예제 #1
0
파일: vstring.c 프로젝트: Akheon23/nvopencc
/* add char to vstring */
vstring
vstr_append (vstring v, const char c)
{
	if (vstr_len(v) + 1 > vstr_max(v)) {
		Realloc_Vstring (&v, vstr_len(v) + 1);
	}
	v.str[v.len] = c;
	v.len++;
	return v;
}
예제 #2
0
파일: vstring.c 프로젝트: Akheon23/nvopencc
/* add string to vstring */
vstring
vstr_concat (vstring v, const char *s)
{
	int slen = strlen(s);
	if (vstr_len(v) + slen > vstr_max(v)) {
		Realloc_Vstring (&v, vstr_len(v) + slen);
	}
	/* may be nulls in vstr, so can't concat from beginning;
	 * instead just copy onto end of string. */
	strcpy(vstr_str(v)+vstr_len(v), s);
	set_vstr_len(v, vstr_len(v) + slen);
	return v;
}
예제 #3
0
파일: string.c 프로젝트: senarukana/ABCLang
static void vstr_append_character(VString *v, ABC_Char ch) {
    int current_len;
    
    current_len = vstr_len(v->string);
    v->string = MEM_realloc(v->string, sizeof(ABC_Char) *current_len + 2);
    v->string[current_len] = ch;
    v->string[current_len+1] = L'\0';
}
예제 #4
0
파일: string.c 프로젝트: senarukana/ABCLang
static void vstr_append_string(VString *v, ABC_Char *str){
    int new_size;
    int old_len;

    old_len = vstr_len(v->string);
    new_size = old_len + abc_wcs_len(str)  + 1;
    v->string = MEM_realloc(v->string, sizeof(ABC_Char) * new_size);
    abc_wcs_cpy(&v->string[old_len], str);
}
예제 #5
0
파일: vstring.c 프로젝트: Akheon23/nvopencc
vstring
vstr_begin (int len) 
{
	if (vstr_max(current_vstring) == 0) {
		set_vstr_str(current_vstring, (char*) malloc(len));
		set_vstr_max(current_vstring, len);
	}
	else if (vstr_len(current_vstring) != 0) {
		DevWarn("vstr_begin before finishing old one?\n");
		set_vstr_len(current_vstring, 0);
	}
	if (vstr_max(current_vstring) < len) {
		Realloc_Vstring (&current_vstring, len);
	}
	return current_vstring;
}
예제 #6
0
mp_obj_t mp_builtin___import__(mp_uint_t n_args, const mp_obj_t *args) {
#if DEBUG_PRINT
    DEBUG_printf("__import__:\n");
    for (mp_uint_t i = 0; i < n_args; i++) {
        DEBUG_printf("  ");
        mp_obj_print(args[i], PRINT_REPR);
        DEBUG_printf("\n");
    }
#endif

    mp_obj_t module_name = args[0];
    mp_obj_t fromtuple = mp_const_none;
    mp_int_t level = 0;
    if (n_args >= 4) {
        fromtuple = args[3];
        if (n_args >= 5) {
            level = MP_OBJ_SMALL_INT_VALUE(args[4]);
        }
    }

    mp_uint_t mod_len;
    const char *mod_str = mp_obj_str_get_data(module_name, &mod_len);

    if (level != 0) {
        // What we want to do here is to take name of current module,
        // chop <level> trailing components, and concatenate with passed-in
        // module name, thus resolving relative import name into absolue.
        // This even appears to be correct per
        // http://legacy.python.org/dev/peps/pep-0328/#relative-imports-and-name
        // "Relative imports use a module's __name__ attribute to determine that
        // module's position in the package hierarchy."
        level--;
        mp_obj_t this_name_q = mp_obj_dict_get(mp_globals_get(), MP_OBJ_NEW_QSTR(MP_QSTR___name__));
        assert(this_name_q != MP_OBJ_NULL);
        #if MICROPY_CPYTHON_COMPAT
        if (MP_OBJ_QSTR_VALUE(this_name_q) == MP_QSTR___main__) {
            // This is a module run by -m command-line switch, get its real name from backup attribute
            this_name_q = mp_obj_dict_get(mp_globals_get(), MP_OBJ_NEW_QSTR(MP_QSTR___main__));
        }
        #endif
        mp_map_t *globals_map = mp_obj_dict_get_map(mp_globals_get());
        mp_map_elem_t *elem = mp_map_lookup(globals_map, MP_OBJ_NEW_QSTR(MP_QSTR___path__), MP_MAP_LOOKUP);
        bool is_pkg = (elem != NULL);

#if DEBUG_PRINT
        DEBUG_printf("Current module/package: ");
        mp_obj_print(this_name_q, PRINT_REPR);
        DEBUG_printf(", is_package: %d", is_pkg);
        DEBUG_printf("\n");
#endif

        mp_uint_t this_name_l;
        const char *this_name = mp_obj_str_get_data(this_name_q, &this_name_l);

        const char *p = this_name + this_name_l;
        if (!is_pkg) {
            // We have module, but relative imports are anchored at package, so
            // go there.
            chop_component(this_name, &p);
        }


        uint dots_seen = 0;
        while (level--) {
            chop_component(this_name, &p);
            dots_seen++;
        }

        if (dots_seen == 0 && level >= 1) {
            // http://legacy.python.org/dev/peps/pep-0328/#relative-imports-and-name
            // "If the module's name does not contain any package information
            // (e.g. it is set to '__main__') then relative imports are
            // resolved as if the module were a top level module, regardless
            // of where the module is actually located on the file system."
            // Supposedly this if catches this condition and resolve it properly
            // TODO: But nobody knows for sure. This condition happens when
            // package's __init__.py does something like "import .submod". So,
            // maybe we should check for package here? But quote above doesn't
            // talk about packages, it talks about dot-less module names.
            DEBUG_printf("Warning: no dots in current module name and level>0\n");
            p = this_name + this_name_l;
        } else if (level != -1) {
            nlr_raise(mp_obj_new_exception_msg(&mp_type_ImportError, "Invalid relative import"));
        }

        uint new_mod_l = (mod_len == 0 ? (size_t)(p - this_name) : (size_t)(p - this_name) + 1 + mod_len);
        char *new_mod = alloca(new_mod_l);
        memcpy(new_mod, this_name, p - this_name);
        if (mod_len != 0) {
            new_mod[p - this_name] = '.';
            memcpy(new_mod + (p - this_name) + 1, mod_str, mod_len);
        }

        qstr new_mod_q = qstr_from_strn(new_mod, new_mod_l);
        DEBUG_printf("Resolved base name for relative import: '%s'\n", qstr_str(new_mod_q));
        if (new_mod_q == MP_QSTR_) {
            // CPython raises SystemError
            nlr_raise(mp_obj_new_exception_msg(&mp_type_ImportError, "cannot perform relative import"));
        }
        module_name = MP_OBJ_NEW_QSTR(new_mod_q);
        mod_str = new_mod;
        mod_len = new_mod_l;
    }

    // check if module already exists
    qstr module_name_qstr = mp_obj_str_get_qstr(module_name);
    mp_obj_t module_obj = mp_module_get(module_name_qstr);
    if (module_obj != MP_OBJ_NULL) {
        DEBUG_printf("Module already loaded\n");
        // If it's not a package, return module right away
        char *p = strchr(mod_str, '.');
        if (p == NULL) {
            return module_obj;
        }
        // If fromlist is not empty, return leaf module
        if (fromtuple != mp_const_none) {
            return module_obj;
        }
        // Otherwise, we need to return top-level package
        qstr pkg_name = qstr_from_strn(mod_str, p - mod_str);
        return mp_module_get(pkg_name);
    }
    DEBUG_printf("Module not yet loaded\n");

    #if MICROPY_MODULE_FROZEN
    mp_lexer_t *lex = mp_find_frozen_module(mod_str, mod_len);
    if (lex != NULL) {
        module_obj = mp_obj_new_module(module_name_qstr);
        // if args[3] (fromtuple) has magic value False, set up
        // this module for command-line "-m" option (set module's
        // name to __main__ instead of real name).
        // TODO: Duplicated below too.
        if (fromtuple == mp_const_false) {
            mp_obj_module_t *o = module_obj;
            mp_obj_dict_store(o->globals, MP_OBJ_NEW_QSTR(MP_QSTR___name__), MP_OBJ_NEW_QSTR(MP_QSTR___main__));
        }
        do_load_from_lexer(module_obj, lex, mod_str);
        return module_obj;
    }
    #endif

    uint last = 0;
    VSTR_FIXED(path, MICROPY_ALLOC_PATH_MAX)
    module_obj = MP_OBJ_NULL;
    mp_obj_t top_module_obj = MP_OBJ_NULL;
    mp_obj_t outer_module_obj = MP_OBJ_NULL;
    uint i;
    for (i = 1; i <= mod_len; i++) {
        if (i == mod_len || mod_str[i] == '.') {
            // create a qstr for the module name up to this depth
            qstr mod_name = qstr_from_strn(mod_str, i);
            DEBUG_printf("Processing module: %s\n", qstr_str(mod_name));
            DEBUG_printf("Previous path: =%.*s=\n", vstr_len(&path), vstr_str(&path));

