move sections
[python/dscho.git] / Modules / parsermodule.c
blob5d1bfb6674eb72f5d0eeb0b773828fcb71654011
1 /* parsermodule.c
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.
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.
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.
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(...)".
28 #include "Python.h" /* general Python API */
29 #include "Python-ast.h" /* mod_ty */
30 #include "graminit.h" /* symbols defined in the grammar */
31 #include "node.h" /* internal parser structure */
32 #include "errcode.h" /* error codes for PyNode_*() */
33 #include "token.h" /* token definitions */
34 #include "grammar.h"
35 #include "parsetok.h"
36 /* ISTERMINAL() / ISNONTERMINAL() */
37 #include "compile.h"
38 #undef Yield
39 #include "ast.h"
40 #include "pyarena.h"
42 extern grammar _PyParser_Grammar; /* From graminit.c */
44 #ifdef lint
45 #include <note.h>
46 #else
47 #define NOTE(x)
48 #endif
50 /* String constants used to initialize module attributes.
53 static char parser_copyright_string[] =
54 "Copyright 1995-1996 by Virginia Polytechnic Institute & State\n\
55 University, Blacksburg, Virginia, USA, and Fred L. Drake, Jr., Reston,\n\
56 Virginia, USA. Portions copyright 1991-1995 by Stichting Mathematisch\n\
57 Centrum, Amsterdam, The Netherlands.";
60 PyDoc_STRVAR(parser_doc_string,
61 "This is an interface to Python's internal parser.");
63 static char parser_version_string[] = "0.5";
66 typedef PyObject* (*SeqMaker) (Py_ssize_t length);
67 typedef int (*SeqInserter) (PyObject* sequence,
68 Py_ssize_t index,
69 PyObject* element);
71 /* The function below is copyrighted by Stichting Mathematisch Centrum. The
72 * original copyright statement is included below, and continues to apply
73 * in full to the function immediately following. All other material is
74 * original, copyrighted by Fred L. Drake, Jr. and Virginia Polytechnic
75 * Institute and State University. Changes were made to comply with the
76 * new naming conventions. Added arguments to provide support for creating
77 * lists as well as tuples, and optionally including the line numbers.
81 static PyObject*
82 node2tuple(node *n, /* node to convert */
83 SeqMaker mkseq, /* create sequence */
84 SeqInserter addelem, /* func. to add elem. in seq. */
85 int lineno, /* include line numbers? */
86 int col_offset) /* include column offsets? */
88 if (n == NULL) {
89 Py_INCREF(Py_None);
90 return (Py_None);
92 if (ISNONTERMINAL(TYPE(n))) {
93 int i;
94 PyObject *v;
95 PyObject *w;
97 v = mkseq(1 + NCH(n) + (TYPE(n) == encoding_decl));
98 if (v == NULL)
99 return (v);
100 w = PyInt_FromLong(TYPE(n));
101 if (w == NULL) {
102 Py_DECREF(v);
103 return ((PyObject*) NULL);
105 (void) addelem(v, 0, w);
106 for (i = 0; i < NCH(n); i++) {
107 w = node2tuple(CHILD(n, i), mkseq, addelem, lineno, col_offset);
108 if (w == NULL) {
109 Py_DECREF(v);
110 return ((PyObject*) NULL);
112 (void) addelem(v, i+1, w);
115 if (TYPE(n) == encoding_decl)
116 (void) addelem(v, i+1, PyString_FromString(STR(n)));
117 return (v);
119 else if (ISTERMINAL(TYPE(n))) {
120 PyObject *result = mkseq(2 + lineno + col_offset);
121 if (result != NULL) {
122 (void) addelem(result, 0, PyInt_FromLong(TYPE(n)));
123 (void) addelem(result, 1, PyString_FromString(STR(n)));
124 if (lineno == 1)
125 (void) addelem(result, 2, PyInt_FromLong(n->n_lineno));
126 if (col_offset == 1)
127 (void) addelem(result, 3, PyInt_FromLong(n->n_col_offset));
129 return (result);
131 else {
132 PyErr_SetString(PyExc_SystemError,
133 "unrecognized parse tree node type");
134 return ((PyObject*) NULL);
138 * End of material copyrighted by Stichting Mathematisch Centrum.
143 /* There are two types of intermediate objects we're interested in:
144 * 'eval' and 'exec' types. These constants can be used in the st_type
145 * field of the object type to identify which any given object represents.
146 * These should probably go in an external header to allow other extensions
147 * to use them, but then, we really should be using C++ too. ;-)
150 #define PyST_EXPR 1
151 #define PyST_SUITE 2
154 /* These are the internal objects and definitions required to implement the
155 * ST type. Most of the internal names are more reminiscent of the 'old'
156 * naming style, but the code uses the new naming convention.
159 static PyObject*
160 parser_error = 0;
163 typedef struct {
164 PyObject_HEAD /* standard object header */
165 node* st_node; /* the node* returned by the parser */
166 int st_type; /* EXPR or SUITE ? */
167 PyCompilerFlags st_flags; /* Parser and compiler flags */
168 } PyST_Object;
171 static void parser_free(PyST_Object *st);
172 static int parser_compare(PyST_Object *left, PyST_Object *right);
173 static PyObject *parser_getattr(PyObject *self, char *name);
176 static
177 PyTypeObject PyST_Type = {
178 PyVarObject_HEAD_INIT(NULL, 0)
179 "parser.st", /* tp_name */
180 (int) sizeof(PyST_Object), /* tp_basicsize */
181 0, /* tp_itemsize */
182 (destructor)parser_free, /* tp_dealloc */
183 0, /* tp_print */
184 parser_getattr, /* tp_getattr */
185 0, /* tp_setattr */
186 (cmpfunc)parser_compare, /* tp_compare */
187 0, /* tp_repr */
188 0, /* tp_as_number */
189 0, /* tp_as_sequence */
190 0, /* tp_as_mapping */
191 0, /* tp_hash */
192 0, /* tp_call */
193 0, /* tp_str */
194 0, /* tp_getattro */
195 0, /* tp_setattro */
197 /* Functions to access object as input/output buffer */
198 0, /* tp_as_buffer */
200 Py_TPFLAGS_DEFAULT, /* tp_flags */
202 /* __doc__ */
203 "Intermediate representation of a Python parse tree."
204 }; /* PyST_Type */
207 static int
208 parser_compare_nodes(node *left, node *right)
210 int j;
212 if (TYPE(left) < TYPE(right))
213 return (-1);
215 if (TYPE(right) < TYPE(left))
216 return (1);
218 if (ISTERMINAL(TYPE(left)))
219 return (strcmp(STR(left), STR(right)));
221 if (NCH(left) < NCH(right))
222 return (-1);
224 if (NCH(right) < NCH(left))
225 return (1);
227 for (j = 0; j < NCH(left); ++j) {
228 int v = parser_compare_nodes(CHILD(left, j), CHILD(right, j));
230 if (v != 0)
231 return (v);
233 return (0);
237 /* int parser_compare(PyST_Object* left, PyST_Object* right)
239 * Comparison function used by the Python operators ==, !=, <, >, <=, >=
240 * This really just wraps a call to parser_compare_nodes() with some easy
241 * checks and protection code.
244 static int
245 parser_compare(PyST_Object *left, PyST_Object *right)
247 if (left == right)
248 return (0);
250 if ((left == 0) || (right == 0))
251 return (-1);
253 return (parser_compare_nodes(left->st_node, right->st_node));
257 /* parser_newstobject(node* st)
259 * Allocates a new Python object representing an ST. This is simply the
260 * 'wrapper' object that holds a node* and allows it to be passed around in
261 * Python code.
264 static PyObject*
265 parser_newstobject(node *st, int type)
267 PyST_Object* o = PyObject_New(PyST_Object, &PyST_Type);
269 if (o != 0) {
270 o->st_node = st;
271 o->st_type = type;
272 o->st_flags.cf_flags = 0;
274 else {
275 PyNode_Free(st);
277 return ((PyObject*)o);
281 /* void parser_free(PyST_Object* st)
283 * This is called by a del statement that reduces the reference count to 0.
286 static void
287 parser_free(PyST_Object *st)
289 PyNode_Free(st->st_node);
290 PyObject_Del(st);
294 /* parser_st2tuple(PyObject* self, PyObject* args, PyObject* kw)
296 * This provides conversion from a node* to a tuple object that can be
297 * returned to the Python-level caller. The ST object is not modified.
300 static PyObject*
301 parser_st2tuple(PyST_Object *self, PyObject *args, PyObject *kw)
303 PyObject *line_option = 0;
304 PyObject *col_option = 0;
305 PyObject *res = 0;
306 int ok;
308 static char *keywords[] = {"ast", "line_info", "col_info", NULL};
310 if (self == NULL) {
311 ok = PyArg_ParseTupleAndKeywords(args, kw, "O!|OO:st2tuple", keywords,
312 &PyST_Type, &self, &line_option,
313 &col_option);
315 else
316 ok = PyArg_ParseTupleAndKeywords(args, kw, "|OO:totuple", &keywords[1],
317 &line_option, &col_option);
318 if (ok != 0) {
319 int lineno = 0;
320 int col_offset = 0;
321 if (line_option != NULL) {
322 lineno = (PyObject_IsTrue(line_option) != 0) ? 1 : 0;
324 if (col_option != NULL) {
325 col_offset = (PyObject_IsTrue(col_option) != 0) ? 1 : 0;
328 * Convert ST into a tuple representation. Use Guido's function,
329 * since it's known to work already.
331 res = node2tuple(((PyST_Object*)self)->st_node,
332 PyTuple_New, PyTuple_SetItem, lineno, col_offset);
334 return (res);
337 static PyObject*
338 parser_ast2tuple(PyST_Object *self, PyObject *args, PyObject *kw)
340 if (PyErr_WarnPy3k("ast2tuple is removed in 3.x; use st2tuple", 1) < 0)
341 return NULL;
342 return parser_st2tuple(self, args, kw);
346 /* parser_st2list(PyObject* self, PyObject* args, PyObject* kw)
348 * This provides conversion from a node* to a list object that can be
349 * returned to the Python-level caller. The ST object is not modified.
352 static PyObject*
353 parser_st2list(PyST_Object *self, PyObject *args, PyObject *kw)
355 PyObject *line_option = 0;
356 PyObject *col_option = 0;
357 PyObject *res = 0;
358 int ok;
360 static char *keywords[] = {"ast", "line_info", "col_info", NULL};
362 if (self == NULL)
363 ok = PyArg_ParseTupleAndKeywords(args, kw, "O!|OO:st2list", keywords,
364 &PyST_Type, &self, &line_option,
365 &col_option);
366 else
367 ok = PyArg_ParseTupleAndKeywords(args, kw, "|OO:tolist", &keywords[1],
368 &line_option, &col_option);
369 if (ok) {
370 int lineno = 0;
371 int col_offset = 0;
372 if (line_option != 0) {
373 lineno = PyObject_IsTrue(line_option) ? 1 : 0;
375 if (col_option != NULL) {
376 col_offset = (PyObject_IsTrue(col_option) != 0) ? 1 : 0;
379 * Convert ST into a tuple representation. Use Guido's function,
380 * since it's known to work already.
382 res = node2tuple(self->st_node,
383 PyList_New, PyList_SetItem, lineno, col_offset);
385 return (res);
388 static PyObject*
389 parser_ast2list(PyST_Object *self, PyObject *args, PyObject *kw)
391 if (PyErr_WarnPy3k("ast2list is removed in 3.x; use st2list", 1) < 0)
392 return NULL;
393 return parser_st2list(self, args, kw);
397 /* parser_compilest(PyObject* self, PyObject* args)
399 * This function creates code objects from the parse tree represented by
400 * the passed-in data object. An optional file name is passed in as well.
403 static PyObject*
404 parser_compilest(PyST_Object *self, PyObject *args, PyObject *kw)
406 PyObject* res = 0;
407 PyArena* arena;
408 mod_ty mod;
409 char* str = "<syntax-tree>";
410 int ok;
412 static char *keywords[] = {"ast", "filename", NULL};
414 if (self == NULL)
415 ok = PyArg_ParseTupleAndKeywords(args, kw, "O!|s:compilest", keywords,
416 &PyST_Type, &self, &str);
417 else
418 ok = PyArg_ParseTupleAndKeywords(args, kw, "|s:compile", &keywords[1],
419 &str);
421 if (ok) {
422 arena = PyArena_New();
423 if (arena) {
424 mod = PyAST_FromNode(self->st_node, &(self->st_flags), str, arena);
425 if (mod) {
426 res = (PyObject *)PyAST_Compile(mod, str, &(self->st_flags), arena);
428 PyArena_Free(arena);
432 return (res);
435 static PyObject*
436 parser_compileast(PyST_Object *self, PyObject *args, PyObject *kw)
438 if (PyErr_WarnPy3k("compileast is removed in 3.x; use compilest", 1) < 0)
439 return NULL;
440 return parser_compilest(self, args, kw);
444 /* PyObject* parser_isexpr(PyObject* self, PyObject* args)
445 * PyObject* parser_issuite(PyObject* self, PyObject* args)
447 * Checks the passed-in ST object to determine if it is an expression or
448 * a statement suite, respectively. The return is a Python truth value.
