mirrors
/
micropython
mirror of https://github.com/micropython/micropython.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity
micropython/py/parse.c

1183 lines
46 KiB
C
Raw Blame History

/*
* This file is part of the MicroPython project, http://micropython.org/
*
* The MIT License (MIT)
*
* Copyright (c) 2013-2017 Damien P. George
*
* Permission is hereby granted, free of charge, to any person obtaining a copy
* of this software and associated documentation files (the "Software"), to deal
* in the Software without restriction, including without limitation the rights
* to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
* copies of the Software, and to permit persons to whom the Software is
* furnished to do so, subject to the following conditions:
*
* The above copyright notice and this permission notice shall be included in
* all copies or substantial portions of the Software.
*
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
* IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
* FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
* AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
* LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
* OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
* THE SOFTWARE.
*/
#include <stdbool.h>
#include <stdint.h>
#include <stdio.h>
#include <unistd.h> // for ssize_t
#include <assert.h>
#include <string.h>
#include "py/lexer.h"
#include "py/parse.h"
#include "py/parsenum.h"
#include "py/runtime.h"
#include "py/objint.h"
#include "py/objstr.h"
#include "py/builtin.h"
#if MICROPY_ENABLE_COMPILER
#define RULE_ACT_ARG_MASK (0x0f)
#define RULE_ACT_KIND_MASK (0x30)
#define RULE_ACT_ALLOW_IDENT (0x40)
#define RULE_ACT_ADD_BLANK (0x80)
#define RULE_ACT_OR (0x10)
#define RULE_ACT_AND (0x20)
#define RULE_ACT_LIST (0x30)
#define RULE_ARG_KIND_MASK (0xf000)
#define RULE_ARG_ARG_MASK (0x0fff)
#define RULE_ARG_TOK (0x1000)
#define RULE_ARG_RULE (0x2000)
#define RULE_ARG_OPT_RULE (0x3000)
// (un)comment to use rule names; for debugging
//#define USE_RULE_NAME (1)
enum {
// define rules with a compile function
#define DEF_RULE(rule, comp, kind, ...) RULE_##rule,
#define DEF_RULE_NC(rule, kind, ...)
#include "py/grammar.h"
#undef DEF_RULE
#undef DEF_RULE_NC
RULE_const_object, // special node for a constant, generic Python object
// define rules without a compile function
#define DEF_RULE(rule, comp, kind, ...)
#define DEF_RULE_NC(rule, kind, ...) RULE_##rule,
#include "py/grammar.h"
#undef DEF_RULE
#undef DEF_RULE_NC
};
// Define an array of actions corresponding to each rule
STATIC const uint8_t rule_act_table[] = {
#define or(n) (RULE_ACT_OR | n)
#define and(n) (RULE_ACT_AND | n)
#define and_ident(n) (RULE_ACT_AND | n | RULE_ACT_ALLOW_IDENT)
#define and_blank(n) (RULE_ACT_AND | n | RULE_ACT_ADD_BLANK)
#define one_or_more (RULE_ACT_LIST | 2)
#define list (RULE_ACT_LIST | 1)
#define list_with_end (RULE_ACT_LIST | 3)
#define DEF_RULE(rule, comp, kind, ...) kind,
#define DEF_RULE_NC(rule, kind, ...)
#include "py/grammar.h"
#undef DEF_RULE
#undef DEF_RULE_NC
0, // RULE_const_object
#define DEF_RULE(rule, comp, kind, ...)
#define DEF_RULE_NC(rule, kind, ...) kind,
#include "py/grammar.h"
#undef DEF_RULE
#undef DEF_RULE_NC
#undef or
#undef and
#undef and_ident
#undef and_blank
#undef one_or_more
#undef list
#undef list_with_end
};
// Define the argument data for each rule, as a combined array
STATIC const uint16_t rule_arg_combined_table[] = {
#define tok(t) (RULE_ARG_TOK | MP_TOKEN_##t)
#define rule(r) (RULE_ARG_RULE | RULE_##r)
#define opt_rule(r) (RULE_ARG_OPT_RULE | RULE_##r)
#define DEF_RULE(rule, comp, kind, ...) __VA_ARGS__,
#define DEF_RULE_NC(rule, kind, ...)
#include "py/grammar.h"
#undef DEF_RULE
#undef DEF_RULE_NC
#define DEF_RULE(rule, comp, kind, ...)
