| 1 | ! ============================================================================== |
| 2 | ! Module: expansion |
| 3 | ! Purpose: Parameter expansion and arithmetic operations |
| 4 | ! ============================================================================== |
| 5 | module expansion |
| 6 | use shell_types |
| 7 | use variables ! includes check_nounset |
| 8 | use command_capture, only: execute_command_and_capture |
| 9 | use glob, only: pattern_matches_no_dotfile_check |
| 10 | use iso_fortran_env, only: output_unit, error_unit |
| 11 | #ifdef USE_C_STRINGS |
| 12 | use iso_c_binding, only: c_char, c_int, c_null_char, c_ptr, c_f_pointer, c_size_t |
| 13 | #endif |
| 14 | implicit none |
| 15 | |
| 16 | ! Recursion depth limits |
| 17 | integer, parameter :: MAX_RECURSION_DEPTH = 1000 |
| 18 | |
| 19 | ! Arithmetic error tracking |
| 20 | logical :: arithmetic_error = .false. |
| 21 | character(len=256) :: arithmetic_error_msg = '' |
| 22 | |
| 23 | #ifdef USE_C_STRINGS |
| 24 | interface |
| 25 | function c_pattern_replace_alloc(input, input_len, pattern, pat_len, & |
| 26 | replacement, repl_len, replace_all, & |
| 27 | result_out) result(out_len) bind(C, name='fortsh_pattern_replace_alloc') |
| 28 | import :: c_char, c_int, c_ptr |
| 29 | character(kind=c_char), intent(in) :: input(*), pattern(*), replacement(*) |
| 30 | integer(c_int), value :: input_len, pat_len, repl_len, replace_all |
| 31 | type(c_ptr), intent(out) :: result_out |
| 32 | integer(c_int) :: out_len |
| 33 | end function |
| 34 | |
| 35 | subroutine c_free_string(ptr) bind(C, name='fortsh_free_string') |
| 36 | import :: c_ptr |
| 37 | type(c_ptr), value :: ptr |
| 38 | end subroutine |
| 39 | |
| 40 | function c_buf_create(capacity) result(handle) bind(C, name='fortsh_buffer_create') |
| 41 | import :: c_ptr, c_size_t |
| 42 | integer(c_size_t), value :: capacity |
| 43 | type(c_ptr) :: handle |
| 44 | end function |
| 45 | |
| 46 | subroutine c_buf_destroy(handle) bind(C, name='fortsh_buffer_destroy') |
| 47 | import :: c_ptr |
| 48 | type(c_ptr), value :: handle |
| 49 | end subroutine |
| 50 | |
| 51 | function c_buf_append_chars(handle, str, slen) result(rc) bind(C, name='fortsh_buffer_append_chars') |
| 52 | import :: c_ptr, c_int, c_char, c_size_t |
| 53 | type(c_ptr), value :: handle |
| 54 | character(kind=c_char), intent(in) :: str(*) |
| 55 | integer(c_size_t), value :: slen |
| 56 | integer(c_int) :: rc |
| 57 | end function |
| 58 | |
| 59 | function c_buf_append_char(handle, ch) result(rc) bind(C, name='fortsh_buffer_append_char') |
| 60 | import :: c_ptr, c_int, c_char |
| 61 | type(c_ptr), value :: handle |
| 62 | character(kind=c_char), value :: ch |
| 63 | integer(c_int) :: rc |
| 64 | end function |
| 65 | |
| 66 | function c_buf_length(handle) result(blen) bind(C, name='fortsh_buffer_length') |
| 67 | import :: c_ptr, c_size_t |
| 68 | type(c_ptr), value :: handle |
| 69 | integer(c_size_t) :: blen |
| 70 | end function |
| 71 | |
| 72 | function c_buf_to_fortran(handle, fstr, flen) result(copied) bind(C, name='fortsh_buffer_to_fortran') |
| 73 | import :: c_ptr, c_size_t, c_char |
| 74 | type(c_ptr), value :: handle |
| 75 | character(kind=c_char) :: fstr(*) |
| 76 | integer(c_size_t), value :: flen |
| 77 | integer(c_size_t) :: copied |
| 78 | end function |
| 79 | |
| 80 | subroutine c_buf_clear(handle) bind(C, name='fortsh_buffer_clear') |
| 81 | import :: c_ptr |
| 82 | type(c_ptr), value :: handle |
| 83 | end subroutine |
| 84 | |
| 85 | function c_buf_pattern_replace(input_buf, pattern, pat_len, replacement, repl_len, & |
| 86 | replace_all, result_out) result(out_len) & |
| 87 | bind(C, name='fortsh_buffer_pattern_replace') |
| 88 | import :: c_ptr, c_int, c_char |
| 89 | type(c_ptr), value :: input_buf |
| 90 | character(kind=c_char), intent(in) :: pattern(*), replacement(*) |
| 91 | integer(c_int), value :: pat_len, repl_len, replace_all |
| 92 | type(c_ptr), intent(out) :: result_out |
| 93 | integer(c_int) :: out_len |
| 94 | end function |
| 95 | end interface |
| 96 | #endif |
| 97 | |
| 98 | contains |
| 99 | |