            // find the file corresponding to the module name
            mp_import_stat_t stat;
            if (vstr_len(&path) == 0) {
                // first module in the dotted-name; search for a directory or file
                stat = find_file(mod_str, i, &path);
            } else {
                // latter module in the dotted-name; append to path
                vstr_add_char(&path, PATH_SEP_CHAR);
                vstr_add_strn(&path, mod_str + last, i - last);
                stat = stat_dir_or_file(&path);
            }
            DEBUG_printf("Current path: %.*s\n", vstr_len(&path), vstr_str(&path));

            if (stat == MP_IMPORT_STAT_NO_EXIST) {
                #if MICROPY_MODULE_WEAK_LINKS
                // check if there is a weak link to this module
                if (i == mod_len) {
                    mp_map_elem_t *el = mp_map_lookup((mp_map_t*)&mp_builtin_module_weak_links_map, MP_OBJ_NEW_QSTR(mod_name), MP_MAP_LOOKUP);
                    if (el == NULL) {
                        goto no_exist;
                    }
                    // found weak linked module
                    module_obj = el->value;
                } else {
                    no_exist:
                #else
                {
                #endif
                    // couldn't find the file, so fail
                    if (MICROPY_ERROR_REPORTING == MICROPY_ERROR_REPORTING_TERSE) {
                        nlr_raise(mp_obj_new_exception_msg(&mp_type_ImportError, "module not found"));
                    } else {
                        nlr_raise(mp_obj_new_exception_msg_varg(&mp_type_ImportError,
                            "no module named '%q'", mod_name));
                    }
                }
            } else {
                // found the file, so get the module
                module_obj = mp_module_get(mod_name);
            }

            if (module_obj == MP_OBJ_NULL) {
                // module not already loaded, so load it!

                module_obj = mp_obj_new_module(mod_name);

                // if args[3] (fromtuple) has magic value False, set up
                // this module for command-line "-m" option (set module's
                // name to __main__ instead of real name).
                if (i == mod_len && fromtuple == mp_const_false) {
                    mp_obj_module_t *o = module_obj;
                    mp_obj_dict_store(o->globals, MP_OBJ_NEW_QSTR(MP_QSTR___name__), MP_OBJ_NEW_QSTR(MP_QSTR___main__));
                    #if MICROPY_CPYTHON_COMPAT
                    // Store real name in "__main__" attribute. Choosen semi-randonly, to reuse existing qstr's.
                    mp_obj_dict_store(o->globals, MP_OBJ_NEW_QSTR(MP_QSTR___main__), MP_OBJ_NEW_QSTR(mod_name));
                    #endif
                }

                if (stat == MP_IMPORT_STAT_DIR) {
                    DEBUG_printf("%.*s is dir\n", vstr_len(&path), vstr_str(&path));
                    // https://docs.python.org/3/reference/import.html
                    // "Specifically, any module that contains a __path__ attribute is considered a package."
                    mp_store_attr(module_obj, MP_QSTR___path__, mp_obj_new_str(vstr_str(&path), vstr_len(&path), false));
                    vstr_add_char(&path, PATH_SEP_CHAR);
                    vstr_add_str(&path, "__init__.py");
                    if (mp_import_stat(vstr_null_terminated_str(&path)) != MP_IMPORT_STAT_FILE) {
                        vstr_cut_tail_bytes(&path, sizeof("/__init__.py") - 1); // cut off /__init__.py
                        mp_warning("%s is imported as namespace package", vstr_str(&path));
                    } else {
                        do_load(module_obj, &path);
                        vstr_cut_tail_bytes(&path, sizeof("/__init__.py") - 1); // cut off /__init__.py
                    }
                } else { // MP_IMPORT_STAT_FILE
                    do_load(module_obj, &path);
                    // TODO: We cannot just break here, at the very least, we must execute
                    // trailer code below. But otherwise if there're remaining components,
                    // that would be (??) object path within module, not modules path within FS.
                    // break;
                }
            }
            if (outer_module_obj != MP_OBJ_NULL) {
                qstr s = qstr_from_strn(mod_str + last, i - last);
                mp_store_attr(outer_module_obj, s, module_obj);
            }
            outer_module_obj = module_obj;
            if (top_module_obj == MP_OBJ_NULL) {
                top_module_obj = module_obj;
            }
            last = i + 1;
        }
    }
예제 #7
0
int pyexec_friendly_repl(void) {
    vstr_t line;
    vstr_init(&line, 32);

#if defined(USE_HOST_MODE) && MICROPY_HW_HAS_LCD
    // in host mode, we enable the LCD for the repl
    mp_obj_t lcd_o = mp_call_function_0(mp_load_name(qstr_from_str("LCD")));
    mp_call_function_1(mp_load_attr(lcd_o, qstr_from_str("light")), mp_const_true);
#endif

friendly_repl_reset:
    mp_hal_stdout_tx_str("MicroPython " MICROPY_GIT_TAG " on " MICROPY_BUILD_DATE "; " MICROPY_HW_BOARD_NAME " with " MICROPY_HW_MCU_NAME "\r\n");
    #if MICROPY_PY_BUILTINS_HELP
    mp_hal_stdout_tx_str("Type \"help()\" for more information.\r\n");
    #endif

    // to test ctrl-C
    /*
    {
        uint32_t x[4] = {0x424242, 0xdeaddead, 0x242424, 0xdeadbeef};
        for (;;) {
            nlr_buf_t nlr;
            printf("pyexec_repl: %p\n", x);
            mp_hal_set_interrupt_char(CHAR_CTRL_C);
            if (nlr_push(&nlr) == 0) {
                for (;;) {
                }
            } else {
                printf("break\n");
            }
        }
    }
    */

    for (;;) {
    input_restart:

        #if defined(USE_DEVICE_MODE)
        if (usb_vcp_is_enabled()) {
            // If the user gets to here and interrupts are disabled then
            // they'll never see the prompt, traceback etc. The USB REPL needs
            // interrupts to be enabled or no transfers occur. So we try to
            // do the user a favor and reenable interrupts.
            if (query_irq() == IRQ_STATE_DISABLED) {
                enable_irq(IRQ_STATE_ENABLED);
                mp_hal_stdout_tx_str("PYB: enabling IRQs\r\n");
            }
        }
        #endif

        vstr_reset(&line);
        int ret = readline(&line, ">>> ");
        mp_parse_input_kind_t parse_input_kind = MP_PARSE_SINGLE_INPUT;

        if (ret == CHAR_CTRL_A) {
            // change to raw REPL
            mp_hal_stdout_tx_str("\r\n");
            vstr_clear(&line);
            pyexec_mode_kind = PYEXEC_MODE_RAW_REPL;
            return 0;
        } else if (ret == CHAR_CTRL_B) {
            // reset friendly REPL
            mp_hal_stdout_tx_str("\r\n");
            goto friendly_repl_reset;
        } else if (ret == CHAR_CTRL_C) {
            // break
            mp_hal_stdout_tx_str("\r\n");
            continue;
        } else if (ret == CHAR_CTRL_D) {
            // exit for a soft reset
            mp_hal_stdout_tx_str("\r\n");
            vstr_clear(&line);
            return PYEXEC_FORCED_EXIT;
        } else if (ret == CHAR_CTRL_E) {
            // paste mode
            mp_hal_stdout_tx_str("\r\npaste mode; Ctrl-C to cancel, Ctrl-D to finish\r\n=== ");
            vstr_reset(&line);
            for (;;) {
                char c = mp_hal_stdin_rx_chr();
                if (c == CHAR_CTRL_C) {
                    // cancel everything
                    mp_hal_stdout_tx_str("\r\n");
                    goto input_restart;
                } else if (c == CHAR_CTRL_D) {
                    // end of input
                    mp_hal_stdout_tx_str("\r\n");
                    break;
                } else {
                    // add char to buffer and echo
                    vstr_add_byte(&line, c);
                    if (c == '\r') {
                        mp_hal_stdout_tx_str("\r\n=== ");
                    } else {
                        mp_hal_stdout_tx_strn(&c, 1);
                    }
                }
            }
            parse_input_kind = MP_PARSE_FILE_INPUT;
        } else if (vstr_len(&line) == 0) {
            continue;
        } else {
            // got a line with non-zero length, see if it needs continuing
            while (mp_repl_continue_with_input(vstr_null_terminated_str(&line))) {
                vstr_add_byte(&line, '\n');
                ret = readline(&line, "... ");
                if (ret == CHAR_CTRL_C) {
                    // cancel everything
                    mp_hal_stdout_tx_str("\r\n");
                    goto input_restart;
                } else if (ret == CHAR_CTRL_D) {
                    // stop entering compound statement
                    break;
                }
            }
        }

        ret = parse_compile_execute(&line, parse_input_kind, EXEC_FLAG_ALLOW_DEBUGGING | EXEC_FLAG_IS_REPL | EXEC_FLAG_SOURCE_IS_VSTR);
        if (ret & PYEXEC_FORCED_EXIT) {
            return ret;
        }
    }
}
예제 #8
0
int pyexec_friendly_repl(void) {
    vstr_t line;
    vstr_init(&line, 32);

#if defined(USE_HOST_MODE) && MICROPY_HW_HAS_LCD
    // in host mode, we enable the LCD for the repl
    mp_obj_t lcd_o = mp_call_function_0(mp_load_name(qstr_from_str("LCD")));
    mp_call_function_1(mp_load_attr(lcd_o, qstr_from_str("light")), mp_const_true);
#endif

friendly_repl_reset:
    mp_hal_stdout_tx_str("Micro Python " MICROPY_GIT_TAG " on " MICROPY_BUILD_DATE "; " MICROPY_HW_BOARD_NAME " with " MICROPY_HW_MCU_NAME "\r\n");
    mp_hal_stdout_tx_str("Type \"help()\" for more information.\r\n");

