, PyObject* args) 358 * 359 * This function creates code objects from the parse tree represented by 360 * the passed-in data object. An optional file name is passed in as well. 361 * 362 */ 363static PyObject* 364parser_compilest(PyST_Object *self, PyObject *args, PyObject *kw) 365{ 366 PyObject* res = 0; 367 char* str = "<syntax-tree>"; 368 int ok; 369 370 static char *keywords[] = {"ast", "filename", NULL}; 371 372 if (self == NULL) 373 ok = PyArg_ParseTupleAndKeywords(args, kw, "O!|s:compilest", keywords, 374 &PyST_Type, &self, &str); 375 else 376 ok = PyArg_ParseTupleAndKeywords(args, kw, "|s:compile", &keywords[1], 377 &str); 378 379 if (ok) 380 res = (PyObject *)PyNode_Compile(self->st_node, str); 381 382 return (res); 383} 384 385 386/* PyObject* parser_isexpr(PyObject* self, PyObject* args) 387 * PyObject* parser_issuite(PyObject* self, PyObject* args) 388 * 389 * Checks the passed-in ST object to determine if it is an expression or 390 * a statement suite, respectively. The return is a Python truth value. 391 * 392 */ 393static PyObject* 394parser_isexpr(PyST_Object *self, PyObject *args, PyObject *kw) 395{ 396 PyObject* res = 0; 397 int ok; 398 399 static char *keywords[] = {"ast", NULL}; 400 401 if (self == NULL) 402 ok = PyArg_ParseTupleAndKeywords(args, kw, "O!:isexpr", keywords, 403 &PyST_Type, &self); 404 else 405 ok = PyArg_ParseTupleAndKeywords(args, kw, ":isexpr", &keywords[1]); 406 407 if (ok) { 408 /* Check to see if the ST represents an expression or not. */ 409 res = (self->st_type == PyST_EXPR) ? Py_True : Py_False; 410 Py_INCREF(res); 411 } 412 return (res); 413} 414 415 416static PyObject* 417parser_issuite(PyST_Object *self, PyObject *args, PyObject *kw) 418{ 419 PyObject* res = 0; 420 int ok; 421 422 static char *keywords[] = {"ast", NULL}; 423 424 if (self == NULL) 425 ok = PyArg_ParseTupleAndKeywords(args, kw, "O!:issuite", keywords, 426 &PyST_Type, &self); 427 else 428 ok = PyArg_ParseTupleAndKeywords(args, kw, ":issuite", &keywords[1]); 429 430 if (ok) { 431 /* Check to see if the ST represents an expression or not. */ 432 res = (self->st_type == PyST_EXPR) ? Py_False : Py_True; 433 Py_INCREF(res); 434 } 435 return (res); 436} 437 438 439#define PUBLIC_METHOD_TYPE (METH_VARARGS|METH_KEYWORDS) 440 441static PyMethodDef 442parser_methods[] = { 443 {"compile", (PyCFunction)parser_compilest, PUBLIC_METHOD_TYPE, 444 PyDoc_STR("Compile this ST object into a code object.")}, 445 {"isexpr", (PyCFunction)parser_isexpr, PUBLIC_METHOD_TYPE, 446 PyDoc_STR("Determines if this ST object was created from an expression.")}, 447 {"issuite", (PyCFunction)parser_issuite, PUBLIC_METHOD_TYPE, 448 PyDoc_STR("Determines if this ST object was created from a suite.")}, 449 {"tolist", (PyCFunction)parser_st2list, PUBLIC_METHOD_TYPE, 450 PyDoc_STR("Creates a list-tree representation of this ST.")}, 451 {"totuple", (PyCFunction)parser_st2tuple, PUBLIC_METHOD_TYPE, 452 PyDoc_STR("Creates a tuple-tree representation of this ST.")}, 453 454 {NULL, NULL, 0, NULL} 455}; 456 457 458static PyObject* 459parser_getattr(PyObject *self, char *name) 460{ 461 return (Py_FindMethod(parser_methods, self, name)); 462} 463 464 465/* err_string(char* message) 466 * 467 * Sets the error string for an exception of type ParserError. 468 * 469 */ 470static void 471err_string(char *message) 472{ 473 PyErr_SetString(parser_error, message); 474} 475 476 477/* PyObject* parser_do_parse(PyObject* args, int type) 478 * 479 * Internal function to actually execute the parse and return the result if 480 * successful or set an exception if not. 481 * 482 */ 483static PyObject* 484parser_do_parse(PyObject *args, PyObject *kw, char *argspec, int type) 485{ 486 char* string = 0; 487 PyObject* res = 0; 488 489 static char *keywords[] = {"source", NULL}; 490 491 if (PyArg_ParseTupleAndKeywords(args, kw, argspec, keywords, &string)) { 492 node* n = PyParser_SimpleParseString(string, 493 (type == PyST_EXPR) 494 ? eval_input : file_input); 495 496 if (n) 497 res = parser_newstobject(n, type); 498 } 499 return (res); 500} 501 502 503/* PyObject* parser_expr(PyObject* self, PyObject* args) 504 * PyObject* parser_suite(PyObject* self, PyObject* args) 505 * 506 * External interfaces to the parser itself. Which is called determines if 507 * the parser attempts to recognize an expression ('eval' form) or statement 508 * suite ('exec' form). The real work is done by parser_do_parse() above. 509 * 510 */ 511static PyObject* 512parser_expr(PyST_Object *self, PyObject *args, PyObject *kw) 513{ 514 NOTE(ARGUNUSED(self)) 515 return (parser_do_parse(args, kw, "s:expr", PyST_EXPR)); 516} 517 518 519static PyObject* 520parser_suite(PyST_Object *self, PyObject *args, PyObject *kw) 521{ 522 NOTE(ARGUNUSED(self)) 523 return (parser_do_parse(args, kw, "s:suite", PyST_SUITE)); 524} 525 526 527 528/* This is the messy part of the code. Conversion from a tuple to an ST 529 * object requires that the input tuple be valid without having to rely on 530 * catching an exception from the compiler. This is done to allow the 531 * compiler itself to remain fast, since most of its input will come from 532 * the parser directly, and therefore be known to be syntactically correct. 533 * This validation is done to ensure that we don't core dump the compile 534 * phase, returning an exception instead. 535 * 536 * Two aspects can be broken out in this code: creating a node tree from 537 * the tuple passed in, and verifying that it is indeed valid. It may be 538 * advantageous to expand the number of ST types to include funcdefs and 539 * lambdadefs to take advantage of the optimizer, recognizing those STs 540 * here. They are not necessary, and not quite as useful in a raw form. 541 * For now, let's get expressions and suites working reliably. 542 */ 543 544 545static node* build_node_tree(PyObject *tuple); 546static int validate_expr_tree(node *tree); 547static int validate_file_input(node *tree); 548static int validate_encoding_decl(node *tree); 549 550/* PyObject* parser_tuple2st(PyObject* self, PyObject* args) 551 * 552 * This is the public function, called from the Python code. It receives a 553 * single tuple object from the caller, and creates an ST object if the 554 * tuple can be validated. It does this by checking the first code of the 555 * tuple, and, if acceptable, builds the internal representation. If this 556 * step succeeds, the internal representation is validated as fully as 557 * possible with the various validate_*() routines defined below. 558 * 559 * This function must be changed if support is to be added for PyST_FRAGMENT 560 * ST objects. 561 * 562 */ 563static PyObject* 564parser_tuple2st(PyST_Object *self, PyObject *args, PyObject *kw) 565{ 566 NOTE(ARGUNUSED(self)) 567 PyObject *st = 0; 568 PyObject *tuple; 569 node *tree; 570 571 static char *keywords[] = {"sequence", NULL}; 572 573 if (!PyArg_ParseTupleAndKeywords(args, kw, "O:sequence2st", keywords, 574 &tuple)) 575 return (0); 576 if (!PySequence_Check(tuple)) { 577 PyErr_SetString(PyExc_ValueError, 578 "sequence2st() requires a single sequence argument"); 579 return (0); 580 } 581 /* 582 * Convert the tree to the internal form before checking it. 583 */ 584 tree = build_node_tree(tuple); 585 if (tree != 0) { 586 int start_sym = TYPE(tree); 587 if (start_sym == eval_input) { 588 /* Might be an eval form. */ 589 if (validate_expr_tree(tree)) 590 st = parser_newstobject(tree, PyST_EXPR); 591 else 592 PyNode_Free(tree); 593 } 594 else if (start_sym == file_input) { 595 /* This looks like an exec form so far. */ 596 if (validate_file_input(tree)) 597 st = parser_newstobject(tree, PyST_SUITE); 598 else 599 PyNode_Free(tree); 600 } 601 else if (start_sym == encoding_decl) { 602 /* This looks like an encoding_decl so far. */ 603 if (validate_encoding_decl(tree)) 604 st = parser_newstobject(tree, PyST_SUITE); 605 else 606 PyNode_Free(tree); 607 } 608 else { 609 /* This is a fragment, at best. */ 610 PyNode_Free(tree); 611 err_string("parse tree does not use a valid start symbol"); 612 } 613 } 614 /* Make sure we throw an exception on all errors. We should never 615 * get this, but we'd do well to be sure something is done. 616 */ 617 if (st == NULL && !PyErr_Occurred()) 618 err_string("unspecified ST error occurred"); 619 620 return st; 621} 622 623 624/* node* build_node_children() 625 * 626 * Iterate across the children of the current non-terminal node and build 627 * their structures. If successful, return the root of this portion of 628 * the tree, otherwise, 0. Any required exception will be specified already, 629 * and no memory will have been deallocated. 630 * 631 */ 632static node* 633build_node_children(PyObject *tuple, node *root, int *line_num) 634{ 635 Py_ssize_t len = PyObject_Size(tuple); 636 Py_ssize_t i; 637 int err; 638 639 for (i = 1; i < len; ++i) { 640 /* elem must always be a sequence, however simple */ 641 PyObject* elem = PySequence_GetItem(tuple, i); 642 int ok = elem != NULL; 643 long type = 0; 644 char *strn = 0; 645 646 if (ok) 647 ok = PySequence_Check(elem); 648 if (ok) { 649 PyObject *temp = PySequence_GetItem(elem, 0); 650 if (temp == NULL) 651 ok = 0; 652 else { 653 ok = PyInt_Check(temp); 654 if (ok) 655 type = PyInt_AS_LONG(temp); 656 Py_DECREF(temp); 657 } 658 } 659 if (!ok) { 660 PyObject *err = Py_BuildValue("os", elem, 661 "Illegal node construct."); 662 PyErr_SetObject(parser_error, err); 663 Py_XDECREF(err); 664 Py_XDECREF(elem); 665 return (0); 666 } 667 if (ISTERMINAL(type)) { 668 Py_ssize_t len = PyObject_Size(elem); 669 PyObject *temp; 670 671 if ((len != 2) && (len != 3)) { 672 err_string("terminal nodes must have 2 or 3 entries"); 673 return 0; 674 } 675 temp = PySequence_GetItem(elem, 1); 676 if (temp == NULL) 677 return 0; 678 if (!PyString_Check(temp)) { 679 PyErr_Format(parser_error, 680 "second item in terminal node must be a string," 681 " found %s", 682 temp->ob_type->tp_name); 683 Py_DECREF(temp); 684 return 0; 685 } 686 if (len == 3) { 687 PyObject *o = PySequence_GetItem(elem, 2); 688 if (o != NULL) { 689 if (PyInt_Check(o)) 690 *line_num = PyInt_AS_LONG(o); 691 else { 692 PyErr_Format(parser_error, 693 "third item in terminal node must be an" 694 " integer, found %s", 695 temp->ob_type->tp_name); 696 Py_DECREF(o); 697 Py_DECREF(temp); 698 return 0; 699 } 700 Py_DECREF(o); 701 } 702 } 703 len = PyString_GET_SIZE(temp) + 1; 704 strn = (char *)PyObject_MALLOC(len); 705 if (strn != NULL) 706 (void) memcpy(strn, PyString_AS_STRING(temp), len); 707 Py_DECREF(temp); 708 } 709 else if (!ISNONTERMINAL(type)) { 710 /* 711 * It has to be one or the other; this is an error. 