#define DEF_RULE_NC(rule, kind, ...) __VA_ARGS__,
#include "py/grammar.h"
#undef DEF_RULE
#undef DEF_RULE_NC
#undef tok
#undef rule
#undef opt_rule
};
// Macro to create a list of N identifiers where N is the number of variable arguments to the macro
#define RULE_EXPAND(x) x
#define RULE_PADDING(rule, ...) RULE_PADDING2(rule, __VA_ARGS__, RULE_PADDING_IDS(rule))
#define RULE_PADDING2(rule, ...) RULE_EXPAND(RULE_PADDING3(rule, __VA_ARGS__))
#define RULE_PADDING3(rule, _1, _2, _3, _4, _5, _6, _7, _8, _9, _10, _11, _12, ...) __VA_ARGS__
#define RULE_PADDING_IDS(r) PAD12_##r, PAD11_##r, PAD10_##r, PAD9_##r, PAD8_##r, PAD7_##r, PAD6_##r, PAD5_##r, PAD4_##r, PAD3_##r, PAD2_##r, PAD1_##r,
// Use an enum to create constants specifying how much room a rule takes in rule_arg_combined_table
enum {
#define DEF_RULE(rule, comp, kind, ...) RULE_PADDING(rule, __VA_ARGS__)
#define DEF_RULE_NC(rule, kind, ...)
#include "py/grammar.h"
#undef DEF_RULE
#undef DEF_RULE_NC
#define DEF_RULE(rule, comp, kind, ...)
#define DEF_RULE_NC(rule, kind, ...) RULE_PADDING(rule, __VA_ARGS__)
#include "py/grammar.h"
#undef DEF_RULE
#undef DEF_RULE_NC
};
// Macro to compute the start of a rule in rule_arg_combined_table
#define RULE_ARG_OFFSET(rule, ...) RULE_ARG_OFFSET2(rule, __VA_ARGS__, RULE_ARG_OFFSET_IDS(rule))
#define RULE_ARG_OFFSET2(rule, ...) RULE_EXPAND(RULE_ARG_OFFSET3(rule, __VA_ARGS__))
#define RULE_ARG_OFFSET3(rule, _1, _2, _3, _4, _5, _6, _7, _8, _9, _10, _11, _12, _13, ...) _13
#define RULE_ARG_OFFSET_IDS(r) PAD12_##r, PAD11_##r, PAD10_##r, PAD9_##r, PAD8_##r, PAD7_##r, PAD6_##r, PAD5_##r, PAD4_##r, PAD3_##r, PAD2_##r, PAD1_##r, PAD0_##r,
// Use the above enum values to create a table of offsets for each rule's arg
// data, which indexes rule_arg_combined_table. The offsets require 9 bits of
// storage but only the lower 8 bits are stored here. The 9th bit is computed
// in get_rule_arg using the FIRST_RULE_WITH_OFFSET_ABOVE_255 constant.
STATIC const uint8_t rule_arg_offset_table[] = {
#define DEF_RULE(rule, comp, kind, ...) RULE_ARG_OFFSET(rule, __VA_ARGS__) & 0xff,
#define DEF_RULE_NC(rule, kind, ...)
#include "py/grammar.h"
#undef DEF_RULE
#undef DEF_RULE_NC
0, // RULE_const_object
#define DEF_RULE(rule, comp, kind, ...)
#define DEF_RULE_NC(rule, kind, ...) RULE_ARG_OFFSET(rule, __VA_ARGS__) & 0xff,
#include "py/grammar.h"
#undef DEF_RULE
#undef DEF_RULE_NC
};
// Define a constant that's used to determine the 9th bit of the values in rule_arg_offset_table
static const size_t FIRST_RULE_WITH_OFFSET_ABOVE_255 =
#define DEF_RULE(rule, comp, kind, ...) RULE_ARG_OFFSET(rule, __VA_ARGS__) >= 0x100 ? RULE_##rule :
#define DEF_RULE_NC(rule, kind, ...)
#include "py/grammar.h"
#undef DEF_RULE
#undef DEF_RULE_NC
#define DEF_RULE(rule, comp, kind, ...)