    // to test ctrl-C
    /*
    {
        uint32_t x[4] = {0x424242, 0xdeaddead, 0x242424, 0xdeadbeef};
        for (;;) {
            nlr_buf_t nlr;
            printf("pyexec_repl: %p\n", x);
            mp_hal_set_interrupt_char(CHAR_CTRL_C);
            if (nlr_push(&nlr) == 0) {
                for (;;) {
                }
            } else {
                printf("break\n");
            }
        }
    }
    */

    for (;;) {
    input_restart:
        vstr_reset(&line);
        int ret = readline(&line, ">>> ");

        if (ret == CHAR_CTRL_A) {
            // change to raw REPL
            mp_hal_stdout_tx_str("\r\n");
            vstr_clear(&line);
            pyexec_mode_kind = PYEXEC_MODE_RAW_REPL;
            return 0;
        } else if (ret == CHAR_CTRL_B) {
            // reset friendly REPL
            mp_hal_stdout_tx_str("\r\n");
            goto friendly_repl_reset;
        } else if (ret == CHAR_CTRL_C) {
            // break
            mp_hal_stdout_tx_str("\r\n");
            continue;
        } else if (ret == CHAR_CTRL_D) {
            // exit for a soft reset
            mp_hal_stdout_tx_str("\r\n");
            vstr_clear(&line);
            return PYEXEC_FORCED_EXIT;
        } else if (vstr_len(&line) == 0) {
            continue;
        }

        while (mp_repl_continue_with_input(vstr_null_terminated_str(&line))) {
            vstr_add_byte(&line, '\n');
            ret = readline(&line, "... ");
            if (ret == CHAR_CTRL_C) {
                // cancel everything
                mp_hal_stdout_tx_str("\r\n");
                goto input_restart;
            } else if (ret == CHAR_CTRL_D) {
                // stop entering compound statement
                break;
            }
        }

        mp_lexer_t *lex = mp_lexer_new_from_str_len(MP_QSTR__lt_stdin_gt_, vstr_str(&line), vstr_len(&line), 0);
        if (lex == NULL) {
            printf("MemoryError\n");
        } else {
            ret = parse_compile_execute(lex, MP_PARSE_SINGLE_INPUT, EXEC_FLAG_ALLOW_DEBUGGING | EXEC_FLAG_IS_REPL);
            if (ret & PYEXEC_FORCED_EXIT) {
                return ret;
            }
        }
    }
}
예제 #9
0
int pyexec_friendly_repl_process_char(int c) {
    int ret = readline_process_char(c);

    if (!repl.cont_line) {

        if (ret == CHAR_CTRL_A) {
            // change to raw REPL
            pyexec_mode_kind = PYEXEC_MODE_RAW_REPL;
            mp_hal_stdout_tx_str("\r\n");
            vstr_clear(&repl.line);
            return PYEXEC_SWITCH_MODE;
        } else if (ret == CHAR_CTRL_B) {
            // reset friendly REPL
            mp_hal_stdout_tx_str("\r\n");
            goto friendly_repl_reset;
        } else if (ret == CHAR_CTRL_C) {
            // break
            mp_hal_stdout_tx_str("\r\n");
            goto input_restart;
        } else if (ret == CHAR_CTRL_D) {
            // exit for a soft reset
            mp_hal_stdout_tx_str("\r\n");
            vstr_clear(&repl.line);
            return PYEXEC_FORCED_EXIT;
        } else if (vstr_len(&repl.line) == 0) {
            //goto input_restart;
        }

        if (ret < 0) {
            return 0;
        }

        if (!mp_repl_continue_with_input(vstr_null_terminated_str(&repl.line))) {
            goto exec;
        }

        vstr_add_byte(&repl.line, '\n');
        repl.cont_line = true;
        mp_hal_stdout_tx_str("... ");
        readline_note_newline();
        return 0;

    } else {

        if (ret == CHAR_CTRL_C) {
                // cancel everything
                mp_hal_stdout_tx_str("\r\n");
                repl.cont_line = false;
                goto input_restart;
        } else if (ret == CHAR_CTRL_D) {
                // stop entering compound statement
                goto exec;
        }

        if (ret < 0) {
            return 0;
        }

        if (mp_repl_continue_with_input(vstr_null_terminated_str(&repl.line))) {
            vstr_add_byte(&repl.line, '\n');
            mp_hal_stdout_tx_str("... ");
            readline_note_newline();
            return 0;
        }

exec: ;
        mp_lexer_t *lex = mp_lexer_new_from_str_len(MP_QSTR__lt_stdin_gt_, vstr_str(&repl.line), vstr_len(&repl.line), 0);
        if (lex == NULL) {
            printf("MemoryError\n");
        } else {
            int ret = parse_compile_execute(lex, MP_PARSE_SINGLE_INPUT, EXEC_FLAG_ALLOW_DEBUGGING | EXEC_FLAG_IS_REPL);
            if (ret & PYEXEC_FORCED_EXIT) {
                return ret;
            }
        }

friendly_repl_reset: // TODO
input_restart:
        pyexec_friendly_repl_reset();
        mp_hal_stdout_tx_str(">>> ");
        return 0;
    }
}
예제 #10
0
파일: lexer.c 프로젝트: un33k/micropython
STATIC void mp_lexer_next_token_into(mp_lexer_t *lex, mp_token_t *tok, bool first_token) {
    // skip white space and comments
    bool had_physical_newline = false;
    while (!is_end(lex)) {
        if (is_physical_newline(lex)) {
            had_physical_newline = true;
            next_char(lex);
        } else if (is_whitespace(lex)) {
            next_char(lex);
        } else if (is_char(lex, '#')) {
            next_char(lex);
            while (!is_end(lex) && !is_physical_newline(lex)) {
                next_char(lex);
            }
            // had_physical_newline will be set on next loop
        } else if (is_char(lex, '\\')) {
            // backslash (outside string literals) must appear just before a physical newline
            next_char(lex);
            if (!is_physical_newline(lex)) {
                // SyntaxError: unexpected character after line continuation character
                tok->src_line = lex->line;
                tok->src_column = lex->column;
                tok->kind = MP_TOKEN_BAD_LINE_CONTINUATION;
                vstr_reset(&lex->vstr);
                tok->str = vstr_str(&lex->vstr);
                tok->len = 0;
                return;
            } else {
                next_char(lex);
            }
        } else {
            break;
        }
    }

    // set token source information
    tok->src_line = lex->line;
    tok->src_column = lex->column;

    // start new token text
    vstr_reset(&lex->vstr);

    if (first_token && lex->line == 1 && lex->column != 1) {
        // check that the first token is in the first column
        // if first token is not on first line, we get a physical newline and
        // this check is done as part of normal indent/dedent checking below
        // (done to get equivalence with CPython)
        tok->kind = MP_TOKEN_INDENT;

    } else if (lex->emit_dent < 0) {
        tok->kind = MP_TOKEN_DEDENT;
        lex->emit_dent += 1;

    } else if (lex->emit_dent > 0) {
        tok->kind = MP_TOKEN_INDENT;
        lex->emit_dent -= 1;

    } else if (had_physical_newline && lex->nested_bracket_level == 0) {
        tok->kind = MP_TOKEN_NEWLINE;

        uint num_spaces = lex->column - 1;
        lex->emit_dent = 0;
        if (num_spaces == indent_top(lex)) {
        } else if (num_spaces > indent_top(lex)) {
            indent_push(lex, num_spaces);
            lex->emit_dent += 1;
        } else {
            while (num_spaces < indent_top(lex)) {
                indent_pop(lex);
                lex->emit_dent -= 1;
            }
            if (num_spaces != indent_top(lex)) {
                tok->kind = MP_TOKEN_DEDENT_MISMATCH;
            }
        }

    } else if (is_end(lex)) {
        if (indent_top(lex) > 0) {
            tok->kind = MP_TOKEN_NEWLINE;
            lex->emit_dent = 0;
            while (indent_top(lex) > 0) {
                indent_pop(lex);
                lex->emit_dent -= 1;
            }
        } else {
            tok->kind = MP_TOKEN_END;
        }

    } else if (is_char_or(lex, '\'', '\"')
               || (is_char_or3(lex, 'r', 'u', 'b') && is_char_following_or(lex, '\'', '\"'))
               || ((is_char_and(lex, 'r', 'b') || is_char_and(lex, 'b', 'r')) && is_char_following_following_or(lex, '\'', '\"'))) {
        // a string or bytes literal

        // parse type codes
        bool is_raw = false;
        bool is_bytes = false;
        if (is_char(lex, 'u')) {
            next_char(lex);
        } else if (is_char(lex, 'b')) {
            is_bytes = true;
            next_char(lex);
            if (is_char(lex, 'r')) {
                is_raw = true;
                next_char(lex);
            }
        } else if (is_char(lex, 'r')) {
            is_raw = true;
            next_char(lex);
            if (is_char(lex, 'b')) {
                is_bytes = true;
                next_char(lex);
            }
        }

        // set token kind
        if (is_bytes) {
            tok->kind = MP_TOKEN_BYTES;
        } else {
            tok->kind = MP_TOKEN_STRING;
        }

        // get first quoting character
        char quote_char = '\'';
        if (is_char(lex, '\"')) {
            quote_char = '\"';
        }
        next_char(lex);

        // work out if it's a single or triple quoted literal
        int num_quotes;
        if (is_char_and(lex, quote_char, quote_char)) {
            // triple quotes
            next_char(lex);
            next_char(lex);
            num_quotes = 3;
        } else {
            // single quotes
            num_quotes = 1;
        }

