STATIC mp_obj_t struct_unpack(mp_obj_t fmt_in, mp_obj_t data_in) {
// TODO: "The buffer must contain exactly the amount of data required by the format (len(bytes) must equal calcsize(fmt))."
const char *fmt = mp_obj_str_get_str(fmt_in);
char fmt_type = get_fmt_type(&fmt);
uint size = calcsize_items(fmt);
mp_obj_tuple_t *res = mp_obj_new_tuple(size, NULL);
mp_buffer_info_t bufinfo;
mp_get_buffer_raise(data_in, &bufinfo, MP_BUFFER_READ);
byte *p = bufinfo.buf;
for (uint i = 0; i < size; i++) {
machine_uint_t sz = 1;
if (unichar_isdigit(*fmt)) {
sz = get_fmt_num(&fmt);
}
if (sz > 1) {
// TODO: size spec support only for string len
assert(*fmt == 's');
}
mp_obj_t item;
if (*fmt == 's') {
item = mp_obj_new_bytes(p, sz);
p += sz;
fmt++;
} else {
item = mp_binary_get_val(fmt_type, *fmt++, &p);
}
res->items[i] = item;
}
return res;
}
STATIC mp_obj_t struct_calcsize(mp_obj_t fmt_in) {
const char *fmt = mp_obj_str_get_str(fmt_in);
char fmt_type = get_fmt_type(&fmt);
machine_uint_t size;
for (size = 0; *fmt; fmt++) {
uint align;
machine_uint_t cnt = 1;
if (unichar_isdigit(*fmt)) {
cnt = get_fmt_num(&fmt);
}
if (cnt > 1) {
// TODO: count spec support only for string len
assert(*fmt == 's');
}
machine_uint_t sz;
if (*fmt == 's') {
sz = cnt;
} else {
sz = (machine_uint_t)mp_binary_get_size(fmt_type, *fmt, &align);
}
// TODO
assert(sz != (machine_uint_t)-1);
// Apply alignment
size = (size + align - 1) & ~(align - 1);
size += sz;
}
return MP_OBJ_NEW_SMALL_INT(size);
}
STATIC uint calcsize_items(const char *fmt) {
uint cnt = 0;
while (*fmt) {
// TODO supports size spec only for "s"
if (!unichar_isdigit(*fmt++)) {
cnt++;
}
}
return cnt;
}
STATIC machine_uint_t get_fmt_num(const char **p) {
const char *num = *p;
uint len = 1;
while (unichar_isdigit(*++num)) {
len++;
}
machine_uint_t val = (machine_uint_t)MP_OBJ_SMALL_INT_VALUE(mp_parse_num_integer(*p, len, 10));
*p = num;
return val;
}
STATIC mp_obj_t struct_pack(uint n_args, mp_obj_t *args) {
// TODO: "The arguments must match the values required by the format exactly."
const char *fmt = mp_obj_str_get_str(args[0]);
char fmt_type = get_fmt_type(&fmt);
int size = MP_OBJ_SMALL_INT_VALUE(struct_calcsize(args[0]));
byte *p;
mp_obj_t res = mp_obj_str_builder_start(&mp_type_bytes, size, &p);
memset(p, 0, size);
for (uint i = 1; i < n_args; i++) {
machine_uint_t sz = 1;
if (unichar_isdigit(*fmt)) {
sz = get_fmt_num(&fmt);
}
if (sz > 1) {
// TODO: size spec support only for string len
assert(*fmt == 's');
}
if (*fmt == 's') {
mp_buffer_info_t bufinfo;
mp_get_buffer_raise(args[i], &bufinfo, MP_BUFFER_READ);
machine_uint_t to_copy = sz;
if (bufinfo.len < to_copy) {
to_copy = bufinfo.len;
}
memcpy(p, bufinfo.buf, to_copy);
memset(p + to_copy, 0, sz - to_copy);
p += sz;
fmt++;
} else {
mp_binary_set_val(fmt_type, *fmt++, args[i], &p);
}
}
return res;
}
int main(int argc, char **argv) {
/*
printf("sizeof(void*)=%u\n", (uint)sizeof(void*));
for (int i = 0; i < sizeof(table); ++i) {
byte *ptr = (void*)&table[0];
printf(" %02x", ptr[i]);
if ((i + 1)%8 == 0) printf("\n");
}
*/
mp_stack_set_limit(40000 * (BYTES_PER_WORD / 4));
pre_process_options(argc, argv);
#if MICROPY_ENABLE_GC
char *heap = malloc(heap_size);
gc_init(heap, heap + heap_size);
#endif
mp_init();
// create keyboard interrupt object
MP_STATE_VM(keyboard_interrupt_obj) = mp_obj_new_exception(&mp_type_KeyboardInterrupt);
char *home = getenv("HOME");
char *path = getenv("MICROPYPATH");
if (path == NULL) {
#ifdef MICROPY_PY_SYS_PATH_DEFAULT
path = MICROPY_PY_SYS_PATH_DEFAULT;
#else
path = "~/.micropython/lib:/usr/lib/micropython";
#endif
}
mp_uint_t path_num = 1; // [0] is for current dir (or base dir of the script)
for (char *p = path; p != NULL; p = strchr(p, PATHLIST_SEP_CHAR)) {
path_num++;
if (p != NULL) {
p++;
}
}
mp_obj_list_init(MP_OBJ_TO_PTR(mp_sys_path), path_num);
mp_obj_t *path_items;
mp_obj_list_get(mp_sys_path, &path_num, &path_items);
path_items[0] = MP_OBJ_NEW_QSTR(MP_QSTR_);
{
char *p = path;
for (mp_uint_t i = 1; i < path_num; i++) {
char *p1 = strchr(p, PATHLIST_SEP_CHAR);
if (p1 == NULL) {
p1 = p + strlen(p);
}
if (p[0] == '~' && p[1] == '/' && home != NULL) {