| 100 | ! Parameter expansion: ${var:offset:length} |
| 101 | ! parameter_expansion — wrapper that strips ${ } and delegates to process_param_expansion |
| 102 | function parameter_expansion(shell, expression) result(expanded) |
| 103 | type(shell_state_t), intent(inout) :: shell |
| 104 | character(len=*), intent(in) :: expression |
| 105 | character(len=:), allocatable :: expanded |
| 106 | |
| 107 | expanded = '' |
| 108 | if (len_trim(expression) < 4) return |
| 109 | call process_param_expansion(expression(3:len_trim(expression)-1), expanded, shell) |
| 110 | end function |
| 111 | |
| 112 | ! process_param_expansion — canonical parameter expansion implementation |
| 113 | ! Takes the content between ${ and }, e.g. for ${var:-default} receives "var:-default" |
| 114 | ! This is the subroutine that parser.f90 will delegate to after consolidation. |
| 115 | recursive subroutine process_param_expansion(param_expr, result_value, shell) |
| 116 | character(len=*), intent(in) :: param_expr |
| 117 | character(len=:), allocatable, intent(out) :: result_value |
| 118 | type(shell_state_t), intent(inout) :: shell |
| 119 | |
| 120 | character(len=256) :: var_name, operation, param1, param2, replacement |
| 121 | character(len=32) :: num_buf |
| 122 | character(len=:), allocatable :: pattern |
| 123 | character(len=:), allocatable :: var_value |
| 124 | integer :: colon_pos, dash_pos, plus_pos, percent_pos, hash_pos, slash_pos, equals_pos, question_pos |
| 125 | integer :: offset, length, i, at_pos |
| 126 | character :: transform_op |
| 127 | logical :: replace_all, greedy, has_colon, var_is_set, var_is_null |
| 128 | |
| 129 | ! Array expansion variables |
| 130 | integer :: bracket_pos, bracket_end, j, num_keys |
| 131 | character(len=256) :: array_name, array_key |
| 132 | character(len=256), allocatable :: keys(:) |
| 133 | logical :: is_keys_expansion, is_length_expansion, is_all_expansion |
| 134 | |
| 135 | result_value = '' |
| 136 | var_name = param_expr |
| 137 | |
| 138 | ! ======================================================================== |
| 139 | ! Check for array bracket syntax FIRST: ${array[key]}, ${!array[@]}, ${#array[@]} |
| 140 | ! ======================================================================== |
| 141 | bracket_pos = index(var_name, '[') |
| 142 | if (bracket_pos > 0) then |
| 143 | bracket_end = index(var_name, ']') |
| 144 | if (bracket_end > bracket_pos) then |
| 145 | ! Extract array name and key |
| 146 | array_name = var_name(:bracket_pos-1) |
| 147 | array_key = var_name(bracket_pos+1:bracket_end-1) |
| 148 | |
| 149 | ! Strip quotes from array subscript (bash strips quotes) |
| 150 | call strip_quotes(array_key) |
| 151 | |
| 152 | ! Check for special prefixes (! for keys, # for length) |
| 153 | is_keys_expansion = .false. |
| 154 | is_length_expansion = .false. |
| 155 | |
| 156 | if (len_trim(array_name) > 0 .and. array_name(1:1) == '!') then |
| 157 | ! ${!array[@]} - get all keys |
| 158 | is_keys_expansion = .true. |
| 159 | array_name = array_name(2:) ! Remove ! prefix |
| 160 | else if (len_trim(array_name) > 0 .and. array_name(1:1) == '#') then |
| 161 | ! ${#array[@]} - get array length |
| 162 | is_length_expansion = .true. |
| 163 | array_name = array_name(2:) ! Remove # prefix |
| 164 | end if |
| 165 | |
| 166 | ! Check for [@] or [*] (all values/keys) |
| 167 | is_all_expansion = (trim(array_key) == '@' .or. trim(array_key) == '*') |
| 168 | |
| 169 | ! Handle associative arrays |
| 170 | if (is_associative_array(shell, trim(array_name))) then |
| 171 | if (is_keys_expansion .and. is_all_expansion) then |
| 172 | ! ${!array[@]} - return all keys |
| 173 | block |
| 174 | character(len=4096) :: kbuf |
| 175 | integer :: kpos, klen |
| 176 | kbuf = ''; kpos = 1 |
| 177 | allocate(keys(500)) |
| 178 | call get_assoc_array_keys(shell, trim(array_name), keys, num_keys) |
| 179 | do j = 1, min(num_keys, 500) |
| 180 | klen = len_trim(keys(j)) |
| 181 | if (klen == 0) cycle |
| 182 | if (kpos > 1) then; kbuf(kpos:kpos) = ' '; kpos = kpos + 1; end if |
| 183 | kbuf(kpos:kpos+klen-1) = keys(j)(1:klen) |
| 184 | kpos = kpos + klen |