        // parse the literal
        int n_closing = 0;
        while (!is_end(lex) && (num_quotes > 1 || !is_char(lex, '\n')) && n_closing < num_quotes) {
            if (is_char(lex, quote_char)) {
                n_closing += 1;
                vstr_add_char(&lex->vstr, CUR_CHAR(lex));
            } else {
                n_closing = 0;
                if (is_char(lex, '\\')) {
                    next_char(lex);
                    unichar c = CUR_CHAR(lex);
                    if (is_raw) {
                        // raw strings allow escaping of quotes, but the backslash is also emitted
                        vstr_add_char(&lex->vstr, '\\');
                    } else {
                        switch (c) {
                            case MP_LEXER_CHAR_EOF: break; // TODO a proper error message?
                            case '\n': c = MP_LEXER_CHAR_EOF; break; // TODO check this works correctly (we are supposed to ignore it
                            case '\\': break;
                            case '\'': break;
                            case '"': break;
                            case 'a': c = 0x07; break;
                            case 'b': c = 0x08; break;
                            case 't': c = 0x09; break;
                            case 'n': c = 0x0a; break;
                            case 'v': c = 0x0b; break;
                            case 'f': c = 0x0c; break;
                            case 'r': c = 0x0d; break;
                            case 'u':
                            case 'U':
                                if (is_bytes) {
                                    // b'\u1234' == b'\\u1234'
                                    vstr_add_char(&lex->vstr, '\\');
                                    break;
                                }
                                // Otherwise fall through.
                            case 'x':
                            {
                                uint num = 0;
                                if (!get_hex(lex, (c == 'x' ? 2 : c == 'u' ? 4 : 8), &num)) {
                                    // TODO error message
                                    assert(0);
                                }
                                c = num;
                                break;
                            }
                            case 'N':
                                // Supporting '\N{LATIN SMALL LETTER A}' == 'a' would require keeping the
                                // entire Unicode name table in the core. As of Unicode 6.3.0, that's nearly
                                // 3MB of text; even gzip-compressed and with minimal structure, it'll take
                                // roughly half a meg of storage. This form of Unicode escape may be added
                                // later on, but it's definitely not a priority right now. -- CJA 20140607
                                assert(!"Unicode name escapes not supported");
                                break;
                            default:
                                if (c >= '0' && c <= '7') {
                                    // Octal sequence, 1-3 chars
                                    int digits = 3;
                                    int num = c - '0';
                                    while (is_following_odigit(lex) && --digits != 0) {
                                        next_char(lex);
                                        num = num * 8 + (CUR_CHAR(lex) - '0');
                                    }
                                    c = num;
                                } else {
                                    // unrecognised escape character; CPython lets this through verbatim as '\' and then the character
                                    vstr_add_char(&lex->vstr, '\\');
                                }
                                break;
                        }
                    }
                    if (c != MP_LEXER_CHAR_EOF) {
                        if (c < 0x110000 && !is_bytes) {
                            vstr_add_char(&lex->vstr, c);
                        } else if (c < 0x100 && is_bytes) {
                            vstr_add_byte(&lex->vstr, c);
                        } else {
                            assert(!"TODO: Throw an error, invalid escape code probably");
                        }
                    }
                } else {
                    vstr_add_char(&lex->vstr, CUR_CHAR(lex));
                }
            }
            next_char(lex);
        }

        // check we got the required end quotes
        if (n_closing < num_quotes) {
            tok->kind = MP_TOKEN_LONELY_STRING_OPEN;
        }

        // cut off the end quotes from the token text
        vstr_cut_tail_bytes(&lex->vstr, n_closing);

    } else if (is_head_of_identifier(lex)) {
        tok->kind = MP_TOKEN_NAME;

        // get first char
        vstr_add_char(&lex->vstr, CUR_CHAR(lex));
        next_char(lex);

        // get tail chars
        while (!is_end(lex) && is_tail_of_identifier(lex)) {
            vstr_add_char(&lex->vstr, CUR_CHAR(lex));
            next_char(lex);
        }

    } else if (is_digit(lex) || (is_char(lex, '.') && is_following_digit(lex))) {
        tok->kind = MP_TOKEN_NUMBER;

        // get first char
        vstr_add_char(&lex->vstr, CUR_CHAR(lex));
        next_char(lex);

        // get tail chars
        while (!is_end(lex)) {
            if (is_char_or(lex, 'e', 'E')) {
                vstr_add_char(&lex->vstr, 'e');
                next_char(lex);
                if (is_char(lex, '+') || is_char(lex, '-')) {
                    vstr_add_char(&lex->vstr, CUR_CHAR(lex));
                    next_char(lex);
                }
            } else if (is_letter(lex) || is_digit(lex) || is_char_or(lex, '_', '.')) {
                vstr_add_char(&lex->vstr, CUR_CHAR(lex));
                next_char(lex);
            } else {
                break;
            }
        }

    } else if (is_char(lex, '.')) {
        // special handling for . and ... operators, because .. is not a valid operator

        // get first char
        vstr_add_char(&lex->vstr, '.');
        next_char(lex);

        if (is_char_and(lex, '.', '.')) {
            vstr_add_char(&lex->vstr, '.');
            vstr_add_char(&lex->vstr, '.');
            next_char(lex);
            next_char(lex);
            tok->kind = MP_TOKEN_ELLIPSIS;
        } else {
            tok->kind = MP_TOKEN_DEL_PERIOD;
        }

    } else {
        // search for encoded delimiter or operator

        const char *t = tok_enc;
        uint tok_enc_index = 0;
        for (; *t != 0 && !is_char(lex, *t); t += 1) {
            if (*t == 'e' || *t == 'c') {
                t += 1;
            } else if (*t == 'E') {
                tok_enc_index -= 1;
                t += 1;
            }
            tok_enc_index += 1;
        }

        next_char(lex);

        if (*t == 0) {
            // didn't match any delimiter or operator characters
            tok->kind = MP_TOKEN_INVALID;

        } else {
            // matched a delimiter or operator character

            // get the maximum characters for a valid token
            t += 1;
            uint t_index = tok_enc_index;
            for (;;) {
                for (; *t == 'e'; t += 1) {
                    t += 1;
                    t_index += 1;
                    if (is_char(lex, *t)) {
                        next_char(lex);
                        tok_enc_index = t_index;
                        break;
                    }
                }

                if (*t == 'E') {
                    t += 1;
                    if (is_char(lex, *t)) {
                        next_char(lex);
                        tok_enc_index = t_index;
                    } else {
                        tok->kind = MP_TOKEN_INVALID;
                        goto tok_enc_no_match;
                    }
                    break;
                }

                if (*t == 'c') {
                    t += 1;
                    t_index += 1;
                    if (is_char(lex, *t)) {
                        next_char(lex);
                        tok_enc_index = t_index;
                        t += 1;
                    } else {
                        break;
                    }
                } else {
                    break;
                }
            }

            // set token kind
            tok->kind = tok_enc_kind[tok_enc_index];

            tok_enc_no_match:

            // compute bracket level for implicit line joining
            if (tok->kind == MP_TOKEN_DEL_PAREN_OPEN || tok->kind == MP_TOKEN_DEL_BRACKET_OPEN || tok->kind == MP_TOKEN_DEL_BRACE_OPEN) {
                lex->nested_bracket_level += 1;
            } else if (tok->kind == MP_TOKEN_DEL_PAREN_CLOSE || tok->kind == MP_TOKEN_DEL_BRACKET_CLOSE || tok->kind == MP_TOKEN_DEL_BRACE_CLOSE) {
                lex->nested_bracket_level -= 1;
            }
        }
    }

    // point token text to vstr buffer
    tok->str = vstr_str(&lex->vstr);
    tok->len = vstr_len(&lex->vstr);

    // check for keywords
    if (tok->kind == MP_TOKEN_NAME) {
        // We check for __debug__ here and convert it to its value.  This is so
        // the parser gives a syntax error on, eg, x.__debug__.  Otherwise, we
        // need to check for this special token in many places in the compiler.
        // TODO improve speed of these string comparisons
        //for (int i = 0; tok_kw[i] != NULL; i++) {
        for (int i = 0; i < MP_ARRAY_SIZE(tok_kw); i++) {
            if (str_strn_equal(tok_kw[i], tok->str, tok->len)) {
                if (i == MP_ARRAY_SIZE(tok_kw) - 1) {
                    // tok_kw[MP_ARRAY_SIZE(tok_kw) - 1] == "__debug__"
                    tok->kind = (mp_optimise_value == 0 ? MP_TOKEN_KW_TRUE : MP_TOKEN_KW_FALSE);
                } else {
                    tok->kind = MP_TOKEN_KW_FALSE + i;
                }
                break;
            }
        }
    }
}
예제 #11
0
mp_obj_t mp_builtin___import__(uint n_args, mp_obj_t *args) {
#if DEBUG_PRINT
    printf("__import__:\n");
    for (int i = 0; i < n_args; i++) {
        printf("  ");
        mp_obj_print(args[i], PRINT_REPR);
        printf("\n");
    }
#endif

    mp_obj_t module_name = args[0];
    mp_obj_t fromtuple = mp_const_none;
    int level = 0;
    if (n_args >= 4) {
        fromtuple = args[3];
        if (n_args >= 5) {
            level = MP_OBJ_SMALL_INT_VALUE(args[4]);
        }
    }

    uint mod_len;
    const char *mod_str = (const char*)mp_obj_str_get_data(module_name, &mod_len);

    if (level != 0) {
        // What we want to do here is to take name of current module,
        // chop <level> trailing components, and concatenate with passed-in
        // module name, thus resolving relative import name into absolue.
        // This even appears to be correct per
        // http://legacy.python.org/dev/peps/pep-0328/#relative-imports-and-name
        // "Relative imports use a module's __name__ attribute to determine that
        // module's position in the package hierarchy."
        mp_obj_t this_name_q = mp_obj_dict_get(mp_globals_get(), MP_OBJ_NEW_QSTR(MP_QSTR___name__));
        assert(this_name_q != MP_OBJ_NULL);
#if DEBUG_PRINT
        printf("Current module: ");
        mp_obj_print(this_name_q, PRINT_REPR);
        printf("\n");
#endif

        uint this_name_l;
        const char *this_name = (const char*)mp_obj_str_get_data(this_name_q, &this_name_l);