// Expand standalone ~ to $HOME
CHECKBUF(buf, PATH_MAX);
CHECKBUF_APPEND(buf, home, strlen(home));
CHECKBUF_APPEND(buf, p + 1, (size_t)(p1 - p - 1));
path_items[i] = MP_OBJ_NEW_QSTR(qstr_from_strn(buf, CHECKBUF_LEN(buf)));
} else {
path_items[i] = MP_OBJ_NEW_QSTR(qstr_from_strn(p, p1 - p));
}
p = p1 + 1;
}
}
mp_obj_list_init(MP_OBJ_TO_PTR(mp_sys_argv), 0);
#if defined(MICROPY_UNIX_COVERAGE)
{
MP_DECLARE_CONST_FUN_OBJ(extra_coverage_obj);
mp_store_global(QSTR_FROM_STR_STATIC("extra_coverage"), (mp_obj_t)&extra_coverage_obj);
}
#endif
// Here is some example code to create a class and instance of that class.
// First is the Python, then the C code.
//
// class TestClass:
// pass
// test_obj = TestClass()
// test_obj.attr = 42
//
// mp_obj_t test_class_type, test_class_instance;
// test_class_type = mp_obj_new_type(QSTR_FROM_STR_STATIC("TestClass"), mp_const_empty_tuple, mp_obj_new_dict(0));
// mp_store_name(QSTR_FROM_STR_STATIC("test_obj"), test_class_instance = mp_call_function_0(test_class_type));
// mp_store_attr(test_class_instance, QSTR_FROM_STR_STATIC("attr"), mp_obj_new_int(42));
/*
printf("bytes:\n");
printf(" total %d\n", m_get_total_bytes_allocated());
printf(" cur %d\n", m_get_current_bytes_allocated());
printf(" peak %d\n", m_get_peak_bytes_allocated());
*/
const int NOTHING_EXECUTED = -2;
int ret = NOTHING_EXECUTED;
for (int a = 1; a < argc; a++) {
if (argv[a][0] == '-') {
if (strcmp(argv[a], "-c") == 0) {
if (a + 1 >= argc) {
return usage(argv);
}
ret = do_str(argv[a + 1]);
if (ret & FORCED_EXIT) {
break;
}
a += 1;
} else if (strcmp(argv[a], "-m") == 0) {
if (a + 1 >= argc) {
return usage(argv);
}
mp_obj_t import_args[4];
import_args[0] = mp_obj_new_str(argv[a + 1], strlen(argv[a + 1]), false);
import_args[1] = import_args[2] = mp_const_none;
// Ask __import__ to handle imported module specially - set its __name__
// to __main__, and also return this leaf module, not top-level package
// containing it.
import_args[3] = mp_const_false;
// TODO: https://docs.python.org/3/using/cmdline.html#cmdoption-m :
// "the first element of sys.argv will be the full path to
// the module file (while the module file is being located,
// the first element will be set to "-m")."
set_sys_argv(argv, argc, a + 1);
mp_obj_t mod;
nlr_buf_t nlr;
if (nlr_push(&nlr) == 0) {
mod = mp_builtin___import__(MP_ARRAY_SIZE(import_args), import_args);
nlr_pop();
} else {
// uncaught exception
return handle_uncaught_exception(nlr.ret_val) & 0xff;
}
if (mp_obj_is_package(mod)) {
// TODO
mp_printf(&mp_stderr_print, "%s: -m for packages not yet implemented\n", argv[0]);
exit(1);
}
ret = 0;
break;
} else if (strcmp(argv[a], "-X") == 0) {
a += 1;
#if MICROPY_DEBUG_PRINTERS
} else if (strcmp(argv[a], "-v") == 0) {
mp_verbose_flag++;
#endif
} else if (strncmp(argv[a], "-O", 2) == 0) {
if (unichar_isdigit(argv[a][2])) {
MP_STATE_VM(mp_optimise_value) = argv[a][2] & 0xf;
} else {
MP_STATE_VM(mp_optimise_value) = 0;
for (char *p = argv[a] + 1; *p && *p == 'O'; p++, MP_STATE_VM(mp_optimise_value)++);
}
} else {
return usage(argv);
}
} else {
char *pathbuf = malloc(PATH_MAX);
char *basedir = realpath(argv[a], pathbuf);
if (basedir == NULL) {
mp_printf(&mp_stderr_print, "%s: can't open file '%s': [Errno %d] %s\n", argv[0], argv[a], errno, strerror(errno));
// CPython exits with 2 in such case
ret = 2;
break;
}
// Set base dir of the script as first entry in sys.path
char *p = strrchr(basedir, '/');
path_items[0] = MP_OBJ_NEW_QSTR(qstr_from_strn(basedir, p - basedir));
free(pathbuf);
set_sys_argv(argv, argc, a);
ret = do_file(argv[a]);
break;
}
}
if (ret == NOTHING_EXECUTED) {
if (isatty(0)) {
prompt_read_history();
ret = do_repl();
prompt_write_history();
} else {
mp_lexer_t *lex = mp_lexer_new_from_fd(MP_QSTR__lt_stdin_gt_, 0, false);
ret = execute_from_lexer(lex, MP_PARSE_FILE_INPUT, false);
}
}
#if MICROPY_PY_MICROPYTHON_MEM_INFO
if (mp_verbose_flag) {
mp_micropython_mem_info(0, NULL);
}
#endif
mp_deinit();
#if MICROPY_ENABLE_GC && !defined(NDEBUG)
// We don't really need to free memory since we are about to exit the
// process, but doing so helps to find memory leaks.