#define DEF_RULE_NC(rule, kind, ...) RULE_ARG_OFFSET(rule, __VA_ARGS__) >= 0x100 ? RULE_##rule :
#include "py/grammar.h"
#undef DEF_RULE
#undef DEF_RULE_NC
0;
#if USE_RULE_NAME
// Define an array of rule names corresponding to each rule
STATIC const char *const rule_name_table[] = {
#define DEF_RULE(rule, comp, kind, ...) #rule,
#define DEF_RULE_NC(rule, kind, ...)
#include "py/grammar.h"
#undef DEF_RULE
#undef DEF_RULE_NC
"", // RULE_const_object
#define DEF_RULE(rule, comp, kind, ...)
#define DEF_RULE_NC(rule, kind, ...) #rule,
#include "py/grammar.h"
#undef DEF_RULE
#undef DEF_RULE_NC
};
#endif
typedef struct _rule_stack_t {
size_t src_line : 8 * sizeof(size_t) - 8; // maximum bits storing source line number
size_t rule_id : 8; // this must be large enough to fit largest rule number
size_t arg_i; // this dictates the maximum nodes in a "list" of things
} rule_stack_t;
typedef struct _mp_parse_chunk_t {
size_t alloc;
union {
size_t used;
struct _mp_parse_chunk_t *next;
} union_;
byte data[];
} mp_parse_chunk_t;
typedef struct _parser_t {
size_t rule_stack_alloc;
size_t rule_stack_top;
rule_stack_t *rule_stack;
size_t result_stack_alloc;
size_t result_stack_top;
mp_parse_node_t *result_stack;
mp_lexer_t *lexer;
mp_parse_tree_t tree;
mp_parse_chunk_t *cur_chunk;
#if MICROPY_COMP_CONST
mp_map_t consts;
#endif
} parser_t;
STATIC const uint16_t *get_rule_arg(uint8_t r_id) {
size_t off = rule_arg_offset_table[r_id];
if (r_id >= FIRST_RULE_WITH_OFFSET_ABOVE_255) {
off |= 0x100;
}
return &rule_arg_combined_table[off];
}
STATIC void *parser_alloc(parser_t *parser, size_t num_bytes) {
// use a custom memory allocator to store parse nodes sequentially in large chunks
mp_parse_chunk_t *chunk = parser->cur_chunk;
if (chunk != NULL && chunk->union_.used + num_bytes > chunk->alloc) {
// not enough room at end of previously allocated chunk so try to grow
mp_parse_chunk_t *new_data = (mp_parse_chunk_t*)m_renew_maybe(byte, chunk,
sizeof(mp_parse_chunk_t) + chunk->alloc,
sizeof(mp_parse_chunk_t) + chunk->alloc + num_bytes, false);
if (new_data == NULL) {
// could not grow existing memory; shrink it to fit previous
(void)m_renew_maybe(byte, chunk, sizeof(mp_parse_chunk_t) + chunk->alloc,
sizeof(mp_parse_chunk_t) + chunk->union_.used, false);
chunk->alloc = chunk->union_.used;
chunk->union_.next = parser->tree.chunk;
parser->tree.chunk = chunk;
chunk = NULL;
} else {
// could grow existing memory
chunk->alloc += num_bytes;
}
}
if (chunk == NULL) {
// no previous chunk, allocate a new chunk
size_t alloc = MICROPY_ALLOC_PARSE_CHUNK_INIT;
if (alloc < num_bytes) {
alloc = num_bytes;
}
chunk = (mp_parse_chunk_t*)m_new(byte, sizeof(mp_parse_chunk_t) + alloc);
chunk->alloc = alloc;
chunk->union_.used = 0;
parser->cur_chunk = chunk;
}
byte *ret = chunk->data + chunk->union_.used;
chunk->union_.used += num_bytes;
return ret;
}
STATIC void push_rule(parser_t *parser, size_t src_line, uint8_t rule_id, size_t arg_i) {
if (parser->rule_stack_top >= parser->rule_stack_alloc) {