451 static PyObject*
452 parser_isexpr(PyST_Object *self, PyObject *args, PyObject *kw)
454 PyObject* res = 0;
455 int ok;
457 static char *keywords[] = {"ast", NULL};
459 if (self == NULL)
460 ok = PyArg_ParseTupleAndKeywords(args, kw, "O!:isexpr", keywords,
461 &PyST_Type, &self);
462 else
463 ok = PyArg_ParseTupleAndKeywords(args, kw, ":isexpr", &keywords[1]);
465 if (ok) {
466 /* Check to see if the ST represents an expression or not. */
467 res = (self->st_type == PyST_EXPR) ? Py_True : Py_False;
468 Py_INCREF(res);
470 return (res);
474 static PyObject*
475 parser_issuite(PyST_Object *self, PyObject *args, PyObject *kw)
477 PyObject* res = 0;
478 int ok;
480 static char *keywords[] = {"ast", NULL};
482 if (self == NULL)
483 ok = PyArg_ParseTupleAndKeywords(args, kw, "O!:issuite", keywords,
484 &PyST_Type, &self);
485 else
486 ok = PyArg_ParseTupleAndKeywords(args, kw, ":issuite", &keywords[1]);
488 if (ok) {
489 /* Check to see if the ST represents an expression or not. */
490 res = (self->st_type == PyST_EXPR) ? Py_False : Py_True;
491 Py_INCREF(res);
493 return (res);
497 #define PUBLIC_METHOD_TYPE (METH_VARARGS|METH_KEYWORDS)
499 static PyMethodDef
500 parser_methods[] = {
501 {"compile", (PyCFunction)parser_compilest, PUBLIC_METHOD_TYPE,
502 PyDoc_STR("Compile this ST object into a code object.")},
503 {"isexpr", (PyCFunction)parser_isexpr, PUBLIC_METHOD_TYPE,
504 PyDoc_STR("Determines if this ST object was created from an expression.")},
505 {"issuite", (PyCFunction)parser_issuite, PUBLIC_METHOD_TYPE,
506 PyDoc_STR("Determines if this ST object was created from a suite.")},
507 {"tolist", (PyCFunction)parser_st2list, PUBLIC_METHOD_TYPE,
508 PyDoc_STR("Creates a list-tree representation of this ST.")},
509 {"totuple", (PyCFunction)parser_st2tuple, PUBLIC_METHOD_TYPE,
510 PyDoc_STR("Creates a tuple-tree representation of this ST.")},
512 {NULL, NULL, 0, NULL}
516 static PyObject*
517 parser_getattr(PyObject *self, char *name)
519 return (Py_FindMethod(parser_methods, self, name));
523 /* err_string(char* message)
525 * Sets the error string for an exception of type ParserError.
528 static void
529 err_string(char *message)
531 PyErr_SetString(parser_error, message);
535 /* PyObject* parser_do_parse(PyObject* args, int type)
537 * Internal function to actually execute the parse and return the result if
538 * successful or set an exception if not.
541 static PyObject*
542 parser_do_parse(PyObject *args, PyObject *kw, char *argspec, int type)
544 char* string = 0;
545 PyObject* res = 0;
546 int flags = 0;
547 perrdetail err;
549 static char *keywords[] = {"source", NULL};
551 if (PyArg_ParseTupleAndKeywords(args, kw, argspec, keywords, &string)) {
552 node* n = PyParser_ParseStringFlagsFilenameEx(string, NULL,
553 &_PyParser_Grammar,
554 (type == PyST_EXPR)
555 ? eval_input : file_input,
556 &err, &flags);
558 if (n) {
559 res = parser_newstobject(n, type);
560 if (res)
561 ((PyST_Object *)res)->st_flags.cf_flags = flags & PyCF_MASK;
563 else
564 PyParser_SetError(&err);
566 return (res);
570 /* PyObject* parser_expr(PyObject* self, PyObject* args)
571 * PyObject* parser_suite(PyObject* self, PyObject* args)
573 * External interfaces to the parser itself. Which is called determines if
574 * the parser attempts to recognize an expression ('eval' form) or statement
575 * suite ('exec' form). The real work is done by parser_do_parse() above.
578 static PyObject*
579 parser_expr(PyST_Object *self, PyObject *args, PyObject *kw)
581 NOTE(ARGUNUSED(self))
582 return (parser_do_parse(args, kw, "s:expr", PyST_EXPR));
586 static PyObject*
587 parser_suite(PyST_Object *self, PyObject *args, PyObject *kw)
589 NOTE(ARGUNUSED(self))
590 return (parser_do_parse(args, kw, "s:suite", PyST_SUITE));
595 /* This is the messy part of the code. Conversion from a tuple to an ST
596 * object requires that the input tuple be valid without having to rely on
597 * catching an exception from the compiler. This is done to allow the
598 * compiler itself to remain fast, since most of its input will come from
599 * the parser directly, and therefore be known to be syntactically correct.
600 * This validation is done to ensure that we don't core dump the compile
601 * phase, returning an exception instead.
603 * Two aspects can be broken out in this code: creating a node tree from
604 * the tuple passed in, and verifying that it is indeed valid. It may be
605 * advantageous to expand the number of ST types to include funcdefs and
606 * lambdadefs to take advantage of the optimizer, recognizing those STs
607 * here. They are not necessary, and not quite as useful in a raw form.
608 * For now, let's get expressions and suites working reliably.
612 static node* build_node_tree(PyObject *tuple);
613 static int validate_expr_tree(node *tree);
614 static int validate_file_input(node *tree);
615 static int validate_encoding_decl(node *tree);
617 /* PyObject* parser_tuple2st(PyObject* self, PyObject* args)
619 * This is the public function, called from the Python code. It receives a
620 * single tuple object from the caller, and creates an ST object if the
621 * tuple can be validated. It does this by checking the first code of the
622 * tuple, and, if acceptable, builds the internal representation. If this
623 * step succeeds, the internal representation is validated as fully as
624 * possible with the various validate_*() routines defined below.
626 * This function must be changed if support is to be added for PyST_FRAGMENT
627 * ST objects.
630 static PyObject*
631 parser_tuple2st(PyST_Object *self, PyObject *args, PyObject *kw)
633 NOTE(ARGUNUSED(self))
634 PyObject *st = 0;
635 PyObject *tuple;
636 node *tree;
638 static char *keywords[] = {"sequence", NULL};
640 if (!PyArg_ParseTupleAndKeywords(args, kw, "O:sequence2st", keywords,
641 &tuple))
642 return (0);
643 if (!PySequence_Check(tuple)) {
644 PyErr_SetString(PyExc_ValueError,
645 "sequence2st() requires a single sequence argument");
646 return (0);
649 * Convert the tree to the internal form before checking it.
651 tree = build_node_tree(tuple);
652 if (tree != 0) {
653 int start_sym = TYPE(tree);
654 if (start_sym == eval_input) {
655 /* Might be an eval form. */
656 if (validate_expr_tree(tree))
657 st = parser_newstobject(tree, PyST_EXPR);
658 else
659 PyNode_Free(tree);
661 else if (start_sym == file_input) {
662 /* This looks like an exec form so far. */
663 if (validate_file_input(tree))
664 st = parser_newstobject(tree, PyST_SUITE);
665 else
666 PyNode_Free(tree);
668 else if (start_sym == encoding_decl) {
669 /* This looks like an encoding_decl so far. */
670 if (validate_encoding_decl(tree))
671 st = parser_newstobject(tree, PyST_SUITE);
672 else
673 PyNode_Free(tree);
675 else {
676 /* This is a fragment, at best. */
677 PyNode_Free(tree);
678 err_string("parse tree does not use a valid start symbol");
681 /* Make sure we throw an exception on all errors. We should never
682 * get this, but we'd do well to be sure something is done.
684 if (st == NULL && !PyErr_Occurred())
685 err_string("unspecified ST error occurred");
687 return st;
690 static PyObject*
691 parser_tuple2ast(PyST_Object *self, PyObject *args, PyObject *kw)
693 if (PyErr_WarnPy3k("tuple2ast is removed in 3.x; use tuple2st", 1) < 0)
694 return NULL;
695 return parser_tuple2st(self, args, kw);
699 /* node* build_node_children()
701 * Iterate across the children of the current non-terminal node and build
702 * their structures. If successful, return the root of this portion of
703 * the tree, otherwise, 0. Any required exception will be specified already,
704 * and no memory will have been deallocated.
707 static node*
708 build_node_children(PyObject *tuple, node *root, int *line_num)
710 Py_ssize_t len = PyObject_Size(tuple);
711 Py_ssize_t i;
712 int err;
714 for (i = 1; i < len; ++i) {
715 /* elem must always be a sequence, however simple */
716 PyObject* elem = PySequence_GetItem(tuple, i);
717 int ok = elem != NULL;
718 long type = 0;
719 char *strn = 0;
721 if (ok)
722 ok = PySequence_Check(elem);
723 if (ok) {
724 PyObject *temp = PySequence_GetItem(elem, 0);
725 if (temp == NULL)
726 ok = 0;
727 else {
728 ok = PyInt_Check(temp);
729 if (ok)
730 type = PyInt_AS_LONG(temp);
731 Py_DECREF(temp);
734 if (!ok) {
735 PyObject *err = Py_BuildValue("os", elem,
736 "Illegal node construct.");
737 PyErr_SetObject(parser_error, err);
738 Py_XDECREF(err);
739 Py_XDECREF(elem);
740 return (0);
742 if (ISTERMINAL(type)) {
743 Py_ssize_t len = PyObject_Size(elem);
744 PyObject *temp;
746 if ((len != 2) && (len != 3)) {
747 err_string("terminal nodes must have 2 or 3 entries");
748 return 0;
750 temp = PySequence_GetItem(elem, 1);
751 if (temp == NULL)
752 return 0;
753 if (!PyString_Check(temp)) {
754 PyErr_Format(parser_error,
755 "second item in terminal node must be a string,"
756 " found %s",
757 Py_TYPE(temp)->tp_name);
758 Py_DECREF(temp);
759 return 0;
761 if (len == 3) {
762 PyObject *o = PySequence_GetItem(elem, 2);
763 if (o != NULL) {
764 if (PyInt_Check(o))
765 *line_num = PyInt_AS_LONG(o);
766 else {
767 PyErr_Format(parser_error,
768 "third item in terminal node must be an"
769 " integer, found %s",
770 Py_TYPE(temp)->tp_name);
771 Py_DECREF(o);
772 Py_DECREF(temp);
773 return 0;
775 Py_DECREF(o);
778 len = PyString_GET_SIZE(temp) + 1;
779 strn = (char *)PyObject_MALLOC(len);
780 if (strn != NULL)
781 (void) memcpy(strn, PyString_AS_STRING(temp), len);
782 Py_DECREF(temp);
784 else if (!ISNONTERMINAL(type)) {
786 * It has to be one or the other; this is an error.
787 * Throw an exception.
789 PyObject *err = Py_BuildValue("os", elem, "unknown node type.");
790 PyErr_SetObject(parser_error, err);
791 Py_XDECREF(err);
792 Py_XDECREF(elem);
793 return (0);
795 err = PyNode_AddChild(root, type, strn, *line_num, 0);
796 if (err == E_NOMEM) {
797 PyObject_FREE(strn);
798 return (node *) PyErr_NoMemory();
800 if (err == E_OVERFLOW) {
801 PyObject_FREE(strn);
802 PyErr_SetString(PyExc_ValueError,
803 "unsupported number of child nodes");
804 return NULL;
807 if (ISNONTERMINAL(type)) {
808 node* new_child = CHILD(root, i - 1);
810 if (new_child != build_node_children(elem, new_child, line_num)) {
811 Py_XDECREF(elem);
812 return (0);
815 else if (type == NEWLINE) { /* It's true: we increment the */
816 ++(*line_num); /* line number *after* the newline! */
818 Py_XDECREF(elem);
820 return root;
824 static node*
825 build_node_tree(PyObject *tuple)
827 node* res = 0;
828 PyObject *temp = PySequence_GetItem(tuple, 0);
829 long num = -1;
831 if (temp != NULL)
832 num = PyInt_AsLong(temp);
833 Py_XDECREF(temp);
834 if (ISTERMINAL(num)) {
836 * The tuple is simple, but it doesn't start with a start symbol.
837 * Throw an exception now and be done with it.
839 tuple = Py_BuildValue("os", tuple,
840 "Illegal syntax-tree; cannot start with terminal symbol.");
841 PyErr_SetObject(parser_error, tuple);
842 Py_XDECREF(tuple);
844 else if (ISNONTERMINAL(num)) {
846 * Not efficient, but that can be handled later.
848 int line_num = 0;
849 PyObject *encoding = NULL;
851 if (num == encoding_decl) {
852 encoding = PySequence_GetItem(tuple, 2);
853 /* tuple isn't borrowed anymore here, need to DECREF */
854 tuple = PySequence_GetSlice(tuple, 0, 2);
856 res = PyNode_New(num);
857 if (res != NULL) {
858 if (res != build_node_children(tuple, res, &line_num)) {
859 PyNode_Free(res);
860 res = NULL;
862 if (res && encoding) {
863 Py_ssize_t len;
864 len = PyString_GET_SIZE(encoding) + 1;
865 res->n_str = (char *)PyObject_MALLOC(len);
866 if (res->n_str != NULL)
867 (void) memcpy(res->n_str, PyString_AS_STRING(encoding), len);
868 Py_DECREF(encoding);
869 Py_DECREF(tuple);
873 else {
874 /* The tuple is illegal -- if the number is neither TERMINAL nor
875 * NONTERMINAL, we can't use it. Not sure the implementation
876 * allows this condition, but the API doesn't preclude it.