712 * Throw an exception. 713 */ 714 PyObject *err = Py_BuildValue("os", elem, "unknown node type."); 715 PyErr_SetObject(parser_error, err); 716 Py_XDECREF(err); 717 Py_XDECREF(elem); 718 return (0); 719 } 720 err = PyNode_AddChild(root, type, strn, *line_num, 0); 721 if (err == E_NOMEM) { 722 PyObject_FREE(strn); 723 return (node *) PyErr_NoMemory(); 724 } 725 if (err == E_OVERFLOW) { 726 PyObject_FREE(strn); 727 PyErr_SetString(PyExc_ValueError, 728 "unsupported number of child nodes"); 729 return NULL; 730 } 731 732 if (ISNONTERMINAL(type)) { 733 node* new_child = CHILD(root, i - 1); 734 735 if (new_child != build_node_children(elem, new_child, line_num)) { 736 Py_XDECREF(elem); 737 return (0); 738 } 739 } 740 else if (type == NEWLINE) { /* It's true: we increment the */ 741 ++(*line_num); /* line number *after* the newline! */ 742 } 743 Py_XDECREF(elem); 744 } 745 return root; 746} 747 748 749static node* 750build_node_tree(PyObject *tuple) 751{ 752 node* res = 0; 753 PyObject *temp = PySequence_GetItem(tuple, 0); 754 long num = -1; 755 756 if (temp != NULL) 757 num = PyInt_AsLong(temp); 758 Py_XDECREF(temp); 759 if (ISTERMINAL(num)) { 760 /* 761 * The tuple is simple, but it doesn't start with a start symbol. 762 * Throw an exception now and be done with it. 763 */ 764 tuple = Py_BuildValue("os", tuple, 765 "Illegal syntax-tree; cannot start with terminal symbol."); 766 PyErr_SetObject(parser_error, tuple); 767 Py_XDECREF(tuple); 768 } 769 else if (ISNONTERMINAL(num)) { 770 /* 771 * Not efficient, but that can be handled later. 772 */ 773 int line_num = 0; 774 PyObject *encoding = NULL; 775 776 if (num == encoding_decl) { 777 encoding = PySequence_GetItem(tuple, 2); 778 /* tuple isn't borrowed anymore here, need to DECREF */ 779 tuple = PySequence_GetSlice(tuple, 0, 2); 780 } 781 res = PyNode_New(num); 782 if (res != NULL) { 783 if (res != build_node_children(tuple, res, &line_num)) { 784 PyNode_Free(res); 785 res = NULL; 786 } 787 if (res && encoding) { 788 Py_ssize_t len; 789 len = PyString_GET_SIZE(encoding) + 1; 790 res->n_str = (char *)PyObject_MALLOC(len); 791 if (res->n_str != NULL) 792 (void) memcpy(res->n_str, PyString_AS_STRING(encoding), len); 793 Py_DECREF(encoding); 794 Py_DECREF(tuple); 795 } 796 } 797 } 798 else { 799 /* The tuple is illegal -- if the number is neither TERMINAL nor 800 * NONTERMINAL, we can't use it. Not sure the implementation 801 * allows this condition, but the API doesn't preclude it. 802 */ 803 PyObject *err = Py_BuildValue("os", tuple, 804 "Illegal component tuple."); 805 PyErr_SetObject(parser_error, err); 806 Py_XDECREF(err); 807 } 808 809 return (res); 810} 811 812 813/* 814 * Validation routines used within the validation section: 815 */ 816static int validate_terminal(node *terminal, int type, char *string); 817 818#define validate_ampersand(ch) validate_terminal(ch, AMPER, "&") 819#define validate_circumflex(ch) validate_terminal(ch, CIRCUMFLEX, "^") 820#define validate_colon(ch) validate_terminal(ch, COLON, ":") 821#define validate_comma(ch) validate_terminal(ch, COMMA, ",") 822#define validate_dedent(ch) validate_terminal(ch, DEDENT, "") 823#define validate_equal(ch) validate_terminal(ch, EQUAL, "=") 824#define validate_indent(ch) validate_terminal(ch, INDENT, (char*)NULL) 825#define validate_lparen(ch) validate_terminal(ch, LPAR, "(") 826#define validate_newline(ch) validate_terminal(ch, NEWLINE, (char*)NULL) 827#define validate_rparen(ch) validate_terminal(ch, RPAR, ")") 828#define validate_semi(ch) validate_terminal(ch, SEMI, ";") 829#define validate_star(ch) validate_terminal(ch, STAR, "*") 830#define validate_vbar(ch) validate_terminal(ch, VBAR, "|") 831#define validate_doublestar(ch) validate_terminal(ch, DOUBLESTAR, "**") 832#define validate_dot(ch) validate_terminal(ch, DOT, ".") 833#define validate_at(ch) validate_terminal(ch, AT, "@") 834#define validate_name(ch, str) validate_terminal(ch, NAME, str) 835 836#define VALIDATER(n) static int validate_##n(node *tree) 837 838VALIDATER(node); VALIDATER(small_stmt); 839VALIDATER(class); VALIDATER(node); 840VALIDATER(parameters); VALIDATER(suite); 841VALIDATER(testlist); VALIDATER(varargslist); 842VALIDATER(fpdef); VALIDATER(fplist); 843VALIDATER(stmt); VALIDATER(simple_stmt); 844VALIDATER(expr_stmt); VALIDATER(power); 845VALIDATER(print_stmt); VALIDATER(del_stmt); 846VALIDATER(return_stmt); VALIDATER(list_iter); 847VALIDATER(raise_stmt); VALIDATER(import_stmt); 848VALIDATER(import_name); VALIDATER(import_from); 849VALIDATER(global_stmt); VALIDATER(list_if); 850VALIDATER(assert_stmt); VALIDATER(list_for); 851VALIDATER(compound_stmt); 852VALIDATER(while); VALIDATER(for); 853VALIDATER(try); VALIDATER(except_clause); 854VALIDATER(test); VALIDATER(and_test); 855VALIDATER(not_test); VALIDATER(comparison); 856VALIDATER(comp_op); VALIDATER(expr); 857VALIDATER(xor_expr); VALIDATER(and_expr); 858VALIDATER(shift_expr); VALIDATER(arith_expr); 859VALIDATER(term); VALIDATER(factor); 860VALIDATER(atom); VALIDATER(lambdef); 861VALIDATER(trailer); VALIDATER(subscript); 862VALIDATER(subscriptlist); VALIDATER(sliceop); 863VALIDATER(exprlist); VALIDATER(dictsetmaker); 864VALIDATER(arglist); VALIDATER(argument); 865VALIDATER(listmaker); VALIDATER(yield_stmt); 866VALIDATER(testlist1); VALIDATER(gen_for); 867VALIDATER(gen_iter); VALIDATER(gen_if); 868VALIDATER(testlist_gexp); VALIDATER(yield_expr); 869VALIDATER(yield_or_testlist); VALIDATER(or_test); 870VALIDATER(old_test); VALIDATER(old_lambdef); 871 872#undef VALIDATER 873 874#define is_even(n) (((n) & 1) == 0) 875#define is_odd(n) (((n) & 1) == 1) 876 877 878static int 879validate_ntype(node *n, int t) 880{ 881 if (TYPE(n) != t) { 882 PyErr_Format(parser_error, "Expected node type %d, got %d.", 883 t, TYPE(n)); 884 return 0; 885 } 886 return 1; 887} 888 889 890/* Verifies that the number of child nodes is exactly 'num', raising 891 * an exception if it isn't. The exception message does not indicate 892 * the exact number of nodes, allowing this to be used to raise the 893 * "right" exception when the wrong number of nodes is present in a 894 * specific variant of a statement's syntax. This is commonly used 895 * in that fashion. 896 */ 897static int 898validate_numnodes(node *n, int num, const char *const name) 899{ 900 if (NCH(n) != num) { 901 PyErr_Format(parser_error, 902 "Illegal number of children for %s node.", name); 903 return 0; 904 } 905 return 1; 906} 907 908 909static int 910validate_terminal(node *terminal, int type, char *string) 911{ 912 int res = (validate_ntype(terminal, type) 913 && ((string == 0) || (strcmp(string, STR(terminal)) == 0))); 914 915 if (!res && !PyErr_Occurred()) { 916 PyErr_Format(parser_error, 917 "Illegal terminal: expected \"%s\"", string); 918 } 919 return (res); 920} 921 922 923/* X (',' X) [','] 924 */ 925static int 926validate_repeating_list(node *tree, int ntype, int (*vfunc)(node *), 927 const char *const name) 928{ 929 int nch = NCH(tree); 930 int res = (nch && validate_ntype(tree, ntype) 931 && vfunc(CHILD(tree, 0))); 932 933 if (!res && !PyErr_Occurred()) 934 (void) validate_numnodes(tree, 1, name); 935 else { 936 if (is_even(nch)) 937 res = validate_comma(CHILD(tree, --nch)); 938 if (res && nch > 1) { 939 int pos = 1; 940 for ( ; res && pos < nch; pos += 2) 941 res = (validate_comma(CHILD(tree, pos)) 942 && vfunc(CHILD(tree, pos + 1))); 943 } 944 } 945 return (res); 946} 947 948 949/* validate_class() 950 * 951 * classdef: 952 * 'class' NAME ['(' testlist ')'] ':' suite 953 */ 954static int 955validate_class(node *tree) 956{ 957 int nch = NCH(tree); 958 int res = (validate_ntype(tree, classdef) && 959 ((nch == 4) || (nch == 6) || (nch == 7))); 960 961 if (res) { 962 res = (validate_name(CHILD(tree, 0), "class") 963 && validate_ntype(CHILD(tree, 1), NAME) 964 && validate_colon(CHILD(tree, nch - 2)) 965 && validate_suite(CHILD(tree, nch - 1))); 966 } 967 else { 968 (void) validate_numnodes(tree, 4, "class"); 969 } 970 971 if (res) { 972 if (nch == 7) { 973 res = ((validate_lparen(CHILD(tree, 2)) && 974 validate_testlist(CHILD(tree, 3)) && 975 validate_rparen(CHILD(tree, 4)))); 976 } 977 else if (nch == 6) { 978 res = (validate_lparen(CHILD(tree,2)) && 979 validate_rparen(CHILD(tree,3))); 980 } 981 } 982 return (res); 983} 984 985 986/* if_stmt: 987 * 'if' test ':' suite ('elif' test ':' suite)* ['else' ':' suite] 988 */ 989static int 990validate_if(node *tree) 991{ 992 int nch = NCH(tree); 993 int res = (validate_ntype(tree, if_stmt) 994 && (nch >= 4) 995 && validate_name(CHILD(tree, 0), "if") 996 && validate_test(CHILD(tree, 1)) 997 && validate_colon(CHILD(tree, 2)) 998 && validate_suite(CHILD(tree, 3))); 999 1000 if (res && ((nch % 4) == 3)) { 1001 /* ... 