free(heap);
#endif
//printf("total bytes = %d\n", m_get_total_bytes_allocated());
return ret & 0xff;
}
bool unichar_isxdigit(unichar c) {
return unichar_isdigit(c) || (c >= 'a' && c <= 'f') || (c >= 'A' && c <= 'F');
}
STATIC bool is_following_digit(mp_lexer_t *lex) {
return unichar_isdigit(lex->chr1);
}
size_t mp_repl_autocomplete(const char *str, size_t len, const mp_print_t *print, const char **compl_str) {
// scan backwards to find start of "a.b.c" chain
const char *org_str = str;
const char *top = str + len;
for (const char *s = top; --s >= str;) {
if (!(unichar_isalpha(*s) || unichar_isdigit(*s) || *s == '_' || *s == '.')) {
++s;
str = s;
break;
}
}
size_t nqstr = QSTR_TOTAL();
// begin search in outer global dict which is accessed from __main__
mp_obj_t obj = MP_OBJ_FROM_PTR(&mp_module___main__);
mp_obj_t dest[2];
for (;;) {
// get next word in string to complete
const char *s_start = str;
while (str < top && *str != '.') {
++str;
}
size_t s_len = str - s_start;
if (str < top) {
// a complete word, lookup in current object
qstr q = qstr_find_strn(s_start, s_len);
if (q == MP_QSTR_NULL) {
// lookup will fail
return 0;
}
mp_load_method_maybe(obj, q, dest);
obj = dest[0]; // attribute, method, or MP_OBJ_NULL if nothing found
if (obj == MP_OBJ_NULL) {
// lookup failed
return 0;
}
// skip '.' to move to next word
++str;
} else {
// end of string, do completion on this partial name
// look for matches
const char *match_str = NULL;
size_t match_len = 0;
qstr q_first = 0, q_last;
for (qstr q = 1; q < nqstr; ++q) {
size_t d_len;
const char *d_str = (const char*)qstr_data(q, &d_len);
if (s_len <= d_len && strncmp(s_start, d_str, s_len) == 0) {
mp_load_method_maybe(obj, q, dest);
if (dest[0] != MP_OBJ_NULL) {
if (match_str == NULL) {
match_str = d_str;
match_len = d_len;
} else {
// search for longest common prefix of match_str and d_str
// (assumes these strings are null-terminated)
for (size_t j = s_len; j <= match_len && j <= d_len; ++j) {
if (match_str[j] != d_str[j]) {
match_len = j;
break;
}
}
}
if (q_first == 0) {
q_first = q;
}
q_last = q;
}
}
}
// nothing found
if (q_first == 0) {
// If there're no better alternatives, and if it's first word
// in the line, try to complete "import".
if (s_start == org_str) {
static const char import_str[] = "import ";
if (memcmp(s_start, import_str, s_len) == 0) {
*compl_str = import_str + s_len;
return sizeof(import_str) - 1 - s_len;
}
}
return 0;
}
// 1 match found, or multiple matches with a common prefix
if (q_first == q_last || match_len > s_len) {
*compl_str = match_str + s_len;
return match_len - s_len;
}
// multiple matches found, print them out
#define WORD_SLOT_LEN (16)
#define MAX_LINE_LEN (4 * WORD_SLOT_LEN)
int line_len = MAX_LINE_LEN; // force a newline for first word
for (qstr q = q_first; q <= q_last; ++q) {
size_t d_len;
const char *d_str = (const char*)qstr_data(q, &d_len);
if (s_len <= d_len && strncmp(s_start, d_str, s_len) == 0) {
mp_load_method_maybe(obj, q, dest);
if (dest[0] != MP_OBJ_NULL) {
int gap = (line_len + WORD_SLOT_LEN - 1) / WORD_SLOT_LEN * WORD_SLOT_LEN - line_len;
if (gap < 2) {
gap += WORD_SLOT_LEN;
}
if (line_len + gap + d_len <= MAX_LINE_LEN) {
// TODO optimise printing of gap?