878 PyObject *err = Py_BuildValue("os", tuple,
879 "Illegal component tuple.");
880 PyErr_SetObject(parser_error, err);
881 Py_XDECREF(err);
884 return (res);
889 * Validation routines used within the validation section:
891 static int validate_terminal(node *terminal, int type, char *string);
893 #define validate_ampersand(ch) validate_terminal(ch, AMPER, "&")
894 #define validate_circumflex(ch) validate_terminal(ch, CIRCUMFLEX, "^")
895 #define validate_colon(ch) validate_terminal(ch, COLON, ":")
896 #define validate_comma(ch) validate_terminal(ch, COMMA, ",")
897 #define validate_dedent(ch) validate_terminal(ch, DEDENT, "")
898 #define validate_equal(ch) validate_terminal(ch, EQUAL, "=")
899 #define validate_indent(ch) validate_terminal(ch, INDENT, (char*)NULL)
900 #define validate_lparen(ch) validate_terminal(ch, LPAR, "(")
901 #define validate_newline(ch) validate_terminal(ch, NEWLINE, (char*)NULL)
902 #define validate_rparen(ch) validate_terminal(ch, RPAR, ")")
903 #define validate_semi(ch) validate_terminal(ch, SEMI, ";")
904 #define validate_star(ch) validate_terminal(ch, STAR, "*")
905 #define validate_vbar(ch) validate_terminal(ch, VBAR, "|")
906 #define validate_doublestar(ch) validate_terminal(ch, DOUBLESTAR, "**")
907 #define validate_dot(ch) validate_terminal(ch, DOT, ".")
908 #define validate_at(ch) validate_terminal(ch, AT, "@")
909 #define validate_name(ch, str) validate_terminal(ch, NAME, str)
911 #define VALIDATER(n) static int validate_##n(node *tree)
913 VALIDATER(node); VALIDATER(small_stmt);
914 VALIDATER(class); VALIDATER(node);
915 VALIDATER(parameters); VALIDATER(suite);
916 VALIDATER(testlist); VALIDATER(varargslist);
917 VALIDATER(fpdef); VALIDATER(fplist);
918 VALIDATER(stmt); VALIDATER(simple_stmt);
919 VALIDATER(expr_stmt); VALIDATER(power);
920 VALIDATER(print_stmt); VALIDATER(del_stmt);
921 VALIDATER(return_stmt); VALIDATER(list_iter);
922 VALIDATER(raise_stmt); VALIDATER(import_stmt);
923 VALIDATER(import_name); VALIDATER(import_from);
924 VALIDATER(global_stmt); VALIDATER(list_if);
925 VALIDATER(assert_stmt); VALIDATER(list_for);
926 VALIDATER(exec_stmt); VALIDATER(compound_stmt);
927 VALIDATER(while); VALIDATER(for);
928 VALIDATER(try); VALIDATER(except_clause);
929 VALIDATER(test); VALIDATER(and_test);
930 VALIDATER(not_test); VALIDATER(comparison);
931 VALIDATER(comp_op); VALIDATER(expr);
932 VALIDATER(xor_expr); VALIDATER(and_expr);
933 VALIDATER(shift_expr); VALIDATER(arith_expr);
934 VALIDATER(term); VALIDATER(factor);
935 VALIDATER(atom); VALIDATER(lambdef);
936 VALIDATER(trailer); VALIDATER(subscript);
937 VALIDATER(subscriptlist); VALIDATER(sliceop);
938 VALIDATER(exprlist); VALIDATER(dictorsetmaker);
939 VALIDATER(arglist); VALIDATER(argument);
940 VALIDATER(listmaker); VALIDATER(yield_stmt);
941 VALIDATER(testlist1); VALIDATER(comp_for);
942 VALIDATER(comp_iter); VALIDATER(comp_if);
943 VALIDATER(testlist_comp); VALIDATER(yield_expr);
944 VALIDATER(yield_or_testlist); VALIDATER(or_test);
945 VALIDATER(old_test); VALIDATER(old_lambdef);
947 #undef VALIDATER
949 #define is_even(n) (((n) & 1) == 0)
950 #define is_odd(n) (((n) & 1) == 1)
953 static int
954 validate_ntype(node *n, int t)
956 if (TYPE(n) != t) {
957 PyErr_Format(parser_error, "Expected node type %d, got %d.",
958 t, TYPE(n));
959 return 0;
961 return 1;
965 /* Verifies that the number of child nodes is exactly 'num', raising
966 * an exception if it isn't. The exception message does not indicate
967 * the exact number of nodes, allowing this to be used to raise the
968 * "right" exception when the wrong number of nodes is present in a
969 * specific variant of a statement's syntax. This is commonly used
970 * in that fashion.
972 static int
973 validate_numnodes(node *n, int num, const char *const name)
975 if (NCH(n) != num) {
976 PyErr_Format(parser_error,
977 "Illegal number of children for %s node.", name);
978 return 0;
980 return 1;
984 static int
985 validate_terminal(node *terminal, int type, char *string)
987 int res = (validate_ntype(terminal, type)
988 && ((string == 0) || (strcmp(string, STR(terminal)) == 0)));
990 if (!res && !PyErr_Occurred()) {
991 PyErr_Format(parser_error,
992 "Illegal terminal: expected \"%s\"", string);
994 return (res);
998 /* X (',' X) [',']
1000 static int
1001 validate_repeating_list(node *tree, int ntype, int (*vfunc)(node *),
1002 const char *const name)
1004 int nch = NCH(tree);
1005 int res = (nch && validate_ntype(tree, ntype)
1006 && vfunc(CHILD(tree, 0)));
1008 if (!res && !PyErr_Occurred())
1009 (void) validate_numnodes(tree, 1, name);
1010 else {
1011 if (is_even(nch))
1012 res = validate_comma(CHILD(tree, --nch));
1013 if (res && nch > 1) {
1014 int pos = 1;
1015 for ( ; res && pos < nch; pos += 2)
1016 res = (validate_comma(CHILD(tree, pos))
1017 && vfunc(CHILD(tree, pos + 1)));
1020 return (res);
1024 /* validate_class()
1026 * classdef:
1027 * 'class' NAME ['(' testlist ')'] ':' suite
1029 static int
1030 validate_class(node *tree)
1032 int nch = NCH(tree);
1033 int res = (validate_ntype(tree, classdef) &&
1034 ((nch == 4) || (nch == 6) || (nch == 7)));
1036 if (res) {
1037 res = (validate_name(CHILD(tree, 0), "class")
1038 && validate_ntype(CHILD(tree, 1), NAME)
1039 && validate_colon(CHILD(tree, nch - 2))
1040 && validate_suite(CHILD(tree, nch - 1)));
1042 else {
1043 (void) validate_numnodes(tree, 4, "class");
1046 if (res) {
1047 if (nch == 7) {
1048 res = ((validate_lparen(CHILD(tree, 2)) &&
1049 validate_testlist(CHILD(tree, 3)) &&
1050 validate_rparen(CHILD(tree, 4))));
1052 else if (nch == 6) {
1053 res = (validate_lparen(CHILD(tree,2)) &&
1054 validate_rparen(CHILD(tree,3)));
1057 return (res);
1061 /* if_stmt:
1062 * 'if' test ':' suite ('elif' test ':' suite)* ['else' ':' suite]
1064 static int
1065 validate_if(node *tree)
1067 int nch = NCH(tree);
1068 int res = (validate_ntype(tree, if_stmt)
1069 && (nch >= 4)
1070 && validate_name(CHILD(tree, 0), "if")
1071 && validate_test(CHILD(tree, 1))
1072 && validate_colon(CHILD(tree, 2))
1073 && validate_suite(CHILD(tree, 3)));
1075 if (res && ((nch % 4) == 3)) {
1076 /* ... 'else' ':' suite */
1077 res = (validate_name(CHILD(tree, nch - 3), "else")
1078 && validate_colon(CHILD(tree, nch - 2))
1079 && validate_suite(CHILD(tree, nch - 1)));
1080 nch -= 3;
1082 else if (!res && !PyErr_Occurred())
1083 (void) validate_numnodes(tree, 4, "if");
1084 if ((nch % 4) != 0)
1085 /* Will catch the case for nch < 4 */
1086 res = validate_numnodes(tree, 0, "if");
1087 else if (res && (nch > 4)) {
1088 /* ... ('elif' test ':' suite)+ ... */
1089 int j = 4;
1090 while ((j < nch) && res) {
1091 res = (validate_name(CHILD(tree, j), "elif")
1092 && validate_colon(CHILD(tree, j + 2))
1093 && validate_test(CHILD(tree, j + 1))
1094 && validate_suite(CHILD(tree, j + 3)));
1095 j += 4;
1098 return (res);
1102 /* parameters:
1103 * '(' [varargslist] ')'
1106 static int
1107 validate_parameters(node *tree)
1109 int nch = NCH(tree);
1110 int res = validate_ntype(tree, parameters) && ((nch == 2) || (nch == 3));
1112 if (res) {
1113 res = (validate_lparen(CHILD(tree, 0))
1114 && validate_rparen(CHILD(tree, nch - 1)));
1115 if (res && (nch == 3))
1116 res = validate_varargslist(CHILD(tree, 1));
1118 else {
1119 (void) validate_numnodes(tree, 2, "parameters");
1121 return (res);
1125 /* validate_suite()
1127 * suite:
1128 * simple_stmt
1129 * | NEWLINE INDENT stmt+ DEDENT
1131 static int
1132 validate_suite(node *tree)
1134 int nch = NCH(tree);
1135 int res = (validate_ntype(tree, suite) && ((nch == 1) || (nch >= 4)));
1137 if (res && (nch == 1))
1138 res = validate_simple_stmt(CHILD(tree, 0));
1139 else if (res) {
1140 /* NEWLINE INDENT stmt+ DEDENT */
1141 res = (validate_newline(CHILD(tree, 0))
1142 && validate_indent(CHILD(tree, 1))
1143 && validate_stmt(CHILD(tree, 2))
1144 && validate_dedent(CHILD(tree, nch - 1)));
1146 if (res && (nch > 4)) {
1147 int i = 3;
1148 --nch; /* forget the DEDENT */
1149 for ( ; res && (i < nch); ++i)
1150 res = validate_stmt(CHILD(tree, i));
1152 else if (nch < 4)
1153 res = validate_numnodes(tree, 4, "suite");
1155 return (res);
1159 static int
1160 validate_testlist(node *tree)
1162 return (validate_repeating_list(tree, testlist,
1163 validate_test, "testlist"));
1167 static int
1168 validate_testlist1(node *tree)
1170 return (validate_repeating_list(tree, testlist1,
1171 validate_test, "testlist1"));
1175 static int
1176 validate_testlist_safe(node *tree)
1178 return (validate_repeating_list(tree, testlist_safe,
1179 validate_old_test, "testlist_safe"));
1183 /* '*' NAME [',' '**' NAME] | '**' NAME
1185 static int
1186 validate_varargslist_trailer(node *tree, int start)
1188 int nch = NCH(tree);
1189 int res = 0;
1190 int sym;
1192 if (nch <= start) {
1193 err_string("expected variable argument trailer for varargslist");
1194 return 0;
1196 sym = TYPE(CHILD(tree, start));
1197 if (sym == STAR) {
1199 * ('*' NAME [',' '**' NAME]
1201 if (nch-start == 2)
1202 res = validate_name(CHILD(tree, start+1), NULL);
1203 else if (nch-start == 5)
1204 res = (validate_name(CHILD(tree, start+1), NULL)
1205 && validate_comma(CHILD(tree, start+2))
1206 && validate_doublestar(CHILD(tree, start+3))
1207 && validate_name(CHILD(tree, start+4), NULL));
1209 else if (sym == DOUBLESTAR) {
1211 * '**' NAME
1213 if (nch-start == 2)
1214 res = validate_name(CHILD(tree, start+1), NULL);
1216 if (!res)
1217 err_string("illegal variable argument trailer for varargslist");
1218 return res;
1222 /* validate_varargslist()
1224 * varargslist:
1225 * (fpdef ['=' test] ',')*
1226 * ('*' NAME [',' '**' NAME]
1227 * | '**' NAME)
1228 * | fpdef ['=' test] (',' fpdef ['=' test])* [',']
1231 static int
1232 validate_varargslist(node *tree)
1234 int nch = NCH(tree);
1235 int res = validate_ntype(tree, varargslist) && (nch != 0);
1236 int sym;
1238 if (!res)
1239 return 0;
1240 if (nch < 1) {
1241 err_string("varargslist missing child nodes");
1242 return 0;
1244 sym = TYPE(CHILD(tree, 0));
1245 if (sym == STAR || sym == DOUBLESTAR)
1246 /* whole thing matches:
1247 * '*' NAME [',' '**' NAME] | '**' NAME
1249 res = validate_varargslist_trailer(tree, 0);
1250 else if (sym == fpdef) {
1251 int i = 0;
1253 sym = TYPE(CHILD(tree, nch-1));
1254 if (sym == NAME) {
1256 * (fpdef ['=' test] ',')+
1257 * ('*' NAME [',' '**' NAME]
1258 * | '**' NAME)
1260 /* skip over (fpdef ['=' test] ',')+ */
1261 while (res && (i+2 <= nch)) {
1262 res = validate_fpdef(CHILD(tree, i));
1263 ++i;
1264 if (res && TYPE(CHILD(tree, i)) == EQUAL && (i+2 <= nch)) {
1265 res = (validate_equal(CHILD(tree, i))
1266 && validate_test(CHILD(tree, i+1)));