rule_stack_t *rs = m_renew(rule_stack_t, parser->rule_stack, parser->rule_stack_alloc, parser->rule_stack_alloc + MICROPY_ALLOC_PARSE_RULE_INC);
parser->rule_stack = rs;
parser->rule_stack_alloc += MICROPY_ALLOC_PARSE_RULE_INC;
}
rule_stack_t *rs = &parser->rule_stack[parser->rule_stack_top++];
rs->src_line = src_line;
rs->rule_id = rule_id;
rs->arg_i = arg_i;
}
STATIC void push_rule_from_arg(parser_t *parser, size_t arg) {
assert((arg & RULE_ARG_KIND_MASK) == RULE_ARG_RULE || (arg & RULE_ARG_KIND_MASK) == RULE_ARG_OPT_RULE);
size_t rule_id = arg & RULE_ARG_ARG_MASK;
push_rule(parser, parser->lexer->tok_line, rule_id, 0);
}
STATIC uint8_t pop_rule(parser_t *parser, size_t *arg_i, size_t *src_line) {
parser->rule_stack_top -= 1;
uint8_t rule_id = parser->rule_stack[parser->rule_stack_top].rule_id;
*arg_i = parser->rule_stack[parser->rule_stack_top].arg_i;
*src_line = parser->rule_stack[parser->rule_stack_top].src_line;
return rule_id;
}
bool mp_parse_node_is_const_false(mp_parse_node_t pn) {
return MP_PARSE_NODE_IS_TOKEN_KIND(pn, MP_TOKEN_KW_FALSE)
|| (MP_PARSE_NODE_IS_SMALL_INT(pn) && MP_PARSE_NODE_LEAF_SMALL_INT(pn) == 0);
}
bool mp_parse_node_is_const_true(mp_parse_node_t pn) {
return MP_PARSE_NODE_IS_TOKEN_KIND(pn, MP_TOKEN_KW_TRUE)
|| (MP_PARSE_NODE_IS_SMALL_INT(pn) && MP_PARSE_NODE_LEAF_SMALL_INT(pn) != 0);
}
bool mp_parse_node_get_int_maybe(mp_parse_node_t pn, mp_obj_t *o) {
if (MP_PARSE_NODE_IS_SMALL_INT(pn)) {
*o = MP_OBJ_NEW_SMALL_INT(MP_PARSE_NODE_LEAF_SMALL_INT(pn));
return true;
} else if (MP_PARSE_NODE_IS_STRUCT_KIND(pn, RULE_const_object)) {
mp_parse_node_struct_t *pns = (mp_parse_node_struct_t*)pn;
#if MICROPY_OBJ_REPR == MICROPY_OBJ_REPR_D
// nodes are 32-bit pointers, but need to extract 64-bit object
*o = (uint64_t)pns->nodes[0] | ((uint64_t)pns->nodes[1] << 32);
#else
*o = (mp_obj_t)pns->nodes[0];
#endif
return MP_OBJ_IS_INT(*o);
} else {
return false;
}
}
int mp_parse_node_extract_list(mp_parse_node_t *pn, size_t pn_kind, mp_parse_node_t **nodes) {
if (MP_PARSE_NODE_IS_NULL(*pn)) {
*nodes = NULL;
return 0;
} else if (MP_PARSE_NODE_IS_LEAF(*pn)) {
*nodes = pn;
return 1;
} else {
mp_parse_node_struct_t *pns = (mp_parse_node_struct_t*)(*pn);
if (MP_PARSE_NODE_STRUCT_KIND(pns) != pn_kind) {
*nodes = pn;
return 1;
} else {
*nodes = pns->nodes;
return MP_PARSE_NODE_STRUCT_NUM_NODES(pns);
}
}
}
#if MICROPY_DEBUG_PRINTERS
void mp_parse_node_print(mp_parse_node_t pn, size_t indent) {
if (MP_PARSE_NODE_IS_STRUCT(pn)) {
printf("[% 4d] ", (int)((mp_parse_node_struct_t*)pn)->source_line);
} else {
printf(" ");
}
for (size_t i = 0; i < indent; i++) {
printf(" ");
}
if (MP_PARSE_NODE_IS_NULL(pn)) {
printf("NULL\n");
} else if (MP_PARSE_NODE_IS_SMALL_INT(pn)) {
mp_int_t arg = MP_PARSE_NODE_LEAF_SMALL_INT(pn);
printf("int(" INT_FMT ")\n", arg);
} else if (MP_PARSE_NODE_IS_LEAF(pn)) {
uintptr_t arg = MP_PARSE_NODE_LEAF_ARG(pn);
switch (MP_PARSE_NODE_LEAF_KIND(pn)) {
case MP_PARSE_NODE_ID: printf("id(%s)\n", qstr_str(arg)); break;
case MP_PARSE_NODE_STRING: printf("str(%s)\n", qstr_str(arg)); break;
case MP_PARSE_NODE_BYTES: printf("bytes(%s)\n", qstr_str(arg)); break;
