| 1 | // This file is part of SmallBASIC |
| 2 | // |
| 3 | // bc module. Bytecode manipulation routines (bytecode segments API) |
| 4 | // |
| 5 | // This program is distributed under the terms of the GPL v2.0 or later |
| 6 | // Download the GNU Public License (GPL) from www.gnu.org |
| 7 | // |
| 8 | // Copyright(C) 2001 Nicholas Christopoulos |
| 9 | |
| 10 | #include "common/bc.h" |
| 11 | #include "common/smbas.h" |
| 12 | #include "common/str.h" |
| 13 | |
| 14 | /* |
| 15 | * string escape codes |
| 16 | */ |
| 17 | const char escapes[][2] = { |
| 18 | {'a', 0x07}, |
| 19 | {'b', 0x08}, |
| 20 | {'e', 0x1b}, |
| 21 | {'f', 0x0c}, |
| 22 | {'n', 0x0a}, |
| 23 | {'r', 0x0d}, |
| 24 | {'t', 0x09}, |
| 25 | {'v', 0x0b}, |
| 26 | }; |
| 27 | |
| 28 | /* |
| 29 | * whether the character is an escape |
| 30 | */ |
| 31 | int bc_is_escape(char c, char *value) { |
| 32 | int len = sizeof(escapes) / sizeof(escapes[0]); |
| 33 | int result = 0; |
| 34 | for (int i = 0; i < len && !result; i++) { |
| 35 | if (escapes[i][0] == c) { |
| 36 | *value = escapes[i][1]; |
| 37 | result = 1; |
| 38 | } |
| 39 | } |
| 40 | return result; |
| 41 | } |
| 42 | |
| 43 | /* |
| 44 | * whether the character is octal escape |
| 45 | */ |
| 46 | int bc_is_octal(char *str, char *output) { |
| 47 | char *next = str + 1; |
| 48 | int result = 0; |
| 49 | int digits = 0; |
| 50 | int value = 0; |
| 51 | |
| 52 | while (isdigit(*next) && digits < 4) { |
| 53 | value = (value << 3) + (*next - '0'); |
| 54 | digits++; |
| 55 | next++; |
| 56 | } |
| 57 | if (digits == 3 && value < 256) { |
| 58 | *output = value; |
| 59 | result = 1; |
| 60 | } |
| 61 | return result; |
| 62 | } |
| 63 | |
| 64 | /* |
| 65 | * Create a bytecode segment |
| 66 | */ |
| 67 | void bc_create(bc_t *bc) { |
| 68 | bc->ptr = malloc(BC_ALLOC_INCR); |
| 69 | bc->size = BC_ALLOC_INCR; |
| 70 | bc->count = 0; |
| 71 | bc->cp = 0; |
| 72 | bc->eoc_position = 0; |
| 73 | bc->line_position = 0; |
| 74 | } |
| 75 | |
| 76 | /* |
| 77 | * Destroy a bytecode segment |
| 78 | */ |
| 79 | void bc_destroy(bc_t *bc) { |
| 80 | free(bc->ptr); |
| 81 | bc->ptr = NULL; |
| 82 | bc->size = 0; |
| 83 | bc->count = 0; |
| 84 | bc->cp = 0; |
| 85 | bc->eoc_position = 0; |
| 86 | bc->line_position = 0; |
| 87 | } |
| 88 | |
| 89 | /* |
| 90 | * Resize a bytecode segment |
| 91 | */ |
| 92 | void bc_resize(bc_t *bc, uint32_t new_size) { |
| 93 | if (new_size != bc->size) { |
| 94 | bc->ptr = realloc(bc->ptr, new_size); |
| 95 | bc->size = new_size; |
| 96 | if (new_size == 1) { |
| 97 | bc->count = 0; |
| 98 | } |
| 99 | } |
| 100 | } |
| 101 | |
| 102 | /* |
| 103 | * add one command |
| 104 | */ |
| 105 | void bc_add1(bc_t *bc, char code) { |
| 106 | if (bc->count + sizeof(byte) >= bc->size) { |
| 107 | bc_resize(bc, bc->size + BC_ALLOC_INCR); |
| 108 | } |
| 109 | bc->ptr[bc->count] = code; |
| 110 | bc->count++; |
| 111 | } |
| 112 | |
| 113 | /* |
| 114 | * add one uint32_t |
| 115 | */ |
| 116 | void bc_add_dword(bc_t *bc, uint32_t p1) { |
| 117 | if (bc->count + sizeof(uint32_t) >= bc->size) { |
| 118 | bc_resize(bc, bc->size + BC_ALLOC_INCR); |
| 119 | } |
| 120 | memcpy(bc->ptr + bc->count, &p1, sizeof(uint32_t)); |
| 121 | bc->count += sizeof(uint32_t); |
| 122 | } |
| 123 | |
| 124 | /* |