| 185 | end do |
| 186 | if (kpos > 1) then |
| 187 | result_value = kbuf(1:kpos-1) |
| 188 | else |
| 189 | result_value = '' |
| 190 | end if |
| 191 | end block |
| 192 | deallocate(keys) |
| 193 | return |
| 194 | else if (is_length_expansion .and. is_all_expansion) then |
| 195 | ! ${#array[@]} - return number of keys |
| 196 | allocate(keys(500)) |
| 197 | call get_assoc_array_keys(shell, trim(array_name), keys, num_keys) |
| 198 | deallocate(keys) |
| 199 | write(num_buf, '(I0)') num_keys |
| 200 | result_value = trim(num_buf) |
| 201 | return |
| 202 | else if (is_all_expansion) then |
| 203 | ! ${array[@]} - return all values |
| 204 | block |
| 205 | character(len=4096) :: vbuf |
| 206 | integer :: vpos, vlen |
| 207 | vbuf = ''; vpos = 1 |
| 208 | allocate(keys(500)) |
| 209 | call get_assoc_array_keys(shell, trim(array_name), keys, num_keys) |
| 210 | do j = 1, min(num_keys, 500) |
| 211 | var_value = get_assoc_array_value(shell, trim(array_name), trim(keys(j))) |
| 212 | vlen = len_trim(var_value) |
| 213 | if (vpos > 1) then; vbuf(vpos:vpos) = ' '; vpos = vpos + 1; end if |
| 214 | if (vlen > 0) then; vbuf(vpos:vpos+vlen-1) = var_value(1:vlen); vpos = vpos + vlen; end if |
| 215 | end do |
| 216 | if (vpos > 1) then |
| 217 | result_value = vbuf(1:vpos-1) |
| 218 | else |
| 219 | result_value = '' |
| 220 | end if |
| 221 | end block |
| 222 | deallocate(keys) |
| 223 | return |
| 224 | else |
| 225 | ! ${array[key]} or ${#array[key]} - get value or length for specific key |
| 226 | var_value = get_assoc_array_value(shell, trim(array_name), trim(array_key)) |
| 227 | if (is_length_expansion) then |
| 228 | write(num_buf, '(I0)') len_trim(var_value) |
| 229 | result_value = trim(num_buf) |
| 230 | else |
| 231 | result_value = trim(var_value) |
| 232 | end if |
| 233 | return |
| 234 | end if |
| 235 | else |
| 236 | ! Handle indexed arrays |
| 237 | if (is_all_expansion) then |
| 238 | ! ${arr[@]} or ${arr[*]} — all elements |
| 239 | if (is_keys_expansion) then |
| 240 | ! ${!arr[@]} — list indices of set elements (sparse-aware) |
| 241 | block |
| 242 | integer :: ki, arr_sz, kpos |
| 243 | character(len=4096) :: kbuf |
| 244 | character(len=:), allocatable :: elem |
| 245 | kbuf = ''; kpos = 1 |
| 246 | arr_sz = get_array_size(shell, trim(array_name)) |
| 247 | do ki = 1, arr_sz |
| 248 | elem = get_array_element(shell, trim(array_name), ki) |
| 249 | if (len_trim(elem) > 0) then |
| 250 | if (kpos > 1) then; kbuf(kpos:kpos) = ' '; kpos = kpos + 1; end if |
| 251 | write(num_buf, '(I0)') ki - 1 ! 0-based index |
| 252 | kbuf(kpos:kpos+len_trim(num_buf)-1) = trim(num_buf) |
| 253 | kpos = kpos + len_trim(num_buf) |
| 254 | end if |
| 255 | end do |
| 256 | if (kpos > 1) then |
| 257 | result_value = kbuf(1:kpos-1) |
| 258 | else |
| 259 | result_value = '' |
| 260 | end if |
| 261 | end block |
| 262 | return |
| 263 | else if (is_length_expansion) then |
| 264 | ! ${#arr[@]} — count of elements |
| 265 | block |
| 266 | integer :: ki, arr_count |
| 267 | character(len=4096) :: all_str |
| 268 | all_str = trim(get_array_all_elements(shell, trim(array_name))) |
| 269 | arr_count = 0 |
| 270 | if (len_trim(all_str) > 0) then |
| 271 | arr_count = 1 |
| 272 | do ki = 1, len_trim(all_str) |
| 273 | if (all_str(ki:ki) == ' ') arr_count = arr_count + 1 |
| 274 | end do |
| 275 | end if |
| 276 | write(num_buf, '(I0)') arr_count |
| 277 | result_value = trim(num_buf) |
| 278 | end block |
| 279 | return |
| 280 | else |
| 281 | ! ${arr[@]} — all values, with optional slicing |
| 282 | block |
| 283 | character(len=4096) :: all_str |
| 284 | character(len=256) :: slice_spec |
| 285 | integer :: colon_after, sc_pos2, ios1, ios2 |
| 286 | integer :: s_offset, s_length, w_count, w_start, w_idx, w_out |
| 287 | logical :: has_slice |
| 288 | all_str = trim(get_array_all_elements(shell, trim(array_name))) |
| 289 | |
| 290 | ! Check for slice syntax after ]: ${arr[@]:offset:length} |
| 291 | has_slice = .false. |
| 292 | if (bracket_end < len_trim(var_name)) then |
| 293 | slice_spec = var_name(bracket_end+1:) |
| 294 | if (len_trim(slice_spec) > 0 .and. slice_spec(1:1) == ':') then |
