/*************************************************************************** gbc_read.c (c) 2000-2012 BenoƮt Minisini This program is free software; you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation; either version 2, or (at your option) any later version. This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details. You should have received a copy of the GNU General Public License along with this program; if not, write to the Free Software Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA. ***************************************************************************/ #define __GBC_READ_C #include #include #include #include #include #include #include #include #include #include "gb_common.h" #include "gb_common_buffer.h" #include "gb_common_case.h" #include "gb_error.h" #include "gb_table.h" #include "gb_file.h" #include "gbc_compile.h" #include "gbc_class.h" #include "gbc_preprocess.h" #include "gbc_read.h" //#define DEBUG //#define BIG_COMMENT 1 static bool is_init = FALSE; static COMPILE *comp; static const char *source_ptr; static int source_length; static bool _begin_line = FALSE; static bool _no_quote = FALSE; static bool _prep = FALSE; static int _prep_index; static char ident_car[256]; static char first_car[256]; char READ_digit_car[256]; static char noop_car[256]; static char canres_car[256]; enum { GOTO_BREAK, GOTO_SPACE, GOTO_COMMENT, GOTO_STRING, GOTO_IDENT, GOTO_QUOTED_IDENT, GOTO_NUMBER, GOTO_ERROR, GOTO_SHARP, GOTO_OTHER }; static void READ_init(void) { unsigned char i; JOB->line = 1; JOB->column = TRUE; if (!is_init) { for (i = 0; i < 255; i++) { ident_car[i] = (i != 0) && ((i >= 'A' && i <= 'Z') || (i >= 'a' && i <= 'z') || (i >= '0' && i <= '9') || strchr("$_?@", i)); READ_digit_car[i] = (i >= '0' && i <= '9'); noop_car[i] = ident_car[i] || READ_digit_car[i] || i <= ' '; canres_car[i] = (i != ':') && (i != '.') && (i != '!') && (i != '('); if (i == 0) first_car[i] = GOTO_BREAK; else if (i <= ' ') first_car[i] = GOTO_SPACE; else if (i == '\'') first_car[i] = GOTO_COMMENT; else if (i == '"') first_car[i] = GOTO_STRING; else if (i == '#') first_car[i] = GOTO_SHARP; else if ((i >= 'A' && i <= 'Z') || (i >= 'a' && i <= 'z') || i == '$' || i == '_') first_car[i] = GOTO_IDENT; else if (i == '{') first_car[i] = GOTO_QUOTED_IDENT; else if (i >= '0' && i <= '9') first_car[i] = GOTO_NUMBER; else if (i >= 127) first_car[i] = GOTO_ERROR; else first_car[i] = GOTO_OTHER; } is_init = TRUE; } } static void READ_exit(void) { char *name = NULL; int len; int index; bool local; local = FALSE; for(;;) { COMPILE_enum_class(&name, &len); if (!len) break; if (len == 1 && *name == '-') { local = TRUE; } else { if (TABLE_find_symbol(JOB->class->table, name, len, &index)) { if (local) CLASS_add_class_unused(JOB->class, index); else CLASS_add_class_exported_unused(JOB->class, index); } } } if (JOB->verbose) printf("\n"); JOB->column = FALSE; } static int get_utf8_length(const char *str, int len) { int ulen = 0; int i; for (i = 0; i < len; i++) { if ((str[i] & 0xC0) != 0x80) ulen++; } return ulen; } int READ_get_column() { const char *p = source_ptr; int col = 0; while (p > comp->source) { if (*p == '\n') { p++; break; } p--; col++; } return get_utf8_length(p, col + 1); } char *READ_get_pattern(PATTERN *pattern) { int type = PATTERN_type(*pattern); int index = PATTERN_index(*pattern); const char *str; const char *before = _no_quote ? "" : "'"; const char *after = _no_quote ? "" : "'"; switch(type) { case RT_RESERVED: str = COMP_res_info[index].name; //TABLE_get_symbol_name(COMP_res_table, index); if (ispunct(*str)) snprintf(COMMON_buffer, COMMON_BUF_MAX, "%s%s%s", before, str, after); else strcpy(COMMON_buffer, str); break; case RT_NUMBER: case RT_IDENTIFIER: case RT_CLASS: snprintf(COMMON_buffer, COMMON_BUF_MAX, "%s%s%s", before, TABLE_get_symbol_name(JOB->class->table, index), after); break; case RT_STRING: case RT_TSTRING: if (_no_quote) snprintf(COMMON_buffer, COMMON_BUF_MAX, "\"%s\"", TABLE_get_symbol_name(JOB->class->string, index)); else strcpy(COMMON_buffer, "string"); break; case RT_COMMAND: snprintf(COMMON_buffer, COMMON_BUF_MAX, "#%d", index); break; case RT_NEWLINE: strcpy(COMMON_buffer, "end of line"); break; case RT_END: strcpy(COMMON_buffer, "end of file"); break; case RT_SUBR: //snprintf(COMMON_buffer, COMMON_BUF_MAX, "%s%s%s", bafore, COMP_subr_info[index].name, after); strcpy(COMMON_buffer, COMP_subr_info[index].name); break; default: sprintf(COMMON_buffer, "%s?%02X.%02X.%d?%s", before, PATTERN_type(*pattern), PATTERN_flag(*pattern), (int)PATTERN_index(*pattern), after); } return COMMON_buffer; } void THROW_UNEXPECTED(PATTERN *pattern) { switch (PATTERN_type(*pattern)) { case RT_NEWLINE: case RT_END: THROW("Unexpected end of line"); case RT_STRING: case RT_TSTRING: THROW("Unexpected string"); default: THROW("Unexpected &1", READ_get_pattern(pattern)); } } void READ_dump_pattern(PATTERN *pattern) { int type = PATTERN_type(*pattern); int index = PATTERN_index(*pattern); /*pos = (int)(pattern - JOB->pattern); if (pos < 0 || pos >= JOB->pattern_count) return; printf("%d ", pos);*/ if (PATTERN_flag(*pattern) & RT_FIRST) printf("!"); else printf(" "); if (PATTERN_flag(*pattern) & RT_POINT) printf("."); else printf(" "); if (PATTERN_flag(*pattern) & RT_CLASS) printf("C"); else printf(" "); printf(" "); _no_quote = TRUE; if (type == RT_RESERVED) printf("RESERVED %s\n", READ_get_pattern(pattern)); else if (type == RT_NUMBER) printf("NUMBER %s\n", READ_get_pattern(pattern)); else if (type == RT_IDENTIFIER) printf("IDENTIFIER %s\n", READ_get_pattern(pattern)); else if (type == RT_CLASS) printf("CLASS %s\n", READ_get_pattern(pattern)); else if (type == RT_STRING) printf("STRING %s\n", READ_get_pattern(pattern)); else if (type == RT_TSTRING) printf("TSTRING %s\n", READ_get_pattern(pattern)); else if (type == RT_NEWLINE) printf("NEWLINE (%d)\n", index); else if (type == RT_END) printf("END\n"); else if (type == RT_PARAM) printf("PARAM %d\n", index); else if (type == RT_SUBR) printf("SUBR %s\n", READ_get_pattern(pattern)); else if (type == RT_COMMAND) printf("COMMAND %d\n", index); else printf("? %d\n", index); _no_quote = FALSE; } #if 0 static inline unsigned char get_char_offset(int offset) { offset += source_ptr; if (offset >= source_length || offset < 0) return 0; else return (unsigned char)(comp->source[offset]); } #endif #if 0 static unsigned char get_char(void) { return get_char_offset(0); } #endif #define get_char_offset(_offset) ((unsigned char)source_ptr[(_offset)]) #define get_char() ((unsigned char)(*source_ptr)) static unsigned char next_char(void) { source_ptr++; return get_char(); } #ifdef DEBUG static void add_pattern(int type, int index) { comp->pattern[comp->pattern_count++] = PATTERN_make(type, index); READ_dump_pattern(&comp->pattern[comp->pattern_count - 1]); } #define add_pattern_no_dump(_type, _index) comp->pattern[comp->pattern_count++] = PATTERN_make((_type), (_index)); #else #define add_pattern(_type, _index) comp->pattern[comp->pattern_count++] = PATTERN_make((_type), (_index)); #define add_pattern_no_dump add_pattern #endif static PATTERN get_last_last_pattern() { if (LIKELY(comp->pattern_count > 1)) return comp->pattern[comp->pattern_count - 2]; else return NULL_PATTERN; } #define get_last_pattern() (comp->pattern[comp->pattern_count - 1]) static void jump_to_next_prep(void) { unsigned char car; const char *line_start; for (;;) { line_start = source_ptr; for(;;) { car = get_char(); if (!car) return; if (car == '\n' || !isspace(car)) break; source_ptr++; } if (car == '#') { source_ptr = line_start; return; } for(;;) { car = get_char(); if (!car) return; source_ptr++; if (car == '\n') break; } add_pattern(RT_NEWLINE, comp->line); comp->line++; } } static void add_newline() { int action = PREP_CONTINUE; if (_prep) { int line = comp->line; add_pattern_no_dump(RT_NEWLINE, comp->line); action = PREP_analyze(&comp->pattern[_prep_index]); _prep = FALSE; comp->pattern_count = _prep_index; comp->line = line; } if (action == PREP_LINE) comp->line = PREP_next_line; add_pattern(RT_NEWLINE, comp->line); comp->line++; if (action == PREP_IGNORE) jump_to_next_prep(); } static void add_end() { add_pattern(RT_END, 0); comp->line++; } static bool add_number() { unsigned char car; const char *start; int index; char sign; PATTERN last_pattern; bool has_digit; start = source_ptr; car = get_char(); if (car == '-' || car == '+') { sign = car; car = next_char(); } else sign = 0; if (car == '&') { car = toupper(next_char()); if (car == 'H') goto READ_HEXA; else if (car == 'X') goto READ_BINARY; else { source_ptr--; goto READ_HEXA; } } else if (car == '%') goto READ_BINARY; else if (isdigit(car)) goto READ_NUMBER; else goto NOT_A_NUMBER; READ_BINARY: has_digit = FALSE; for (;;) { car = next_char(); if (UNLIKELY(car != '0' && car != '1')) break; has_digit = TRUE; } goto END_BINARY_HEXA; READ_HEXA: has_digit = FALSE; for (;;) { car = next_char(); if (UNLIKELY(!isxdigit(car))) break; has_digit = TRUE; } END_BINARY_HEXA: if (!has_digit) goto NOT_A_NUMBER; if (car == '&') car = next_char(); else if (first_car[car] == GOTO_IDENT) goto NOT_A_NUMBER; goto END; READ_NUMBER: while (isdigit(car)) car = next_char(); if (car == '.') { do { car = next_char(); } while (isdigit(car)); } if (toupper(car) == 'E') { car = next_char(); if (car == '+' || car == '-') car = next_char(); while (isdigit(car)) car = next_char(); } goto END; END: last_pattern = get_last_pattern(); if (sign && !PATTERN_is_null(last_pattern) && (!PATTERN_is_reserved(last_pattern) || PATTERN_is(last_pattern, RS_RBRA) || PATTERN_is(last_pattern, RS_RSQR))) { add_pattern(RT_RESERVED, RESERVED_find_word(&sign, 1)); TABLE_add_symbol(comp->class->table, start + 1, source_ptr - start - 1, &index); add_pattern(RT_NUMBER, index); } else { TABLE_add_symbol(comp->class->table, start, source_ptr - start, &index); add_pattern(RT_NUMBER, index); } return FALSE; NOT_A_NUMBER: source_ptr = start; return TRUE; } static void add_identifier() { unsigned char car; const char *start; int len; int index; int type; int flag; PATTERN last_pattern, last_last_pattern; bool not_first; bool can_be_reserved; bool last_identifier, last_type, last_class, last_pub; last_pattern = get_last_pattern(); if (PATTERN_is_reserved(last_pattern)) { flag = RES_get_ident_flag(PATTERN_index(last_pattern)); if (flag & RSF_PREV) { last_last_pattern = get_last_last_pattern(); if (PATTERN_is_reserved(last_last_pattern)) flag = RES_get_ident_flag(PATTERN_index(last_last_pattern)); else flag = 0; } } else flag = 0; type = RT_IDENTIFIER; start = source_ptr; len = 1; last_class = (flag & RSF_CLASS) != 0; last_type = (flag & RSF_AS) != 0; if (last_type) { for(;;) { source_ptr++; len++; car = get_char(); if (ident_car[car]) continue; if (car == '[') { car = get_char_offset(1); if (car == ']') { source_ptr++; len++; TABLE_add_symbol(comp->class->table, start, len - 2, &index); continue; } } len--; break; } } else { for(;;) { source_ptr++; car = get_char(); if (!ident_car[car]) break; len++; } } not_first = (flag & RSF_POINT) != 0; car = get_char(); //can_be_reserved = !not_first && TABLE_find_symbol(COMP_res_table, &comp->source[start], len, NULL, &index); can_be_reserved = !not_first && !last_class; if (can_be_reserved) { index = RESERVED_find_word(start, len); can_be_reserved = (index >= 0); } if (can_be_reserved) { static void *jump[] = { &&__OTHERS, &&__ME_NEW_LAST_SUPER, &&__CLASS, &&__STRUCT, &&__SUB_PROCEDURE_FUNCTION, &&__CONST_EXTERN, &&__ENUM, &&__READ, &&__DATATYPE }; last_identifier = (flag & RSF_IDENT) != 0; last_pub = (flag & RSF_PUB) != 0; goto *jump[RES_get_read_switch(index)]; do { __ME_NEW_LAST_SUPER: can_be_reserved = !last_identifier; break; __CLASS: can_be_reserved = canres_car[car] && _begin_line; break; __STRUCT: can_be_reserved = canres_car[car] && (_begin_line || last_pub || PATTERN_is(last_pattern, RS_AS) || PATTERN_is(last_pattern, RS_END) || PATTERN_is(last_pattern, RS_NEW)); break; __SUB_PROCEDURE_FUNCTION: can_be_reserved = canres_car[car] && (_begin_line || last_pub || PATTERN_is(last_pattern, RS_END)); break; __CONST_EXTERN: can_be_reserved = canres_car[car] && (_begin_line || last_pub); break; __ENUM: can_be_reserved = canres_car[car] && (_begin_line || last_pub); break; __READ: can_be_reserved = canres_car[car] && (!last_identifier || PATTERN_is(last_pattern, RS_PROPERTY)); break; __DATATYPE: if (car == '[' && get_char_offset(1) == ']') { len += 2; source_ptr += 2; can_be_reserved = FALSE; } else { if (last_type) can_be_reserved = TRUE; else can_be_reserved = FALSE; } break; __OTHERS: if (last_type || last_identifier) can_be_reserved = FALSE; else can_be_reserved = canres_car[car]; break; } while (0); } if (can_be_reserved) { type = RT_RESERVED; goto __ADD_PATTERN; } if ((flag == 0) && car != '.' && car != '!') { index = RESERVED_find_subr(start, len); if (index >= 0) { type = RT_SUBR; goto __ADD_PATTERN; } } if (last_type) type = RT_CLASS; if (flag & RSF_EVENT) { start--; len++; *((char *)start) = ':'; } if (PATTERN_is(last_pattern, RS_EXCL)) { TABLE_add_symbol(comp->class->string, start, len, &index); type = RT_STRING; } else TABLE_add_symbol(comp->class->table, start, len, &index); __ADD_PATTERN: add_pattern(type, index); } static void add_quoted_identifier(void) { unsigned char car; const char *start; int len; int index; int type; PATTERN last_pattern; last_pattern = get_last_pattern(); type = RT_IDENTIFIER; start = source_ptr; len = 1; for(;;) { source_ptr++; car = get_char(); if (!ident_car[car]) break; len++; } if (get_char() == '}') source_ptr++; if (PATTERN_is(last_pattern, RS_EVENT) || PATTERN_is(last_pattern, RS_RAISE)) { start--; len++; *((char *)start) = ':'; } if (PATTERN_is(last_pattern, RS_EXCL)) { TABLE_add_symbol(comp->class->string, start, len, &index); type = RT_STRING; } else TABLE_add_symbol(comp->class->table, start, len, &index); add_pattern(type, index); } static void add_operator() { unsigned char car; const char *start; const char *end; int len; int op = NO_SYMBOL; int index; start = source_ptr; end = start; len = 1; for(;;) { source_ptr++; index = RESERVED_find_word(start, len); if (index >= 0) { op = index; end = source_ptr; } car = get_char(); //if (!isascii(car) || !ispunct(car)) if (noop_car[car]) break; len++; } source_ptr = end; if (op < 0) THROW("Unknown operator"); add_pattern(RT_RESERVED, op); } static int xdigit_val(unsigned char c) { if (c >= '0' && c <= '9') return (c - '0'); else if (c >= 'a' && c <= 'f') return (c - 'a' + 10); else if (c >= 'A' && c <= 'F') return (c - 'A' + 10); else return (-1); } static void add_string() { unsigned char car; const char *start; int len; int index; int newline; bool jump; char *p; const char *end; int i; start = end = source_ptr; len = 0; newline = 0; jump = FALSE; p = (char *)source_ptr; for(;;) { source_ptr++; car = get_char(); if (jump) { if (car == '\n') newline++; else if (car == '"') jump = FALSE; else if (!car || !isspace(car)) break; } else { p++; len++; if (car == '\n') THROW("Non terminated string"); if (car == '\\') { source_ptr++; car = get_char(); if (car == 'n') *p = '\n'; else if (car == 't') *p = '\t'; else if (car == 'r') *p = '\r'; else if (car == 'b') *p = '\b'; else if (car == 'v') *p = '\v'; else if (car == 'f') *p = '\f'; else if (car == 'e') *p = '\x1B'; else if (car == '0') *p = 0; else if (car == '\"' || car == '\'' || car == '\\') *p = car; else { if (car == 'x') { i = xdigit_val(get_char_offset(1)); if (i >= 0) { car = i; i = xdigit_val(get_char_offset(2)); if (i >= 0) { car = (car << 4) | i; *p = car; source_ptr += 2; continue; } } } THROW("Bad character constant in string"); } } else if (car == '"') { p--; len--; end = source_ptr; comp->line += newline; newline = 0; jump = TRUE; } else *p = car; } } if (len > 0) { TABLE_add_symbol(comp->class->string, start + 1, len, &index); add_pattern(RT_STRING, index); } else add_pattern(RT_STRING, VOID_STRING); source_ptr = end + 1; //for (i = 0; i < newline; i++) // add_newline(); } #if 0 static void add_command() { unsigned char car; const char *start; int len; start = source_ptr; len = 0; for(;;) { source_ptr++; car = get_char(); if (car == '\n' || !car) break; len++; } if (len) { //TABLE_add_symbol(comp->class->string, start + 1, len, NULL, &index); if (len == 7 && !strncasecmp(start + 1, "SECTION", 7)) add_pattern(RT_COMMAND, RC_SECTION); } add_newline(); } #endif void READ_do(void) { static void *jump_char[] = { &&__BREAK, &&__SPACE, &&__COMMENT, &&__STRING, &&__IDENT, &&__QUOTED_IDENT, &&__NUMBER, &&__ERROR, &&__SHARP, &&__OTHER }; unsigned char car; comp = JOB; READ_init(); PREP_init(); //add_pattern(RT_NEWLINE, 0); source_ptr = comp->source; source_length = BUFFER_length(comp->source); _begin_line = TRUE; _prep = FALSE; //while (source_ptr < source_length) for(;;) { car = get_char(); goto *jump_char[(int)first_car[car]]; __ERROR: THROW("Syntax error"); __SPACE: source_ptr++; if (car == '\n') { add_newline(); _begin_line = TRUE; } continue; __COMMENT: do { source_ptr++; car = get_char(); } while (car != '\n'); _begin_line = FALSE; continue; __STRING: add_string(); _begin_line = FALSE; continue; __IDENT: add_identifier(); _begin_line = FALSE; continue; __QUOTED_IDENT: source_ptr++; add_quoted_identifier(); _begin_line = FALSE; continue; __NUMBER: add_number(); _begin_line = FALSE; continue; __SHARP: if (_begin_line) { _prep = TRUE; _prep_index = comp->pattern_count; add_identifier(); _begin_line = FALSE; continue; } else goto __OTHER; __OTHER: #if BIG_COMMENT if (car == '/' && get_char_offset(1) == '*') { for(;;) { source_ptr++; car = get_char(); if (car == 0) break; if (car == '*' && get_char_offset(1) == '/') { source_ptr += 2; break; } if (car == '\n') add_newline(); } _begin_line = FALSE; continue; } #endif if (add_number()) add_operator(); _begin_line = FALSE; } __BREAK: // We add end markers to simplify the compiler job, when it needs to look // at many patterns in one shot. add_newline(); add_end(); add_end(); add_end(); add_end(); PREP_exit(); READ_exit(); }