default:
assert(MP_PARSE_NODE_LEAF_KIND(pn) == MP_PARSE_NODE_TOKEN);
printf("tok(%u)\n", (uint)arg); break;
}
} else {
// node must be a mp_parse_node_struct_t
mp_parse_node_struct_t *pns = (mp_parse_node_struct_t*)pn;
if (MP_PARSE_NODE_STRUCT_KIND(pns) == RULE_const_object) {
#if MICROPY_OBJ_REPR == MICROPY_OBJ_REPR_D
printf("literal const(%016llx)\n", (uint64_t)pns->nodes[0] | ((uint64_t)pns->nodes[1] << 32));
#else
printf("literal const(%p)\n", (mp_obj_t)pns->nodes[0]);
#endif
} else {
size_t n = MP_PARSE_NODE_STRUCT_NUM_NODES(pns);
#if USE_RULE_NAME
printf("%s(%u) (n=%u)\n", rule_name_table[MP_PARSE_NODE_STRUCT_KIND(pns)], (uint)MP_PARSE_NODE_STRUCT_KIND(pns), (uint)n);
#else
printf("rule(%u) (n=%u)\n", (uint)MP_PARSE_NODE_STRUCT_KIND(pns), (uint)n);
#endif
for (size_t i = 0; i < n; i++) {
mp_parse_node_print(pns->nodes[i], indent + 2);
}
}
}
}
#endif // MICROPY_DEBUG_PRINTERS
/*
STATIC void result_stack_show(parser_t *parser) {
printf("result stack, most recent first\n");
for (ssize_t i = parser->result_stack_top - 1; i >= 0; i--) {
mp_parse_node_print(parser->result_stack[i], 0);
}
}
*/
STATIC mp_parse_node_t pop_result(parser_t *parser) {
assert(parser->result_stack_top > 0);
return parser->result_stack[--parser->result_stack_top];
}
STATIC mp_parse_node_t peek_result(parser_t *parser, size_t pos) {
assert(parser->result_stack_top > pos);
return parser->result_stack[parser->result_stack_top - 1 - pos];
}
STATIC void push_result_node(parser_t *parser, mp_parse_node_t pn) {
if (parser->result_stack_top >= parser->result_stack_alloc) {
mp_parse_node_t *stack = m_renew(mp_parse_node_t, parser->result_stack, parser->result_stack_alloc, parser->result_stack_alloc + MICROPY_ALLOC_PARSE_RESULT_INC);
parser->result_stack = stack;
parser->result_stack_alloc += MICROPY_ALLOC_PARSE_RESULT_INC;
}
parser->result_stack[parser->result_stack_top++] = pn;
}
STATIC mp_parse_node_t make_node_const_object(parser_t *parser, size_t src_line, mp_obj_t obj) {
mp_parse_node_struct_t *pn = parser_alloc(parser, sizeof(mp_parse_node_struct_t) + sizeof(mp_obj_t));
pn->source_line = src_line;
#if MICROPY_OBJ_REPR == MICROPY_OBJ_REPR_D
// nodes are 32-bit pointers, but need to store 64-bit object
pn->kind_num_nodes = RULE_const_object | (2 << 8);
pn->nodes[0] = (uint64_t)obj;
pn->nodes[1] = (uint64_t)obj >> 32;
#else
pn->kind_num_nodes = RULE_const_object | (1 << 8);
pn->nodes[0] = (uintptr_t)obj;
#endif
return (mp_parse_node_t)pn;
}
STATIC mp_parse_node_t mp_parse_node_new_small_int_checked(parser_t *parser, mp_obj_t o_val) {
(void)parser;
mp_int_t val = MP_OBJ_SMALL_INT_VALUE(o_val);
#if MICROPY_OBJ_REPR == MICROPY_OBJ_REPR_D
// A parse node is only 32-bits and the small-int value must fit in 31-bits
if (((val ^ (val << 1)) & 0xffffffff80000000) != 0) {
return make_node_const_object(parser, 0, o_val);
}
#endif
return mp_parse_node_new_small_int(val);
}
STATIC void push_result_token(parser_t *parser, uint8_t rule_id) {
mp_parse_node_t pn;
mp_lexer_t *lex = parser->lexer;
if (lex->tok_kind == MP_TOKEN_NAME) {