'else' ':' suite */ 1002 res = (validate_name(CHILD(tree, nch - 3), "else") 1003 && validate_colon(CHILD(tree, nch - 2)) 1004 && validate_suite(CHILD(tree, nch - 1))); 1005 nch -= 3; 1006 } 1007 else if (!res && !PyErr_Occurred()) 1008 (void) validate_numnodes(tree, 4, "if"); 1009 if ((nch % 4) != 0) 1010 /* Will catch the case for nch < 4 */ 1011 res = validate_numnodes(tree, 0, "if"); 1012 else if (res && (nch > 4)) { 1013 /* ... ('elif' test ':' suite)+ ... */ 1014 int j = 4; 1015 while ((j < nch) && res) { 1016 res = (validate_name(CHILD(tree, j), "elif") 1017 && validate_colon(CHILD(tree, j + 2)) 1018 && validate_test(CHILD(tree, j + 1)) 1019 && validate_suite(CHILD(tree, j + 3))); 1020 j += 4; 1021 } 1022 } 1023 return (res); 1024} 1025 1026 1027/* parameters: 1028 * '(' [varargslist] ')' 1029 * 1030 */ 1031static int 1032validate_parameters(node *tree) 1033{ 1034 int nch = NCH(tree); 1035 int res = validate_ntype(tree, parameters) && ((nch == 2) || (nch == 3)); 1036 1037 if (res) { 1038 res = (validate_lparen(CHILD(tree, 0)) 1039 && validate_rparen(CHILD(tree, nch - 1))); 1040 if (res && (nch == 3)) 1041 res = validate_varargslist(CHILD(tree, 1)); 1042 } 1043 else { 1044 (void) validate_numnodes(tree, 2, "parameters"); 1045 } 1046 return (res); 1047} 1048 1049 1050/* validate_suite() 1051 * 1052 * suite: 1053 * simple_stmt 1054 * | NEWLINE INDENT stmt+ DEDENT 1055 */ 1056static int 1057validate_suite(node *tree) 1058{ 1059 int nch = NCH(tree); 1060 int res = (validate_ntype(tree, suite) && ((nch == 1) || (nch >= 4))); 1061 1062 if (res && (nch == 1)) 1063 res = validate_simple_stmt(CHILD(tree, 0)); 1064 else if (res) { 1065 /* NEWLINE INDENT stmt+ DEDENT */ 1066 res = (validate_newline(CHILD(tree, 0)) 1067 && validate_indent(CHILD(tree, 1)) 1068 && validate_stmt(CHILD(tree, 2)) 1069 && validate_dedent(CHILD(tree, nch - 1))); 1070 1071 if (res && (nch > 4)) { 1072 int i = 3; 1073 --nch; /* forget the DEDENT */ 1074 for ( ; res && (i < nch); ++i) 1075 res = validate_stmt(CHILD(tree, i)); 1076 } 1077 else if (nch < 4) 1078 res = validate_numnodes(tree, 4, "suite"); 1079 } 1080 return (res); 1081} 1082 1083 1084static int 1085validate_testlist(node *tree) 1086{ 1087 return (validate_repeating_list(tree, testlist, 1088 validate_test, "testlist")); 1089} 1090 1091 1092static int 1093validate_testlist1(node *tree) 1094{ 1095 return (validate_repeating_list(tree, testlist1, 1096 validate_test, "testlist1")); 1097} 1098 1099 1100static int 1101validate_testlist_safe(node *tree) 1102{ 1103 return (validate_repeating_list(tree, testlist_safe, 1104 validate_old_test, "testlist_safe")); 1105} 1106 1107 1108/* '*' NAME [',' '**' NAME] | '**' NAME 1109 */ 1110static int 1111validate_varargslist_trailer(node *tree, int start) 1112{ 1113 int nch = NCH(tree); 1114 int res = 0; 1115 int sym; 1116 1117 if (nch <= start) { 1118 err_string("expected variable argument trailer for varargslist"); 1119 return 0; 1120 } 1121 sym = TYPE(CHILD(tree, start)); 1122 if (sym == STAR) { 1123 /* 1124 * ('*' NAME [',' '**' NAME] 1125 */ 1126 if (nch-start == 2) 1127 res = validate_name(CHILD(tree, start+1), NULL); 1128 else if (nch-start == 5) 1129 res = (validate_name(CHILD(tree, start+1), NULL) 1130 && validate_comma(CHILD(tree, start+2)) 1131 && validate_doublestar(CHILD(tree, start+3)) 1132 && validate_name(CHILD(tree, start+4), NULL)); 1133 } 1134 else if (sym == DOUBLESTAR) { 1135 /* 1136 * '**' NAME 1137 */ 1138 if (nch-start == 2) 1139 res = validate_name(CHILD(tree, start+1), NULL); 1140 } 1141 if (!res) 1142 err_string("illegal variable argument trailer for varargslist"); 1143 return res; 1144} 1145 1146 1147/* validate_varargslist() 1148 * 1149 * varargslist: 1150 * (fpdef ['=' test] ',')* 1151 * ('*' NAME [',' '**' NAME] 1152 * | '**' NAME) 1153 * | fpdef ['=' test] (',' fpdef ['=' test])* [','] 1154 * 1155 */ 1156static int 1157validate_varargslist(node *tree) 1158{ 1159 int nch = NCH(tree); 1160 int res = validate_ntype(tree, varargslist) && (nch != 0); 1161 int sym; 1162 1163 if (!res) 1164 return 0; 1165 if (nch < 1) { 1166 err_string("varargslist missing child nodes"); 1167 return 0; 1168 } 1169 sym = TYPE(CHILD(tree, 0)); 1170 if (sym == STAR || sym == DOUBLESTAR) 1171 /* whole thing matches: 1172 * '*' NAME [',' '**' NAME] | '**' NAME 1173 */ 1174 res = validate_varargslist_trailer(tree, 0); 1175 else if (sym == fpdef) { 1176 int i = 0; 1177 1178 sym = TYPE(CHILD(tree, nch-1)); 1179 if (sym == NAME) { 1180 /* 1181 * (fpdef ['=' test] ',')+ 1182 * ('*' NAME [',' '**' NAME] 1183 * | '**' NAME) 1184 */ 1185 /* skip over (fpdef ['=' test] ',')+ */ 1186 while (res && (i+2 <= nch)) { 1187 res = validate_fpdef(CHILD(tree, i)); 1188 ++i; 1189 if (res && TYPE(CHILD(tree, i)) == EQUAL && (i+2 <= nch)) { 1190 res = (validate_equal(CHILD(tree, i)) 1191 && validate_test(CHILD(tree, i+1))); 1192 if (res) 1193 i += 2; 1194 } 1195 if (res && i < nch) { 1196 res = validate_comma(CHILD(tree, i)); 1197 ++i; 1198 if (res && i < nch 1199 && (TYPE(CHILD(tree, i)) == DOUBLESTAR 1200 || TYPE(CHILD(tree, i)) == STAR)) 1201 break; 1202 } 1203 } 1204 /* ... '*' NAME [',' '**' NAME] | '**' NAME 1205 * i --^^^ 1206 */ 1207 if (res) 1208 res = validate_varargslist_trailer(tree, i); 1209 } 1210 else { 1211 /* 1212 * fpdef ['=' test] (',' fpdef ['=' test])* [','] 1213 */ 1214 /* strip trailing comma node */ 1215 if (sym == COMMA) { 1216 res = validate_comma(CHILD(tree, nch-1)); 1217 if (!res) 1218 return 0; 1219 --nch; 1220 } 1221 /* 1222 * fpdef ['=' test] (',' fpdef ['=' test])* 1223 */ 1224 res = validate_fpdef(CHILD(tree, 0)); 1225 ++i; 1226 if (res && (i+2 <= nch) && TYPE(CHILD(tree, i)) == EQUAL) { 1227 res = (validate_equal(CHILD(tree, i)) 1228 && validate_test(CHILD(tree, i+1))); 1229 i += 2; 1230 } 1231 /* 1232 * ... (',' fpdef ['=' test])* 1233 * i ---^^^ 1234 */ 1235 while (res && (nch - i) >= 2) { 1236 res = (validate_comma(CHILD(tree, i)) 1237 && validate_fpdef(CHILD(tree, i+1))); 1238 i += 2; 1239 if (res && (nch - i) >= 2 && TYPE(CHILD(tree, i)) == EQUAL) { 1240 res = (validate_equal(CHILD(tree, i)) 1241 && validate_test(CHILD(tree, i+1))); 1242 i += 2; 1243 } 1244 } 1245 if (res && nch - i != 0) { 1246 res = 0; 1247 err_string("illegal formation for varargslist"); 1248 } 1249 } 1250 } 1251 return res; 1252} 1253 1254 1255/* list_iter: list_for | list_if 1256 */ 1257static int 1258validate_list_iter(node *tree) 1259{ 1260 int res = (validate_ntype(tree, list_iter) 1261 && validate_numnodes(tree, 1, "list_iter")); 1262 if (res && TYPE(CHILD(tree, 0)) == list_for) 1263 res = validate_list_for(CHILD(tree, 0)); 1264 else 1265 res = validate_list_if(CHILD(tree, 0)); 1266 1267 return res; 1268} 1269 1270/* gen_iter: gen_for | gen_if 1271 */ 1272static int 1273validate_gen_iter(node *tree) 1274{ 1275 int res = (validate_ntype(tree, gen_iter) 1276 && validate_numnodes(tree, 1, "gen_iter")); 1277 if (res && TYPE(CHILD(tree, 0)) == gen_for) 1278 res = validate_gen_for(CHILD(tree, 0)); 1279 else 1280 res = validate_gen_if(CHILD(tree, 0)); 1281 1282 return res; 1283} 1284 1285/* list_for: 'for' exprlist 'in' testlist [list_iter] 1286 */ 1287static int 1288validate_list_for(node *tree) 1289{ 1290 int nch = NCH(tree); 1291 int res; 1292 1293 if (nch == 5) 1294 res = validate_list_iter(CHILD(tree, 4)); 1295 else 1296 res = validate_numnodes(tree, 4, "list_for"); 1297 1298 if (res) 1299 res = (validate_name(CHILD(tree, 0), "for") 1300 && validate_exprlist(CHILD(tree, 1)) 1301 && validate_name(CHILD(tree, 2), "in") 1302 && validate_testlist_safe(CHILD(tree, 3))); 1303 1304 return res; 1305} 1306 1307/* gen_for: 'for' exprlist 'in' test [gen_iter] 1308 */ 1309static int 1310validate_gen_for(node *tree) 1311{ 1312 int nch = NCH(tree); 1313 int res; 1314 1315 if (nch == 5) 1316 res = validate_gen_iter(CHILD(tree, 4)); 1317 else 1318 res = validate_numnodes(tree, 4, "gen_for"); 1319 1320 if (res) 1321 res = (validate_name(CHILD(tree, 0), "for") 1322 && validate_exprlist(CHILD(tree, 1)) 1323 && validate_name(CHILD(tree, 2), "in") 1324 && validate_or_test(CHILD(tree, 3))); 1325 1326 return res; 1327} 1328 1329/* list_if: 'if' old_test [list_iter] 1330 */ 1331static int 1332validate_list_if(node *tree) 1333{ 1334 int nch = NCH(tree); 1335 int res; 1336 1337 if (nch == 3) 1338 res = validate_list_iter(CHILD(tree, 2)); 1339 else 1340 res = validate_numnodes(tree, 2, "list_if"); 1341 1342 if (res) 1343 res = (validate_name(CHILD(tree, 0), "if") 1344 && validate_old_test(CHILD(tree, 1))); 1345 1346 return res; 1347} 1348 1349/* gen_if: 'if' old_test [gen_iter] 1350 */ 1351static int 1352validate_gen_if(node *tree) 1353{ 1354 int nch = NCH(tree); 1355 int res; 1356 1357 if (nch == 3) 1358 res = validate_gen_iter(CHILD(tree, 2)); 1359 else 1360 res = validate_numnodes(tree, 2, "gen_if"); 1361 1362 if (res) 1363 res = (validate_name(CHILD(tree, 0), "if") 1364 && validate_old_test(CHILD(tree, 1))); 1365 1366 return res; 1367} 1368 1369/* validate_fpdef() 1370 * 1371 * fpdef: 1372 * NAME 1373 * | '(' fplist ')' 1374 */ 1375static int 1376validate_fpdef(node *tree) 1377{ 1378 int nch = NCH(tree); 1379 int res = validate_ntype(tree, fpdef); 1380 1381 if (res) { 1382 if (nch == 1) 1383 res = validate_ntype(CHILD(tree, 0), NAME); 1384 else if (nch == 3) 1385 res = (validate_lparen(CHILD(tree, 0)) 1386 && validate_fplist(CHILD(tree, 1)) 1387 && validate_rparen(CHILD(tree, 2))); 1388 else 1389 res = validate_numnodes(tree, 1, "fpdef"); 1390 } 1391 return (res); 1392} 1393 1394 1395static int 1396validate_fplist(node *tree) 1397{ 1398 return (validate_repeating_list(tree, fplist, 1399 