| 295 | has_slice = .true. |
| 296 | end if |
| 297 | end if |
| 298 | |
| 299 | if (.not. has_slice) then |
| 300 | result_value = trim(all_str) |
| 301 | return |
| 302 | end if |
| 303 | |
| 304 | ! Parse :offset or :offset:length |
| 305 | slice_spec = slice_spec(2:) ! skip leading : |
| 306 | sc_pos2 = index(trim(slice_spec), ':') |
| 307 | if (sc_pos2 > 0) then |
| 308 | read(slice_spec(:sc_pos2-1), *, iostat=ios1) s_offset |
| 309 | read(slice_spec(sc_pos2+1:), *, iostat=ios2) s_length |
| 310 | if (ios2 /= 0) s_length = 9999 |
| 311 | else |
| 312 | read(slice_spec, *, iostat=ios1) s_offset |
| 313 | s_length = 9999 |
| 314 | end if |
| 315 | if (ios1 /= 0) then |
| 316 | result_value = trim(all_str) |
| 317 | return |
| 318 | end if |
| 319 | |
| 320 | ! Split into words and select slice |
| 321 | result_value = '' |
| 322 | w_count = 0; w_start = 1; w_out = 0 |
| 323 | do w_idx = 1, len_trim(all_str) + 1 |
| 324 | if (w_idx > len_trim(all_str) .or. all_str(w_idx:w_idx) == ' ') then |
| 325 | if (w_idx > w_start) then |
| 326 | if (w_count >= s_offset .and. w_out < s_length) then |
| 327 | if (w_out > 0) result_value = result_value // ' ' |
| 328 | result_value = result_value // all_str(w_start:w_idx-1) |
| 329 | w_out = w_out + 1 |
| 330 | end if |
| 331 | w_count = w_count + 1 |
| 332 | end if |
| 333 | w_start = w_idx + 1 |
| 334 | end if |
| 335 | end do |
| 336 | end block |
| 337 | return |
| 338 | end if |
| 339 | else |
| 340 | ! ${arr[i]} — single element (supports negative indices) |
| 341 | block |
| 342 | integer :: arr_idx, arr_ios, arr_size |
| 343 | character(len=:), allocatable :: arr_val |
| 344 | read(array_key, *, iostat=arr_ios) arr_idx |
| 345 | if (arr_ios == 0) then |
| 346 | if (arr_idx < 0) then |
| 347 | arr_size = get_array_size(shell, trim(array_name)) |
| 348 | arr_idx = arr_size + arr_idx ! -1 → last element (0-based) |
| 349 | end if |
| 350 | arr_val = get_array_element(shell, trim(array_name), arr_idx + 1) |
| 351 | if (is_length_expansion) then |
| 352 | write(num_buf, '(I0)') len_trim(arr_val) |
| 353 | result_value = trim(num_buf) |
| 354 | else |
| 355 | result_value = trim(arr_val) |
| 356 | end if |
| 357 | return |
| 358 | end if |
| 359 | end block |
| 360 | end if |
| 361 | end if |
| 362 | end if |
| 363 | end if |
| 364 | |
| 365 | ! ======================================================================== |
| 366 | ! Handle indirect expansion prefix: ${!ref...} |
| 367 | ! Resolve !name to the variable name it references, then continue with |
| 368 | ! normal operator handling so ${!ref:-default} works correctly. |
| 369 | ! NOTE: Only one level of indirection is resolved (no recursion). |
| 370 | ! Circular references (x->y->x) safely stop after one resolution. |
| 371 | ! ======================================================================== |
| 372 | if (len_trim(var_name) > 1 .and. var_name(1:1) == '!' .and. index(var_name, '[') == 0) then |
| 373 | block |
| 374 | integer :: ref_end |
| 375 | character(len=4096) :: ref_name, resolved_name |
| 376 | ! Extract reference variable name (alphanumeric/underscore chars after !) |
| 377 | ref_end = 2 |
| 378 | do while (ref_end <= len_trim(var_name)) |
| 379 | if (.not. (var_name(ref_end:ref_end) >= 'a' .and. var_name(ref_end:ref_end) <= 'z') .and. & |
| 380 | .not. (var_name(ref_end:ref_end) >= 'A' .and. var_name(ref_end:ref_end) <= 'Z') .and. & |
| 381 | .not. (var_name(ref_end:ref_end) >= '0' .and. var_name(ref_end:ref_end) <= '9') .and. & |
| 382 | var_name(ref_end:ref_end) /= '_') exit |
| 383 | ref_end = ref_end + 1 |
| 384 | end do |
| 385 | ref_name = var_name(2:ref_end-1) |
| 386 | resolved_name = get_shell_variable(shell, trim(ref_name)) |
| 387 | if (len_trim(resolved_name) > 0) then |
| 388 | ! Replace !ref with resolved name, keep any trailing operators |
| 389 | if (ref_end <= len_trim(var_name)) then |
| 390 | var_name = trim(resolved_name) // var_name(ref_end:len_trim(var_name)) |
| 391 | else |
| 392 | var_name = trim(resolved_name) |
| 393 | end if |
| 394 | else |
| 395 | ! Reference variable is unset — error + empty (matches bash) |