qstr id = qstr_from_strn(lex->vstr.buf, lex->vstr.len);
#if MICROPY_COMP_CONST
// if name is a standalone identifier, look it up in the table of dynamic constants
mp_map_elem_t *elem;
if (rule_id == RULE_atom
&& (elem = mp_map_lookup(&parser->consts, MP_OBJ_NEW_QSTR(id), MP_MAP_LOOKUP)) != NULL) {
if (MP_OBJ_IS_SMALL_INT(elem->value)) {
pn = mp_parse_node_new_small_int_checked(parser, elem->value);
} else {
pn = make_node_const_object(parser, lex->tok_line, elem->value);
}
} else {
pn = mp_parse_node_new_leaf(MP_PARSE_NODE_ID, id);
}
#else
(void)rule_id;
pn = mp_parse_node_new_leaf(MP_PARSE_NODE_ID, id);
#endif
} else if (lex->tok_kind == MP_TOKEN_INTEGER) {
mp_obj_t o = mp_parse_num_integer(lex->vstr.buf, lex->vstr.len, 0, lex);
if (MP_OBJ_IS_SMALL_INT(o)) {
pn = mp_parse_node_new_small_int_checked(parser, o);
} else {
pn = make_node_const_object(parser, lex->tok_line, o);
}
} else if (lex->tok_kind == MP_TOKEN_FLOAT_OR_IMAG) {
mp_obj_t o = mp_parse_num_decimal(lex->vstr.buf, lex->vstr.len, true, false, lex);
pn = make_node_const_object(parser, lex->tok_line, o);
} else if (lex->tok_kind == MP_TOKEN_STRING || lex->tok_kind == MP_TOKEN_BYTES) {
// Don't automatically intern all strings/bytes. doc strings (which are usually large)
// will be discarded by the compiler, and so we shouldn't intern them.
qstr qst = MP_QSTR_NULL;
if (lex->vstr.len <= MICROPY_ALLOC_PARSE_INTERN_STRING_LEN) {
// intern short strings
qst = qstr_from_strn(lex->vstr.buf, lex->vstr.len);
} else {
// check if this string is already interned
qst = qstr_find_strn(lex->vstr.buf, lex->vstr.len);
}
if (qst != MP_QSTR_NULL) {
// qstr exists, make a leaf node
pn = mp_parse_node_new_leaf(lex->tok_kind == MP_TOKEN_STRING ? MP_PARSE_NODE_STRING : MP_PARSE_NODE_BYTES, qst);
} else {
// not interned, make a node holding a pointer to the string/bytes object
mp_obj_t o = mp_obj_new_str_copy(
lex->tok_kind == MP_TOKEN_STRING ? &mp_type_str : &mp_type_bytes,
(const byte*)lex->vstr.buf, lex->vstr.len);
pn = make_node_const_object(parser, lex->tok_line, o);
}
} else {
pn = mp_parse_node_new_leaf(MP_PARSE_NODE_TOKEN, lex->tok_kind);
}
push_result_node(parser, pn);
}
#if MICROPY_COMP_MODULE_CONST
STATIC const mp_rom_map_elem_t mp_constants_table[] = {
#if MICROPY_PY_UERRNO
{ MP_ROM_QSTR(MP_QSTR_errno), MP_ROM_PTR(&mp_module_uerrno) },
#endif
#if MICROPY_PY_UCTYPES
{ MP_ROM_QSTR(MP_QSTR_uctypes), MP_ROM_PTR(&mp_module_uctypes) },
#endif
// Extra constants as defined by a port
MICROPY_PORT_CONSTANTS
};
STATIC MP_DEFINE_CONST_MAP(mp_constants_map, mp_constants_table);
#endif
STATIC void push_result_rule(parser_t *parser, size_t src_line, uint8_t rule_id, size_t num_args);
#if MICROPY_COMP_CONST_FOLDING
STATIC bool fold_logical_constants(parser_t *parser, uint8_t rule_id, size_t *num_args) {
if (rule_id == RULE_or_test
|| rule_id == RULE_and_test) {
// folding for binary logical ops: or and
size_t copy_to = *num_args;
for (size_t i = copy_to; i > 0;) {
mp_parse_node_t pn = peek_result(parser, --i);
parser->result_stack[parser->result_stack_top - copy_to] = pn;
if (i == 0) {