validate_fpdef, "fplist")); 1400} 1401 1402 1403/* simple_stmt | compound_stmt 1404 * 1405 */ 1406static int 1407validate_stmt(node *tree) 1408{ 1409 int res = (validate_ntype(tree, stmt) 1410 && validate_numnodes(tree, 1, "stmt")); 1411 1412 if (res) { 1413 tree = CHILD(tree, 0); 1414 1415 if (TYPE(tree) == simple_stmt) 1416 res = validate_simple_stmt(tree); 1417 else 1418 res = validate_compound_stmt(tree); 1419 } 1420 return (res); 1421} 1422 1423 1424/* small_stmt (';' small_stmt)* [';'] NEWLINE 1425 * 1426 */ 1427static int 1428validate_simple_stmt(node *tree) 1429{ 1430 int nch = NCH(tree); 1431 int res = (validate_ntype(tree, simple_stmt) 1432 && (nch >= 2) 1433 && validate_small_stmt(CHILD(tree, 0)) 1434 && validate_newline(CHILD(tree, nch - 1))); 1435 1436 if (nch < 2) 1437 res = validate_numnodes(tree, 2, "simple_stmt"); 1438 --nch; /* forget the NEWLINE */ 1439 if (res && is_even(nch)) 1440 res = validate_semi(CHILD(tree, --nch)); 1441 if (res && (nch > 2)) { 1442 int i; 1443 1444 for (i = 1; res && (i < nch); i += 2) 1445 res = (validate_semi(CHILD(tree, i)) 1446 && validate_small_stmt(CHILD(tree, i + 1))); 1447 } 1448 return (res); 1449} 1450 1451 1452static int 1453validate_small_stmt(node *tree) 1454{ 1455 int nch = NCH(tree); 1456 int res = validate_numnodes(tree, 1, "small_stmt"); 1457 1458 if (res) { 1459 int ntype = TYPE(CHILD(tree, 0)); 1460 1461 if ( (ntype == expr_stmt) 1462 || (ntype == print_stmt) 1463 || (ntype == del_stmt) 1464 || (ntype == pass_stmt) 1465 || (ntype == flow_stmt) 1466 || (ntype == import_stmt) 1467 || (ntype == global_stmt) 1468 || (ntype == assert_stmt)) 1469 res = validate_node(CHILD(tree, 0)); 1470 else { 1471 res = 0; 1472 err_string("illegal small_stmt child type"); 1473 } 1474 } 1475 else if (nch == 1) { 1476 res = 0; 1477 PyErr_Format(parser_error, 1478 "Unrecognized child node of small_stmt: %d.", 1479 TYPE(CHILD(tree, 0))); 1480 } 1481 return (res); 1482} 1483 1484 1485/* compound_stmt: 1486 * if_stmt | while_stmt | for_stmt | try_stmt | funcdef | classdef 1487 */ 1488static int 1489validate_compound_stmt(node *tree) 1490{ 1491 int res = (validate_ntype(tree, compound_stmt) 1492 && validate_numnodes(tree, 1, "compound_stmt")); 1493 int ntype; 1494 1495 if (!res) 1496 return (0); 1497 1498 tree = CHILD(tree, 0); 1499 ntype = TYPE(tree); 1500 if ( (ntype == if_stmt) 1501 || (ntype == while_stmt) 1502 || (ntype == for_stmt) 1503 || (ntype == try_stmt) 1504 || (ntype == funcdef) 1505 || (ntype == classdef)) 1506 res = validate_node(tree); 1507 else { 1508 res = 0; 1509 PyErr_Format(parser_error, 1510 "Illegal compound statement type: %d.", TYPE(tree)); 1511 } 1512 return (res); 1513} 1514 1515 1516static int 1517validate_yield_or_testlist(node *tree) 1518{ 1519 if (TYPE(tree) == yield_expr) 1520 return validate_yield_expr(tree); 1521 else 1522 return validate_testlist(tree); 1523} 1524 1525static int 1526validate_expr_stmt(node *tree) 1527{ 1528 int j; 1529 int nch = NCH(tree); 1530 int res = (validate_ntype(tree, expr_stmt) 1531 && is_odd(nch) 1532 && validate_testlist(CHILD(tree, 0))); 1533 1534 if (res && nch == 3 1535 && TYPE(CHILD(tree, 1)) == augassign) { 1536 res = validate_numnodes(CHILD(tree, 1), 1, "augassign") 1537 && validate_yield_or_testlist(CHILD(tree, 2)); 1538 1539 if (res) { 1540 char *s = STR(CHILD(CHILD(tree, 1), 0)); 1541 1542 res = (strcmp(s, "+=") == 0 1543 || strcmp(s, "-=") == 0 1544 || strcmp(s, "*=") == 0 1545 || strcmp(s, "/=") == 0 1546 || strcmp(s, "//=") == 0 1547 || strcmp(s, "%=") == 0 1548 || strcmp(s, "&=") == 0 1549 || strcmp(s, "|=") == 0 1550 || strcmp(s, "^=") == 0 1551 || strcmp(s, "<<=") == 0 1552 || strcmp(s, ">>=") == 0 1553 || strcmp(s, "**=") == 0); 1554 if (!res) 1555 err_string("illegal augmmented assignment operator"); 1556 } 1557 } 1558 else { 1559 for (j = 1; res && (j < nch); j += 2) 1560 res = validate_equal(CHILD(tree, j)) 1561 && validate_yield_or_testlist(CHILD(tree, j + 1)); 1562 } 1563 return (res); 1564} 1565 1566 1567/* print_stmt: 1568 * 1569 * 'print' ( [ test (',' test)* [','] ] 1570 * | '>>' test [ (',' test)+ [','] ] ) 1571 */ 1572static int 1573validate_print_stmt(node *tree) 1574{ 1575 int nch = NCH(tree); 1576 int res = (validate_ntype(tree, print_stmt) 1577 && (nch > 0) 1578 && validate_name(CHILD(tree, 0), "print")); 1579 1580 if (res && nch > 1) { 1581 int sym = TYPE(CHILD(tree, 1)); 1582 int i = 1; 1583 int allow_trailing_comma = 1; 1584 1585 if (sym == test) 1586 res = validate_test(CHILD(tree, i++)); 1587 else { 1588 if (nch < 3) 1589 res = validate_numnodes(tree, 3, "print_stmt"); 1590 else { 1591 res = (validate_ntype(CHILD(tree, i), RIGHTSHIFT) 1592 && validate_test(CHILD(tree, i+1))); 1593 i += 2; 1594 allow_trailing_comma = 0; 1595 } 1596 } 1597 if (res) { 1598 /* ... (',' test)* [','] */ 1599 while (res && i+2 <= nch) { 1600 res = (validate_comma(CHILD(tree, i)) 1601 && validate_test(CHILD(tree, i+1))); 1602 allow_trailing_comma = 1; 1603 i += 2; 1604 } 1605 if (res && !allow_trailing_comma) 1606 res = validate_numnodes(tree, i, "print_stmt"); 1607 else if (res && i < nch) 1608 res = validate_comma(CHILD(tree, i)); 1609 } 1610 } 1611 return (res); 1612} 1613 1614 1615static int 1616validate_del_stmt(node *tree) 1617{ 1618 return (validate_numnodes(tree, 2, "del_stmt") 1619 && validate_name(CHILD(tree, 0), "del") 1620 && validate_exprlist(CHILD(tree, 1))); 1621} 1622 1623 1624static int 1625validate_return_stmt(node *tree) 1626{ 1627 int nch = NCH(tree); 1628 int res = (validate_ntype(tree, return_stmt) 1629 && ((nch == 1) || (nch == 2)) 1630 && validate_name(CHILD(tree, 0), "return")); 1631 1632 if (res && (nch == 2)) 1633 res = validate_testlist(CHILD(tree, 1)); 1634 1635 return (res); 1636} 1637 1638 1639static int 1640validate_raise_stmt(node *tree) 1641{ 1642 int nch = NCH(tree); 1643 int res = (validate_ntype(tree, raise_stmt) 1644 && ((nch == 1) || (nch == 2) || (nch == 4) || (nch == 6))); 1645 1646 if (res) { 1647 res = validate_name(CHILD(tree, 0), "raise"); 1648 if (res && (nch >= 2)) 1649 res = validate_test(CHILD(tree, 1)); 1650 if (res && nch > 2) { 1651 res = (validate_comma(CHILD(tree, 2)) 1652 && validate_test(CHILD(tree, 3))); 1653 if (res && (nch > 4)) 1654 res = (validate_comma(CHILD(tree, 4)) 1655 && validate_test(CHILD(tree, 5))); 1656 } 1657 } 1658 else 1659 (void) validate_numnodes(tree, 2, "raise"); 1660 if (res && (nch == 4)) 1661 res = (validate_comma(CHILD(tree, 2)) 1662 && validate_test(CHILD(tree, 3))); 1663 1664 return (res); 1665} 1666 1667 1668/* yield_expr: 'yield' [testlist] 1669 */ 1670static int 1671validate_yield_expr(node *tree) 1672{ 1673 int nch = NCH(tree); 1674 int res = (validate_ntype(tree, yield_expr) 1675 && ((nch == 1) || (nch == 2)) 1676 && validate_name(CHILD(tree, 0), "yield")); 1677 1678 if (res && (nch == 2)) 1679 res = validate_testlist(CHILD(tree, 1)); 1680 1681 return (res); 1682} 1683 1684 1685/* yield_stmt: yield_expr 1686 */ 1687static int 1688validate_yield_stmt(node *tree) 1689{ 1690 return (validate_ntype(tree, yield_stmt) 1691 && validate_numnodes(tree, 1, "yield_stmt") 1692 && validate_yield_expr(CHILD(tree, 0))); 1693} 1694 1695 1696static int 1697validate_import_as_name(node *tree) 1698{ 1699 int nch = NCH(tree); 1700 int ok = validate_ntype(tree, import_as_name); 1701 1702 if (ok) { 1703 if (nch == 1) 1704 ok = validate_name(CHILD(tree, 0), NULL); 1705 else if (nch == 3) 1706 ok = (validate_name(CHILD(tree, 0), NULL) 1707 && validate_name(CHILD(tree, 1), "as") 1708 && validate_name(CHILD(tree, 2), NULL)); 1709 else 1710 ok = validate_numnodes(tree, 3, "import_as_name"); 1711 } 1712 return ok; 1713} 1714 1715 1716/* dotted_name: NAME ("." NAME)* 1717 */ 1718static int 1719validate_dotted_name(node *tree) 1720{ 1721 int nch = NCH(tree); 1722 int res = (validate_ntype(tree, dotted_name) 1723 && is_odd(nch) 1724 && validate_name(CHILD(tree, 0), NULL)); 1725 int i; 1726 1727 for (i = 1; res && (i < nch); i += 2) { 1728 res = (validate_dot(CHILD(tree, i)) 1729 && validate_name(CHILD(tree, i+1), NULL)); 1730 } 1731 return res; 1732} 1733 1734 1735/* dotted_as_name: dotted_name [NAME NAME] 1736 */ 1737static int 1738validate_dotted_as_name(node *tree) 1739{ 1740 int nch = NCH(tree); 1741 int res = validate_ntype(tree, dotted_as_name); 1742 1743 if (res) { 1744 if (nch == 1) 1745 res = validate_dotted_name(CHILD(tree, 0)); 1746 else if (nch == 3) 1747 res = (validate_dotted_name(CHILD(tree, 0)) 1748 && validate_name(CHILD(tree, 1), "as") 1749 && validate_name(CHILD(tree, 2), NULL)); 1750 else { 1751 res = 0; 1752 err_string("illegal number of children for dotted_as_name"); 1753 } 1754 } 1755 return res; 1756} 1757 1758 1759/* dotted_as_name (',' dotted_as_name)* */ 1760static int 1761validate_dotted_as_names(node *tree) 1762{ 1763 int nch = NCH(tree); 1764 int res = is_odd(nch) && validate_dotted_as_name(CHILD(tree, 0)); 1765 int i; 1766 1767 for (i = 1; res && (i < nch); i += 2) 1768 res = (validate_comma(CHILD(tree, i)) 1769 && validate_dotted_as_name(CHILD(tree, i + 1))); 1770 return (res); 1771} 1772 1773 1774/* import_as_name (',' import_as_name)* [','] */ 1775static int 1776validate_import_as_names(node *tree) 1777{ 1778 int nch = NCH(tree); 1779 int res = validate_import_as_name(CHILD(tree, 0)); 1780 int i; 1781 1782 for (i = 1; res && (i + 1 < nch); i += 2) 1783 res = (validate_comma(CHILD(tree, i)) 1784 && validate_import_as_name(CHILD(tree, i + 1))); 1785 return (res); 1786} 1787 1788 1789/* 'import' dotted_as_names */ 1790static int 1791validate_import_name(node *tree) 1792{ 1793 return (validate_ntype(tree, import_name) 1794 && validate_numnodes(tree, 2, "import_name") 1795 && validate_name(CHILD(tree, 0), "import") 1796 && validate_dotted_as_names(CHILD(tree, 1))); 1797} 1798 1799/* Helper function to count the number of leading dots in 1800 * 'from ...module import name' 1801 */ 1802static int 1803count_from_dots(node *tree) 1804{ 1805 int i; 1806 for (i = 0; i < NCH(tree); i++) 1807 if (TYPE(CHILD(tree, i)) != DOT) 1808 break; 1809 return i; 1810} 1811 1812/* 'from' ('.'