        uint dots_seen = 0;
        const char *p = this_name + this_name_l - 1;
        while (p > this_name) {
            if (*p == '.') {
                dots_seen++;
                if (--level == 0) {
                    break;
                }
            }
            p--;
        }

        if (dots_seen == 0 && level == 1) {
            // http://legacy.python.org/dev/peps/pep-0328/#relative-imports-and-name
            // "If the module's name does not contain any package information
            // (e.g. it is set to '__main__') then relative imports are
            // resolved as if the module were a top level module, regardless
            // of where the module is actually located on the file system."
            // Supposedly this if catches this condition and resolve it properly
            // TODO: But nobody knows for sure. This condition happens when
            // package's __init__.py does something like "import .submod". So,
            // maybe we should check for package here? But quote above doesn't
            // talk about packages, it talks about dot-less module names.
            p = this_name + this_name_l;
        } else if (level != 0) {
            nlr_raise(mp_obj_new_exception_msg(&mp_type_ImportError, "Invalid relative import"));
        }

        uint new_mod_l = (mod_len == 0 ? p - this_name : p - this_name + 1 + mod_len);
        char *new_mod = alloca(new_mod_l);
        memcpy(new_mod, this_name, p - this_name);
        if (mod_len != 0) {
            new_mod[p - this_name] = '.';
            memcpy(new_mod + (p - this_name) + 1, mod_str, mod_len);
        }

        qstr new_mod_q = qstr_from_strn(new_mod, new_mod_l);
        DEBUG_printf("Resolved relative name: %s\n", qstr_str(new_mod_q));
        module_name = MP_OBJ_NEW_QSTR(new_mod_q);
        mod_str = new_mod;
        mod_len = new_mod_l;
    }

    // check if module already exists
    mp_obj_t module_obj = mp_module_get(mp_obj_str_get_qstr(module_name));
    if (module_obj != MP_OBJ_NULL) {
        DEBUG_printf("Module already loaded\n");
        // If it's not a package, return module right away
        char *p = strchr(mod_str, '.');
        if (p == NULL) {
            return module_obj;
        }
        // If fromlist is not empty, return leaf module
        if (fromtuple != mp_const_none) {
            return module_obj;
        }
        // Otherwise, we need to return top-level package
        qstr pkg_name = qstr_from_strn(mod_str, p - mod_str);
        return mp_module_get(pkg_name);
    }
    DEBUG_printf("Module not yet loaded\n");

    uint last = 0;
    VSTR_FIXED(path, MICROPY_ALLOC_PATH_MAX)
    module_obj = MP_OBJ_NULL;
    mp_obj_t top_module_obj = MP_OBJ_NULL;
    mp_obj_t outer_module_obj = MP_OBJ_NULL;
    uint i;
    for (i = 1; i <= mod_len; i++) {
        if (i == mod_len || mod_str[i] == '.') {
            // create a qstr for the module name up to this depth
            qstr mod_name = qstr_from_strn(mod_str, i);
            DEBUG_printf("Processing module: %s\n", qstr_str(mod_name));
            DEBUG_printf("Previous path: %s\n", vstr_str(&path));

            // find the file corresponding to the module name
            mp_import_stat_t stat;
            if (vstr_len(&path) == 0) {
                // first module in the dotted-name; search for a directory or file
                stat = find_file(mod_str, i, &path);
            } else {
                // latter module in the dotted-name; append to path
                vstr_add_char(&path, PATH_SEP_CHAR);
                vstr_add_strn(&path, mod_str + last, i - last);
                stat = stat_dir_or_file(&path);
            }
            DEBUG_printf("Current path: %s\n", vstr_str(&path));

            // fail if we couldn't find the file
            if (stat == MP_IMPORT_STAT_NO_EXIST) {
                nlr_raise(mp_obj_new_exception_msg_varg(&mp_type_ImportError, "No module named '%s'", qstr_str(mod_name)));
            }

            module_obj = mp_module_get(mod_name);
            if (module_obj == MP_OBJ_NULL) {
                // module not already loaded, so load it!

                module_obj = mp_obj_new_module(mod_name);

                if (stat == MP_IMPORT_STAT_DIR) {
                    DEBUG_printf("%s is dir\n", vstr_str(&path));
                    // https://docs.python.org/3/reference/import.html
                    // "Specifically, any module that contains a __path__ attribute is considered a package."
                    mp_store_attr(module_obj, MP_QSTR___path__, mp_obj_new_str(vstr_str(&path), vstr_len(&path), false));
                    vstr_add_char(&path, PATH_SEP_CHAR);
                    vstr_add_str(&path, "__init__.py");
                    if (mp_import_stat(vstr_str(&path)) != MP_IMPORT_STAT_FILE) {
                        vstr_cut_tail_bytes(&path, sizeof("/__init__.py") - 1); // cut off /__init__.py
                        printf("Notice: %s is imported as namespace package\n", vstr_str(&path));
                    } else {
                        do_load(module_obj, &path);
                        vstr_cut_tail_bytes(&path, sizeof("/__init__.py") - 1); // cut off /__init__.py
                    }
                } else { // MP_IMPORT_STAT_FILE
                    do_load(module_obj, &path);
                    // TODO: We cannot just break here, at the very least, we must execute
                    // trailer code below. But otherwise if there're remaining components,
                    // that would be (??) object path within module, not modules path within FS.
                    // break;
                }
            }
            if (outer_module_obj != MP_OBJ_NULL) {
                qstr s = qstr_from_strn(mod_str + last, i - last);
                mp_store_attr(outer_module_obj, s, module_obj);
            }
            outer_module_obj = module_obj;
            if (top_module_obj == MP_OBJ_NULL) {
                top_module_obj = module_obj;
            }
            last = i + 1;
        }
    }

    if (i < mod_len) {
        // we loaded a package, now need to load objects from within that package
        // TODO
        assert(0);
    }

    // If fromlist is not empty, return leaf module
    if (fromtuple != mp_const_none) {
        return module_obj;
    }
    // Otherwise, we need to return top-level package
    return top_module_obj;
}
예제 #12
0
mp_obj_t mp_builtin___import__(size_t n_args, const mp_obj_t *args) {
#if DEBUG_PRINT
    DEBUG_printf("__import__:\n");
    for (size_t i = 0; i < n_args; i++) {
        DEBUG_printf("  ");
        mp_obj_print(args[i], PRINT_REPR);
        DEBUG_printf("\n");
    }
#endif

    mp_obj_t module_name = args[0];
    mp_obj_t fromtuple = mp_const_none;
    mp_int_t level = 0;
    if (n_args >= 4) {
        fromtuple = args[3];
        if (n_args >= 5) {
            level = MP_OBJ_SMALL_INT_VALUE(args[4]);
            if (level < 0) {
                mp_raise_ValueError(NULL);
            }
        }
    }

    size_t mod_len;
    const char *mod_str = mp_obj_str_get_data(module_name, &mod_len);

    if (level != 0) {
        // What we want to do here is to take name of current module,
        // chop <level> trailing components, and concatenate with passed-in
        // module name, thus resolving relative import name into absolute.
        // This even appears to be correct per
        // http://legacy.python.org/dev/peps/pep-0328/#relative-imports-and-name
        // "Relative imports use a module's __name__ attribute to determine that
        // module's position in the package hierarchy."
        level--;
        mp_obj_t this_name_q = mp_obj_dict_get(MP_OBJ_FROM_PTR(mp_globals_get()), MP_OBJ_NEW_QSTR(MP_QSTR___name__));
        assert(this_name_q != MP_OBJ_NULL);
        #if MICROPY_CPYTHON_COMPAT
        if (MP_OBJ_QSTR_VALUE(this_name_q) == MP_QSTR___main__) {
            // This is a module run by -m command-line switch, get its real name from backup attribute
            this_name_q = mp_obj_dict_get(MP_OBJ_FROM_PTR(mp_globals_get()), MP_OBJ_NEW_QSTR(MP_QSTR___main__));
        }
        #endif
        mp_map_t *globals_map = &mp_globals_get()->map;
        mp_map_elem_t *elem = mp_map_lookup(globals_map, MP_OBJ_NEW_QSTR(MP_QSTR___path__), MP_MAP_LOOKUP);
        bool is_pkg = (elem != NULL);

#if DEBUG_PRINT
        DEBUG_printf("Current module/package: ");
        mp_obj_print(this_name_q, PRINT_REPR);
        DEBUG_printf(", is_package: %d", is_pkg);
        DEBUG_printf("\n");
#endif

        size_t this_name_l;
        const char *this_name = mp_obj_str_get_data(this_name_q, &this_name_l);

        const char *p = this_name + this_name_l;
        if (!is_pkg) {
            // We have module, but relative imports are anchored at package, so
            // go there.
            chop_component(this_name, &p);
        }

        while (level--) {
            chop_component(this_name, &p);
        }

        // We must have some component left over to import from
        if (p == this_name) {
            mp_raise_ValueError("cannot perform relative import");
        }

        uint new_mod_l = (mod_len == 0 ? (size_t)(p - this_name) : (size_t)(p - this_name) + 1 + mod_len);
        char *new_mod = mp_local_alloc(new_mod_l);
        memcpy(new_mod, this_name, p - this_name);
        if (mod_len != 0) {
            new_mod[p - this_name] = '.';
            memcpy(new_mod + (p - this_name) + 1, mod_str, mod_len);
        }

        qstr new_mod_q = qstr_from_strn(new_mod, new_mod_l);
        mp_local_free(new_mod);
        DEBUG_printf("Resolved base name for relative import: '%s'\n", qstr_str(new_mod_q));
        module_name = MP_OBJ_NEW_QSTR(new_mod_q);
        mod_str = qstr_str(new_mod_q);
        mod_len = new_mod_l;
    }