| 396 | call write_stderr('fortsh: ' // trim(ref_name) // ': invalid indirect expansion') |
| 397 | result_value = '' |
| 398 | return |
| 399 | end if |
| 400 | end block |
| 401 | end if |
| 402 | |
| 403 | ! Check for various expansion operations (need to check in right order!) |
| 404 | |
| 405 | ! Check for @ transformations first: ${var@U}, ${var@L}, ${var@u}, ${var@Q}, ${var@E} |
| 406 | at_pos = index(var_name, '@') |
| 407 | if (at_pos > 0 .and. at_pos < len_trim(var_name)) then |
| 408 | ! Extract variable name and transformation operator |
| 409 | operation = var_name(:at_pos-1) |
| 410 | transform_op = var_name(at_pos+1:at_pos+1) |
| 411 | var_value = get_shell_variable(shell, trim(operation)) |
| 412 | |
| 413 | select case (transform_op) |
| 414 | case ('U') |
| 415 | ! ${var@U} - convert to uppercase |
| 416 | result_value = to_upper(trim(var_value)) |
| 417 | return |
| 418 | case ('L') |
| 419 | ! ${var@L} - convert to lowercase |
| 420 | result_value = to_lower(trim(var_value)) |
| 421 | return |
| 422 | case ('u') |
| 423 | ! ${var@u} - capitalize first character |
| 424 | if (len_trim(var_value) > 0) then |
| 425 | result_value = to_upper(var_value(1:1)) |
| 426 | if (len_trim(var_value) > 1) result_value = trim(result_value) // var_value(2:) |
| 427 | end if |
| 428 | return |
| 429 | case ('l') |
| 430 | ! ${var@l} - lowercase first character |
| 431 | if (len_trim(var_value) > 0) then |
| 432 | result_value = to_lower(var_value(1:1)) |
| 433 | if (len_trim(var_value) > 1) result_value = trim(result_value) // var_value(2:) |
| 434 | end if |
| 435 | return |
| 436 | case ('Q') |
| 437 | ! ${var@Q} - shell-quote value (wrap in single quotes, escape embedded quotes) |
| 438 | result_value = quote_value(trim(var_value)) |
| 439 | return |
| 440 | case ('E') |
| 441 | ! ${var@E} - expand escape sequences |
| 442 | result_value = expand_escape_sequences(trim(var_value)) |
| 443 | return |
| 444 | end select |
| 445 | end if |
| 446 | |
| 447 | ! Check for case conversion first (^, ^^, ,, ,,) |
| 448 | ! Find the position of ^ or , to determine if it's case conversion |
| 449 | i = len_trim(var_name) |
| 450 | if (i > 1) then |
| 451 | if (var_name(i:i) == '^') then |
| 452 | ! Check for ^^ (uppercase all) or ^ (uppercase first) |
| 453 | if (i > 1 .and. var_name(i-1:i-1) == '^') then |
| 454 | ! ${var^^} - uppercase all |
| 455 | operation = var_name(:i-2) |
| 456 | var_value = get_shell_variable(shell, trim(operation)) |
| 457 | result_value = to_upper(trim(var_value)) |
| 458 | else |
| 459 | ! ${var^} - uppercase first |
| 460 | operation = var_name(:i-1) |
| 461 | var_value = get_shell_variable(shell, trim(operation)) |
| 462 | if (len_trim(var_value) > 0) then |
| 463 | result_value = to_upper(var_value(1:1)) |
| 464 | if (len_trim(var_value) > 1) result_value = trim(result_value) // var_value(2:) |
| 465 | end if |
| 466 | end if |
| 467 | return |
| 468 | else if (var_name(i:i) == ',') then |
| 469 | ! Check for ,, (lowercase all) or , (lowercase first) |
| 470 | if (i > 1 .and. var_name(i-1:i-1) == ',') then |
| 471 | ! ${var,,} - lowercase all |
| 472 | operation = var_name(:i-2) |
| 473 | var_value = get_shell_variable(shell, trim(operation)) |
| 474 | result_value = to_lower(trim(var_value)) |
| 475 | else |
| 476 | ! ${var,} - lowercase first |
| 477 | operation = var_name(:i-1) |
| 478 | var_value = get_shell_variable(shell, trim(operation)) |
| 479 | if (len_trim(var_value) > 0) then |
| 480 | result_value = to_lower(var_value(1:1)) |
| 481 | if (len_trim(var_value) > 1) result_value = trim(result_value) // var_value(2:) |
| 482 | end if |
| 483 | end if |
| 484 | return |
| 485 | end if |
| 486 | end if |
| 487 | |
| 488 | ! Find operator positions, skipping characters inside nested ${...} |
| 489 | colon_pos = 0; dash_pos = 0; plus_pos = 0; equals_pos = 0 |
| 490 | question_pos = 0; percent_pos = 0; hash_pos = 0; slash_pos = 0 |
| 491 | block |
| 492 | integer :: scan_i, scan_depth |
| 493 | scan_depth = 0 |
| 494 | do scan_i = 1, len_trim(var_name) |
| 495 | ! Track brace nesting: skip everything inside ${...} |