for (int j = 0; j < gap; ++j) {
mp_print_str(print, " ");
}
mp_print_str(print, d_str);
line_len += gap + d_len;
} else {
mp_printf(print, "\n%s", d_str);
line_len = d_len;
}
}
}
}
mp_print_str(print, "\n");
return (size_t)(-1); // indicate many matches
}
}
}
MP_NOINLINE int main_(int argc, char **argv) {
mp_stack_set_limit(40000 * (BYTES_PER_WORD / 4));
pre_process_options(argc, argv);
char *heap = malloc(heap_size);
gc_init(heap, heap + heap_size);
mp_init();
#ifdef _WIN32
set_fmode_binary();
#endif
mp_obj_list_init(mp_sys_path, 0);
mp_obj_list_init(mp_sys_argv, 0);
// set default compiler configuration
mp_dynamic_compiler.small_int_bits = 31;
mp_dynamic_compiler.opt_cache_map_lookup_in_bytecode = 0;
mp_dynamic_compiler.py_builtins_str_unicode = 1;
const char *input_file = NULL;
const char *output_file = NULL;
const char *source_file = NULL;
// parse main options
for (int a = 1; a < argc; a++) {
if (argv[a][0] == '-') {
if (strcmp(argv[a], "-X") == 0) {
a += 1;
} else if (strcmp(argv[a], "-v") == 0) {
mp_verbose_flag++;
} else if (strncmp(argv[a], "-O", 2) == 0) {
if (unichar_isdigit(argv[a][2])) {
MP_STATE_VM(mp_optimise_value) = argv[a][2] & 0xf;
} else {
MP_STATE_VM(mp_optimise_value) = 0;
for (char *p = argv[a] + 1; *p && *p == 'O'; p++, MP_STATE_VM(mp_optimise_value)++);
}
} else if (strcmp(argv[a], "-o") == 0) {
if (a + 1 >= argc) {
exit(usage(argv));
}
a += 1;
output_file = argv[a];
} else if (strcmp(argv[a], "-s") == 0) {
if (a + 1 >= argc) {
exit(usage(argv));
}
a += 1;
source_file = argv[a];
} else if (strncmp(argv[a], "-msmall-int-bits=", sizeof("-msmall-int-bits=") - 1) == 0) {
char *end;
mp_dynamic_compiler.small_int_bits =
strtol(argv[a] + sizeof("-msmall-int-bits=") - 1, &end, 0);
if (*end) {
return usage(argv);
}
// TODO check that small_int_bits is within range of host's capabilities
} else if (strcmp(argv[a], "-mno-cache-lookup-bc") == 0) {
mp_dynamic_compiler.opt_cache_map_lookup_in_bytecode = 0;
} else if (strcmp(argv[a], "-mcache-lookup-bc") == 0) {
mp_dynamic_compiler.opt_cache_map_lookup_in_bytecode = 1;
} else if (strcmp(argv[a], "-mno-unicode") == 0) {
mp_dynamic_compiler.py_builtins_str_unicode = 0;
} else if (strcmp(argv[a], "-municode") == 0) {
mp_dynamic_compiler.py_builtins_str_unicode = 1;
} else {
return usage(argv);
}
} else {
if (input_file != NULL) {
mp_printf(&mp_stderr_print, "multiple input files\n");
exit(1);
}
input_file = argv[a];
}
}
if (input_file == NULL) {
mp_printf(&mp_stderr_print, "no input file\n");
exit(1);
}
int ret = compile_and_save(input_file, output_file, source_file);
#if MICROPY_PY_MICROPYTHON_MEM_INFO
if (mp_verbose_flag) {
mp_micropython_mem_info(0, NULL);
}
#endif
mp_deinit();
return ret & 0xff;
}
STATIC bool is_digit(mp_lexer_t *lex) {
return unichar_isdigit(lex->chr0);
}
static void push_result_token(parser_t *parser, const mp_lexer_t *lex) {
const mp_token_t *tok = mp_lexer_cur(lex);
mp_parse_node_t pn;
if (tok->kind == MP_TOKEN_NAME) {
pn = mp_parse_node_new_leaf(MP_PARSE_NODE_ID, qstr_from_strn_copy(tok->str, tok->len));
} else if (tok->kind == MP_TOKEN_NUMBER) {
bool dec = false;
bool small_int = true;
machine_int_t int_val = 0;
int len = tok->len;
const char *str = tok->str;
int base = 10;
int i = 0;
if (len >= 3 && str[0] == '0') {
if (str[1] == 'o' || str[1] == 'O') {
// octal
base = 8;
i = 2;
} else if (str[1] == 'x' || str[1] == 'X') {
// hexadecimal
base = 16;
i = 2;
} else if (str[1] == 'b' || str[1] == 'B') {
// binary
base = 2;
i = 2;
}
}
bool overflow = false;
for (; i < len; i++) {
machine_int_t old_val = int_val;
if (unichar_isdigit(str[i]) && str[i] - '0' < base) {
int_val = base * int_val + str[i] - '0';
} else if (base == 16 && 'a' <= str[i] && str[i] <= 'f') {
int_val = base * int_val + str[i] - 'a' + 10;
} else if (base == 16 && 'A' <= str[i] && str[i] <= 'F') {