    // check if module already exists
    qstr module_name_qstr = mp_obj_str_get_qstr(module_name);
    mp_obj_t module_obj = mp_module_get(module_name_qstr);
    if (module_obj != MP_OBJ_NULL) {
        DEBUG_printf("Module already loaded\n");
        // If it's not a package, return module right away
        char *p = strchr(mod_str, '.');
        if (p == NULL) {
            return module_obj;
        }
        // If fromlist is not empty, return leaf module
        if (fromtuple != mp_const_none) {
            return module_obj;
        }
        // Otherwise, we need to return top-level package
        qstr pkg_name = qstr_from_strn(mod_str, p - mod_str);
        return mp_module_get(pkg_name);
    }
    DEBUG_printf("Module not yet loaded\n");

    uint last = 0;
    VSTR_FIXED(path, MICROPY_ALLOC_PATH_MAX)
    module_obj = MP_OBJ_NULL;
    mp_obj_t top_module_obj = MP_OBJ_NULL;
    mp_obj_t outer_module_obj = MP_OBJ_NULL;
    uint i;
    for (i = 1; i <= mod_len; i++) {
        if (i == mod_len || mod_str[i] == '.') {
            // create a qstr for the module name up to this depth
            qstr mod_name = qstr_from_strn(mod_str, i);
            DEBUG_printf("Processing module: %s\n", qstr_str(mod_name));
            DEBUG_printf("Previous path: =%.*s=\n", vstr_len(&path), vstr_str(&path));

            // find the file corresponding to the module name
            mp_import_stat_t stat;
            if (vstr_len(&path) == 0) {
                // first module in the dotted-name; search for a directory or file
                stat = find_file(mod_str, i, &path);
            } else {
                // latter module in the dotted-name; append to path
                vstr_add_char(&path, PATH_SEP_CHAR);
                vstr_add_strn(&path, mod_str + last, i - last);
                stat = stat_dir_or_file(&path);
            }
            DEBUG_printf("Current path: %.*s\n", vstr_len(&path), vstr_str(&path));

            if (stat == MP_IMPORT_STAT_NO_EXIST) {
                #if MICROPY_MODULE_WEAK_LINKS
                // check if there is a weak link to this module
                if (i == mod_len) {
                    mp_map_elem_t *el = mp_map_lookup((mp_map_t*)&mp_builtin_module_weak_links_map, MP_OBJ_NEW_QSTR(mod_name), MP_MAP_LOOKUP);
                    if (el == NULL) {
                        goto no_exist;
                    }
                    // found weak linked module
                    module_obj = el->value;
                    mp_module_call_init(mod_name, module_obj);
                } else {
                    no_exist:
                #else
                {
                #endif
                    // couldn't find the file, so fail
                    if (MICROPY_ERROR_REPORTING == MICROPY_ERROR_REPORTING_TERSE) {
                        mp_raise_msg(&mp_type_ImportError, "module not found");
                    } else {
                        nlr_raise(mp_obj_new_exception_msg_varg(&mp_type_ImportError,
                            "no module named '%q'", mod_name));
                    }
                }
            } else {
                // found the file, so get the module
                module_obj = mp_module_get(mod_name);
            }

            if (module_obj == MP_OBJ_NULL) {
                // module not already loaded, so load it!

                module_obj = mp_obj_new_module(mod_name);

                // if args[3] (fromtuple) has magic value False, set up
                // this module for command-line "-m" option (set module's
                // name to __main__ instead of real name). Do this only
                // for *modules* however - packages never have their names
                // replaced, instead they're -m'ed using a special __main__
                // submodule in them. (This all apparently is done to not
                // touch package name itself, which is important for future
                // imports).
                if (i == mod_len && fromtuple == mp_const_false && stat != MP_IMPORT_STAT_DIR) {
                    mp_obj_module_t *o = MP_OBJ_TO_PTR(module_obj);
                    mp_obj_dict_store(MP_OBJ_FROM_PTR(o->globals), MP_OBJ_NEW_QSTR(MP_QSTR___name__), MP_OBJ_NEW_QSTR(MP_QSTR___main__));
                    #if MICROPY_CPYTHON_COMPAT
                    // Store module as "__main__" in the dictionary of loaded modules (returned by sys.modules).
                    mp_obj_dict_store(MP_OBJ_FROM_PTR(&MP_STATE_VM(mp_loaded_modules_dict)), MP_OBJ_NEW_QSTR(MP_QSTR___main__), module_obj);
                    // Store real name in "__main__" attribute. Chosen semi-randonly, to reuse existing qstr's.
                    mp_obj_dict_store(MP_OBJ_FROM_PTR(o->globals), MP_OBJ_NEW_QSTR(MP_QSTR___main__), MP_OBJ_NEW_QSTR(mod_name));
                    #endif
                }

                if (stat == MP_IMPORT_STAT_DIR) {
                    DEBUG_printf("%.*s is dir\n", vstr_len(&path), vstr_str(&path));
                    // https://docs.python.org/3/reference/import.html
                    // "Specifically, any module that contains a __path__ attribute is considered a package."
                    mp_store_attr(module_obj, MP_QSTR___path__, mp_obj_new_str(vstr_str(&path), vstr_len(&path)));
                    size_t orig_path_len = path.len;
                    vstr_add_char(&path, PATH_SEP_CHAR);
                    vstr_add_str(&path, "__init__.py");
                    if (stat_file_py_or_mpy(&path) != MP_IMPORT_STAT_FILE) {
                        //mp_warning("%s is imported as namespace package", vstr_str(&path));
                    } else {
                        do_load(module_obj, &path);
                    }
                    path.len = orig_path_len;
                } else { // MP_IMPORT_STAT_FILE
                    do_load(module_obj, &path);
                    // This should be the last component in the import path.  If there are
                    // remaining components then it's an ImportError because the current path
                    // (the module that was just loaded) is not a package.  This will be caught
                    // on the next iteration because the file will not exist.
                }
            }
            if (outer_module_obj != MP_OBJ_NULL) {
                qstr s = qstr_from_strn(mod_str + last, i - last);
                mp_store_attr(outer_module_obj, s, module_obj);
            }
            outer_module_obj = module_obj;
            if (top_module_obj == MP_OBJ_NULL) {
                top_module_obj = module_obj;
            }
            last = i + 1;
        }
    }
예제 #13
0
STATIC int pyexec_friendly_repl_process_char(int c) {
    int ret = readline_process_char(c);

    if (!repl.cont_line) {

        if (ret == CHAR_CTRL_A) {
            // change to raw REPL
            pyexec_mode_kind = PYEXEC_MODE_RAW_REPL;
            mp_hal_stdout_tx_str("\r\n");
            pyexec_raw_repl_process_char(CHAR_CTRL_A);
            return 0;
        } else if (ret == CHAR_CTRL_B) {
            // reset friendly REPL
            mp_hal_stdout_tx_str("\r\n");
            mp_hal_stdout_tx_str("Micro Python " MICROPY_GIT_TAG " on " MICROPY_BUILD_DATE "; " MICROPY_HW_BOARD_NAME " with " MICROPY_HW_MCU_NAME "\r\n");
            mp_hal_stdout_tx_str("Type \"help()\" for more information.\r\n");
            goto input_restart;
        } else if (ret == CHAR_CTRL_C) {
            // break
            mp_hal_stdout_tx_str("\r\n");
            goto input_restart;
        } else if (ret == CHAR_CTRL_D) {
            // exit for a soft reset
            mp_hal_stdout_tx_str("\r\n");
            vstr_clear(&repl.line);
            return PYEXEC_FORCED_EXIT;
        }

        if (ret < 0) {
            return 0;
        }

        if (!mp_repl_continue_with_input(vstr_null_terminated_str(&repl.line))) {
            goto exec;
        }

        vstr_add_byte(&repl.line, '\n');
        repl.cont_line = true;
        readline_note_newline("... ");
        return 0;

    } else {

        if (ret == CHAR_CTRL_C) {
           // cancel everything
           mp_hal_stdout_tx_str("\r\n");
           repl.cont_line = false;
           goto input_restart;
        } else if (ret == CHAR_CTRL_D) {
            // stop entering compound statement
            goto exec;
        }

        if (ret < 0) {
            return 0;
        }

        if (mp_repl_continue_with_input(vstr_null_terminated_str(&repl.line))) {
            vstr_add_byte(&repl.line, '\n');
            readline_note_newline("... ");
            return 0;
        }

exec: ;
        mp_lexer_t *lex = mp_lexer_new_from_str_len(MP_QSTR__lt_stdin_gt_, vstr_str(&repl.line), vstr_len(&repl.line), 0);
        if (lex == NULL) {
            printf("MemoryError\n");
        } else {
            int ret = parse_compile_execute(lex, MP_PARSE_SINGLE_INPUT, EXEC_FLAG_ALLOW_DEBUGGING | EXEC_FLAG_IS_REPL);
            if (ret & PYEXEC_FORCED_EXIT) {
                return ret;
            }
        }

input_restart:
        vstr_reset(&repl.line);
        repl.cont_line = false;
        readline_init(&repl.line, ">>> ");
        return 0;
    }
}
mp_obj_t mp_builtin___import__(uint n_args, mp_obj_t *args) {
    /*
    printf("import:\n");
    for (int i = 0; i < n_args; i++) {
        printf("  ");
        mp_obj_print(args[i], PRINT_REPR);
        printf("\n");
    }
    */

    mp_obj_t fromtuple = mp_const_none;
    int level = 0;
    if (n_args >= 4) {
        fromtuple = args[3];
        if (n_args >= 5) {
            level = MP_OBJ_SMALL_INT_VALUE(args[4]);
        }
    }

    if (level != 0) {
        nlr_jump(mp_obj_new_exception_msg(&mp_type_NotImplementedError,
            "Relative import is not implemented"));
    }

    uint mod_len;
    const char *mod_str = (const char*)mp_obj_str_get_data(args[0], &mod_len);