* dotted_name | '.') 'import' ('*' | '(' import_as_names ')' | 1813 * import_as_names 1814 */ 1815static int 1816validate_import_from(node *tree) 1817{ 1818 int nch = NCH(tree); 1819 int ndots = count_from_dots(tree); 1820 int havename = (TYPE(CHILD(tree, ndots + 1)) == dotted_name); 1821 int offset = ndots + havename; 1822 int res = validate_ntype(tree, import_from) 1823 && (nch >= 4 + ndots) 1824 && validate_name(CHILD(tree, 0), "from") 1825 && (!havename || validate_dotted_name(CHILD(tree, ndots + 1))) 1826 && validate_name(CHILD(tree, offset + 1), "import"); 1827 1828 if (res && TYPE(CHILD(tree, offset + 2)) == LPAR) 1829 res = ((nch == offset + 5) 1830 && validate_lparen(CHILD(tree, offset + 2)) 1831 && validate_import_as_names(CHILD(tree, offset + 3)) 1832 && validate_rparen(CHILD(tree, offset + 4))); 1833 else if (res && TYPE(CHILD(tree, offset + 2)) != STAR) 1834 res = validate_import_as_names(CHILD(tree, offset + 2)); 1835 return (res); 1836} 1837 1838 1839/* import_stmt: import_name | import_from */ 1840static int 1841validate_import_stmt(node *tree) 1842{ 1843 int nch = NCH(tree); 1844 int res = validate_numnodes(tree, 1, "import_stmt"); 1845 1846 if (res) { 1847 int ntype = TYPE(CHILD(tree, 0)); 1848 1849 if (ntype == import_name || ntype == import_from) 1850 res = validate_node(CHILD(tree, 0)); 1851 else { 1852 res = 0; 1853 err_string("illegal import_stmt child type"); 1854 } 1855 } 1856 else if (nch == 1) { 1857 res = 0; 1858 PyErr_Format(parser_error, 1859 "Unrecognized child node of import_stmt: %d.", 1860 TYPE(CHILD(tree, 0))); 1861 } 1862 return (res); 1863} 1864 1865 1866 1867 1868static int 1869validate_global_stmt(node *tree) 1870{ 1871 int j; 1872 int nch = NCH(tree); 1873 int res = (validate_ntype(tree, global_stmt) 1874 && is_even(nch) && (nch >= 2)); 1875 1876 if (!res && !PyErr_Occurred()) 1877 err_string("illegal global statement"); 1878 1879 if (res) 1880 res = (validate_name(CHILD(tree, 0), "global") 1881 && validate_ntype(CHILD(tree, 1), NAME)); 1882 for (j = 2; res && (j < nch); j += 2) 1883 res = (validate_comma(CHILD(tree, j)) 1884 && validate_ntype(CHILD(tree, j + 1), NAME)); 1885 1886 return (res); 1887} 1888 1889 1890/* assert_stmt: 1891 * 1892 * 'assert' test [',' test] 1893 */ 1894static int 1895validate_assert_stmt(node *tree) 1896{ 1897 int nch = NCH(tree); 1898 int res = (validate_ntype(tree, assert_stmt) 1899 && ((nch == 2) || (nch == 4)) 1900 && (validate_name(CHILD(tree, 0), "assert")) 1901 && validate_test(CHILD(tree, 1))); 1902 1903 if (!res && !PyErr_Occurred()) 1904 err_string("illegal assert statement"); 1905 if (res && (nch > 2)) 1906 res = (validate_comma(CHILD(tree, 2)) 1907 && validate_test(CHILD(tree, 3))); 1908 1909 return (res); 1910} 1911 1912 1913static int 1914validate_while(node *tree) 1915{ 1916 int nch = NCH(tree); 1917 int res = (validate_ntype(tree, while_stmt) 1918 && ((nch == 4) || (nch == 7)) 1919 && validate_name(CHILD(tree, 0), "while") 1920 && validate_test(CHILD(tree, 1)) 1921 && validate_colon(CHILD(tree, 2)) 1922 && validate_suite(CHILD(tree, 3))); 1923 1924 if (res && (nch == 7)) 1925 res = (validate_name(CHILD(tree, 4), "else") 1926 && validate_colon(CHILD(tree, 5)) 1927 && validate_suite(CHILD(tree, 6))); 1928 1929 return (res); 1930} 1931 1932 1933static int 1934validate_for(node *tree) 1935{ 1936 int nch = NCH(tree); 1937 int res = (validate_ntype(tree, for_stmt) 1938 && ((nch == 6) || (nch == 9)) 1939 && validate_name(CHILD(tree, 0), "for") 1940 && validate_exprlist(CHILD(tree, 1)) 1941 && validate_name(CHILD(tree, 2), "in") 1942 && validate_testlist(CHILD(tree, 3)) 1943 && validate_colon(CHILD(tree, 4)) 1944 && validate_suite(CHILD(tree, 5))); 1945 1946 if (res && (nch == 9)) 1947 res = (validate_name(CHILD(tree, 6), "else") 1948 && validate_colon(CHILD(tree, 7)) 1949 && validate_suite(CHILD(tree, 8))); 1950 1951 return (res); 1952} 1953 1954 1955/* try_stmt: 1956 * 'try' ':' suite (except_clause ':' suite)+ ['else' ':' suite] 1957 * | 'try' ':' suite 'finally' ':' suite 1958 * 1959 */ 1960static int 1961validate_try(node *tree) 1962{ 1963 int nch = NCH(tree); 1964 int pos = 3; 1965 int res = (validate_ntype(tree, try_stmt) 1966 && (nch >= 6) && ((nch % 3) == 0)); 1967 1968 if (res) 1969 res = (validate_name(CHILD(tree, 0), "try") 1970 && validate_colon(CHILD(tree, 1)) 1971 && validate_suite(CHILD(tree, 2)) 1972 && validate_colon(CHILD(tree, nch - 2)) 1973 && validate_suite(CHILD(tree, nch - 1))); 1974 else if (!PyErr_Occurred()) { 1975 const char* name = "except"; 1976 if (TYPE(CHILD(tree, nch - 3)) != except_clause) 1977 name = STR(CHILD(tree, nch - 3)); 1978 1979 PyErr_Format(parser_error, 1980 "Illegal number of children for try/%s node.", name); 1981 } 1982 /* Skip past except_clause sections: */ 1983 while (res && (TYPE(CHILD(tree, pos)) == except_clause)) { 1984 res = (validate_except_clause(CHILD(tree, pos)) 1985 && validate_colon(CHILD(tree, pos + 1)) 1986 && validate_suite(CHILD(tree, pos + 2))); 1987 pos += 3; 1988 } 1989 if (res && (pos < nch)) { 1990 res = validate_ntype(CHILD(tree, pos), NAME); 1991 if (res && (strcmp(STR(CHILD(tree, pos)), "finally") == 0)) 1992 res = (validate_numnodes(tree, 6, "try/finally") 1993 && validate_colon(CHILD(tree, 4)) 1994 && validate_suite(CHILD(tree, 5))); 1995 else if (res) { 1996 if (nch == (pos + 3)) { 1997 res = ((strcmp(STR(CHILD(tree, pos)), "except") == 0) 1998 || (strcmp(STR(CHILD(tree, pos)), "else") == 0)); 1999 if (!res) 2000 err_string("illegal trailing triple in try statement"); 2001 } 2002 else if (nch == (pos + 6)) { 2003 res = (validate_name(CHILD(tree, pos), "except") 2004 && validate_colon(CHILD(tree, pos + 1)) 2005 && validate_suite(CHILD(tree, pos + 2)) 2006 && validate_name(CHILD(tree, pos + 3), "else")); 2007 } 2008 else 2009 res = validate_numnodes(tree, pos + 3, "try/except"); 2010 } 2011 } 2012 return (res); 2013} 2014 2015 2016static int 2017validate_except_clause(node *tree) 2018{ 2019 int nch = NCH(tree); 2020 int res = (validate_ntype(tree, except_clause) 2021 && ((nch == 1) || (nch == 2) || (nch == 4)) 2022 && validate_name(CHILD(tree, 0), "except")); 2023 2024 if (res && (nch > 1)) 2025 res = validate_test(CHILD(tree, 1)); 2026 if (res && (nch == 4)) 2027 res = (validate_comma(CHILD(tree, 2)) 2028 && validate_test(CHILD(tree, 3))); 2029 2030 return (res); 2031} 2032 2033 2034static int 2035validate_test(node *tree) 2036{ 2037 int nch = NCH(tree); 2038 int res = validate_ntype(tree, test) && is_odd(nch); 2039 2040 if (res && (TYPE(CHILD(tree, 0)) == lambdef)) 2041 res = ((nch == 1) 2042 && validate_lambdef(CHILD(tree, 0))); 2043 else if (res) { 2044 res = validate_or_test(CHILD(tree, 0)); 2045 res = (res && (nch == 1 || (nch == 5 && 2046 validate_name(CHILD(tree, 1), "if") && 2047 validate_or_test(CHILD(tree, 2)) && 2048 validate_name(CHILD(tree, 3), "else") && 2049 validate_test(CHILD(tree, 4))))); 2050 } 2051 return (res); 2052} 2053 2054static int 2055validate_old_test(node *tree) 2056{ 2057 int nch = NCH(tree); 2058 int res = validate_ntype(tree, old_test) && (nch == 1); 2059 2060 if (res && (TYPE(CHILD(tree, 0)) == old_lambdef)) 2061 res = (validate_old_lambdef(CHILD(tree, 0))); 2062 else if (res) { 2063 res = (validate_or_test(CHILD(tree, 0))); 2064 } 2065 return (res); 2066} 2067 2068static int 2069validate_or_test(node *tree) 2070{ 2071 int nch = NCH(tree); 2072 int res = validate_ntype(tree, or_test) && is_odd(nch); 2073 2074 if (res) { 2075 int pos; 2076 res = validate_and_test(CHILD(tree, 0)); 2077 for (pos = 1; res && (pos < nch); pos += 2) 2078 res = (validate_name(CHILD(tree, pos), "or") 2079 && validate_and_test(CHILD(tree, pos + 1))); 2080 } 2081 return (res); 2082} 2083 2084 2085static int 2086validate_and_test(node *tree) 2087{ 2088 int pos; 2089 int nch = NCH(tree); 2090 int res = (validate_ntype(tree, and_test) 2091 && is_odd(nch) 2092 && validate_not_test(CHILD(tree, 0))); 2093 2094 for (pos = 1; res && (pos < nch); pos += 2) 2095 res = (validate_name(CHILD(tree, pos), "and") 2096 && validate_not_test(CHILD(tree, 0))); 2097 2098 return (res); 2099} 2100 2101 2102static int 2103validate_not_test(node *tree) 2104{ 2105 int nch = NCH(tree); 2106 int res = validate_ntype(tree, not_test) && ((nch == 1) || (nch == 2)); 2107 2108 if (res) { 2109 if (nch == 2) 2110 res = (validate_name(CHILD(tree, 0), "not") 2111 && validate_not_test(CHILD(tree, 1))); 2112 else if (nch == 1) 2113 res = validate_comparison(CHILD(tree, 0)); 2114 } 2115 return (res); 2116} 2117 2118 2119static int 2120validate_comparison(node *tree) 2121{ 2122 int pos; 2123 int nch = NCH(tree); 2124 int res = (validate_ntype(tree, comparison) 2125 && is_odd(nch) 2126 && validate_expr(CHILD(tree, 0))); 2127 2128 for (pos = 1; res && (pos < nch); pos += 2) 2129 res = (validate_comp_op(CHILD(tree, pos)) 2130 && validate_expr(CHILD(tree, pos + 1))); 2131 2132 return (res); 2133} 2134 2135 2136static int 2137validate_comp_op(node *tree) 2138{ 2139 int res = 0; 2140 int nch = NCH(tree); 2141 2142 if (!validate_ntype(tree, comp_op)) 2143 return (0); 2144 if (nch == 1) { 2145 /* 2146 * Only child will be a terminal with a well-defined symbolic name 2147 * or a NAME with a string of either 'is' or 'in' 2148 */ 2149 tree = CHILD(tree, 0); 2150 switch (TYPE(tree)) { 2151 case LESS: 2152 case GREATER: 2153 case EQEQUAL: 2154 case EQUAL: 2155 case LESSEQUAL: 2156 case GREATEREQUAL: 2157 case NOTEQUAL: 2158 res = 1; 2159 break; 2160 case NAME: 2161 res = ((strcmp(STR(tree), "in") == 0) 2162 || (strcmp(STR(tree), "is") == 0)); 2163 if (!res) { 2164 PyErr_Format(parser_error, 2165 "illegal operator '%s'", STR(tree)); 2166 } 2167 break; 2168 default: 2169 err_string("illegal comparison operator type"); 2170 break; 2171 } 2172 } 2173 else if ((res = validate_numnodes(tree, 2, "comp_op")) != 0) { 2174 res = (validate_ntype(CHILD(tree, 0), NAME) 2175 && validate_ntype(CHILD(tree, 1), NAME) 2176 && (((strcmp(STR(CHILD(tree, 0)), "is") == 0) 2177 && (strcmp(STR(CHILD(tree, 1)), "not") == 0)) 2178 || ((strcmp(STR(CHILD(tree, 0)), "not") == 0) 2179 && (strcmp(STR(CHILD(tree, 1)), "in") == 0)))); 2180 if (!res && !PyErr_Occurred()) 2181 err_string("unknown comparison operator"); 2182 } 2183 return (res); 2184} 2185 2186 2187static int 2188validate_expr(node *tree) 2189{ 2190 int j; 2191 int nch = NCH(tree); 2192 int res = (validate_ntype(tree, expr) 2193 && is_odd(nch) 2194 && validate_xor_expr(CHILD(tree, 0))); 2195 2196 for (j = 2; res && (j < nch); j += 2) 2197 res = (validate_xor_expr(CHILD(tree, j)) 2198 && validate_vbar(CHILD(tree, j - 1))); 2199 2200 return (res); 2201} 2202 2203 2204static int 2205validate_xor_expr(node *tree) 2206{ 2207 int j; 2208 int nch = NCH(tree); 2209 int res = (validate_ntype(tree, xor_expr) 2210 && is_odd(nch) 2211 && validate_and_expr(CHILD(tree, 0))); 2212 2213 for (j = 2; res && (j < nch); j += 2) 2214 res = (validate_circumflex(CHILD(tree, j - 1)) 2215 && validate_and_expr(CHILD(tree, j))); 2216 2217 return (res); 2218} 2219 2220 2221static int 2222validate_and_expr(node *tree) 2223{ 2224 int pos; 2225 int nch = NCH(tree); 2226 int res = (validate_ntype(tree, and_expr) 2227 && is_odd(nch) 2228 && validate_shift_expr(CHILD(tree, 0))); 2229 2230 for (pos = 1; res && (pos < nch); pos += 2) 2231 res = (validate_ampersand(CHILD(tree, pos)) 2232 && validate_shift_expr(CHILD(tree, pos + 1))); 2233 2234 return (res); 2235} 2236 2237 2238static int 2239validate_chain_two_ops(node *tree, int (*termvalid)(node *), int op1, int op2) 2240 { 2241 int pos = 1; 2242 int nch = NCH(tree); 2243 int res = (is_odd(nch) 2244 && (*termvalid)(CHILD(tree, 0))); 2245 2246 for ( ; res && (pos < nch); pos += 2) { 2247 if (TYPE(CHILD(tree, pos)) != op1) 2248 res = validate_ntype(CHILD(tree, pos), op2); 2249 if (res) 2250 res = (*termvalid)(CHILD(tree, pos + 1)); 2251 } 2252 return (res); 2253} 2254 2255 2256static int 2257validate_shift_expr(node *tree) 2258{ 2259 return (validate_ntype(tree, shift_expr) 2260 && validate_chain_two_ops(tree, validate_arith_expr, 2261 LEFTSHIFT, RIGHTSHIFT)); 2262} 2263 2264 2265static int 2266validate_arith_expr(node *tree) 2267{ 2268 return (validate_ntype(tree, arith_expr) 2269 && validate_chain_two_ops(tree, validate_term, PLUS, MINUS)); 2270} 2271 2272 2273static int 2274validate_term(node *tree) 2275{ 2276 int pos = 1; 2277 int nch = NCH(tree); 2278 int res = (validate_ntype(tree, term) 2279 && is_odd(nch) 2280 && validate_factor(CHILD(tree, 0))); 2281 2282 for ( ; res && (pos < nch); pos += 2) 2283 res = (((TYPE(CHILD(tree, pos)) == STAR) 2284 || (TYPE(CHILD(tree, pos)) == SLASH) 2285 || (TYPE(CHILD(tree, pos)) == DOUBLESLASH) 2286 || (TYPE(CHILD(tree, pos)) == PERCENT)) 2287 && validate_factor(CHILD(tree, pos + 1))); 2288 2289 return (res); 2290} 2291 2292 2293/* factor: 2294 * 2295 * factor: ('+'|'-'|'~') factor | power 2296 */ 2297static int 2298validate_factor(node *tree) 2299{ 2300 int nch = NCH(tree); 2301 int res = (validate_ntype(tree, factor) 2302 && (((nch == 2) 2303 && ((TYPE(CHILD(tree, 0)) == PLUS) 2304 || (TYPE(CHILD(tree, 0)) == MINUS) 2305 || (TYPE(CHILD(tree, 0)) == TILDE)) 2306 && validate_factor(CHILD(tree, 1))) 2307 || ((nch == 1) 2308 && validate_power(CHILD(tree, 0))))); 2309 return (res); 2310} 2311 2312 2313/* power: 2314 * 2315 * power: atom trailer* ('**' factor)* 2316 */ 2317static int 2318validate_power(node *tree) 2319{ 2320 int pos = 1; 2321 int nch = NCH(tree); 2322 int res = (validate_ntype(tree, power) && (nch >= 1) 2323 && validate_atom(CHILD(tree, 0))); 2324 2325 while (res && (pos < nch) && (TYPE(CHILD(tree, pos)) == trailer)) 2326 res = validate_trailer(CHILD(tree, pos++)); 2327 if (res && (pos < nch)) { 2328 if (!is_even(nch - pos)) { 2329 err_string("illegal number of nodes for 'power'"); 2330 return (0); 2331 } 2332 for ( ; res && (pos < (nch - 1)); pos += 2) 2333 res = (validate_doublestar(CHILD(tree, pos)) 2334 && validate_factor(CHILD(tree, pos + 1))); 2335 } 2336 return (res); 2337} 2338 2339 2340static int 2341validate_atom(node *tree) 2342{ 2343 int pos; 2344 int nch = NCH(tree); 2345 int res = validate_ntype(tree, atom); 2346 2347 if (res && nch < 1) 2348 res = validate_numnodes(tree, nch+1, "atom"); 2349 if (res) { 2350 switch (TYPE(CHILD(tree, 0))) { 2351 case LPAR: 2352 res = ((nch <= 3) 2353 && (validate_rparen(CHILD(tree, nch - 1)))); 2354 2355 if (res && (nch == 3)) { 2356 if (TYPE(CHILD(tree, 1))==yield_expr) 2357 res = validate_yield_expr(CHILD(tree, 1)); 2358 else 2359 res = validate_testlist_gexp(CHILD(tree, 1)); 2360 } 2361 break; 2362 case LSQB: 2363 if (nch == 2) 2364 res = validate_ntype(CHILD(tree, 1), RSQB); 2365 else if (nch == 3) 2366 res = (validate_listmaker(CHILD(tree, 1)) 2367 && validate_ntype(CHILD(tree, 2), RSQB)); 2368 else { 2369 res = 0; 2370 err_string("illegal list display atom"); 2371 } 2372 break; 2373 case LBRACE: 2374 res = ((nch <= 3) 2375 && validate_ntype(CHILD(tree, nch - 1), RBRACE)); 2376 2377 if (res && (nch == 3)) 2378 res = validate_dictsetmaker(CHILD(tree, 1)); 2379 break; 2380 case NAME: 2381 case NUMBER: 2382 res = (nch == 1); 2383 break; 2384 case STRING: 2385 for (pos = 1; res && (pos < nch); ++pos) 2386 res = validate_ntype(CHILD(tree, pos), STRING); 2387 break; 2388 case DOT: 2389 res = (nch == 3 && 2390 validate_ntype(CHILD(tree, 1), DOT) && 2391 validate_ntype(CHILD(tree, 2), DOT)); 2392 break; 2393 default: 2394 res = 0; 2395 break; 2396 } 2397 } 2398 return (res); 2399} 2400 2401 2402/* listmaker: 2403 * test ( list_for | (',' test)* [','] ) 2404 */ 2405static int 2406validate_listmaker(node *tree) 2407{ 2408 int nch = NCH(tree); 2409 int ok = nch; 2410 2411 if (nch == 0) 2412 err_string("missing child nodes of listmaker"); 2413 else 2414 ok = validate_test(CHILD(tree, 0)); 2415 2416 /* 2417 * list_for | (',' test)* [','] 2418 */ 2419 if (nch == 2 && TYPE(CHILD(tree, 1)) == list_for) 2420 ok = validate_list_for(CHILD(tree, 1)); 2421 else { 2422 /* (',' test)* [','] */ 2423 int i = 1; 2424 while (ok && nch - i >= 2) { 2425 ok = (validate_comma(CHILD(tree, i)) 2426 && validate_test(CHILD(tree, i+1))); 2427 i += 2; 2428 } 2429 if (ok && i == nch-1) 2430 ok = validate_comma(CHILD(tree, i)); 2431 else if (i != nch) { 2432 ok = 0; 2433 err_string("illegal trailing nodes for listmaker"); 2434 } 2435 } 2436 return ok; 2437} 2438 2439/* testlist_gexp: 2440 * test ( gen_for | (',' test)* [','] ) 2441 */ 2442static int 2443validate_testlist_gexp(node *tree) 2444{ 2445 int nch = NCH(tree); 2446 int ok = nch; 2447 2448 if (nch == 0) 2449 err_string("missing child nodes of testlist_gexp"); 2450 else { 2451 ok = validate_test(CHILD(tree, 0)); 2452 } 2453 2454 /* 2455 * gen_for | (',' test)* [','] 2456 */ 2457 if (nch == 2 && TYPE(CHILD(tree, 1)) == gen_for) 2458 ok = validate_gen_for(CHILD(tree, 1)); 2459 else { 2460 /* (',' test)* [','] */ 2461 int i = 1; 2462 while (ok && nch - i >= 2) { 2463 ok = (validate_comma(CHILD(tree, i)) 2464 && validate_test(CHILD(tree, i+1))); 2465 i += 2; 2466 } 2467 if (ok && i == nch-1) 2468 ok = validate_comma(CHILD(tree, i)); 2469 else if (i != nch) { 2470 ok = 0; 2471 err_string("illegal trailing nodes for testlist_gexp"); 2472 } 2473 } 2474 return ok; 2475} 2476 2477/* decorator: 2478 * '@' dotted_name [ '(' [arglist] ')' ] NEWLINE 2479 */ 2480static int 2481validate_decorator(node *tree) 2482{ 2483 int ok; 2484 int nch = NCH(tree); 2485 ok = (validate_ntype(tree, decorator) && 2486 (nch == 3 || nch == 5 || nch == 6) && 2487 validate_at(CHILD(tree, 0)) && 2488 validate_dotted_name(CHILD(tree, 1)) && 2489 validate_newline(RCHILD(tree, -1))); 2490 2491 if (ok && nch != 3) { 2492 ok = (validate_lparen(CHILD(tree, 2)) && 2493 validate_rparen(RCHILD(tree, -2))); 2494 2495 if (ok && nch == 6) 2496 ok = validate_arglist(CHILD(tree, 3)); 2497 } 2498 2499 return ok; 2500} 2501 2502/* decorators: 2503 * decorator+ 2504 */ 2505static int 2506validate_decorators(node *tree) 2507{ 2508 int i, nch, ok; 2509 nch = NCH(tree); 2510 ok = validate_ntype(tree, decorators) && nch >= 1; 2511 2512 for (i = 0; ok && i < nch; ++i) 2513 ok = validate_decorator(CHILD(tree, i)); 2514 2515 return ok; 2516} 2517 2518/* funcdef: 2519 * 2520 * -6 -5 -4 -3 -2 -1 2521 * [decorators] 'def' NAME parameters ':' suite 2522 */ 2523static int 2524validate_funcdef(node *tree) 2525{ 2526 int nch = NCH(tree); 2527 int ok = (validate_ntype(tree, funcdef) 2528 && ((nch == 5) || (nch == 6)) 2529 && validate_name(RCHILD(tree, -5), "def") 2530 && validate_ntype(RCHILD(tree, -4), NAME) 2531 && validate_colon(RCHILD(tree, -2)) 2532 && validate_parameters(RCHILD(tree, -3)) 2533 && validate_suite(RCHILD(tree, -1))); 2534 2535 if (ok && (nch == 6)) 2536 ok = validate_decorators(CHILD(tree, 0)); 2537 2538 return ok; 2539} 2540 2541 2542static int 2543validate_lambdef(node *tree) 2544{ 2545 int nch = NCH(tree); 2546 int res = (validate_ntype(tree, lambdef) 2547 && ((nch == 3) || (nch == 4)) 2548 && validate_name(CHILD(tree, 0), "lambda") 2549 && validate_colon(CHILD(tree, nch - 2)) 2550 && validate_test(CHILD(tree, nch - 1))); 2551 2552 if (res && (nch == 4)) 2553 res = validate_varargslist(CHILD(tree, 1)); 2554 else if (!res && !PyErr_Occurred()) 2555 (void) validate_numnodes(tree, 3, "lambdef"); 2556 2557 return (res); 2558} 2559 2560 2561static int 2562validate_old_lambdef(node *tree) 2563{ 2564 int nch = NCH(tree); 2565 int res = (validate_ntype(tree, old_lambdef) 2566 && ((nch == 3) || (nch == 4)) 2567 && validate_name(CHILD(tree, 0), "lambda") 2568 && validate_colon(CHILD(tree, nch - 2)) 2569 && validate_test(CHILD(tree, nch - 1))); 2570 2571 if (res && (nch == 4)) 2572 res = validate_varargslist(CHILD(tree, 1)); 2573 else if (!res && !PyErr_Occurred()) 2574 (void) validate_numnodes(tree, 3, "old_lambdef"); 2575 2576 return (res); 2577} 2578 2579 2580/* arglist: 2581 * 2582 * (argument ',')* (argument [','] | '*' test [',' '**' test] | '**' test) 2583 */ 2584static int 2585validate_arglist(node *tree) 2586{ 2587 int nch = NCH(tree); 2588 int i = 0; 2589 int ok = 1; 2590 2591 if (nch <= 0) 2592 /* raise the right error from having an invalid number of children */ 2593 return validate_numnodes(tree, nch + 1, "arglist"); 2594 2595 if (nch > 1) { 2596 for (i=0; i<nch; i++) { 2597 if (TYPE(CHILD(tree, i)) == argument) { 2598 node *ch = CHILD(tree, i); 2599 if (NCH(ch) == 2 && TYPE(CHILD(ch, 1)) == gen_for) { 2600 err_string("need '(', ')' for generator expression"); 2601 return 0; 2602 } 2603 } 2604 } 2605 } 2606 2607 while (ok && nch-i >= 2) { 2608 /* skip leading (argument ',') */ 2609 ok = (validate_argument(CHILD(tree, i)) 2610 && validate_comma(CHILD(tree, i+1))); 2611 if (ok) 2612 i += 2; 2613 else 2614 PyErr_Clear(); 2615 } 2616 ok = 1; 2617 if (nch-i > 0) { 2618 /* 2619 * argument | '*' test [',' '**' test] | '**' test 2620 */ 2621 int sym = TYPE(CHILD(tree, i)); 2622 2623 if (sym == argument) { 2624 ok = validate_argument(CHILD(tree, i)); 2625 if (ok && i+1 != nch) { 2626 err_string("illegal arglist specification" 2627 " (extra stuff on end)"); 2628 ok = 0; 2629 } 2630 } 2631 else if (sym == STAR) { 2632 ok = validate_star(CHILD(tree, i)); 2633 if (ok && (nch-i == 2)) 2634 ok = validate_test(CHILD(tree, i+1)); 2635 else if (ok && (nch-i == 5)) 2636 ok = (validate_test(CHILD(tree, i+1)) 2637 && validate_comma(CHILD(tree, i+2)) 2638 && validate_doublestar(CHILD(tree, i+3)) 2639 && validate_test(CHILD(tree, i+4))); 2640 else { 2641 err_string("illegal use of '*' in arglist"); 2642 ok = 0; 2643 } 2644 } 2645 else if (sym == DOUBLESTAR) { 2646 if (nch-i == 2) 2647 ok = (validate_doublestar(CHILD(tree, i)) 2648 && validate_test(CHILD(tree, i+1))); 2649 else { 2650 err_string("illegal use of '**' in arglist"); 2651 ok = 0; 2652 } 2653 } 2654 else { 2655 err_string("illegal arglist specification"); 2656 ok = 0; 2657 } 2658 } 2659 return (ok); 2660} 2661 2662 2663 2664/* argument: 2665 * 2666 * [test '='] test [gen_for] 2667 */ 2668static int 2669validate_argument(node *tree) 2670{ 2671 int nch = NCH(tree); 2672 int res = (validate_ntype(tree, argument) 2673 && ((nch == 1) || (nch == 2) || (nch == 3)) 2674 && validate_test(CHILD(tree, 0))); 2675 2676 if (res && (nch == 2)) 2677 res = validate_gen_for(CHILD(tree, 1)); 2678 else if (res && (nch == 3)) 2679 res = (validate_equal(CHILD(tree, 1)) 2680 && validate_test(CHILD(tree, 2))); 2681 2682 return (res); 2683} 2684 2685 2686 2687/* trailer: 2688 * 2689 * '(' [arglist] ')' | '[' subscriptlist ']' | '.' NAME 2690 */ 2691static int 2692validate_trailer(node *tree) 2693{ 2694 int nch = NCH(tree); 2695 int res = validate_ntype(tree, trailer) && ((nch == 2) || (nch == 3)); 2696 2697 if (res) { 2698 switch (TYPE(CHILD(tree, 0))) { 2699 case LPAR: 2700 res = validate_rparen(CHILD(tree, nch - 1)); 2701 if (res && (nch == 3)) 2702 res = validate_arglist(CHILD(tree, 1)); 2703 break; 2704 case LSQB: 2705 res = (validate_numnodes(tree, 3, "trailer") 2706 && validate_subscriptlist(CHILD(tree, 1)) 2707 && validate_ntype(CHILD(tree, 2), RSQB)); 2708 break; 2709 case DOT: 2710 res = (validate_numnodes(tree, 2, "trailer") 2711 && validate_ntype(CHILD(tree, 1), NAME)); 2712 break; 2713 default: 2714 res = 0; 2715 break; 2716 } 2717 } 2718 else { 2719 (void) validate_numnodes(tree, 2, "trailer"); 2720 } 2721 return (res); 2722} 2723 2724 2725/* subscriptlist: 2726 * 2727 * subscript (',' subscript)* [','] 2728 */ 2729static int 2730validate_subscriptlist(node *tree) 2731{ 2732 return (validate_repeating_list(tree, subscriptlist, 2733 validate_subscript, "subscriptlist")); 2734} 2735 2736 2737/* subscript: 2738 * 2739 * '.' '.' '.' | test | [test] ':' [test] [sliceop] 2740 */ 2741static int 2742validate_subscript(node *tree) 2743{ 2744 int offset = 0; 2745 int nch = NCH(tree); 2746 int res = validate_ntype(tree, subscript) && (nch >= 1) && (nch <= 4); 2747 2748 if (!res) { 2749 if (!PyErr_Occurred()) 2750 err_string("invalid number of arguments for subscript node"); 2751 return (0); 2752 } 2753 if (TYPE(CHILD(tree, 0)) == DOT) 2754 /* take care of ('.' '.' '.') possibility */ 2755 return (validate_numnodes(tree, 3, "subscript") 2756 && validate_dot(CHILD(tree, 0)) 2757 && validate_dot(CHILD(tree, 1)) 2758 && validate_dot(CHILD(tree, 2))); 2759 if (nch == 1) { 2760 if (TYPE(CHILD(tree, 0)) == test) 2761 res = validate_test(CHILD(tree, 0)); 2762 else 2763 res = validate_colon(CHILD(tree, 0)); 2764 return (res); 2765 } 2766 /* Must be [test] ':' [test] [sliceop], 2767 * but at least one of the optional components will 2768 * be present, but we don't know which yet. 2769 */ 2770 if ((TYPE(CHILD(tree, 0)) != COLON) || (nch == 4)) { 2771 res = validate_test(CHILD(tree, 0)); 2772 offset = 1; 2773 } 2774 if (res) 2775 res = validate_colon(CHILD(tree, offset)); 2776 if (res) { 2777 int rem = nch - ++offset; 2778 if (rem) { 2779 if (TYPE(CHILD(tree, offset)) == test) { 2780 res = validate_test(CHILD(tree, offset)); 2781 ++offset; 2782 --rem; 2783 } 2784 if (res && rem) 2785 res = validate_sliceop(CHILD(tree, offset)); 2786 } 2787 } 2788 return (res); 2789} 2790 2791 2792static int 2793validate_sliceop(node *tree) 2794{ 2795 int nch = NCH(tree); 2796 int res = ((nch == 1) || validate_numnodes(tree, 2, "sliceop")) 2797 && validate_ntype(tree, sliceop); 2798 if (!res && !PyErr_Occurred()) { 2799 res = validate_numnodes(tree, 1, "sliceop"); 2800 } 2801 if (res) 2802 res = validate_colon(CHILD(tree, 0)); 2803 if (res && (nch == 2)) 2804 res = validate_test(CHILD(tree, 1)); 2805 2806 return (res); 2807} 2808 2809 2810static int 2811validate_exprlist(node *tree) 2812{ 2813 return (validate_repeating_list(tree, exprlist, 2814 validate_expr, "exprlist")); 2815} 2816 2817 2818static int 2819validate_dictsetmaker(node *tree) 2820{ 2821 int nch = NCH(tree); 2822 int res = (validate_ntype(tree, dictsetmaker) 2823 && (nch >= 3) 2824 && validate_test(CHILD(tree, 0)) 2825 && validate_colon(CHILD(tree, 1)) 2826 && validate_test(CHILD(tree, 2))); 2827 2828 if (res && ((nch % 4) == 0)) 2829 res = validate_comma(CHILD(tree, --nch)); 2830 else if (res) 2831 res = ((nch % 4) == 3); 2832 2833 if (res && (nch > 3)) { 2834 int pos = 3; 2835 /* ( ',' test ':' test )* */ 2836 while (res && (pos < nch)) { 2837 res = (validate_comma(CHILD(tree, pos)) 2838 && validate_test(CHILD(tree, pos + 1)) 2839 && validate_colon(CHILD(tree, pos + 2)) 2840 && validate_test(CHILD(tree, pos + 3))); 2841 pos += 4; 2842 } 2843 } 2844 return (res); 2845} 2846 2847 2848static int 2849validate_eval_input(node *tree) 2850{ 2851 int pos; 2852 int nch = NCH(tree); 2853 int res = (validate_ntype(tree, eval_input) 2854 && (nch >= 2) 2855 && validate_testlist(CHILD(tree, 0)) 2856 && validate_ntype(CHILD(tree, nch - 1), ENDMARKER)); 2857 2858 for (pos = 1; res && (pos < (nch - 1)); ++pos) 2859 res = validate_ntype(CHILD(tree, pos), NEWLINE); 2860 2861 return (res); 2862} 2863 2864 2865static int 2866validate_node(node *tree) 2867{ 2868 int nch = 0; /* num. children on current node */ 2869 int res = 1; /* result value */ 2870 node* next = 0; /* node to process after this one */ 2871 2872 while (res && (tree != 0)) { 2873 nch = NCH(tree); 2874 next = 0; 2875 switch (TYPE(tree)) { 2876 /* 2877 * Definition nodes. 2878 */ 2879 case funcdef: 2880 res = validate_funcdef(tree); 2881 break; 2882 case classdef: 2883 res = validate_class(tree); 2884 break; 2885 /* 2886 * "Trivial" parse tree nodes. 2887 * (Why did I call these trivial?) 2888 */ 2889 case stmt: 2890 res = validate_stmt(tree); 2891 break; 2892 case small_stmt: 2893 /* 2894 * expr_stmt | print_stmt | del_stmt | pass_stmt | flow_stmt 2895 * | import_stmt | global_stmt | assert_stmt 2896 */ 2897 res = validate_small_stmt(tree); 2898 break; 2899 case flow_stmt: 2900 res = (validate_numnodes(tree, 1, "flow_stmt") 2901 && ((TYPE(CHILD(tree, 0)) == break_stmt) 2902 || (TYPE(CHILD(tree, 0)) == continue_stmt) 2903 || (TYPE(CHILD(tree, 0)) == yield_stmt) 2904 || (TYPE(CHILD(tree, 0)) == return_stmt) 2905 || (TYPE(CHILD(tree, 0)) == raise_stmt))); 2906 if (res) 2907 next = CHILD(tree, 0); 2908 else if (nch == 1) 2909 err_string("illegal flow_stmt type"); 2910 break; 2911 case yield_stmt: 2912 res = validate_yield_stmt(tree); 2913 break; 2914 /* 2915 * Compound statements. 2916 */ 2917 case simple_stmt: 2918 res = validate_simple_stmt(tree); 2919 break; 2920 case compound_stmt: 2921 res = validate_compound_stmt(tree); 2922 break; 2923 /* 2924 * Fundamental statements. 