1267 if (res)
1268 i += 2;
1270 if (res && i < nch) {
1271 res = validate_comma(CHILD(tree, i));
1272 ++i;
1273 if (res && i < nch
1274 && (TYPE(CHILD(tree, i)) == DOUBLESTAR
1275 || TYPE(CHILD(tree, i)) == STAR))
1276 break;
1279 /* ... '*' NAME [',' '**' NAME] | '**' NAME
1280 * i --^^^
1282 if (res)
1283 res = validate_varargslist_trailer(tree, i);
1285 else {
1287 * fpdef ['=' test] (',' fpdef ['=' test])* [',']
1289 /* strip trailing comma node */
1290 if (sym == COMMA) {
1291 res = validate_comma(CHILD(tree, nch-1));
1292 if (!res)
1293 return 0;
1294 --nch;
1297 * fpdef ['=' test] (',' fpdef ['=' test])*
1299 res = validate_fpdef(CHILD(tree, 0));
1300 ++i;
1301 if (res && (i+2 <= nch) && TYPE(CHILD(tree, i)) == EQUAL) {
1302 res = (validate_equal(CHILD(tree, i))
1303 && validate_test(CHILD(tree, i+1)));
1304 i += 2;
1307 * ... (',' fpdef ['=' test])*
1308 * i ---^^^
1310 while (res && (nch - i) >= 2) {
1311 res = (validate_comma(CHILD(tree, i))
1312 && validate_fpdef(CHILD(tree, i+1)));
1313 i += 2;
1314 if (res && (nch - i) >= 2 && TYPE(CHILD(tree, i)) == EQUAL) {
1315 res = (validate_equal(CHILD(tree, i))
1316 && validate_test(CHILD(tree, i+1)));
1317 i += 2;
1320 if (res && nch - i != 0) {
1321 res = 0;
1322 err_string("illegal formation for varargslist");
1326 return res;
1330 /* list_iter: list_for | list_if
1332 static int
1333 validate_list_iter(node *tree)
1335 int res = (validate_ntype(tree, list_iter)
1336 && validate_numnodes(tree, 1, "list_iter"));
1337 if (res && TYPE(CHILD(tree, 0)) == list_for)
1338 res = validate_list_for(CHILD(tree, 0));
1339 else
1340 res = validate_list_if(CHILD(tree, 0));
1342 return res;
1345 /* comp_iter: comp_for | comp_if
1347 static int
1348 validate_comp_iter(node *tree)
1350 int res = (validate_ntype(tree, comp_iter)
1351 && validate_numnodes(tree, 1, "comp_iter"));
1352 if (res && TYPE(CHILD(tree, 0)) == comp_for)
1353 res = validate_comp_for(CHILD(tree, 0));
1354 else
1355 res = validate_comp_if(CHILD(tree, 0));
1357 return res;
1360 /* list_for: 'for' exprlist 'in' testlist [list_iter]
1362 static int
1363 validate_list_for(node *tree)
1365 int nch = NCH(tree);
1366 int res;
1368 if (nch == 5)
1369 res = validate_list_iter(CHILD(tree, 4));
1370 else
1371 res = validate_numnodes(tree, 4, "list_for");
1373 if (res)
1374 res = (validate_name(CHILD(tree, 0), "for")
1375 && validate_exprlist(CHILD(tree, 1))
1376 && validate_name(CHILD(tree, 2), "in")
1377 && validate_testlist_safe(CHILD(tree, 3)));
1379 return res;
1382 /* comp_for: 'for' exprlist 'in' test [comp_iter]
1384 static int
1385 validate_comp_for(node *tree)
1387 int nch = NCH(tree);
1388 int res;
1390 if (nch == 5)
1391 res = validate_comp_iter(CHILD(tree, 4));
1392 else
1393 res = validate_numnodes(tree, 4, "comp_for");
1395 if (res)
1396 res = (validate_name(CHILD(tree, 0), "for")
1397 && validate_exprlist(CHILD(tree, 1))
1398 && validate_name(CHILD(tree, 2), "in")
1399 && validate_or_test(CHILD(tree, 3)));
1401 return res;
1404 /* list_if: 'if' old_test [list_iter]
1406 static int
1407 validate_list_if(node *tree)
1409 int nch = NCH(tree);
1410 int res;
1412 if (nch == 3)
1413 res = validate_list_iter(CHILD(tree, 2));
1414 else
1415 res = validate_numnodes(tree, 2, "list_if");
1417 if (res)
1418 res = (validate_name(CHILD(tree, 0), "if")
1419 && validate_old_test(CHILD(tree, 1)));
1421 return res;
1424 /* comp_if: 'if' old_test [comp_iter]
1426 static int
1427 validate_comp_if(node *tree)
1429 int nch = NCH(tree);
1430 int res;
1432 if (nch == 3)
1433 res = validate_comp_iter(CHILD(tree, 2));
1434 else
1435 res = validate_numnodes(tree, 2, "comp_if");
1437 if (res)
1438 res = (validate_name(CHILD(tree, 0), "if")
1439 && validate_old_test(CHILD(tree, 1)));
1441 return res;
1444 /* validate_fpdef()
1446 * fpdef:
1447 * NAME
1448 * | '(' fplist ')'
1450 static int
1451 validate_fpdef(node *tree)
1453 int nch = NCH(tree);
1454 int res = validate_ntype(tree, fpdef);
1456 if (res) {
1457 if (nch == 1)
1458 res = validate_ntype(CHILD(tree, 0), NAME);
1459 else if (nch == 3)
1460 res = (validate_lparen(CHILD(tree, 0))
1461 && validate_fplist(CHILD(tree, 1))
1462 && validate_rparen(CHILD(tree, 2)));
1463 else
1464 res = validate_numnodes(tree, 1, "fpdef");
1466 return (res);
1470 static int
1471 validate_fplist(node *tree)
1473 return (validate_repeating_list(tree, fplist,
1474 validate_fpdef, "fplist"));
1478 /* simple_stmt | compound_stmt
1481 static int
1482 validate_stmt(node *tree)
1484 int res = (validate_ntype(tree, stmt)
1485 && validate_numnodes(tree, 1, "stmt"));
1487 if (res) {
1488 tree = CHILD(tree, 0);
1490 if (TYPE(tree) == simple_stmt)
1491 res = validate_simple_stmt(tree);
1492 else
1493 res = validate_compound_stmt(tree);
1495 return (res);
1499 /* small_stmt (';' small_stmt)* [';'] NEWLINE
1502 static int
1503 validate_simple_stmt(node *tree)
1505 int nch = NCH(tree);
1506 int res = (validate_ntype(tree, simple_stmt)
1507 && (nch >= 2)
1508 && validate_small_stmt(CHILD(tree, 0))
1509 && validate_newline(CHILD(tree, nch - 1)));
1511 if (nch < 2)
1512 res = validate_numnodes(tree, 2, "simple_stmt");
1513 --nch; /* forget the NEWLINE */
1514 if (res && is_even(nch))
1515 res = validate_semi(CHILD(tree, --nch));
1516 if (res && (nch > 2)) {
1517 int i;
1519 for (i = 1; res && (i < nch); i += 2)
1520 res = (validate_semi(CHILD(tree, i))
1521 && validate_small_stmt(CHILD(tree, i + 1)));
1523 return (res);
1527 static int
1528 validate_small_stmt(node *tree)
1530 int nch = NCH(tree);
1531 int res = validate_numnodes(tree, 1, "small_stmt");
1533 if (res) {
1534 int ntype = TYPE(CHILD(tree, 0));
1536 if ( (ntype == expr_stmt)
1537 || (ntype == print_stmt)
1538 || (ntype == del_stmt)
1539 || (ntype == pass_stmt)
1540 || (ntype == flow_stmt)
1541 || (ntype == import_stmt)
1542 || (ntype == global_stmt)
1543 || (ntype == assert_stmt)
1544 || (ntype == exec_stmt))
1545 res = validate_node(CHILD(tree, 0));
1546 else {
1547 res = 0;
1548 err_string("illegal small_stmt child type");
1551 else if (nch == 1) {
1552 res = 0;
1553 PyErr_Format(parser_error,
1554 "Unrecognized child node of small_stmt: %d.",
1555 TYPE(CHILD(tree, 0)));
1557 return (res);
1561 /* compound_stmt:
1562 * if_stmt | while_stmt | for_stmt | try_stmt | with_stmt | funcdef | classdef | decorated
1564 static int
1565 validate_compound_stmt(node *tree)
1567 int res = (validate_ntype(tree, compound_stmt)
1568 && validate_numnodes(tree, 1, "compound_stmt"));
1569 int ntype;
1571 if (!res)
1572 return (0);
1574 tree = CHILD(tree, 0);
1575 ntype = TYPE(tree);
1576 if ( (ntype == if_stmt)
1577 || (ntype == while_stmt)
1578 || (ntype == for_stmt)
1579 || (ntype == try_stmt)
1580 || (ntype == with_stmt)
1581 || (ntype == funcdef)
1582 || (ntype == classdef)
1583 || (ntype == decorated))
1584 res = validate_node(tree);
1585 else {
1586 res = 0;
1587 PyErr_Format(parser_error,
1588 "Illegal compound statement type: %d.", TYPE(tree));
1590 return (res);
1593 static int
1594 validate_yield_or_testlist(node *tree)
1596 if (TYPE(tree) == yield_expr)
1597 return validate_yield_expr(tree);
1598 else
1599 return validate_testlist(tree);
1602 static int
1603 validate_expr_stmt(node *tree)
1605 int j;
1606 int nch = NCH(tree);
1607 int res = (validate_ntype(tree, expr_stmt)
1608 && is_odd(nch)
1609 && validate_testlist(CHILD(tree, 0)));
1611 if (res && nch == 3
1612 && TYPE(CHILD(tree, 1)) == augassign) {
1613 res = validate_numnodes(CHILD(tree, 1), 1, "augassign")
1614 && validate_yield_or_testlist(CHILD(tree, 2));
1616 if (res) {
1617 char *s = STR(CHILD(CHILD(tree, 1), 0));
1619 res = (strcmp(s, "+=") == 0
1620 || strcmp(s, "-=") == 0
1621 || strcmp(s, "*=") == 0
1622 || strcmp(s, "/=") == 0
1623 || strcmp(s, "//=") == 0
1624 || strcmp(s, "%=") == 0
1625 || strcmp(s, "&=") == 0
1626 || strcmp(s, "|=") == 0
1627 || strcmp(s, "^=") == 0
1628 || strcmp(s, "<<=") == 0
1629 || strcmp(s, ">>=") == 0
1630 || strcmp(s, "**=") == 0);
1631 if (!res)
1632 err_string("illegal augmmented assignment operator");
1635 else {
1636 for (j = 1; res && (j < nch); j += 2)
1637 res = validate_equal(CHILD(tree, j))
1638 && validate_yield_or_testlist(CHILD(tree, j + 1));
1640 return (res);
1644 /* print_stmt:
1646 * 'print' ( [ test (',' test)* [','] ]
1647 * | '>>' test [ (',' test)+ [','] ] )
1649 static int
1650 validate_print_stmt(node *tree)
1652 int nch = NCH(tree);
1653 int res = (validate_ntype(tree, print_stmt)
1654 && (nch > 0)
1655 && validate_name(CHILD(tree, 0), "print"));
1657 if (res && nch > 1) {
1658 int sym = TYPE(CHILD(tree, 1));
1659 int i = 1;
1660 int allow_trailing_comma = 1;
1662 if (sym == test)
1663 res = validate_test(CHILD(tree, i++));
1664 else {
1665 if (nch < 3)
1666 res = validate_numnodes(tree, 3, "print_stmt");
1667 else {
1668 res = (validate_ntype(CHILD(tree, i), RIGHTSHIFT)
1669 && validate_test(CHILD(tree, i+1)));
1670 i += 2;
1671 allow_trailing_comma = 0;
1674 if (res) {
1675 /* ... (',' test)* [','] */
1676 while (res && i+2 <= nch) {
1677 res = (validate_comma(CHILD(tree, i))
1678 && validate_test(CHILD(tree, i+1)));
1679 allow_trailing_comma = 1;
1680 i += 2;
1682 if (res && !allow_trailing_comma)
1683 res = validate_numnodes(tree, i, "print_stmt");
1684 else if (res && i < nch)
1685 res = validate_comma(CHILD(tree, i));
1688 return (res);
1692 static int
1693 validate_del_stmt(node *tree)
1695 return (validate_numnodes(tree, 2, "del_stmt")
1696 && validate_name(CHILD(tree, 0), "del")
1697 && validate_exprlist(CHILD(tree, 1)));
1701 static int
1702 validate_return_stmt(node *tree)
1704 int nch = NCH(tree);
1705 int res = (validate_ntype(tree, return_stmt)
1706 && ((nch == 1) || (nch == 2))
1707 && validate_name(CHILD(tree, 0), "return"));
1709 if (res && (nch == 2))
1710 res = validate_testlist(CHILD(tree, 1));
1712 return (res);
1716 static int
1717 validate_raise_stmt(node *tree)
1719 int nch = NCH(tree);
1720 int res = (validate_ntype(tree, raise_stmt)
1721 && ((nch == 1) || (nch == 2) || (nch == 4) || (nch == 6)));
1723 if (res) {
1724 res = validate_name(CHILD(tree, 0), "raise");
1725 if (res && (nch >= 2))
1726 res = validate_test(CHILD(tree, 1));
1727 if (res && nch > 2) {
1728 res = (validate_comma(CHILD(tree, 2))
1729 && validate_test(CHILD(tree, 3)));
1730 if (res && (nch > 4))
1731 res = (validate_comma(CHILD(tree, 4))
1732 && validate_test(CHILD(tree, 5)));
1735 else
1736 (void) validate_numnodes(tree, 2, "raise");
1737 if (res && (nch == 4))
1738 res = (validate_comma(CHILD(tree, 2))
1739 && validate_test(CHILD(tree, 3)));
1741 return (res);
1745 /* yield_expr: 'yield' [testlist]