| 125 | * add one command and one long int (32 bits) |
| 126 | */ |
| 127 | void bc_add2l(bc_t *bc, byte code, bid_t p1) { |
| 128 | if (bc->count + sizeof(bid_t) >= bc->size) { |
| 129 | bc_resize(bc, bc->size + BC_ALLOC_INCR); |
| 130 | } |
| 131 | bc->ptr[bc->count] = code; |
| 132 | bc->count++; |
| 133 | memcpy(bc->ptr + bc->count, &p1, sizeof(bid_t)); |
| 134 | bc->count += sizeof(bid_t); |
| 135 | } |
| 136 | |
| 137 | /* |
| 138 | * add buildin function call |
| 139 | */ |
| 140 | void bc_add_fcode(bc_t *bc, bid_t idx) { |
| 141 | bc_add2l(bc, kwTYPE_CALLF, idx); |
| 142 | } |
| 143 | |
| 144 | /* |
| 145 | * add buildin procedure call |
| 146 | */ |
| 147 | void bc_add_pcode(bc_t *bc, bid_t idx) { |
| 148 | bc_add2l(bc, kwTYPE_CALLP, idx); |
| 149 | } |
| 150 | |
| 151 | /* |
| 152 | * add an external function-call |
| 153 | */ |
| 154 | void bc_add_extfcode(bc_t *bc, uint32_t lib, uint32_t idx) { |
| 155 | bc_add_code(bc, kwTYPE_CALLEXTF); |
| 156 | bc_add_dword(bc, lib); |
| 157 | bc_add_dword(bc, idx); |
| 158 | } |
| 159 | |
| 160 | /* |
| 161 | * add an external procedure-call |
| 162 | */ |
| 163 | void bc_add_extpcode(bc_t *bc, uint32_t lib, uint32_t idx) { |
| 164 | bc_add_code(bc, kwTYPE_CALLEXTP); |
| 165 | bc_add_dword(bc, lib); |
| 166 | bc_add_dword(bc, idx); |
| 167 | } |
| 168 | |
| 169 | /* |
| 170 | * add an address |
| 171 | */ |
| 172 | void bc_add_addr(bc_t *bc, bcip_t idx) { |
| 173 | if (bc->count + sizeof(bcip_t) >= bc->size) { |
| 174 | bc_resize(bc, bc->size + BC_ALLOC_INCR); |
| 175 | } |
| 176 | memcpy(bc->ptr + bc->count, &idx, sizeof(bcip_t)); |
| 177 | bc->count += sizeof(bcip_t); |
| 178 | } |
| 179 | |
| 180 | /* |
| 181 | * add a control code |
| 182 | * |
| 183 | * Control codes are followed by 2 bcip_t elements (jump on true, jump on false) |
| 184 | */ |
| 185 | void bc_add_ctrl(bc_t *bc, code_t code, bcip_t true_ip, bcip_t false_ip) { |
| 186 | bc_add1(bc, code); |
| 187 | bc_add_addr(bc, true_ip); |
| 188 | bc_add_addr(bc, false_ip); |
| 189 | } |
| 190 | |
| 191 | /* |
| 192 | * add an integer |
| 193 | */ |
| 194 | void bc_add_cint(bc_t *bc, var_int_t v) { |
| 195 | if (bc->count + sizeof(var_int_t) >= bc->size) { |
| 196 | bc_resize(bc, bc->size + BC_ALLOC_INCR); |
| 197 | } |
| 198 | bc->ptr[bc->count] = kwTYPE_INT; |
| 199 | bc->count++; |
| 200 | memcpy(bc->ptr + bc->count, &v, sizeof(var_int_t)); |
| 201 | bc->count += sizeof(var_int_t); |
| 202 | } |
| 203 | |
| 204 | /* |
| 205 | * add an real |
| 206 | */ |
| 207 | void bc_add_creal(bc_t *bc, var_num_t v) { |
| 208 | if (bc->count + sizeof(var_num_t) >= bc->size) { |
| 209 | bc_resize(bc, bc->size + BC_ALLOC_INCR); |
| 210 | } |
| 211 | bc->ptr[bc->count] = kwTYPE_NUM; |
| 212 | bc->count++; |
| 213 | memcpy(bc->ptr + bc->count, &v, sizeof(var_num_t)); |
| 214 | bc->count += sizeof(var_num_t); |
| 215 | } |
| 216 | |
| 217 | /* |
| 218 | * add one command and one string (see: bc_store_string) |
| 219 | */ |
| 220 | void bc_add_strn(bc_t *bc, const char *str, int len) { |
| 221 | bc_add_code(bc, kwTYPE_STR); |
| 222 | bc_add_dword(bc, len + 1); |
| 223 | if (bc->count + len >= bc->size) { |
| 224 | bc_resize(bc, bc->size + BC_ALLOC_INCR + len); |
| 225 | } |
| 226 | memcpy(bc->ptr + bc->count, str, len); |
| 227 | bc->count += len; |
| 228 | bc_add1(bc, 0); |
| 229 | } |