| 496 | if (scan_i + 1 <= len_trim(var_name) .and. & |
| 497 | var_name(scan_i:scan_i) == '$' .and. var_name(scan_i+1:scan_i+1) == '{') then |
| 498 | scan_depth = scan_depth + 1 |
| 499 | else if (var_name(scan_i:scan_i) == '}' .and. scan_depth > 0) then |
| 500 | scan_depth = scan_depth - 1 |
| 501 | end if |
| 502 | if (scan_depth > 0) cycle ! inside nested ${...}, skip |
| 503 | |
| 504 | ! Record first occurrence of each operator |
| 505 | select case (var_name(scan_i:scan_i)) |
| 506 | case (':') |
| 507 | if (colon_pos == 0) colon_pos = scan_i |
| 508 | case ('-') |
| 509 | if (dash_pos == 0) dash_pos = scan_i |
| 510 | case ('+') |
| 511 | if (plus_pos == 0) plus_pos = scan_i |
| 512 | case ('=') |
| 513 | if (equals_pos == 0) equals_pos = scan_i |
| 514 | case ('?') |
| 515 | if (question_pos == 0) question_pos = scan_i |
| 516 | case ('%') |
| 517 | if (percent_pos == 0) percent_pos = scan_i |
| 518 | case ('#') |
| 519 | if (hash_pos == 0) hash_pos = scan_i |
| 520 | case ('/') |
| 521 | if (slash_pos == 0) slash_pos = scan_i |
| 522 | end select |
| 523 | end do |
| 524 | end block |
| 525 | ! write(error_unit, '(A,A,A,I0)') 'DEBUG AFTER OPS: var_name=[', trim(var_name), '] dash_pos=', dash_pos |
| 526 | |
| 527 | ! Pattern replacement: ${var/pattern/replacement} or ${var//pattern/replacement} |
| 528 | ! Only if / appears before # and % (otherwise /*/ in ${var#/*/} is a pattern, not replacement) |
| 529 | if (slash_pos > 0 .and. & |
| 530 | (hash_pos == 0 .or. slash_pos < hash_pos) .and. & |
| 531 | (percent_pos == 0 .or. slash_pos < percent_pos)) then |
| 532 | ! Check if it's replace all (//) — look AFTER slash_pos for second / |
| 533 | operation = var_name(:slash_pos-1) |
| 534 | if (slash_pos < len_trim(var_name) .and. var_name(slash_pos+1:slash_pos+1) == '/') then |
| 535 | replace_all = .true. |
| 536 | ! Skip both slashes: pattern starts after // |
| 537 | i = index(var_name(slash_pos+2:), '/') |
| 538 | if (i > 0) then |
| 539 | i = slash_pos + 1 + i |
| 540 | pattern = var_name(slash_pos+2:i-1) |
| 541 | replacement = var_name(i+1:) |
| 542 | else |
| 543 | pattern = var_name(slash_pos+2:) |
| 544 | replacement = '' |
| 545 | end if |
| 546 | else |
| 547 | replace_all = .false. |
| 548 | ! Single /: find separator slash after pattern |
| 549 | i = index(var_name(slash_pos+1:), '/') |
| 550 | if (i > 0) then |
| 551 | i = slash_pos + i |
| 552 | pattern = var_name(slash_pos+1:i-1) |
| 553 | replacement = var_name(i+1:) |
| 554 | else |
| 555 | pattern = var_name(slash_pos+1:) |
| 556 | replacement = '' |
| 557 | end if |
| 558 | end if |
| 559 | |
| 560 | var_value = get_shell_variable(shell, trim(operation)) |
| 561 | |
| 562 | ! Check for anchor prefix in pattern |
| 563 | if (len_trim(pattern) > 0 .and. pattern(1:1) == '#') then |
| 564 | ! Anchored at start: ${var/#pat/repl} |
| 565 | pattern = pattern(2:) |
| 566 | if (len_trim(pattern) == 0) then |
| 567 | result_value = trim(replacement) // trim(var_value) |
| 568 | else if (len_trim(var_value) >= len_trim(pattern) .and. & |
| 569 | var_value(1:len_trim(pattern)) == trim(pattern)) then |
| 570 | result_value = trim(replacement) // var_value(len_trim(pattern)+1:len_trim(var_value)) |
| 571 | else |
| 572 | result_value = var_value |
| 573 | end if |
| 574 | else if (len_trim(pattern) > 0 .and. pattern(1:1) == '%') then |
| 575 | ! Anchored at end: ${var/%pat/repl} |
| 576 | pattern = pattern(2:) |
| 577 | if (len_trim(pattern) == 0) then |
| 578 | result_value = trim(var_value) // trim(replacement) |
| 579 | else if (len_trim(var_value) >= len_trim(pattern) .and. & |
| 580 | var_value(len_trim(var_value)-len_trim(pattern)+1:len_trim(var_value)) == & |
| 581 | trim(pattern)) then |
| 582 | result_value = var_value(1:len_trim(var_value)-len_trim(pattern)) // trim(replacement) |
| 583 | else |
| 584 | result_value = var_value |
| 585 | end if |
| 586 | else |
| 587 | call pattern_replace(trim(var_value), trim(pattern), trim(replacement), & |
| 588 | replace_all, result_value) |
| 589 | end if |
| 590 | return |
| 591 | end if |
| 592 | |