2925 */ 2926 case expr_stmt: 2927 res = validate_expr_stmt(tree); 2928 break; 2929 case print_stmt: 2930 res = validate_print_stmt(tree); 2931 break; 2932 case del_stmt: 2933 res = validate_del_stmt(tree); 2934 break; 2935 case pass_stmt: 2936 res = (validate_numnodes(tree, 1, "pass") 2937 && validate_name(CHILD(tree, 0), "pass")); 2938 break; 2939 case break_stmt: 2940 res = (validate_numnodes(tree, 1, "break") 2941 && validate_name(CHILD(tree, 0), "break")); 2942 break; 2943 case continue_stmt: 2944 res = (validate_numnodes(tree, 1, "continue") 2945 && validate_name(CHILD(tree, 0), "continue")); 2946 break; 2947 case return_stmt: 2948 res = validate_return_stmt(tree); 2949 break; 2950 case raise_stmt: 2951 res = validate_raise_stmt(tree); 2952 break; 2953 case import_stmt: 2954 res = validate_import_stmt(tree); 2955 break; 2956 case import_name: 2957 res = validate_import_name(tree); 2958 break; 2959 case import_from: 2960 res = validate_import_from(tree); 2961 break; 2962 case global_stmt: 2963 res = validate_global_stmt(tree); 2964 break; 2965 case assert_stmt: 2966 res = validate_assert_stmt(tree); 2967 break; 2968 case if_stmt: 2969 res = validate_if(tree); 2970 break; 2971 case while_stmt: 2972 res = validate_while(tree); 2973 break; 2974 case for_stmt: 2975 res = validate_for(tree); 2976 break; 2977 case try_stmt: 2978 res = validate_try(tree); 2979 break; 2980 case suite: 2981 res = validate_suite(tree); 2982 break; 2983 /* 2984 * Expression nodes. 2985 */ 2986 case testlist: 2987 res = validate_testlist(tree); 2988 break; 2989 case yield_expr: 2990 res = validate_yield_expr(tree); 2991 break; 2992 case testlist1: 2993 res = validate_testlist1(tree); 2994 break; 2995 case test: 2996 res = validate_test(tree); 2997 break; 2998 case and_test: 2999 res = validate_and_test(tree); 3000 break; 3001 case not_test: 3002 res = validate_not_test(tree); 3003 break; 3004 case comparison: 3005 res = validate_comparison(tree); 3006 break; 3007 case exprlist: 3008 res = validate_exprlist(tree); 3009 break; 3010 case comp_op: 3011 res = validate_comp_op(tree); 3012 break; 3013 case expr: 3014 res = validate_expr(tree); 3015 break; 3016 case xor_expr: 3017 res = validate_xor_expr(tree); 3018 break; 3019 case and_expr: 3020 res = validate_and_expr(tree); 3021 break; 3022 case shift_expr: 3023 res = validate_shift_expr(tree); 3024 break; 3025 case arith_expr: 3026 res = validate_arith_expr(tree); 3027 break; 3028 case term: 3029 res = validate_term(tree); 3030 break; 3031 case factor: 3032 res = validate_factor(tree); 3033 break; 3034 case power: 3035 res = validate_power(tree); 3036 break; 3037 case atom: 3038 res = validate_atom(tree); 3039 break; 3040 3041 default: 3042 /* Hopefully never reached! */ 3043 err_string("unrecognized node type"); 3044 res = 0; 3045 break; 3046 } 3047 tree = next; 3048 } 3049 return (res); 3050} 3051 3052 3053static int 3054validate_expr_tree(node *tree) 3055{ 3056 int res = validate_eval_input(tree); 3057 3058 if (!res && !PyErr_Occurred()) 3059 err_string("could not validate expression tuple"); 3060 3061 return (res); 3062} 3063 3064 3065/* file_input: 3066 * (NEWLINE | stmt)* ENDMARKER 3067 */ 3068static int 3069validate_file_input(node *tree) 3070{ 3071 int j; 3072 int nch = NCH(tree) - 1; 3073 int res = ((nch >= 0) 3074 && validate_ntype(CHILD(tree, nch), ENDMARKER)); 3075 3076 for (j = 0; res && (j < nch); ++j) { 3077 if (TYPE(CHILD(tree, j)) == stmt) 3078 res = validate_stmt(CHILD(tree, j)); 3079 else 3080 res = validate_newline(CHILD(tree, j)); 3081 } 3082 /* This stays in to prevent any internal failures from getting to the 3083 * user. Hopefully, this won't be needed. If a user reports getting 3084 * this, we have some debugging to do. 3085 */ 3086 if (!res && !PyErr_Occurred()) 3087 err_string("VALIDATION FAILURE: report this to the maintainer!"); 3088 3089 return (res); 3090} 3091 3092static int 3093validate_encoding_decl(node *tree) 3094{ 3095 int nch = NCH(tree); 3096 int res = ((nch == 1) 3097 && validate_file_input(CHILD(tree, 0))); 3098 3099 if (!res && !PyErr_Occurred()) 3100 err_string("Error Parsing encoding_decl"); 3101 3102 return res; 3103} 3104 3105static PyObject* 3106pickle_constructor = NULL; 3107 3108 3109static PyObject* 3110parser__pickler(PyObject *self, PyObject *args) 3111{ 3112 NOTE(ARGUNUSED(self)) 3113 PyObject *result = NULL; 3114 PyObject *st = NULL; 3115 PyObject *empty_dict = NULL; 3116 3117 if (PyArg_ParseTuple(args, "O!:_pickler", &PyST_Type, &st)) { 3118 PyObject *newargs; 3119 PyObject *tuple; 3120 3121 if ((empty_dict = PyDict_New()) == NULL) 3122 goto finally; 3123 if ((newargs = Py_BuildValue("Oi", st, 1)) == NULL) 3124 goto finally; 3125 tuple = parser_st2tuple((PyST_Object*)NULL, newargs, empty_dict); 3126 if (tuple != NULL) { 3127 result = Py_BuildValue("O(O)", pickle_constructor, tuple); 3128 Py_DECREF(tuple); 3129 } 3130 Py_DECREF(empty_dict); 3131 Py_DECREF(newargs); 3132 } 3133 finally: 3134 Py_XDECREF(empty_dict); 3135 3136 return (result); 3137} 3138 3139 3140/* Functions exported by this module. Most of this should probably 3141 * be converted into an ST object with methods, but that is better 3142 * done directly in Python, allowing subclasses to be created directly. 3143 * We'd really have to write a wrapper around it all anyway to allow 3144 * inheritance. 3145 */ 3146static PyMethodDef parser_functions[] = { 3147 {"ast2tuple", (PyCFunction)parser_st2tuple, PUBLIC_METHOD_TYPE, 3148 PyDoc_STR("Creates a tuple-tree representation of an ST.")}, 3149 {"ast2list", (PyCFunction)parser_st2list, PUBLIC_METHOD_TYPE, 3150 PyDoc_STR("Creates a list-tree representation of an ST.")}, 3151 {"compileast", (PyCFunction)parser_compilest, PUBLIC_METHOD_TYPE, 3152 PyDoc_STR("Compiles an ST object into a code object.")}, 3153 {"compilest", (PyCFunction)parser_compilest, PUBLIC_METHOD_TYPE, 3154 PyDoc_STR("Compiles an ST object into a code object.")}, 3155 {"expr", (PyCFunction)parser_expr, PUBLIC_METHOD_TYPE, 3156 PyDoc_STR("Creates an ST object from an expression.")}, 3157 {"isexpr", (PyCFunction)parser_isexpr, PUBLIC_METHOD_TYPE, 3158 PyDoc_STR("Determines if an ST object was created from an expression.")}, 3159 {"issuite", (PyCFunction)parser_issuite, PUBLIC_METHOD_TYPE, 3160 PyDoc_STR("Determines if an ST object was created from a suite.")}, 3161 {"suite", (PyCFunction)parser_suite, PUBLIC_METHOD_TYPE, 3162 PyDoc_STR("Creates an ST object from a suite.")}, 3163 {"sequence2ast", (PyCFunction)parser_tuple2st, PUBLIC_METHOD_TYPE, 3164 PyDoc_STR("Creates an ST object from a tree representation.")}, 3165 {"sequence2st", (PyCFunction)parser_tuple2st, PUBLIC_METHOD_TYPE, 3166 PyDoc_STR("Creates an ST object from a tree representation.")}, 3167 {"st2tuple", (PyCFunction)parser_st2tuple, PUBLIC_METHOD_TYPE, 3168 PyDoc_STR("Creates a tuple-tree representation of an ST.")}, 3169 {"st2list", (PyCFunction)parser_st2list, PUBLIC_METHOD_TYPE, 3170 PyDoc_STR("Creates a list-tree representation of an ST.")}, 3171 {"tuple2ast", (PyCFunction)parser_tuple2st, PUBLIC_METHOD_TYPE, 3172 PyDoc_STR("Creates an ST object from a tree representation.")}, 3173 {"tuple2st", (PyCFunction)parser_tuple2st, PUBLIC_METHOD_TYPE, 3174 PyDoc_STR("Creates an ST object from a tree representation.")}, 3175 3176 /* private stuff: support pickle module */ 3177 {"_pickler", (PyCFunction)parser__pickler, METH_VARARGS, 3178 PyDoc_STR("Returns the pickle magic to allow ST objects to be pickled.")}, 3179 3180 {NULL, NULL, 0, NULL} 3181 }; 3182 3183 3184PyMODINIT_FUNC initparser(void); /* supply a prototype */ 3185 3186PyMODINIT_FUNC 3187initparser(void) 3188{ 3189 PyObject *module, *copyreg; 3190 3191 PyST_Type.ob_type = &PyType_Type; 3192 module = Py_InitModule("parser", parser_functions); 3193 if (module == NULL) 3194 return; 3195 3196 if (parser_error == 0) 3197 parser_error = PyErr_NewException("parser.ParserError", NULL, NULL); 3198 3199 if (parser_error == 0) 3200 /* caller will check PyErr_Occurred() */ 3201 return; 3202 /* CAUTION: The code next used to skip bumping the refcount on 3203 * parser_error. That's a disaster if initparser() gets called more 3204 * than once. By incref'ing, we ensure that each module dict that 3205 * gets created owns its reference to the shared parser_error object, 3206 * and the file static parser_error vrbl owns a reference too. 3207 */ 3208 Py_INCREF(parser_error); 3209 if (PyModule_AddObject(module, "ParserError", parser_error) != 0) 3210 return; 3211 3212 Py_INCREF(&PyST_Type); 3213 PyModule_AddObject(module, "ASTType", (PyObject*)&PyST_Type); 3214 Py_INCREF(&PyST_Type); 3215 PyModule_AddObject(module, "STType", (PyObject*)&PyST_Type); 3216 3217 PyModule_AddStringConstant(module, "__copyright__", 3218 parser_copyright_string); 3219 PyModule_AddStringConstant(module, "__doc__", 3220 parser_doc_string); 3221 PyModule_AddStringConstant(module, "__version__", 3222 parser_version_string); 3223 3224 /* Register to support pickling. 3225 * If this fails, the import of this module will fail because an 3226 * exception will be raised here; should we clear the exception? 3227 */ 3228 copyreg = PyImport_ImportModule("copy_reg"); 3229 if (copyreg != NULL) { 3230 PyObject *func, *pickler; 3231 3232 func = PyObject_GetAttrString(copyreg, "pickle"); 3233 pickle_constructor = PyObject_GetAttrString(module, "sequence2st"); 3234 pickler = PyObject_GetAttrString(module, "_pickler"); 3235 Py_XINCREF(pickle_constructor); 3236 if ((func != NULL) && (pickle_constructor != NULL) 3237 && (pickler != NULL)) { 3238 PyObject *res; 3239 3240 res = PyObject_CallFunctionObjArgs(func, &PyST_Type, pickler, 3241 pickle_constructor, NULL); 3242 Py_XDECREF(res); 3243 } 3244 Py_XDECREF(func); 3245 Py_XDECREF(pickle_constructor); 3246 Py_XDECREF(pickler); 3247 Py_DECREF(copyreg); 3248 } 3249} 3250