int_val = base * int_val + str[i] - 'A' + 10;
} else if (str[i] == '.' || str[i] == 'e' || str[i] == 'E' || str[i] == 'j' || str[i] == 'J') {
dec = true;
break;
} else {
small_int = false;
break;
}
if (int_val < old_val) {
// If new value became less than previous, it's overflow
overflow = true;
} else if ((old_val ^ int_val) & WORD_MSBIT_HIGH) {
// If signed number changed sign - it's overflow
overflow = true;
}
}
if (dec) {
pn = mp_parse_node_new_leaf(MP_PARSE_NODE_DECIMAL, qstr_from_strn_copy(str, len));
} else if (small_int && !overflow && MP_FIT_SMALL_INT(int_val)) {
pn = mp_parse_node_new_leaf(MP_PARSE_NODE_SMALL_INT, int_val);
} else {
pn = mp_parse_node_new_leaf(MP_PARSE_NODE_INTEGER, qstr_from_strn_copy(str, len));
}
} else if (tok->kind == MP_TOKEN_STRING) {
pn = mp_parse_node_new_leaf(MP_PARSE_NODE_STRING, qstr_from_strn_copy(tok->str, tok->len));
} else if (tok->kind == MP_TOKEN_BYTES) {
pn = mp_parse_node_new_leaf(MP_PARSE_NODE_BYTES, qstr_from_strn_copy(tok->str, tok->len));
} else {
pn = mp_parse_node_new_leaf(MP_PARSE_NODE_TOKEN, tok->kind);
}
push_result_node(parser, pn);
}
STATIC mp_obj_t mod_ujson_load(mp_obj_t stream_obj) {
const mp_stream_p_t *stream_p = mp_get_stream_raise(stream_obj, MP_STREAM_OP_READ);
ujson_stream_t s = {stream_obj, stream_p->read, 0, 0};
vstr_t vstr;
vstr_init(&vstr, 8);
mp_obj_list_t stack; // we use a list as a simple stack for nested JSON
stack.len = 0;
stack.items = NULL;
mp_obj_t stack_top = MP_OBJ_NULL;
mp_obj_type_t *stack_top_type = NULL;
mp_obj_t stack_key = MP_OBJ_NULL;
S_NEXT(s);
for (;;) {
cont:
if (S_END(s)) {
break;
}
mp_obj_t next = MP_OBJ_NULL;
bool enter = false;
byte cur = S_CUR(s);
S_NEXT(s);
switch (cur) {
case ',':
case ':':
case ' ':
case '\t':
case '\n':
case '\r':
goto cont;
case 'n':
if (S_CUR(s) == 'u' && S_NEXT(s) == 'l' && S_NEXT(s) == 'l') {
S_NEXT(s);
next = mp_const_none;
} else {
goto fail;
}
break;
case 'f':
if (S_CUR(s) == 'a' && S_NEXT(s) == 'l' && S_NEXT(s) == 's' && S_NEXT(s) == 'e') {
S_NEXT(s);
next = mp_const_false;
} else {
goto fail;
}
break;
case 't':
if (S_CUR(s) == 'r' && S_NEXT(s) == 'u' && S_NEXT(s) == 'e') {
S_NEXT(s);
next = mp_const_true;
} else {
goto fail;
}
break;
case '"':
vstr_reset(&vstr);
for (; !S_END(s) && S_CUR(s) != '"';) {
byte c = S_CUR(s);
if (c == '\\') {
c = S_NEXT(s);
switch (c) {
case 'b': c = 0x08; break;
case 'f': c = 0x0c; break;
case 'n': c = 0x0a; break;
case 'r': c = 0x0d; break;
case 't': c = 0x09; break;
case 'u': {
mp_uint_t num = 0;
for (int i = 0; i < 4; i++) {
c = (S_NEXT(s) | 0x20) - '0';
if (c > 9) {
c -= ('a' - ('9' + 1));
}
num = (num << 4) | c;
}
vstr_add_char(&vstr, num);
goto str_cont;
}
}
}
vstr_add_byte(&vstr, c);
str_cont:
S_NEXT(s);
}
if (S_END(s)) {
goto fail;
}
S_NEXT(s);
next = mp_obj_new_str(vstr.buf, vstr.len, false);
break;
case '-':
case '0': case '1': case '2': case '3': case '4': case '5': case '6': case '7': case '8': case '9': {
bool flt = false;
vstr_reset(&vstr);
for (;;) {
vstr_add_byte(&vstr, cur);
cur = S_CUR(s);
if (cur == '.' || cur == 'E' || cur == 'e') {
flt = true;
} else if (cur == '-' || unichar_isdigit(cur)) {
// pass
} else {
break;
}
S_NEXT(s);
}
if (flt) {
next = mp_parse_num_decimal(vstr.buf, vstr.len, false, false, NULL);
} else {
next = mp_parse_num_integer(vstr.buf, vstr.len, 10, NULL);
}
break;
}
case '[':
next = mp_obj_new_list(0, NULL);
enter = true;
break;
case '{':
next = mp_obj_new_dict(0);
enter = true;
break;
case '}':
case ']': {
if (stack_top == MP_OBJ_NULL) {
// no object at all
goto fail;
}
if (stack.len == 0) {
// finished; compound object
goto success;
}
stack.len -= 1;
stack_top = stack.items[stack.len];
stack_top_type = mp_obj_get_type(stack_top);
goto cont;
}
default:
goto fail;
}
if (stack_top == MP_OBJ_NULL) {
stack_top = next;