    // check if module already exists
    mp_obj_t module_obj = mp_module_get(mp_obj_str_get_qstr(args[0]));
    if (module_obj != MP_OBJ_NULL) {
        // If it's not a package, return module right away
        char *p = strchr(mod_str, '.');
        if (p == NULL) {
            return module_obj;
        }
        // If fromlist is not empty, return leaf module
        if (fromtuple != mp_const_none) {
            return module_obj;
        }
        // Otherwise, we need to return top-level package
        qstr pkg_name = qstr_from_strn(mod_str, p - mod_str);
        return mp_module_get(pkg_name);
    }

    uint last = 0;
    VSTR_FIXED(path, MICROPY_PATH_MAX)
    module_obj = MP_OBJ_NULL;
    mp_obj_t top_module_obj = MP_OBJ_NULL;
    mp_obj_t outer_module_obj = MP_OBJ_NULL;
    uint i;
    for (i = 1; i <= mod_len; i++) {
        if (i == mod_len || mod_str[i] == '.') {
            // create a qstr for the module name up to this depth
            qstr mod_name = qstr_from_strn(mod_str, i);

            // find the file corresponding to the module name
            mp_import_stat_t stat;
            if (vstr_len(&path) == 0) {
                // first module in the dotted-name; search for a directory or file
                stat = find_file(mod_str, i, &path);
            } else {
                // latter module in the dotted-name; append to path
                vstr_add_char(&path, PATH_SEP_CHAR);
                vstr_add_strn(&path, mod_str + last, i - last);
                stat = stat_dir_or_file(&path);
            }

            // fail if we couldn't find the file
            if (stat == MP_IMPORT_STAT_NO_EXIST) {
                nlr_jump(mp_obj_new_exception_msg_varg(&mp_type_ImportError, "ImportError: No module named '%s'", qstr_str(mod_name)));
            }

            module_obj = mp_module_get(mod_name);
            if (module_obj == MP_OBJ_NULL) {
                // module not already loaded, so load it!

                module_obj = mp_obj_new_module(mod_name);

                if (stat == MP_IMPORT_STAT_DIR) {
                    vstr_add_char(&path, PATH_SEP_CHAR);
                    vstr_add_str(&path, "__init__.py");
                    if (mp_import_stat(vstr_str(&path)) != MP_IMPORT_STAT_FILE) {
                        vstr_cut_tail_bytes(&path, sizeof("/__init__.py") - 1); // cut off /__init__.py
                        nlr_jump(mp_obj_new_exception_msg_varg(&mp_type_ImportError,
                            "Per PEP-420 a dir without __init__.py (%s) is a namespace package; "
                            "namespace packages are not supported", vstr_str(&path)));
                    }
                    do_load(module_obj, &path);
                    vstr_cut_tail_bytes(&path, sizeof("/__init__.py") - 1); // cut off /__init__.py
                } else { // MP_IMPORT_STAT_FILE
                    do_load(module_obj, &path);
                    // TODO: We cannot just break here, at the very least, we must execute
                    // trailer code below. But otherwise if there're remaining components,
                    // that would be (??) object path within module, not modules path within FS.
                    // break;
                }
            }
            if (outer_module_obj != MP_OBJ_NULL) {
                qstr s = qstr_from_strn(mod_str + last, i - last);
                mp_store_attr(outer_module_obj, s, module_obj);
            }
            outer_module_obj = module_obj;
            if (top_module_obj == MP_OBJ_NULL) {
                top_module_obj = module_obj;
            }
            last = i + 1;
        }
    }

    if (i < mod_len) {
        // we loaded a package, now need to load objects from within that package
        // TODO
        assert(0);
    }

    // If fromlist is not empty, return leaf module
    if (fromtuple != mp_const_none) {
        return module_obj;
    }
    // Otherwise, we need to return top-level package
    return top_module_obj;
}
예제 #15
0
파일: main.c 프로젝트: QYQ11544/openmv
int main(void)
{
    // Stack limit should be less than real stack size, so we
    // had chance to recover from limit hit.
    mp_stack_set_limit((char*)&_ram_end - (char*)&_heap_end - 1024);

    /* STM32F4xx HAL library initialization:
       - Configure the Flash prefetch, instruction and Data caches
       - Configure the Systick to generate an interrupt each 1 msec
       - Set NVIC Group Priority to 4
       - Global MSP (MCU Support Package) initialization
    */
    HAL_Init();

    // basic sub-system init
    pendsv_init();
    timer_tim3_init();
    led_init();

soft_reset:
    // check if user switch held to select the reset mode
    led_state(LED_RED, 1);
    led_state(LED_GREEN, 1);
    led_state(LED_BLUE, 1);

#if MICROPY_HW_ENABLE_RTC
    rtc_init();
#endif

    // GC init
    gc_init(&_heap_start, &_heap_end);

    // Micro Python init
    mp_init();
    mp_obj_list_init(mp_sys_path, 0);
    mp_obj_list_init(mp_sys_argv, 0);

    readline_init0();
    pin_init0();
    extint_init0();
    timer_init0();
    rng_init0();
    i2c_init0();
    spi_init0();
    uart_init0();
    pyb_usb_init0();

    usbdbg_init();

    if (sensor_init() != 0) {
        __fatal_error("Failed to init sensor");
    }

    /* Export functions to the global python namespace */
    mp_store_global(qstr_from_str("randint"),           (mp_obj_t)&py_randint_obj);
    mp_store_global(qstr_from_str("cpu_freq"),          (mp_obj_t)&py_cpu_freq_obj);
    mp_store_global(qstr_from_str("Image"),             (mp_obj_t)&py_image_load_image_obj);
    mp_store_global(qstr_from_str("HaarCascade"),       (mp_obj_t)&py_image_load_cascade_obj);
    mp_store_global(qstr_from_str("FreakDesc"),         (mp_obj_t)&py_image_load_descriptor_obj);
    mp_store_global(qstr_from_str("FreakDescSave"),     (mp_obj_t)&py_image_save_descriptor_obj);
    mp_store_global(qstr_from_str("LBPDesc"),           (mp_obj_t)&py_image_load_lbp_obj);
    mp_store_global(qstr_from_str("vcp_is_connected"),  (mp_obj_t)&py_vcp_is_connected_obj);

    if (sdcard_is_present()) {
        sdcard_init();
        FRESULT res = f_mount(&fatfs, "1:", 1);
        if (res != FR_OK) {
            __fatal_error("could not mount SD\n");
        }
        // Set CWD and USB medium to SD
        f_chdrive("1:");
        pyb_usb_storage_medium = PYB_USB_STORAGE_MEDIUM_SDCARD;
    } else {
        storage_init();
        // try to mount the flash
        FRESULT res = f_mount(&fatfs, "0:", 1);
        if (res == FR_NO_FILESYSTEM) {
            // create a fresh fs
            make_flash_fs();
        } else if (res != FR_OK) {
            __fatal_error("could not access LFS\n");
        }

        // Set CWD and USB medium to flash
        f_chdrive("0:");
        pyb_usb_storage_medium = PYB_USB_STORAGE_MEDIUM_FLASH;
    }

    // turn boot-up LEDs off
    led_state(LED_RED, 0);
    led_state(LED_GREEN, 0);
    led_state(LED_BLUE, 0);

    // init USB device to default setting if it was not already configured
    if (!(pyb_usb_flags & PYB_USB_FLAG_USB_MODE_CALLED)) {
        pyb_usb_dev_init(USBD_VID, USBD_PID_CDC_MSC, USBD_MODE_CDC_MSC, NULL);
    }

    // Run the main script from the current directory.
    FRESULT res = f_stat("main.py", NULL);
    if (res == FR_OK) {
        if (!pyexec_file("main.py")) {
            nlr_buf_t nlr;
            if (nlr_push(&nlr) == 0) {
                flash_error(3);
                nlr_pop();
            }
        }
    }

    // Enter REPL
    nlr_buf_t nlr;
    for (;;) {
        if (nlr_push(&nlr) == 0) {
            while (usbdbg_script_ready()) {
                nlr_buf_t nlr;
                vstr_t *script_buf = usbdbg_get_script();
                // clear script flag
                usbdbg_clr_script();

                // execute the script
                if (nlr_push(&nlr) == 0) {
                    pyexec_push_scope();

                    // parse and compile script
                    mp_lexer_t *lex = mp_lexer_new_from_str_len(MP_QSTR__lt_stdin_gt_,
                            vstr_str(script_buf), vstr_len(script_buf), 0);
                    mp_parse_node_t pn = mp_parse(lex, MP_PARSE_FILE_INPUT);
                    mp_obj_t script = mp_compile(pn, lex->source_name, MP_EMIT_OPT_NONE, false);

                    // execute the script
                    mp_call_function_0(script);
                    nlr_pop();
                } else {
                    mp_obj_print_exception(&mp_plat_print, (mp_obj_t)nlr.ret_val);
                }
                pyexec_pop_scope();
            }

            // clear script flag
            usbdbg_clr_script();