| 593 | ! Suffix removal: ${var%pattern} or ${var%%pattern} |
| 594 | if (percent_pos > 0) then |
| 595 | ! Check for %% |
| 596 | if (percent_pos < len_trim(var_name) .and. var_name(percent_pos+1:percent_pos+1) == '%') then |
| 597 | ! ${var%%pattern} - remove largest matching suffix |
| 598 | greedy = .true. |
| 599 | operation = var_name(:percent_pos-1) |
| 600 | pattern = var_name(percent_pos+2:) |
| 601 | else |
| 602 | ! ${var%pattern} - remove smallest matching suffix |
| 603 | greedy = .false. |
| 604 | operation = var_name(:percent_pos-1) |
| 605 | pattern = var_name(percent_pos+1:) |
| 606 | end if |
| 607 | var_value = get_shell_variable(shell, trim(operation)) |
| 608 | ! Expand $var or ${...} in pattern |
| 609 | if (index(trim(pattern), '$') > 0) then |
| 610 | pattern = expand_word_operand(trim(pattern), shell) |
| 611 | end if |
| 612 | call remove_suffix(trim(var_value), trim(pattern), greedy, result_value) |
| 613 | return |
| 614 | end if |
| 615 | |
| 616 | ! Prefix removal: ${var#pattern} or ${var##pattern} |
| 617 | ! But first check if it's ${#var} (length) |
| 618 | if (hash_pos == 1) then |
| 619 | ! Check if this is just ${#} (number of positional params) |
| 620 | if (len_trim(var_name) == 1) then |
| 621 | ! ${#} alone - return number of positional parameters |
| 622 | write(num_buf, '(I0)') shell%num_positional |
| 623 | result_value = trim(num_buf) |
| 624 | return |
| 625 | else if (len_trim(var_name) > 1) then |
| 626 | ! ${#var} length expansion |
| 627 | operation = var_name(2:) |
| 628 | |
| 629 | ! Check for special parameters |
| 630 | if (trim(operation) == '@' .or. trim(operation) == '*') then |
| 631 | ! ${#@} or ${#*} - return number of positional parameters |
| 632 | write(num_buf, '(I0)') shell%num_positional |
| 633 | result_value = trim(num_buf) |
| 634 | return |
| 635 | else if (len(trim(operation)) > 0) then |
| 636 | ! Check if it's a positional parameter (digit) |
| 637 | read(operation, *, iostat=i) j |
| 638 | if (i == 0 .and. j > 0) then |
| 639 | ! ${#1}, ${#2}, etc. - return length of specific positional parameter |
| 640 | if (j <= shell%num_positional) then |
| 641 | write(num_buf, '(I0)') len_trim(shell%positional_params(j)%str) |
| 642 | result_value = trim(num_buf) |
| 643 | else |
| 644 | result_value = '0' |
| 645 | end if |
| 646 | return |
| 647 | else |
| 648 | ! Regular variable length — use stored length to preserve whitespace |
| 649 | block |
| 650 | integer :: stored_len |
| 651 | call get_variable_length(shell, trim(operation), stored_len) |
| 652 | if (stored_len >= 0) then |
| 653 | write(num_buf, '(I0)') stored_len |
| 654 | else |
| 655 | var_value = get_shell_variable(shell, trim(operation)) |
| 656 | write(num_buf, '(I0)') len_trim(var_value) |
| 657 | end if |
| 658 | result_value = trim(num_buf) |
| 659 | end block |
| 660 | return |
| 661 | end if |
| 662 | end if |
| 663 | end if |
| 664 | else if (hash_pos > 1) then |
| 665 | ! Check for ## |
| 666 | if (hash_pos < len_trim(var_name) .and. var_name(hash_pos+1:hash_pos+1) == '#') then |
| 667 | ! ${var##pattern} - remove largest matching prefix |
| 668 | greedy = .true. |
| 669 | operation = var_name(:hash_pos-1) |
| 670 | pattern = var_name(hash_pos+2:) |
| 671 | else |
| 672 | ! ${var#pattern} - remove smallest matching prefix |
| 673 | greedy = .false. |
| 674 | operation = var_name(:hash_pos-1) |
| 675 | pattern = var_name(hash_pos+1:) |
| 676 | end if |
| 677 | var_value = get_shell_variable(shell, trim(operation)) |
| 678 | ! Expand $var or ${...} in pattern |
| 679 | if (index(trim(pattern), '$') > 0) then |
| 680 | pattern = expand_word_operand(trim(pattern), shell) |
| 681 | end if |
| 682 | call remove_prefix(trim(var_value), trim(pattern), greedy, result_value) |
| 683 | return |
| 684 | end if |
| 685 | |
| 686 | ! Check if colon is for substring expansion (followed by digit) or parameter expansion (followed by operator) |
| 687 | if (colon_pos > 0) then |
| 688 | ! Check what follows the colon |
| 689 | if (colon_pos < len_trim(var_name)) then |
| 690 | ! If followed by an operator (-+?=), it's parameter expansion, not substring |