stack_top_type = mp_obj_get_type(stack_top);
if (!enter) {
// finished; single primitive only
goto success;
}
} else {
// append to list or dict
if (stack_top_type == &mp_type_list) {
mp_obj_list_append(stack_top, next);
} else {
if (stack_key == MP_OBJ_NULL) {
stack_key = next;
if (enter) {
goto fail;
}
} else {
mp_obj_dict_store(stack_top, stack_key, next);
stack_key = MP_OBJ_NULL;
}
}
if (enter) {
if (stack.items == NULL) {
mp_obj_list_init(&stack, 1);
stack.items[0] = stack_top;
} else {
mp_obj_list_append(MP_OBJ_FROM_PTR(&stack), stack_top);
}
stack_top = next;
stack_top_type = mp_obj_get_type(stack_top);
}
}
}
success:
// eat trailing whitespace
while (unichar_isspace(S_CUR(s))) {
S_NEXT(s);
}
if (!S_END(s)) {
// unexpected chars
goto fail;
}
if (stack_top == MP_OBJ_NULL || stack.len != 0) {
// not exactly 1 object
goto fail;
}
vstr_clear(&vstr);
return stack_top;
fail:
nlr_raise(mp_obj_new_exception_msg(&mp_type_ValueError, "syntax error in JSON"));
}
// This function implements a simple non-recursive JSON parser.
//
// The JSON specification is at http://www.ietf.org/rfc/rfc4627.txt
// The parser here will parse any valid JSON and return the correct
// corresponding Python object. It allows through a superset of JSON, since
// it treats commas and colons as "whitespace", and doesn't care if
// brackets/braces are correctly paired. It will raise a ValueError if the
// input is outside it's specs.
//
// Most of the work is parsing the primitives (null, false, true, numbers,
// strings). It does 1 pass over the input string and so is easily extended to
// being able to parse from a non-seekable stream. It tries to be fast and
// small in code size, while not using more RAM than necessary.
STATIC mp_obj_t mod_ujson_loads(mp_obj_t obj) {
mp_uint_t len;
const char *s = mp_obj_str_get_data(obj, &len);
const char *top = s + len;
vstr_t vstr;
vstr_init(&vstr, 8);
mp_obj_list_t stack; // we use a list as a simple stack for nested JSON
stack.len = 0;
stack.items = NULL;
mp_obj_t stack_top = MP_OBJ_NULL;
mp_obj_type_t *stack_top_type = NULL;
mp_obj_t stack_key = MP_OBJ_NULL;
for (;;) {
cont:
if (s == top) {
break;
}
mp_obj_t next = MP_OBJ_NULL;
bool enter = false;
switch (*s) {
case ',':
case ':':
case ' ':
case '\t':
case '\n':
case '\r':
s += 1;
goto cont;
case 'n':
if (s + 3 < top && s[1] == 'u' && s[2] == 'l' && s[3] == 'l') {
s += 4;
next = mp_const_none;
} else {
goto fail;
}
break;
case 'f':
if (s + 4 < top && s[1] == 'a' && s[2] == 'l' && s[3] == 's' && s[4] == 'e') {
s += 5;
next = mp_const_false;
} else {
goto fail;
}
break;
case 't':
if (s + 3 < top && s[1] == 'r' && s[2] == 'u' && s[3] == 'e') {
s += 4;
next = mp_const_true;
} else {
goto fail;
}
break;
case '"':
vstr_reset(&vstr);
for (s++; s < top && *s != '"';) {
byte c = *s;
if (c == '\\') {
s++;
c = *s;
switch (c) {
case 'b': c = 0x08; break;
case 'f': c = 0x0c; break;
case 'n': c = 0x0a; break;
case 'r': c = 0x0d; break;
case 't': c = 0x09; break;
case 'u': {
if (s + 4 >= top) { goto fail; }
mp_uint_t num = 0;
for (int i = 0; i < 4; i++) {
c = (*++s | 0x20) - '0';
if (c > 9) {
c -= ('a' - ('9' + 1));
}
num = (num << 4) | c;
}
vstr_add_char(&vstr, num);
goto str_cont;
}
}
}
vstr_add_byte(&vstr, c);
str_cont:
s++;
}
if (s == top) {
goto fail;
}
s++;
next = mp_obj_new_str(vstr.buf, vstr.len, false);
break;
case '-':
case '0': case '1': case '2': case '3': case '4': case '5': case '6': case '7': case '8': case '9': {
bool flt = false;
vstr_reset(&vstr);
for (; s < top; s++) {
if (*s == '.' || *s == 'E' || *s == 'e') {
flt = true;
} else if (*s == '-' || unichar_isdigit(*s)) {
// pass
} else {
break;
}
vstr_add_byte(&vstr, *s);
}
if (flt) {
next = mp_parse_num_decimal(vstr.buf, vstr.len, false, false, NULL);
} else {
next = mp_parse_num_integer(vstr.buf, vstr.len, 10, NULL);
}