            // no script run REPL
            pyexec_friendly_repl();

            nlr_pop();
        }

    }

    printf("PYB: sync filesystems\n");
    storage_flush();

    printf("PYB: soft reboot\n");

    goto soft_reset;
}
예제 #16
0
static void mp_lexer_next_token_into(mp_lexer_t *lex, mp_token_t *tok, bool first_token) {
    // skip white space and comments
    bool had_physical_newline = false;
    while (!is_end(lex)) {
        if (is_physical_newline(lex)) {
            had_physical_newline = true;
            next_char(lex);
        } else if (is_whitespace(lex)) {
            next_char(lex);
        } else if (is_char(lex, '#')) {
            next_char(lex);
            while (!is_end(lex) && !is_physical_newline(lex)) {
                next_char(lex);
            }
            // had_physical_newline will be set on next loop
        } else if (is_char(lex, '\\')) {
            // backslash (outside string literals) must appear just before a physical newline
            next_char(lex);
            if (!is_physical_newline(lex)) {
                // TODO SyntaxError
                assert(0);
            } else {
                next_char(lex);
            }
        } else {
            break;
        }
    }

    // set token source information
    tok->src_name = lex->name;
    tok->src_line = lex->line;
    tok->src_column = lex->column;

    // start new token text
    vstr_reset(&lex->vstr);

    if (first_token && lex->line == 1 && lex->column != 1) {
        // check that the first token is in the first column
        // if first token is not on first line, we get a physical newline and
        // this check is done as part of normal indent/dedent checking below
        // (done to get equivalence with CPython)
        tok->kind = MP_TOKEN_INDENT;

    } else if (lex->emit_dent < 0) {
        tok->kind = MP_TOKEN_DEDENT;
        lex->emit_dent += 1;

    } else if (lex->emit_dent > 0) {
        tok->kind = MP_TOKEN_INDENT;
        lex->emit_dent -= 1;

    } else if (had_physical_newline && lex->nested_bracket_level == 0) {
        tok->kind = MP_TOKEN_NEWLINE;

        uint num_spaces = lex->column - 1;
        lex->emit_dent = 0;
        if (num_spaces == indent_top(lex)) {
        } else if (num_spaces > indent_top(lex)) {
            indent_push(lex, num_spaces);
            lex->emit_dent += 1;
        } else {
            while (num_spaces < indent_top(lex)) {
                indent_pop(lex);
                lex->emit_dent -= 1;
            }
            if (num_spaces != indent_top(lex)) {
                tok->kind = MP_TOKEN_DEDENT_MISMATCH;
            }
        }

    } else if (is_end(lex)) {
        if (indent_top(lex) > 0) {
            tok->kind = MP_TOKEN_NEWLINE;
            lex->emit_dent = 0;
            while (indent_top(lex) > 0) {
                indent_pop(lex);
                lex->emit_dent -= 1;
            }
        } else {
            tok->kind = MP_TOKEN_END;
        }

    } else if (is_char_or(lex, '\'', '\"')
               || (is_char_or3(lex, 'r', 'u', 'b') && is_char_following_or(lex, '\'', '\"'))
               || ((is_char_and(lex, 'r', 'b') || is_char_and(lex, 'b', 'r')) && is_char_following_following_or(lex, '\'', '\"'))) {
        // a string or bytes literal

        // parse type codes
        bool is_raw = false;
        bool is_bytes = false;
        if (is_char(lex, 'u')) {
            next_char(lex);
        } else if (is_char(lex, 'b')) {
            is_bytes = true;
            next_char(lex);
            if (is_char(lex, 'r')) {
                is_raw = true;
                next_char(lex);
            }
        } else if (is_char(lex, 'r')) {
            is_raw = true;
            next_char(lex);
            if (is_char(lex, 'b')) {
                is_bytes = true;
                next_char(lex);
            }
        }

        // set token kind
        if (is_bytes) {
            tok->kind = MP_TOKEN_BYTES;
        } else {
            tok->kind = MP_TOKEN_STRING;
        }

        // get first quoting character
        char quote_char = '\'';
        if (is_char(lex, '\"')) {
            quote_char = '\"';
        }
        next_char(lex);

        // work out if it's a single or triple quoted literal
        int num_quotes;
        if (is_char_and(lex, quote_char, quote_char)) {
            // triple quotes
            next_char(lex);
            next_char(lex);
            num_quotes = 3;
        } else {
            // single quotes
            num_quotes = 1;
        }

        // parse the literal
        int n_closing = 0;
        while (!is_end(lex) && (num_quotes > 1 || !is_char(lex, '\n')) && n_closing < num_quotes) {
            if (is_char(lex, quote_char)) {
                n_closing += 1;
                vstr_add_char(&lex->vstr, CUR_CHAR(lex));
            } else {
                n_closing = 0;
                if (!is_raw && is_char(lex, '\\')) {
                    next_char(lex);
                    unichar c = CUR_CHAR(lex);
                    switch (c) {
                        case MP_LEXER_CHAR_EOF: break; // TODO a proper error message?
                        case '\n': c = MP_LEXER_CHAR_EOF; break; // TODO check this works correctly (we are supposed to ignore it
                        case '\\': break;
                        case '\'': break;
                        case '"': break;
                        case 'a': c = 0x07; break;
                        case 'b': c = 0x08; break;
                        case 't': c = 0x09; break;
                        case 'n': c = 0x0a; break;
                        case 'v': c = 0x0b; break;
                        case 'f': c = 0x0c; break;
                        case 'r': c = 0x0d; break;
                        // TODO \ooo octal
                        case 'x': // TODO \xhh
                        case 'N': // TODO \N{name} only in strings
                        case 'u': // TODO \uxxxx only in strings
                        case 'U': // TODO \Uxxxxxxxx only in strings
                        default: break; // TODO error message
                    }
                    if (c != MP_LEXER_CHAR_EOF) {
                        vstr_add_char(&lex->vstr, c);
                    }
                } else {
                    vstr_add_char(&lex->vstr, CUR_CHAR(lex));
                }
            }
            next_char(lex);
        }

        // check we got the required end quotes
        if (n_closing < num_quotes) {
            tok->kind = MP_TOKEN_LONELY_STRING_OPEN;
        }

        // cut off the end quotes from the token text
        vstr_cut_tail(&lex->vstr, n_closing);

    } else if (is_head_of_identifier(lex)) {
        tok->kind = MP_TOKEN_NAME;

        // get first char
        vstr_add_char(&lex->vstr, CUR_CHAR(lex));
        next_char(lex);

        // get tail chars
        while (!is_end(lex) && is_tail_of_identifier(lex)) {
            vstr_add_char(&lex->vstr, CUR_CHAR(lex));
            next_char(lex);
        }

    } else if (is_digit(lex) || (is_char(lex, '.') && is_following_digit(lex))) {
        tok->kind = MP_TOKEN_NUMBER;

        // get first char
        vstr_add_char(&lex->vstr, CUR_CHAR(lex));
        next_char(lex);

        // get tail chars
        while (!is_end(lex)) {
            if (is_char_or(lex, 'e', 'E')) {
                vstr_add_char(&lex->vstr, 'e');
                next_char(lex);
                if (is_char(lex, '+') || is_char(lex, '-')) {
                    vstr_add_char(&lex->vstr, CUR_CHAR(lex));
                    next_char(lex);
                }
            } else if (is_letter(lex) || is_digit(lex) || is_char_or(lex, '_', '.')) {
                vstr_add_char(&lex->vstr, CUR_CHAR(lex));
                next_char(lex);
            } else {
                break;
            }
        }

    } else {
        // search for encoded delimiter or operator

        const char *t = tok_enc;
        uint tok_enc_index = 0;
        for (; *t != 0 && !is_char(lex, *t); t += 1) {
            if (*t == 'e' || *t == 'c') {
                t += 1;
            } else if (*t == 'E') {
                tok_enc_index -= 1;
                t += 1;
            }
            tok_enc_index += 1;
        }

        next_char(lex);

        if (*t == 0) {
            // didn't match any delimiter or operator characters
            tok->kind = MP_TOKEN_INVALID;

        } else {
            // matched a delimiter or operator character

            // get the maximum characters for a valid token
            t += 1;
            uint t_index = tok_enc_index;
            for (;;) {
                for (; *t == 'e'; t += 1) {
                    t += 1;
                    t_index += 1;
                    if (is_char(lex, *t)) {
                        next_char(lex);
                        tok_enc_index = t_index;
                        break;
                    }
                }

                if (*t == 'E') {
                    t += 1;
                    if (is_char(lex, *t)) {
                        next_char(lex);
                        tok_enc_index = t_index;
                    } else {
                        tok->kind = MP_TOKEN_INVALID;
                    }
                    break;
                }

                if (*t == 'c') {
                    t += 1;
                    t_index += 1;
                    if (is_char(lex, *t)) {
                        next_char(lex);
                        tok_enc_index = t_index;
                        t += 1;
                    } else {
                        break;
                    }
                } else {
                    break;
                }
            }

            // set token kind
            tok->kind = tok_enc_kind[tok_enc_index];

            // compute bracket level for implicit line joining
            if (tok->kind == MP_TOKEN_DEL_PAREN_OPEN || tok->kind == MP_TOKEN_DEL_BRACKET_OPEN || tok->kind == MP_TOKEN_DEL_BRACE_OPEN) {
                lex->nested_bracket_level += 1;
            } else if (tok->kind == MP_TOKEN_DEL_PAREN_CLOSE || tok->kind == MP_TOKEN_DEL_BRACKET_CLOSE || tok->kind == MP_TOKEN_DEL_BRACE_CLOSE) {
                lex->nested_bracket_level -= 1;
            }
        }
    }

    // point token text to vstr buffer
    tok->str = vstr_str(&lex->vstr);
    tok->len = vstr_len(&lex->vstr);

    // check for keywords
    if (tok->kind == MP_TOKEN_NAME) {
        for (int i = 0; tok_kw[i] != NULL; i++) {
            if (str_strn_equal(tok_kw[i], tok->str, tok->len)) {
                tok->kind = MP_TOKEN_KW_FALSE + i;
                break;
            }
        }
    }
}