1747 static int
1748 validate_yield_expr(node *tree)
1750 int nch = NCH(tree);
1751 int res = (validate_ntype(tree, yield_expr)
1752 && ((nch == 1) || (nch == 2))
1753 && validate_name(CHILD(tree, 0), "yield"));
1755 if (res && (nch == 2))
1756 res = validate_testlist(CHILD(tree, 1));
1758 return (res);
1762 /* yield_stmt: yield_expr
1764 static int
1765 validate_yield_stmt(node *tree)
1767 return (validate_ntype(tree, yield_stmt)
1768 && validate_numnodes(tree, 1, "yield_stmt")
1769 && validate_yield_expr(CHILD(tree, 0)));
1773 static int
1774 validate_import_as_name(node *tree)
1776 int nch = NCH(tree);
1777 int ok = validate_ntype(tree, import_as_name);
1779 if (ok) {
1780 if (nch == 1)
1781 ok = validate_name(CHILD(tree, 0), NULL);
1782 else if (nch == 3)
1783 ok = (validate_name(CHILD(tree, 0), NULL)
1784 && validate_name(CHILD(tree, 1), "as")
1785 && validate_name(CHILD(tree, 2), NULL));
1786 else
1787 ok = validate_numnodes(tree, 3, "import_as_name");
1789 return ok;
1793 /* dotted_name: NAME ("." NAME)*
1795 static int
1796 validate_dotted_name(node *tree)
1798 int nch = NCH(tree);
1799 int res = (validate_ntype(tree, dotted_name)
1800 && is_odd(nch)
1801 && validate_name(CHILD(tree, 0), NULL));
1802 int i;
1804 for (i = 1; res && (i < nch); i += 2) {
1805 res = (validate_dot(CHILD(tree, i))
1806 && validate_name(CHILD(tree, i+1), NULL));
1808 return res;
1812 /* dotted_as_name: dotted_name [NAME NAME]
1814 static int
1815 validate_dotted_as_name(node *tree)
1817 int nch = NCH(tree);
1818 int res = validate_ntype(tree, dotted_as_name);
1820 if (res) {
1821 if (nch == 1)
1822 res = validate_dotted_name(CHILD(tree, 0));
1823 else if (nch == 3)
1824 res = (validate_dotted_name(CHILD(tree, 0))
1825 && validate_name(CHILD(tree, 1), "as")
1826 && validate_name(CHILD(tree, 2), NULL));
1827 else {
1828 res = 0;
1829 err_string("illegal number of children for dotted_as_name");
1832 return res;
1836 /* dotted_as_name (',' dotted_as_name)* */
1837 static int
1838 validate_dotted_as_names(node *tree)
1840 int nch = NCH(tree);
1841 int res = is_odd(nch) && validate_dotted_as_name(CHILD(tree, 0));
1842 int i;
1844 for (i = 1; res && (i < nch); i += 2)
1845 res = (validate_comma(CHILD(tree, i))
1846 && validate_dotted_as_name(CHILD(tree, i + 1)));
1847 return (res);
1851 /* import_as_name (',' import_as_name)* [','] */
1852 static int
1853 validate_import_as_names(node *tree)
1855 int nch = NCH(tree);
1856 int res = validate_import_as_name(CHILD(tree, 0));
1857 int i;
1859 for (i = 1; res && (i + 1 < nch); i += 2)
1860 res = (validate_comma(CHILD(tree, i))
1861 && validate_import_as_name(CHILD(tree, i + 1)));
1862 return (res);
1866 /* 'import' dotted_as_names */
1867 static int
1868 validate_import_name(node *tree)
1870 return (validate_ntype(tree, import_name)
1871 && validate_numnodes(tree, 2, "import_name")
1872 && validate_name(CHILD(tree, 0), "import")
1873 && validate_dotted_as_names(CHILD(tree, 1)));
1876 /* Helper function to count the number of leading dots in
1877 * 'from ...module import name'
1879 static int
1880 count_from_dots(node *tree)
1882 int i;
1883 for (i = 1; i < NCH(tree); i++)
1884 if (TYPE(CHILD(tree, i)) != DOT)
1885 break;
1886 return i-1;
1889 /* 'from' ('.'* dotted_name | '.') 'import' ('*' | '(' import_as_names ')' |
1890 * import_as_names
1892 static int
1893 validate_import_from(node *tree)
1895 int nch = NCH(tree);
1896 int ndots = count_from_dots(tree);
1897 int havename = (TYPE(CHILD(tree, ndots + 1)) == dotted_name);
1898 int offset = ndots + havename;
1899 int res = validate_ntype(tree, import_from)
1900 && (nch >= 4 + ndots)
1901 && validate_name(CHILD(tree, 0), "from")
1902 && (!havename || validate_dotted_name(CHILD(tree, ndots + 1)))
1903 && validate_name(CHILD(tree, offset + 1), "import");
1905 if (res && TYPE(CHILD(tree, offset + 2)) == LPAR)
1906 res = ((nch == offset + 5)
1907 && validate_lparen(CHILD(tree, offset + 2))
1908 && validate_import_as_names(CHILD(tree, offset + 3))
1909 && validate_rparen(CHILD(tree, offset + 4)));
1910 else if (res && TYPE(CHILD(tree, offset + 2)) != STAR)
1911 res = validate_import_as_names(CHILD(tree, offset + 2));
1912 return (res);
1916 /* import_stmt: import_name | import_from */
1917 static int
1918 validate_import_stmt(node *tree)
1920 int nch = NCH(tree);
1921 int res = validate_numnodes(tree, 1, "import_stmt");
1923 if (res) {
1924 int ntype = TYPE(CHILD(tree, 0));
1926 if (ntype == import_name || ntype == import_from)
1927 res = validate_node(CHILD(tree, 0));
1928 else {
1929 res = 0;
1930 err_string("illegal import_stmt child type");
1933 else if (nch == 1) {
1934 res = 0;
1935 PyErr_Format(parser_error,
1936 "Unrecognized child node of import_stmt: %d.",
1937 TYPE(CHILD(tree, 0)));
1939 return (res);
1945 static int
1946 validate_global_stmt(node *tree)
1948 int j;
1949 int nch = NCH(tree);
1950 int res = (validate_ntype(tree, global_stmt)
1951 && is_even(nch) && (nch >= 2));
1953 if (!res && !PyErr_Occurred())
1954 err_string("illegal global statement");
1956 if (res)
1957 res = (validate_name(CHILD(tree, 0), "global")
1958 && validate_ntype(CHILD(tree, 1), NAME));
1959 for (j = 2; res && (j < nch); j += 2)
1960 res = (validate_comma(CHILD(tree, j))
1961 && validate_ntype(CHILD(tree, j + 1), NAME));
1963 return (res);
1967 /* exec_stmt:
1969 * 'exec' expr ['in' test [',' test]]
1971 static int
1972 validate_exec_stmt(node *tree)
1974 int nch = NCH(tree);
1975 int res = (validate_ntype(tree, exec_stmt)
1976 && ((nch == 2) || (nch == 4) || (nch == 6))
1977 && validate_name(CHILD(tree, 0), "exec")
1978 && validate_expr(CHILD(tree, 1)));
1980 if (!res && !PyErr_Occurred())
1981 err_string("illegal exec statement");
1982 if (res && (nch > 2))
1983 res = (validate_name(CHILD(tree, 2), "in")
1984 && validate_test(CHILD(tree, 3)));
1985 if (res && (nch == 6))
1986 res = (validate_comma(CHILD(tree, 4))
1987 && validate_test(CHILD(tree, 5)));
1989 return (res);
1993 /* assert_stmt:
1995 * 'assert' test [',' test]
1997 static int
1998 validate_assert_stmt(node *tree)
2000 int nch = NCH(tree);
2001 int res = (validate_ntype(tree, assert_stmt)
2002 && ((nch == 2) || (nch == 4))
2003 && (validate_name(CHILD(tree, 0), "assert"))
2004 && validate_test(CHILD(tree, 1)));
2006 if (!res && !PyErr_Occurred())
2007 err_string("illegal assert statement");
2008 if (res && (nch > 2))
2009 res = (validate_comma(CHILD(tree, 2))
2010 && validate_test(CHILD(tree, 3)));
2012 return (res);
2016 static int
2017 validate_while(node *tree)
2019 int nch = NCH(tree);
2020 int res = (validate_ntype(tree, while_stmt)
2021 && ((nch == 4) || (nch == 7))
2022 && validate_name(CHILD(tree, 0), "while")
2023 && validate_test(CHILD(tree, 1))
2024 && validate_colon(CHILD(tree, 2))
2025 && validate_suite(CHILD(tree, 3)));
2027 if (res && (nch == 7))
2028 res = (validate_name(CHILD(tree, 4), "else")
2029 && validate_colon(CHILD(tree, 5))
2030 && validate_suite(CHILD(tree, 6)));
2032 return (res);
2036 static int
2037 validate_for(node *tree)
2039 int nch = NCH(tree);
2040 int res = (validate_ntype(tree, for_stmt)
2041 && ((nch == 6) || (nch == 9))
2042 && validate_name(CHILD(tree, 0), "for")
2043 && validate_exprlist(CHILD(tree, 1))
2044 && validate_name(CHILD(tree, 2), "in")
2045 && validate_testlist(CHILD(tree, 3))
2046 && validate_colon(CHILD(tree, 4))
2047 && validate_suite(CHILD(tree, 5)));
2049 if (res && (nch == 9))
2050 res = (validate_name(CHILD(tree, 6), "else")
2051 && validate_colon(CHILD(tree, 7))
2052 && validate_suite(CHILD(tree, 8)));
2054 return (res);
2058 /* try_stmt:
2059 * 'try' ':' suite (except_clause ':' suite)+ ['else' ':' suite]
2060 ['finally' ':' suite]
2061 * | 'try' ':' suite 'finally' ':' suite
2064 static int
2065 validate_try(node *tree)
2067 int nch = NCH(tree);
2068 int pos = 3;
2069 int res = (validate_ntype(tree, try_stmt)
2070 && (nch >= 6) && ((nch % 3) == 0));
2072 if (res)
2073 res = (validate_name(CHILD(tree, 0), "try")
2074 && validate_colon(CHILD(tree, 1))
2075 && validate_suite(CHILD(tree, 2))
2076 && validate_colon(CHILD(tree, nch - 2))
2077 && validate_suite(CHILD(tree, nch - 1)));
2078 else if (!PyErr_Occurred()) {
2079 const char* name = "except";
2080 if (TYPE(CHILD(tree, nch - 3)) != except_clause)
2081 name = STR(CHILD(tree, nch - 3));
2083 PyErr_Format(parser_error,
2084 "Illegal number of children for try/%s node.", name);
2086 /* Handle try/finally statement */
2087 if (res && (TYPE(CHILD(tree, pos)) == NAME) &&
2088 (strcmp(STR(CHILD(tree, pos)), "finally") == 0)) {
2089 res = (validate_numnodes(tree, 6, "try/finally")
2090 && validate_colon(CHILD(tree, 4))
2091 && validate_suite(CHILD(tree, 5)));
2092 return (res);
2094 /* try/except statement: skip past except_clause sections */
2095 while (res && pos < nch && (TYPE(CHILD(tree, pos)) == except_clause)) {
2096 res = (validate_except_clause(CHILD(tree, pos))
2097 && validate_colon(CHILD(tree, pos + 1))
2098 && validate_suite(CHILD(tree, pos + 2)));
2099 pos += 3;
2101 /* skip else clause */
2102 if (res && pos < nch && (TYPE(CHILD(tree, pos)) == NAME) &&
2103 (strcmp(STR(CHILD(tree, pos)), "else") == 0)) {
2104 res = (validate_colon(CHILD(tree, pos + 1))
2105 && validate_suite(CHILD(tree, pos + 2)));
2106 pos += 3;
2108 if (res && pos < nch) {
2109 /* last clause must be a finally */
2110 res = (validate_name(CHILD(tree, pos), "finally")
2111 && validate_numnodes(tree, pos + 3, "try/except/finally")
2112 && validate_colon(CHILD(tree, pos + 1))
2113 && validate_suite(CHILD(tree, pos + 2)));
2115 return (res);
2119 static int
2120 validate_except_clause(node *tree)
2122 int nch = NCH(tree);
2123 int res = (validate_ntype(tree, except_clause)
2124 && ((nch == 1) || (nch == 2) || (nch == 4))
2125 && validate_name(CHILD(tree, 0), "except"));
2127 if (res && (nch > 1))
2128 res = validate_test(CHILD(tree, 1));
2129 if (res && (nch == 4)) {
2130 if (TYPE(CHILD(tree, 2)) == NAME)
2131 res = validate_name(CHILD(tree, 2), "as");
2132 else
2133 res = validate_comma(CHILD(tree, 2));
2134 res = res && validate_test(CHILD(tree, 3));
2136 return (res);
2140 static int
2141 validate_test(node *tree)
2143 int nch = NCH(tree);
2144 int res = validate_ntype(tree, test) && is_odd(nch);
2146 if (res && (TYPE(CHILD(tree, 0)) == lambdef))
2147 res = ((nch == 1)
2148 && validate_lambdef(CHILD(tree, 0)));
2149 else if (res) {
2150 res = validate_or_test(CHILD(tree, 0));
2151 res = (res && (nch == 1 || (nch == 5 &&
2152 validate_name(CHILD(tree, 1), "if") &&
2153 validate_or_test(CHILD(tree, 2)) &&
2154 validate_name(CHILD(tree, 3), "else") &&