break;
}
case '[':
next = mp_obj_new_list(0, NULL);
enter = true;
s += 1;
break;
case '{':
next = mp_obj_new_dict(0);
enter = true;
s += 1;
break;
case '}':
case ']': {
s += 1;
if (stack_top == MP_OBJ_NULL) {
// no object at all
goto fail;
}
if (stack.len == 0) {
// finished; compound object
goto success;
}
stack.len -= 1;
stack_top = stack.items[stack.len];
stack_top_type = mp_obj_get_type(stack_top);
goto cont;
}
default:
goto fail;
}
if (stack_top == MP_OBJ_NULL) {
stack_top = next;
stack_top_type = mp_obj_get_type(stack_top);
if (!enter) {
// finished; single primitive only
goto success;
}
} else {
// append to list or dict
if (stack_top_type == &mp_type_list) {
mp_obj_list_append(stack_top, next);
} else {
if (stack_key == MP_OBJ_NULL) {
stack_key = next;
if (enter) {
goto fail;
}
} else {
mp_obj_dict_store(stack_top, stack_key, next);
stack_key = MP_OBJ_NULL;
}
}
if (enter) {
if (stack.items == NULL) {
mp_obj_list_init(&stack, 1);
stack.items[0] = stack_top;
} else {
mp_obj_list_append(MP_OBJ_FROM_PTR(&stack), stack_top);
}
stack_top = next;
stack_top_type = mp_obj_get_type(stack_top);
}
}
}
success:
// eat trailing whitespace
while (s < top && unichar_isspace(*s)) {
s++;
}
if (s < top) {
// unexpected chars
goto fail;
}
if (stack_top == MP_OBJ_NULL || stack.len != 0) {
// not exactly 1 object
goto fail;
}
vstr_clear(&vstr);
return stack_top;
fail:
nlr_raise(mp_obj_new_exception_msg(&mp_type_ValueError, "syntax error in JSON"));
}
mp_uint_t mp_repl_autocomplete(const char *str, mp_uint_t len, const mp_print_t *print, const char **compl_str) {
// scan backwards to find start of "a.b.c" chain
const char *org_str = str;
const char *top = str + len;
for (const char *s = top; --s >= str;) {
if (!(unichar_isalpha(*s) || unichar_isdigit(*s) || *s == '_' || *s == '.')) {
++s;
str = s;
break;
}
}
// begin search in locals dict
mp_obj_dict_t *dict = mp_locals_get();
for (;;) {
// get next word in string to complete
const char *s_start = str;
while (str < top && *str != '.') {
++str;
}
mp_uint_t s_len = str - s_start;
if (str < top) {
// a complete word, lookup in current dict
mp_obj_t obj = MP_OBJ_NULL;
for (mp_uint_t i = 0; i < dict->map.alloc; i++) {
if (MP_MAP_SLOT_IS_FILLED(&dict->map, i)) {
size_t d_len;
const char *d_str = mp_obj_str_get_data(dict->map.table[i].key, &d_len);
if (s_len == d_len && strncmp(s_start, d_str, d_len) == 0) {
obj = dict->map.table[i].value;
break;
}
}
}
if (obj == MP_OBJ_NULL) {
// lookup failed
return 0;
}
// found an object of this name; try to get its dict
if (MP_OBJ_IS_TYPE(obj, &mp_type_module)) {
dict = mp_obj_module_get_globals(obj);
} else {
mp_obj_type_t *type;
if (MP_OBJ_IS_TYPE(obj, &mp_type_type)) {
type = MP_OBJ_TO_PTR(obj);
} else {
type = mp_obj_get_type(obj);
}
if (type->locals_dict != NULL && type->locals_dict->base.type == &mp_type_dict) {
dict = type->locals_dict;
} else {
// obj has no dict
return 0;
}
}
// skip '.' to move to next word
++str;
} else {
// end of string, do completion on this partial name
// look for matches
int n_found = 0;
const char *match_str = NULL;
mp_uint_t match_len = 0;
for (mp_uint_t i = 0; i < dict->map.alloc; i++) {
if (MP_MAP_SLOT_IS_FILLED(&dict->map, i)) {
size_t d_len;
const char *d_str = mp_obj_str_get_data(dict->map.table[i].key, &d_len);
if (s_len <= d_len && strncmp(s_start, d_str, s_len) == 0) {
if (match_str == NULL) {
match_str = d_str;
match_len = d_len;
} else {
// search for longest common prefix of match_str and d_str
// (assumes these strings are null-terminated)
for (mp_uint_t j = s_len; j <= match_len && j <= d_len; ++j) {
if (match_str[j] != d_str[j]) {
match_len = j;
break;
}
}
}
++n_found;
}
}
}
// nothing found
if (n_found == 0) {
// If there're no better alternatives, and if it's first word
// in the line, try to complete "import".