// always need to keep the last value
break;
}
if (rule_id == RULE_or_test) {
if (mp_parse_node_is_const_true(pn)) {
//
break;
} else if (!mp_parse_node_is_const_false(pn)) {
copy_to -= 1;
}
} else {
// RULE_and_test
if (mp_parse_node_is_const_false(pn)) {
break;
} else if (!mp_parse_node_is_const_true(pn)) {
copy_to -= 1;
}
}
}
copy_to -= 1; // copy_to now contains number of args to pop
// pop and discard all the short-circuited expressions
for (size_t i = 0; i < copy_to; ++i) {
pop_result(parser);
}
*num_args -= copy_to;
// we did a complete folding if there's only 1 arg left
return *num_args == 1;
} else if (rule_id == RULE_not_test_2) {
// folding for unary logical op: not
mp_parse_node_t pn = peek_result(parser, 0);
if (mp_parse_node_is_const_false(pn)) {
pn = mp_parse_node_new_leaf(MP_PARSE_NODE_TOKEN, MP_TOKEN_KW_TRUE);
} else if (mp_parse_node_is_const_true(pn)) {
pn = mp_parse_node_new_leaf(MP_PARSE_NODE_TOKEN, MP_TOKEN_KW_FALSE);
} else {
return false;
}
pop_result(parser);
push_result_node(parser, pn);
return true;
}
return false;
}
STATIC bool fold_constants(parser_t *parser, uint8_t rule_id, size_t num_args) {
// this code does folding of arbitrary integer expressions, eg 1 + 2 * 3 + 4
// it does not do partial folding, eg 1 + 2 + x -> 3 + x
mp_obj_t arg0;
if (rule_id == RULE_expr
|| rule_id == RULE_xor_expr
|| rule_id == RULE_and_expr) {
// folding for binary ops: | ^ &
mp_parse_node_t pn = peek_result(parser, num_args - 1);
if (!mp_parse_node_get_int_maybe(pn, &arg0)) {
return false;
}
mp_binary_op_t op;
if (rule_id == RULE_expr) {
op = MP_BINARY_OP_OR;
} else if (rule_id == RULE_xor_expr) {
op = MP_BINARY_OP_XOR;
} else {
op = MP_BINARY_OP_AND;
}
for (ssize_t i = num_args - 2; i >= 0; --i) {
pn = peek_result(parser, i);
mp_obj_t arg1;
if (!mp_parse_node_get_int_maybe(pn, &arg1)) {
return false;
}
arg0 = mp_binary_op(op, arg0, arg1);
}
} else if (rule_id == RULE_shift_expr
|| rule_id == RULE_arith_expr
|| rule_id == RULE_term) {
// folding for binary ops: << >> + - * / % //
mp_parse_node_t pn = peek_result(parser, num_args - 1);
if (!mp_parse_node_get_int_maybe(pn, &arg0)) {
return false;
}
for (ssize_t i = num_args - 2; i >= 1; i -= 2) {
pn = peek_result(parser, i - 1);
mp_obj_t arg1;
if (!mp_parse_node_get_int_maybe(pn, &arg1)) {
return false;
}
mp_token_kind_t tok = MP_PARSE_NODE_LEAF_ARG(peek_result(parser, i));
static const uint8_t token_to_op[] = {
MP_BINARY_OP_ADD,
MP_BINARY_OP_SUBTRACT,
MP_BINARY_OP_MULTIPLY,
255,//MP_BINARY_OP_POWER,
255,//MP_BINARY_OP_TRUE_DIVIDE,
MP_BINARY_OP_FLOOR_DIVIDE,
MP_BINARY_OP_MODULO,
255,//MP_BINARY_OP_LESS
MP_BINARY_OP_LSHIFT,
255,//MP_BINARY_OP_MORE
MP_BINARY_OP_RSHIFT,
};
mp_binary_op_t op = token_to_op[tok - MP_TOKEN_OP_PLUS];
if (op == (mp_binary_op_t)255) {
return false;
}
int rhs_sign = mp_obj_int_sign(arg1);
if (op <= MP_BINARY_OP_RSHIFT) {
// << and >> can't have negative rhs
if (rhs_sign < 0) {
return false;
}
} else if (op >= MP_BINARY_OP_FLOOR_DIVIDE) {
// % and // can't have zero rhs
if (rhs_sign == 0) {
return false;
}
}
arg0 = mp_binary_op(op, arg0, arg1);
}