2155 validate_test(CHILD(tree, 4)))));
2157 return (res);
2160 static int
2161 validate_old_test(node *tree)
2163 int nch = NCH(tree);
2164 int res = validate_ntype(tree, old_test) && (nch == 1);
2166 if (res && (TYPE(CHILD(tree, 0)) == old_lambdef))
2167 res = (validate_old_lambdef(CHILD(tree, 0)));
2168 else if (res) {
2169 res = (validate_or_test(CHILD(tree, 0)));
2171 return (res);
2174 static int
2175 validate_or_test(node *tree)
2177 int nch = NCH(tree);
2178 int res = validate_ntype(tree, or_test) && is_odd(nch);
2180 if (res) {
2181 int pos;
2182 res = validate_and_test(CHILD(tree, 0));
2183 for (pos = 1; res && (pos < nch); pos += 2)
2184 res = (validate_name(CHILD(tree, pos), "or")
2185 && validate_and_test(CHILD(tree, pos + 1)));
2187 return (res);
2191 static int
2192 validate_and_test(node *tree)
2194 int pos;
2195 int nch = NCH(tree);
2196 int res = (validate_ntype(tree, and_test)
2197 && is_odd(nch)
2198 && validate_not_test(CHILD(tree, 0)));
2200 for (pos = 1; res && (pos < nch); pos += 2)
2201 res = (validate_name(CHILD(tree, pos), "and")
2202 && validate_not_test(CHILD(tree, 0)));
2204 return (res);
2208 static int
2209 validate_not_test(node *tree)
2211 int nch = NCH(tree);
2212 int res = validate_ntype(tree, not_test) && ((nch == 1) || (nch == 2));
2214 if (res) {
2215 if (nch == 2)
2216 res = (validate_name(CHILD(tree, 0), "not")
2217 && validate_not_test(CHILD(tree, 1)));
2218 else if (nch == 1)
2219 res = validate_comparison(CHILD(tree, 0));
2221 return (res);
2225 static int
2226 validate_comparison(node *tree)
2228 int pos;
2229 int nch = NCH(tree);
2230 int res = (validate_ntype(tree, comparison)
2231 && is_odd(nch)
2232 && validate_expr(CHILD(tree, 0)));
2234 for (pos = 1; res && (pos < nch); pos += 2)
2235 res = (validate_comp_op(CHILD(tree, pos))
2236 && validate_expr(CHILD(tree, pos + 1)));
2238 return (res);
2242 static int
2243 validate_comp_op(node *tree)
2245 int res = 0;
2246 int nch = NCH(tree);
2248 if (!validate_ntype(tree, comp_op))
2249 return (0);
2250 if (nch == 1) {
2252 * Only child will be a terminal with a well-defined symbolic name
2253 * or a NAME with a string of either 'is' or 'in'
2255 tree = CHILD(tree, 0);
2256 switch (TYPE(tree)) {
2257 case LESS:
2258 case GREATER:
2259 case EQEQUAL:
2260 case EQUAL:
2261 case LESSEQUAL:
2262 case GREATEREQUAL:
2263 case NOTEQUAL:
2264 res = 1;
2265 break;
2266 case NAME:
2267 res = ((strcmp(STR(tree), "in") == 0)
2268 || (strcmp(STR(tree), "is") == 0));
2269 if (!res) {
2270 PyErr_Format(parser_error,
2271 "illegal operator '%s'", STR(tree));
2273 break;
2274 default:
2275 err_string("illegal comparison operator type");
2276 break;
2279 else if ((res = validate_numnodes(tree, 2, "comp_op")) != 0) {
2280 res = (validate_ntype(CHILD(tree, 0), NAME)
2281 && validate_ntype(CHILD(tree, 1), NAME)
2282 && (((strcmp(STR(CHILD(tree, 0)), "is") == 0)
2283 && (strcmp(STR(CHILD(tree, 1)), "not") == 0))
2284 || ((strcmp(STR(CHILD(tree, 0)), "not") == 0)
2285 && (strcmp(STR(CHILD(tree, 1)), "in") == 0))));
2286 if (!res && !PyErr_Occurred())
2287 err_string("unknown comparison operator");
2289 return (res);
2293 static int
2294 validate_expr(node *tree)
2296 int j;
2297 int nch = NCH(tree);
2298 int res = (validate_ntype(tree, expr)
2299 && is_odd(nch)
2300 && validate_xor_expr(CHILD(tree, 0)));
2302 for (j = 2; res && (j < nch); j += 2)
2303 res = (validate_xor_expr(CHILD(tree, j))
2304 && validate_vbar(CHILD(tree, j - 1)));
2306 return (res);
2310 static int
2311 validate_xor_expr(node *tree)
2313 int j;
2314 int nch = NCH(tree);
2315 int res = (validate_ntype(tree, xor_expr)
2316 && is_odd(nch)
2317 && validate_and_expr(CHILD(tree, 0)));
2319 for (j = 2; res && (j < nch); j += 2)
2320 res = (validate_circumflex(CHILD(tree, j - 1))
2321 && validate_and_expr(CHILD(tree, j)));
2323 return (res);
2327 static int
2328 validate_and_expr(node *tree)
2330 int pos;
2331 int nch = NCH(tree);
2332 int res = (validate_ntype(tree, and_expr)
2333 && is_odd(nch)
2334 && validate_shift_expr(CHILD(tree, 0)));
2336 for (pos = 1; res && (pos < nch); pos += 2)
2337 res = (validate_ampersand(CHILD(tree, pos))
2338 && validate_shift_expr(CHILD(tree, pos + 1)));
2340 return (res);
2344 static int
2345 validate_chain_two_ops(node *tree, int (*termvalid)(node *), int op1, int op2)
2347 int pos = 1;
2348 int nch = NCH(tree);
2349 int res = (is_odd(nch)
2350 && (*termvalid)(CHILD(tree, 0)));
2352 for ( ; res && (pos < nch); pos += 2) {
2353 if (TYPE(CHILD(tree, pos)) != op1)
2354 res = validate_ntype(CHILD(tree, pos), op2);
2355 if (res)
2356 res = (*termvalid)(CHILD(tree, pos + 1));
2358 return (res);
2362 static int
2363 validate_shift_expr(node *tree)
2365 return (validate_ntype(tree, shift_expr)
2366 && validate_chain_two_ops(tree, validate_arith_expr,
2367 LEFTSHIFT, RIGHTSHIFT));
2371 static int
2372 validate_arith_expr(node *tree)
2374 return (validate_ntype(tree, arith_expr)
2375 && validate_chain_two_ops(tree, validate_term, PLUS, MINUS));
2379 static int
2380 validate_term(node *tree)
2382 int pos = 1;
2383 int nch = NCH(tree);
2384 int res = (validate_ntype(tree, term)
2385 && is_odd(nch)
2386 && validate_factor(CHILD(tree, 0)));
2388 for ( ; res && (pos < nch); pos += 2)
2389 res = (((TYPE(CHILD(tree, pos)) == STAR)
2390 || (TYPE(CHILD(tree, pos)) == SLASH)
2391 || (TYPE(CHILD(tree, pos)) == DOUBLESLASH)
2392 || (TYPE(CHILD(tree, pos)) == PERCENT))
2393 && validate_factor(CHILD(tree, pos + 1)));
2395 return (res);
2399 /* factor:
2401 * factor: ('+'|'-'|'~') factor | power
2403 static int
2404 validate_factor(node *tree)
2406 int nch = NCH(tree);
2407 int res = (validate_ntype(tree, factor)
2408 && (((nch == 2)
2409 && ((TYPE(CHILD(tree, 0)) == PLUS)
2410 || (TYPE(CHILD(tree, 0)) == MINUS)
2411 || (TYPE(CHILD(tree, 0)) == TILDE))
2412 && validate_factor(CHILD(tree, 1)))
2413 || ((nch == 1)
2414 && validate_power(CHILD(tree, 0)))));
2415 return (res);
2419 /* power:
2421 * power: atom trailer* ('**' factor)*
2423 static int
2424 validate_power(node *tree)
2426 int pos = 1;
2427 int nch = NCH(tree);
2428 int res = (validate_ntype(tree, power) && (nch >= 1)
2429 && validate_atom(CHILD(tree, 0)));
2431 while (res && (pos < nch) && (TYPE(CHILD(tree, pos)) == trailer))
2432 res = validate_trailer(CHILD(tree, pos++));
2433 if (res && (pos < nch)) {
2434 if (!is_even(nch - pos)) {
2435 err_string("illegal number of nodes for 'power'");
2436 return (0);
2438 for ( ; res && (pos < (nch - 1)); pos += 2)
2439 res = (validate_doublestar(CHILD(tree, pos))
2440 && validate_factor(CHILD(tree, pos + 1)));
2442 return (res);
2446 static int
2447 validate_atom(node *tree)
2449 int pos;
2450 int nch = NCH(tree);
2451 int res = validate_ntype(tree, atom);
2453 if (res && nch < 1)
2454 res = validate_numnodes(tree, nch+1, "atom");
2455 if (res) {
2456 switch (TYPE(CHILD(tree, 0))) {
2457 case LPAR:
2458 res = ((nch <= 3)
2459 && (validate_rparen(CHILD(tree, nch - 1))));
2461 if (res && (nch == 3)) {
2462 if (TYPE(CHILD(tree, 1))==yield_expr)
2463 res = validate_yield_expr(CHILD(tree, 1));
2464 else
2465 res = validate_testlist_comp(CHILD(tree, 1));
2467 break;
2468 case LSQB:
2469 if (nch == 2)
2470 res = validate_ntype(CHILD(tree, 1), RSQB);
2471 else if (nch == 3)
2472 res = (validate_listmaker(CHILD(tree, 1))
2473 && validate_ntype(CHILD(tree, 2), RSQB));
2474 else {
2475 res = 0;
2476 err_string("illegal list display atom");
2478 break;
2479 case LBRACE:
2480 res = ((nch <= 3)
2481 && validate_ntype(CHILD(tree, nch - 1), RBRACE));
2483 if (res && (nch == 3))
2484 res = validate_dictorsetmaker(CHILD(tree, 1));
2485 break;
2486 case BACKQUOTE:
2487 res = ((nch == 3)
2488 && validate_testlist1(CHILD(tree, 1))
2489 && validate_ntype(CHILD(tree, 2), BACKQUOTE));
2490 break;
2491 case NAME:
2492 case NUMBER:
2493 res = (nch == 1);
2494 break;
2495 case STRING:
2496 for (pos = 1; res && (pos < nch); ++pos)
2497 res = validate_ntype(CHILD(tree, pos), STRING);
2498 break;
2499 default:
2500 res = 0;
2501 break;
2504 return (res);
2508 /* listmaker:
2509 * test ( list_for | (',' test)* [','] )
2511 static int
2512 validate_listmaker(node *tree)
2514 int nch = NCH(tree);
2515 int ok = nch;
2517 if (nch == 0)
2518 err_string("missing child nodes of listmaker");
2519 else
2520 ok = validate_test(CHILD(tree, 0));
2523 * list_for | (',' test)* [',']
2525 if (nch == 2 && TYPE(CHILD(tree, 1)) == list_for)
2526 ok = validate_list_for(CHILD(tree, 1));
2527 else {
2528 /* (',' test)* [','] */
2529 int i = 1;
2530 while (ok && nch - i >= 2) {
2531 ok = (validate_comma(CHILD(tree, i))
2532 && validate_test(CHILD(tree, i+1)));
2533 i += 2;
2535 if (ok && i == nch-1)
2536 ok = validate_comma(CHILD(tree, i));
2537 else if (i != nch) {
2538 ok = 0;
2539 err_string("illegal trailing nodes for listmaker");
2542 return ok;
2545 /* testlist_comp:
2546 * test ( comp_for | (',' test)* [','] )
2548 static int
2549 validate_testlist_comp(node *tree)
2551 int nch = NCH(tree);
2552 int ok = nch;
2554 if (nch == 0)
2555 err_string("missing child nodes of testlist_comp");
2556 else {
2557 ok = validate_test(CHILD(tree, 0));
2561 * comp_for | (',' test)* [',']
2563 if (nch == 2 && TYPE(CHILD(tree, 1)) == comp_for)
2564 ok = validate_comp_for(CHILD(tree, 1));
2565 else {
2566 /* (',' test)* [','] */
2567 int i = 1;
2568 while (ok && nch - i >= 2) {
2569 ok = (validate_comma(CHILD(tree, i))
2570 && validate_test(CHILD(tree, i+1)));
2571 i += 2;
2573 if (ok && i == nch-1)
2574 ok = validate_comma(CHILD(tree, i));
2575 else if (i != nch) {
2576 ok = 0;
2577 err_string("illegal trailing nodes for testlist_comp");
2580 return ok;
2583 /* decorator:
2584 * '@' dotted_name [ '(' [arglist] ')' ] NEWLINE
2586 static int
2587 validate_decorator(node *tree)
2589 int ok;
2590 int nch = NCH(tree);
2591 ok = (validate_ntype(tree, decorator) &&
2592 (nch == 3 || nch == 5 || nch == 6) &&
2593 validate_at(CHILD(tree, 0)) &&
2594 validate_dotted_name(CHILD(tree, 1)) &&
2595 validate_newline(RCHILD(tree, -1)));
2597 if (ok && nch != 3) {
2598 ok = (validate_lparen(CHILD(tree, 2)) &&
2599 validate_rparen(RCHILD(tree, -2)));
2601 if (ok && nch == 6)
2602 ok = validate_arglist(CHILD(tree, 3));
2605 return ok;
2608 /* decorators:
2609 * decorator+
2611 static int
2612 validate_decorators(node *tree)
2614 int i, nch, ok;
2615 nch = NCH(tree);
2616 ok = validate_ntype(tree, decorators) && nch >= 1;
2618 for (i = 0; ok && i < nch; ++i)
2619 ok = validate_decorator(CHILD(tree, i));
2621 return ok;
2624 /* with_item:
2625 * test ['as' expr]
2627 static int
2628 validate_with_item(node *tree)
2630 int nch = NCH(tree);