if (s_start == org_str) {
static const char import_str[] = "import ";
if (memcmp(s_start, import_str, s_len) == 0) {
*compl_str = import_str + s_len;
return sizeof(import_str) - 1 - s_len;
}
}
return 0;
}
// 1 match found, or multiple matches with a common prefix
if (n_found == 1 || match_len > s_len) {
*compl_str = match_str + s_len;
return match_len - s_len;
}
// multiple matches found, print them out
#define WORD_SLOT_LEN (16)
#define MAX_LINE_LEN (4 * WORD_SLOT_LEN)
int line_len = MAX_LINE_LEN; // force a newline for first word
for (mp_uint_t i = 0; i < dict->map.alloc; i++) {
if (MP_MAP_SLOT_IS_FILLED(&dict->map, i)) {
size_t d_len;
const char *d_str = mp_obj_str_get_data(dict->map.table[i].key, &d_len);
if (s_len <= d_len && strncmp(s_start, d_str, s_len) == 0) {
int gap = (line_len + WORD_SLOT_LEN - 1) / WORD_SLOT_LEN * WORD_SLOT_LEN - line_len;
if (gap < 2) {
gap += WORD_SLOT_LEN;
}
if (line_len + gap + d_len <= MAX_LINE_LEN) {
// TODO optimise printing of gap?
for (int j = 0; j < gap; ++j) {
mp_print_str(print, " ");
}
mp_print_str(print, d_str);
line_len += gap + d_len;
} else {
mp_printf(print, "\n%s", d_str);
line_len = d_len;
}
}
}
}
mp_print_str(print, "\n");
return (mp_uint_t)(-1); // indicate many matches
}
}
}
STATIC void push_result_token(parser_t *parser, const mp_lexer_t *lex) {
const mp_token_t *tok = mp_lexer_cur(lex);
mp_parse_node_t pn;
if (tok->kind == MP_TOKEN_NAME) {
pn = mp_parse_node_new_leaf(MP_PARSE_NODE_ID, qstr_from_strn(tok->str, tok->len));
} else if (tok->kind == MP_TOKEN_NUMBER) {
bool dec = false;
bool small_int = true;
machine_int_t int_val = 0;
int len = tok->len;
const char *str = tok->str;
int base = 0;
int i = mp_parse_num_base(str, len, &base);
bool overflow = false;
for (; i < len; i++) {
int dig;
if (unichar_isdigit(str[i]) && str[i] - '0' < base) {
dig = str[i] - '0';
} else if (base == 16 && 'a' <= str[i] && str[i] <= 'f') {
dig = str[i] - 'a' + 10;
} else if (base == 16 && 'A' <= str[i] && str[i] <= 'F') {
dig = str[i] - 'A' + 10;
} else if (str[i] == '.' || str[i] == 'e' || str[i] == 'E' || str[i] == 'j' || str[i] == 'J') {
dec = true;
break;
} else {
small_int = false;
break;
}
// add next digi and check for overflow
if (mp_small_int_mul_overflow(int_val, base)) {
overflow = true;
}
int_val = int_val * base + dig;
if (!MP_SMALL_INT_FITS(int_val)) {
overflow = true;
}
}
if (dec) {
pn = mp_parse_node_new_leaf(MP_PARSE_NODE_DECIMAL, qstr_from_strn(str, len));
} else if (small_int && !overflow && MP_SMALL_INT_FITS(int_val)) {
pn = mp_parse_node_new_leaf(MP_PARSE_NODE_SMALL_INT, int_val);
} else {
pn = mp_parse_node_new_leaf(MP_PARSE_NODE_INTEGER, qstr_from_strn(str, len));
}
} else if (tok->kind == MP_TOKEN_STRING) {
// Don't automatically intern all strings. 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 (tok->len <= MICROPY_ALLOC_PARSE_INTERN_STRING_LEN) {
// intern short strings
qst = qstr_from_strn(tok->str, tok->len);
} else {
// check if this string is already interned
qst = qstr_find_strn(tok->str, tok->len);
}
if (qst != MP_QSTR_NULL) {
// qstr exists, make a leaf node
pn = mp_parse_node_new_leaf(MP_PARSE_NODE_STRING, qst);
} else {
// not interned, make a node holding a pointer to the string data
push_result_string(parser, mp_lexer_cur(lex)->src_line, tok->str, tok->len);
return;
}
} else if (tok->kind == MP_TOKEN_BYTES) {
pn = mp_parse_node_new_leaf(MP_PARSE_NODE_BYTES, qstr_from_strn(tok->str, tok->len));
} else {
pn = mp_parse_node_new_leaf(MP_PARSE_NODE_TOKEN, tok->kind);
}
push_result_node(parser, pn);
}