| 230 | |
| 231 | /* |
| 232 | * adds a string. |
| 233 | * returns a pointer of src to the next "element" |
| 234 | */ |
| 235 | char *bc_store_string(bc_t *bc, char *src) { |
| 236 | // skip past opening quotes |
| 237 | char *p = src + 1; |
| 238 | char *base = src + 1; |
| 239 | char escape = 0; |
| 240 | cstr cs; |
| 241 | cstr_init(&cs, 5); |
| 242 | |
| 243 | while (*p) { |
| 244 | if ((*p == '\\' && ((*(p + 1) == '\"') || *(p + 1) == '\\')) |
| 245 | || *p == V_JOIN_LINE) { |
| 246 | // escaped quote " or escaped escape |
| 247 | cstr_append_i(&cs, base, p - base); |
| 248 | |
| 249 | if (*p == V_JOIN_LINE) { |
| 250 | // skip null newline |
| 251 | comp_line++; |
| 252 | if (*(p+1) == '\"') { |
| 253 | base = ++p; |
| 254 | continue; |
| 255 | } |
| 256 | } |
| 257 | |
| 258 | // include " (or \ ) in next segment |
| 259 | base = ++p; |
| 260 | } else if (*p == '\\' && bc_is_escape(*(p + 1), &escape)) { |
| 261 | char code[] = {escape, '\0'}; |
| 262 | cstr_append_i(&cs, base, p - base); |
| 263 | cstr_append_i(&cs, code, 1); |
| 264 | // skip single escape character |
| 265 | base = (++p) + 1; |
| 266 | } else if (*p == '\\' && bc_is_octal(p, &escape)) { |
| 267 | char code[] = {escape, '\0'}; |
| 268 | cstr_append_i(&cs, base, p - base); |
| 269 | cstr_append_i(&cs, code, 1); |
| 270 | // skip octal digits |
| 271 | p += 3; |
| 272 | base = p + 1; |
| 273 | } else if (*p == V_QUOTE) { |
| 274 | // revert hidden quote |
| 275 | *p = '\"'; |
| 276 | } else if (*p == V_LINE) { |
| 277 | // revert hidden newline |
| 278 | comp_line++; |
| 279 | *p = '\n'; |
| 280 | } else if (*p == '\"') { |
| 281 | // end of string detected |
| 282 | cstr_append_i(&cs, base, p - base); |
| 283 | bc_add_strn(bc, cs.buf, cs.length); |
| 284 | p++; |
| 285 | break; |
| 286 | } |
| 287 | p++; |
| 288 | } |
| 289 | free(cs.buf); |
| 290 | return p; |
| 291 | } |
| 292 | |
| 293 | /* |
| 294 | * adds an EOC mark at the current position |
| 295 | */ |
| 296 | void bc_eoc(bc_t *bc) { |
| 297 | if (bc && bc->count && |
| 298 | (bc->eoc_position == 0 || bc->eoc_position != bc->count - 1)) { |
| 299 | // avoid appending multiple kwTYPE_EOCs (or kwTYPE_LINE) |
| 300 | bc->eoc_position = bc->count; |
| 301 | bc_add1(bc, kwTYPE_EOC); |
| 302 | } |
| 303 | } |
| 304 | |
| 305 | /* |
| 306 | * pops any EOC mark at the current position |
| 307 | */ |
| 308 | int bc_pop_eoc(bc_t *bc) { |
| 309 | int result; |
| 310 | if (bc->eoc_position > 0 && bc->eoc_position == bc->count - 1) { |
| 311 | bc->eoc_position = 0; |
| 312 | result = (bc->ptr[--bc->count] == kwTYPE_EOC); |
| 313 | } else { |
| 314 | result = 0; |
| 315 | } |
| 316 | return result; |
| 317 | } |
| 318 | |
| 319 | /* |
| 320 | * appends the src to dst |
| 321 | */ |
| 322 | void bc_append(bc_t *dst, bc_t *src) { |
| 323 | bc_resize(dst, dst->count + src->count + 4); |
| 324 | memcpy(dst->ptr + dst->count, src->ptr, src->count); |
| 325 | dst->count += src->count; |
| 326 | } |
| 327 | |
| 328 | /* |
| 329 | * appends n bytes from src to dst |
| 330 | */ |
| 331 | void bc_add_n(bc_t *dst, byte *src, uint32_t n) { |
| 332 | if (dst->count + n >= dst->size) { |
| 333 | bc_resize(dst, dst->size + BC_ALLOC_INCR + n); |
| 334 | } |
| 335 | memcpy(dst->ptr + dst->count, src, n); |
| 336 | dst->count += n; |
| 337 | } |
| 338 | |