2631 int ok = (validate_ntype(tree, with_item)
2632 && (nch == 1 || nch == 3)
2633 && validate_test(CHILD(tree, 0)));
2634 if (ok && nch == 3)
2635 ok = (validate_name(CHILD(tree, 1), "as")
2636 && validate_expr(CHILD(tree, 2)));
2637 return ok;
2640 /* with_stmt:
2641 * 0 1 ... -2 -1
2642 * 'with' with_item (',' with_item)* ':' suite
2644 static int
2645 validate_with_stmt(node *tree)
2647 int i;
2648 int nch = NCH(tree);
2649 int ok = (validate_ntype(tree, with_stmt)
2650 && (nch % 2 == 0)
2651 && validate_name(CHILD(tree, 0), "with")
2652 && validate_colon(RCHILD(tree, -2))
2653 && validate_suite(RCHILD(tree, -1)));
2654 for (i = 1; ok && i < nch - 2; i += 2)
2655 ok = validate_with_item(CHILD(tree, i));
2656 return ok;
2659 /* funcdef:
2661 * -5 -4 -3 -2 -1
2662 * 'def' NAME parameters ':' suite
2664 static int
2665 validate_funcdef(node *tree)
2667 int nch = NCH(tree);
2668 int ok = (validate_ntype(tree, funcdef)
2669 && (nch == 5)
2670 && validate_name(RCHILD(tree, -5), "def")
2671 && validate_ntype(RCHILD(tree, -4), NAME)
2672 && validate_colon(RCHILD(tree, -2))
2673 && validate_parameters(RCHILD(tree, -3))
2674 && validate_suite(RCHILD(tree, -1)));
2675 return ok;
2679 /* decorated
2680 * decorators (classdef | funcdef)
2682 static int
2683 validate_decorated(node *tree)
2685 int nch = NCH(tree);
2686 int ok = (validate_ntype(tree, decorated)
2687 && (nch == 2)
2688 && validate_decorators(RCHILD(tree, -2))
2689 && (validate_funcdef(RCHILD(tree, -1))
2690 || validate_class(RCHILD(tree, -1)))
2692 return ok;
2695 static int
2696 validate_lambdef(node *tree)
2698 int nch = NCH(tree);
2699 int res = (validate_ntype(tree, lambdef)
2700 && ((nch == 3) || (nch == 4))
2701 && validate_name(CHILD(tree, 0), "lambda")
2702 && validate_colon(CHILD(tree, nch - 2))
2703 && validate_test(CHILD(tree, nch - 1)));
2705 if (res && (nch == 4))
2706 res = validate_varargslist(CHILD(tree, 1));
2707 else if (!res && !PyErr_Occurred())
2708 (void) validate_numnodes(tree, 3, "lambdef");
2710 return (res);
2714 static int
2715 validate_old_lambdef(node *tree)
2717 int nch = NCH(tree);
2718 int res = (validate_ntype(tree, old_lambdef)
2719 && ((nch == 3) || (nch == 4))
2720 && validate_name(CHILD(tree, 0), "lambda")
2721 && validate_colon(CHILD(tree, nch - 2))
2722 && validate_test(CHILD(tree, nch - 1)));
2724 if (res && (nch == 4))
2725 res = validate_varargslist(CHILD(tree, 1));
2726 else if (!res && !PyErr_Occurred())
2727 (void) validate_numnodes(tree, 3, "old_lambdef");
2729 return (res);
2733 /* arglist:
2735 * (argument ',')* (argument [','] | '*' test [',' '**' test] | '**' test)
2737 static int
2738 validate_arglist(node *tree)
2740 int nch = NCH(tree);
2741 int i = 0;
2742 int ok = 1;
2744 if (nch <= 0)
2745 /* raise the right error from having an invalid number of children */
2746 return validate_numnodes(tree, nch + 1, "arglist");
2748 if (nch > 1) {
2749 for (i=0; i<nch; i++) {
2750 if (TYPE(CHILD(tree, i)) == argument) {
2751 node *ch = CHILD(tree, i);
2752 if (NCH(ch) == 2 && TYPE(CHILD(ch, 1)) == comp_for) {
2753 err_string("need '(', ')' for generator expression");
2754 return 0;
2760 while (ok && nch-i >= 2) {
2761 /* skip leading (argument ',') */
2762 ok = (validate_argument(CHILD(tree, i))
2763 && validate_comma(CHILD(tree, i+1)));
2764 if (ok)
2765 i += 2;
2766 else
2767 PyErr_Clear();
2769 ok = 1;
2770 if (nch-i > 0) {
2772 * argument | '*' test [',' '**' test] | '**' test
2774 int sym = TYPE(CHILD(tree, i));
2776 if (sym == argument) {
2777 ok = validate_argument(CHILD(tree, i));
2778 if (ok && i+1 != nch) {
2779 err_string("illegal arglist specification"
2780 " (extra stuff on end)");
2781 ok = 0;
2784 else if (sym == STAR) {
2785 ok = validate_star(CHILD(tree, i));
2786 if (ok && (nch-i == 2))
2787 ok = validate_test(CHILD(tree, i+1));
2788 else if (ok && (nch-i == 5))
2789 ok = (validate_test(CHILD(tree, i+1))
2790 && validate_comma(CHILD(tree, i+2))
2791 && validate_doublestar(CHILD(tree, i+3))
2792 && validate_test(CHILD(tree, i+4)));
2793 else {
2794 err_string("illegal use of '*' in arglist");
2795 ok = 0;
2798 else if (sym == DOUBLESTAR) {
2799 if (nch-i == 2)
2800 ok = (validate_doublestar(CHILD(tree, i))
2801 && validate_test(CHILD(tree, i+1)));
2802 else {
2803 err_string("illegal use of '**' in arglist");
2804 ok = 0;
2807 else {
2808 err_string("illegal arglist specification");
2809 ok = 0;
2812 return (ok);
2817 /* argument:
2819 * [test '='] test [comp_for]
2821 static int
2822 validate_argument(node *tree)
2824 int nch = NCH(tree);
2825 int res = (validate_ntype(tree, argument)
2826 && ((nch == 1) || (nch == 2) || (nch == 3))
2827 && validate_test(CHILD(tree, 0)));
2829 if (res && (nch == 2))
2830 res = validate_comp_for(CHILD(tree, 1));
2831 else if (res && (nch == 3))
2832 res = (validate_equal(CHILD(tree, 1))
2833 && validate_test(CHILD(tree, 2)));
2835 return (res);
2840 /* trailer:
2842 * '(' [arglist] ')' | '[' subscriptlist ']' | '.' NAME
2844 static int
2845 validate_trailer(node *tree)
2847 int nch = NCH(tree);
2848 int res = validate_ntype(tree, trailer) && ((nch == 2) || (nch == 3));
2850 if (res) {
2851 switch (TYPE(CHILD(tree, 0))) {
2852 case LPAR:
2853 res = validate_rparen(CHILD(tree, nch - 1));
2854 if (res && (nch == 3))
2855 res = validate_arglist(CHILD(tree, 1));
2856 break;
2857 case LSQB:
2858 res = (validate_numnodes(tree, 3, "trailer")
2859 && validate_subscriptlist(CHILD(tree, 1))
2860 && validate_ntype(CHILD(tree, 2), RSQB));
2861 break;
2862 case DOT:
2863 res = (validate_numnodes(tree, 2, "trailer")
2864 && validate_ntype(CHILD(tree, 1), NAME));
2865 break;
2866 default:
2867 res = 0;
2868 break;
2871 else {
2872 (void) validate_numnodes(tree, 2, "trailer");
2874 return (res);
2878 /* subscriptlist:
2880 * subscript (',' subscript)* [',']
2882 static int
2883 validate_subscriptlist(node *tree)
2885 return (validate_repeating_list(tree, subscriptlist,
2886 validate_subscript, "subscriptlist"));
2890 /* subscript:
2892 * '.' '.' '.' | test | [test] ':' [test] [sliceop]
2894 static int
2895 validate_subscript(node *tree)
2897 int offset = 0;
2898 int nch = NCH(tree);
2899 int res = validate_ntype(tree, subscript) && (nch >= 1) && (nch <= 4);
2901 if (!res) {
2902 if (!PyErr_Occurred())
2903 err_string("invalid number of arguments for subscript node");
2904 return (0);
2906 if (TYPE(CHILD(tree, 0)) == DOT)
2907 /* take care of ('.' '.' '.') possibility */
2908 return (validate_numnodes(tree, 3, "subscript")
2909 && validate_dot(CHILD(tree, 0))
2910 && validate_dot(CHILD(tree, 1))
2911 && validate_dot(CHILD(tree, 2)));
2912 if (nch == 1) {
2913 if (TYPE(CHILD(tree, 0)) == test)
2914 res = validate_test(CHILD(tree, 0));
2915 else
2916 res = validate_colon(CHILD(tree, 0));
2917 return (res);
2919 /* Must be [test] ':' [test] [sliceop],
2920 * but at least one of the optional components will
2921 * be present, but we don't know which yet.
2923 if ((TYPE(CHILD(tree, 0)) != COLON) || (nch == 4)) {
2924 res = validate_test(CHILD(tree, 0));
2925 offset = 1;
2927 if (res)
2928 res = validate_colon(CHILD(tree, offset));
2929 if (res) {
2930 int rem = nch - ++offset;
2931 if (rem) {
2932 if (TYPE(CHILD(tree, offset)) == test) {
2933 res = validate_test(CHILD(tree, offset));
2934 ++offset;
2935 --rem;
2937 if (res && rem)
2938 res = validate_sliceop(CHILD(tree, offset));
2941 return (res);
2945 static int
2946 validate_sliceop(node *tree)
2948 int nch = NCH(tree);
2949 int res = ((nch == 1) || validate_numnodes(tree, 2, "sliceop"))
2950 && validate_ntype(tree, sliceop);
2951 if (!res && !PyErr_Occurred()) {
2952 res = validate_numnodes(tree, 1, "sliceop");
2954 if (res)
2955 res = validate_colon(CHILD(tree, 0));
2956 if (res && (nch == 2))
2957 res = validate_test(CHILD(tree, 1));
2959 return (res);
2963 static int
2964 validate_exprlist(node *tree)
2966 return (validate_repeating_list(tree, exprlist,
2967 validate_expr, "exprlist"));
2972 * dictorsetmaker:
2974 * (test ':' test (comp_for | (',' test ':' test)* [','])) |
2975 * (test (comp_for | (',' test)* [',']))
2977 static int
2978 validate_dictorsetmaker(node *tree)
2980 int nch = NCH(tree);
2981 int ok = validate_ntype(tree, dictorsetmaker);
2982 int i = 0;
2983 int check_trailing_comma = 0;
2985 assert(nch > 0);
2987 if (ok && (nch == 1 || TYPE(CHILD(tree, 1)) == COMMA)) {
2988 /* We got a set:
2989 * test (',' test)* [',']
2991 ok = validate_test(CHILD(tree, i++));
2992 while (ok && nch - i >= 2) {
2993 ok = (validate_comma(CHILD(tree, i))
2994 && validate_test(CHILD(tree, i+1)));
2995 i += 2;
2997 check_trailing_comma = 1;
2999 else if (ok && TYPE(CHILD(tree, 1)) == comp_for) {
3000 /* We got a set comprehension:
3001 * test comp_for
3003 ok = (validate_test(CHILD(tree, 0))
3004 && validate_comp_for(CHILD(tree, 1)));
3006 else if (ok && NCH(tree) > 3 && TYPE(CHILD(tree, 3)) == comp_for) {
3007 /* We got a dict comprehension:
3008 * test ':' test comp_for
3010 ok = (validate_test(CHILD(tree, 0))
3011 && validate_colon(CHILD(tree, 1))
3012 && validate_test(CHILD(tree, 2))
3013 && validate_comp_for(CHILD(tree, 3)));
3015 else if (ok) {
3016 /* We got a dict:
3017 * test ':' test (',' test ':' test)* [',']
3019 if (nch >= 3) {
3020 ok = (validate_test(CHILD(tree, i))
3021 && validate_colon(CHILD(tree, i+1))
3022 && validate_test(CHILD(tree, i+2)));
3023 i += 3;
3025 else {
3026 ok = 0;
3027 err_string("illegal number of nodes for dictorsetmaker");
3030 while (ok && nch - i >= 4) {
3031 ok = (validate_comma(CHILD(tree, i))
3032 && validate_test(CHILD(tree, i+1))
3033 && validate_colon(CHILD(tree, i+2))
3034 && validate_test(CHILD(tree, i+3)));
3035 i += 4;
3037 check_trailing_comma = 1;
3039 if (ok && check_trailing_comma) {
3040 if (i == nch-1)
3041 ok = validate_comma(CHILD(tree, i));
3042 else if (i != nch) {
3043 ok = 0;
3044 err_string("illegal trailing nodes for dictorsetmaker");
3048 return ok;
3052 static int
3053 validate_eval_input(node *tree)
3055 int pos;
3056 int nch = NCH(tree);
3057 int res = (validate_ntype(tree, eval_input)
3058 && (nch >= 2)
3059 && validate_testlist(CHILD(tree, 0))
3060 && validate_ntype(CHILD(tree, nch - 1), ENDMARKER));
3062 for (pos = 1; res && (pos < (nch - 1)); ++pos)
3063 res = validate_ntype(CHILD(tree, pos), NEWLINE);
3065 return (res);
3069 static int
3070 validate_node(node *tree)
3072 int nch = 0; /* num. children on current node */
3073 int res = 1; /* result value */
3074 node* next = 0; /* node to process after this one */
3076 while (res && (tree != 0)) {
3077 nch = NCH(tree);
3078 next = 0;
3079 switch (TYPE(tree)) {
3081 * Definition nodes.
3083 case funcdef:
3084 res = validate_funcdef(tree);
3085 break;
3086 case with_stmt:
3087 res = validate_with_stmt(tree);
3088 break;
3089 case classdef:
3090 res = validate_class(tree);
3091 break;
3092 case decorated:
3093 res = validate_decorated(tree);
3094 break;
3096 * "Trivial" parse tree nodes.
3097 * (Why did I call these trivial?)
3099 case stmt:
3100 res = validate_stmt(tree);
3101 break;
3102 case small_stmt:
3104 * expr_stmt | print_stmt | del_stmt | pass_stmt | flow_stmt
3105 * | import_stmt | global_stmt | exec_stmt | assert_stmt
3107 res = validate_small_stmt(tree);
3108 break;
3109 case flow_stmt:
3110 res = (validate_numnodes(tree, 1, "flow_stmt")
3111 && ((TYPE(CHILD(tree, 0)) == break_stmt)
3112 || (TYPE(CHILD(tree, 0)) == continue_stmt)
3113 || (TYPE(CHILD(tree, 0)) == yield_stmt)
3114 || (TYPE(CHILD(tree, 0)) == return_stmt)
3115 || (TYPE(CHILD(tree, 0)) == raise_stmt)));
3116 if (res)
3117 next = CHILD(tree, 0);
3118 else if (nch == 1)
3119 err_string("illegal flow_stmt type");
3120 break;
3121 case yield_stmt:
3122 res = validate_yield_stmt(tree);
3123 break;
3125 * Compound statements.
3127 case simple_stmt:
3128 res = validate_simple_stmt(tree);
3129 break;
3130 case compound_stmt:
3131 res = validate_compound_stmt(tree);
3132 break;
3134 * Fundamental statements.
3136 case expr_stmt:
3137 res = validate_expr_stmt(tree);
3138 break;
3139 case print_stmt:
3140 res = validate_print_stmt(tree);
3141 break;
3142 case del_stmt:
3143 res = validate_del_stmt(tree);
3144 break;
3145 case pass_stmt:
3146 res = (validate_numnodes(tree, 1, "pass")
3147 && validate_name(CHILD(tree, 0), "pass"));
3148 break;
3149 case break_stmt:
3150 res = (validate_numnodes(tree, 1, "break")
3151 && validate_name(CHILD(tree, 0), "break"));
3152 break;
3153 case continue_stmt:
3154 res = (validate_numnodes(tree, 1, "continue")
3155 && validate_name(CHILD(tree, 0), "continue"));
3156 break;
3157 case return_stmt:
3158 res = validate_return_stmt(tree);
3159 break;
3160 case raise_stmt:
3161 res = validate_raise_stmt(tree);
3162 break;
3163 case import_stmt:
3164 res = validate_import_stmt(tree);
3165 break;
3166 case import_name:
3167 res = validate_import_name(tree);
3168 break;
3169 case import_from:
3170 res = validate_import_from(tree);
3171 break;
3172 case global_stmt:
3173 res = validate_global_stmt(tree);
3174 break;
3175 case exec_stmt:
3176 res = validate_exec_stmt(tree);
3177 break;
3178 case assert_stmt:
3179 res = validate_assert_stmt(tree);
3180 break;
3181 case if_stmt:
3182 res = validate_if(tree);
3183 break;
3184 case while_stmt:
3185 res = validate_while(tree);
3186 break;
3187 case for_stmt:
3188 res = validate_for(tree);
3189 break;
3190 case try_stmt:
3191 res = validate_try(tree);
3192 break;
3193 case suite:
3194 res = validate_suite(tree);
3195 break;
3197 * Expression nodes.
3199 case testlist:
3200 res = validate_testlist(tree);
3201 break;
3202 case yield_expr:
3203 res = validate_yield_expr(tree);
3204 break;
3205 case testlist1:
3206 res = validate_testlist1(tree);
3207 break;
3208 case test:
3209 res = validate_test(tree);
3210 break;
3211 case and_test:
3212 res = validate_and_test(tree);
3213 break;
3214 case not_test:
3215 res = validate_not_test(tree);
3216 break;
3217 case comparison:
3218 res = validate_comparison(tree);
3219 break;
3220 case exprlist:
3221 res = validate_exprlist(tree);
3222 break;
3223 case comp_op:
3224 res = validate_comp_op(tree);
3225 break;
3226 case expr:
3227 res = validate_expr(tree);
3228 break;
3229 case xor_expr:
3230 res = validate_xor_expr(tree);
3231 break;
3232 case and_expr:
3233 res = validate_and_expr(tree);
3234 break;
3235 case shift_expr:
3236 res = validate_shift_expr(tree);
3237 break;
3238 case arith_expr:
3239 res = validate_arith_expr(tree);
3240 break;
3241 case term:
3242 res = validate_term(tree);
3243 break;
3244 case factor:
3245 res = validate_factor(tree);
3246 break;
3247 case power:
3248 res = validate_power(tree);
3249 break;
3250 case atom:
3251 res = validate_atom(tree);
3252 break;
3254 default:
3255 /* Hopefully never reached! */
3256 err_string("unrecognized node type");
3257 res = 0;
3258 break;
3260 tree = next;
3262 return (res);
3266 static int
3267 validate_expr_tree(node *tree)
3269 int res = validate_eval_input(tree);
3271 if (!res && !PyErr_Occurred())
3272 err_string("could not validate expression tuple");
3274 return (res);
3278 /* file_input:
3279 * (NEWLINE | stmt)* ENDMARKER
3281 static int
3282 validate_file_input(node *tree)
3284 int j;
3285 int nch = NCH(tree) - 1;
3286 int res = ((nch >= 0)
3287 && validate_ntype(CHILD(tree, nch), ENDMARKER));
3289 for (j = 0; res && (j < nch); ++j) {
3290 if (TYPE(CHILD(tree, j)) == stmt)
3291 res = validate_stmt(CHILD(tree, j));
3292 else
3293 res = validate_newline(CHILD(tree, j));
3295 /* This stays in to prevent any internal failures from getting to the
3296 * user. Hopefully, this won't be needed. If a user reports getting
3297 * this, we have some debugging to do.
3299 if (!res && !PyErr_Occurred())
3300 err_string("VALIDATION FAILURE: report this to the maintainer!");
3302 return (res);
3305 static int
3306 validate_encoding_decl(node *tree)
3308 int nch = NCH(tree);
3309 int res = ((nch == 1)
3310 && validate_file_input(CHILD(tree, 0)));
3312 if (!res && !PyErr_Occurred())
3313 err_string("Error Parsing encoding_decl");
3315 return res;
3318 static PyObject*
3319 pickle_constructor = NULL;
3322 static PyObject*
3323 parser__pickler(PyObject *self, PyObject *args)
3325 NOTE(ARGUNUSED(self))
3326 PyObject *result = NULL;
3327 PyObject *st = NULL;
3328 PyObject *empty_dict = NULL;
3330 if (PyArg_ParseTuple(args, "O!:_pickler", &PyST_Type, &st)) {
3331 PyObject *newargs;
3332 PyObject *tuple;
3334 if ((empty_dict = PyDict_New()) == NULL)
3335 goto finally;
3336 if ((newargs = Py_BuildValue("Oi", st, 1)) == NULL)
3337 goto finally;
3338 tuple = parser_st2tuple((PyST_Object*)NULL, newargs, empty_dict);
3339 if (tuple != NULL) {
3340 result = Py_BuildValue("O(O)", pickle_constructor, tuple);
3341 Py_DECREF(tuple);
3343 Py_DECREF(empty_dict);
3344 Py_DECREF(newargs);
3346 finally:
3347 Py_XDECREF(empty_dict);
3349 return (result);
3353 /* Functions exported by this module. Most of this should probably
3354 * be converted into an ST object with methods, but that is better
3355 * done directly in Python, allowing subclasses to be created directly.
3356 * We'd really have to write a wrapper around it all anyway to allow
3357 * inheritance.
3359 static PyMethodDef parser_functions[] = {
3360 {"ast2tuple", (PyCFunction)parser_ast2tuple, PUBLIC_METHOD_TYPE,
3361 PyDoc_STR("Creates a tuple-tree representation of an ST.")},
3362 {"ast2list", (PyCFunction)parser_ast2list, PUBLIC_METHOD_TYPE,
3363 PyDoc_STR("Creates a list-tree representation of an ST.")},
3364 {"compileast", (PyCFunction)parser_compileast,PUBLIC_METHOD_TYPE,
3365 PyDoc_STR("Compiles an ST object into a code object.")},
3366 {"compilest", (PyCFunction)parser_compilest, PUBLIC_METHOD_TYPE,
3367 PyDoc_STR("Compiles an ST object into a code object.")},
3368 {"expr", (PyCFunction)parser_expr, PUBLIC_METHOD_TYPE,
3369 PyDoc_STR("Creates an ST object from an expression.")},
3370 {"isexpr", (PyCFunction)parser_isexpr, PUBLIC_METHOD_TYPE,
3371 PyDoc_STR("Determines if an ST object was created from an expression.")},
3372 {"issuite", (PyCFunction)parser_issuite, PUBLIC_METHOD_TYPE,
3373 PyDoc_STR("Determines if an ST object was created from a suite.")},
3374 {"suite", (PyCFunction)parser_suite, PUBLIC_METHOD_TYPE,
3375 PyDoc_STR("Creates an ST object from a suite.")},
3376 {"sequence2ast", (PyCFunction)parser_tuple2ast, PUBLIC_METHOD_TYPE,
3377 PyDoc_STR("Creates an ST object from a tree representation.")},
3378 {"sequence2st", (PyCFunction)parser_tuple2st, PUBLIC_METHOD_TYPE,
3379 PyDoc_STR("Creates an ST object from a tree representation.")},
3380 {"st2tuple", (PyCFunction)parser_st2tuple, PUBLIC_METHOD_TYPE,
3381 PyDoc_STR("Creates a tuple-tree representation of an ST.")},
3382 {"st2list", (PyCFunction)parser_st2list, PUBLIC_METHOD_TYPE,
3383 PyDoc_STR("Creates a list-tree representation of an ST.")},
3384 {"tuple2ast", (PyCFunction)parser_tuple2ast, PUBLIC_METHOD_TYPE,
3385 PyDoc_STR("Creates an ST object from a tree representation.")},
3386 {"tuple2st", (PyCFunction)parser_tuple2st, PUBLIC_METHOD_TYPE,
3387 PyDoc_STR("Creates an ST object from a tree representation.")},
3389 /* private stuff: support pickle module */
3390 {"_pickler", (PyCFunction)parser__pickler, METH_VARARGS,
3391 PyDoc_STR("Returns the pickle magic to allow ST objects to be pickled.")},
3393 {NULL, NULL, 0, NULL}
3397 PyMODINIT_FUNC initparser(void); /* supply a prototype */
3399 PyMODINIT_FUNC
3400 initparser(void)
3402 PyObject *module, *copyreg;
3404 Py_TYPE(&PyST_Type) = &PyType_Type;
3405 module = Py_InitModule("parser", parser_functions);
3406 if (module == NULL)
3407 return;
3409 if (parser_error == 0)
3410 parser_error = PyErr_NewException("parser.ParserError", NULL, NULL);
3412 if (parser_error == 0)
3413 /* caller will check PyErr_Occurred() */
3414 return;
3415 /* CAUTION: The code next used to skip bumping the refcount on
3416 * parser_error. That's a disaster if initparser() gets called more
3417 * than once. By incref'ing, we ensure that each module dict that
3418 * gets created owns its reference to the shared parser_error object,
3419 * and the file static parser_error vrbl owns a reference too.
3421 Py_INCREF(parser_error);
3422 if (PyModule_AddObject(module, "ParserError", parser_error) != 0)
3423 return;
3425 Py_INCREF(&PyST_Type);
3426 PyModule_AddObject(module, "ASTType", (PyObject*)&PyST_Type);
3427 Py_INCREF(&PyST_Type);
3428 PyModule_AddObject(module, "STType", (PyObject*)&PyST_Type);
3430 PyModule_AddStringConstant(module, "__copyright__",
3431 parser_copyright_string);
3432 PyModule_AddStringConstant(module, "__doc__",
3433 parser_doc_string);
3434 PyModule_AddStringConstant(module, "__version__",
3435 parser_version_string);
3437 /* Register to support pickling.
3438 * If this fails, the import of this module will fail because an
3439 * exception will be raised here; should we clear the exception?
3441 copyreg = PyImport_ImportModuleNoBlock("copy_reg");
3442 if (copyreg != NULL) {
3443 PyObject *func, *pickler;
3445 func = PyObject_GetAttrString(copyreg, "pickle");
3446 pickle_constructor = PyObject_GetAttrString(module, "sequence2st");
3447 pickler = PyObject_GetAttrString(module, "_pickler");
3448 Py_XINCREF(pickle_constructor);
3449 if ((func != NULL) && (pickle_constructor != NULL)
3450 && (pickler != NULL)) {
3451 PyObject *res;
3453 res = PyObject_CallFunctionObjArgs(func, &PyST_Type, pickler,
3454 pickle_constructor, NULL);
3455 Py_XDECREF(res);
3457 Py_XDECREF(func);
3458 Py_XDECREF(pickle_constructor);
3459 Py_XDECREF(pickler);
3460 Py_DECREF(copyreg);