| 691 | if (var_name(colon_pos+1:colon_pos+1) == '-' .or. & |
| 692 | var_name(colon_pos+1:colon_pos+1) == '+' .or. & |
| 693 | var_name(colon_pos+1:colon_pos+1) == '?' .or. & |
| 694 | var_name(colon_pos+1:colon_pos+1) == '=') then |
| 695 | ! This is parameter expansion like ${var:-word}, handle it below |
| 696 | else |
| 697 | ! This is substring expansion ${var:offset:length} |
| 698 | call parse_substring_expansion(var_name, operation, param1, param2) |
| 699 | var_value = get_shell_variable(shell, trim(operation)) |
| 700 | |
| 701 | if (len_trim(param1) > 0) then |
| 702 | read(param1, *, iostat=i) offset |
| 703 | if (i /= 0) offset = 0 |
| 704 | ! Handle negative offsets (count from end) |
| 705 | if (offset < 0) offset = len_trim(var_value) + offset |
| 706 | if (offset < 0) offset = 0 |
| 707 | if (len_trim(param2) > 0) then |
| 708 | read(param2, *, iostat=i) length |
| 709 | if (i /= 0) length = 0 |
| 710 | if (offset < len_trim(var_value)) then |
| 711 | i = min(length, len_trim(var_value) - offset) |
| 712 | result_value = var_value(offset+1:offset+i) |
| 713 | end if |
| 714 | else |
| 715 | if (offset < len_trim(var_value)) then |
| 716 | result_value = var_value(offset+1:len_trim(var_value)) |
| 717 | end if |
| 718 | end if |
| 719 | else |
| 720 | result_value = var_value |
| 721 | end if |
| 722 | return |
| 723 | end if |
| 724 | end if |
| 725 | end if |
| 726 | |
| 727 | if (dash_pos > 0 .or. plus_pos > 0 .or. equals_pos > 0 .or. question_pos > 0) then |
| 728 | ! ${var-word}, ${var:-word}, ${var+word}, ${var:+word}, ${var=word}, ${var:=word}, ${var?word}, ${var:?word} |
| 729 | ! write(error_unit, '(A,I0,A,I0,A,I0,A,I0)') 'DEBUG: dash=', dash_pos, & |
| 730 | ! ' plus=', plus_pos, ' eq=', equals_pos, ' q=', question_pos |
| 731 | |
| 732 | ! Determine which operator we have |
| 733 | if (dash_pos > 0 .and. (plus_pos == 0 .or. dash_pos < plus_pos) .and. & |
| 734 | (equals_pos == 0 .or. dash_pos < equals_pos) .and. (question_pos == 0 .or. dash_pos < question_pos)) then |
| 735 | ! Dash operator |
| 736 | ! write(error_unit, '(A)') 'DEBUG: Entering dash operator handler' |
| 737 | has_colon = (dash_pos > 1 .and. var_name(dash_pos-1:dash_pos-1) == ':') |
| 738 | if (has_colon) then |
| 739 | operation = var_name(:dash_pos-2) |
| 740 | param1 = var_name(dash_pos+1:) |
| 741 | else |
| 742 | operation = var_name(:dash_pos-1) |
| 743 | param1 = var_name(dash_pos+1:) |
| 744 | end if |
| 745 | |
| 746 | ! write(error_unit, '(A,L1,A,A,A,A,A)') 'DEBUG: has_colon=', has_colon, & |
| 747 | ! ' op=', trim(operation), ' param1=', trim(param1) |
| 748 | var_is_set = is_shell_variable_set(shell, trim(operation)) |
| 749 | var_value = get_shell_variable(shell, trim(operation)) |
| 750 | var_is_null = (len_trim(var_value) == 0) |
| 751 | ! write(error_unit, '(A,L1,A,A,A,L1)') 'DEBUG: var_is_set=', var_is_set, & |
| 752 | ! ' val=', trim(var_value), ' null=', var_is_null |
| 753 | |
| 754 | ! ${var-word}: use word if var is unset |
| 755 | ! ${var:-word}: use word if var is unset or null |
| 756 | if (has_colon) then |
| 757 | if (.not. var_is_set .or. var_is_null) then |
| 758 | result_value = expand_word_operand(trim(param1), shell) |
| 759 | else |
| 760 | result_value = trim(var_value) |
| 761 | end if |
| 762 | else |
| 763 | if (.not. var_is_set) then |
| 764 | result_value = expand_word_operand(trim(param1), shell) |
| 765 | else |
| 766 | result_value = trim(var_value) |
| 767 | end if |
| 768 | end if |
| 769 | ! write(error_unit, '(A,A,A)') 'DEBUG: result_value=', trim(result_value), '|' |
| 770 | |
| 771 | else if (plus_pos > 0 .and. (equals_pos == 0 .or. plus_pos < equals_pos) .and. & |
| 772 | (question_pos == 0 .or. plus_pos < question_pos)) then |
| 773 | ! Plus operator |
| 774 | has_colon = (plus_pos > 1 .and. var_name(plus_pos-1:plus_pos-1) == ':') |
| 775 | if (has_colon) then |
| 776 | operation = var_name(:plus_pos-2) |
| 777 | param1 = var_name(plus_pos+1:) |
| 778 | else |
| 779 | operation = var_name(:plus_pos-1) |
| 780 | param1 = var_name(plus_pos+1:) |
| 781 | end if |
| 782 | |
| 783 | var_is_set = is_shell_variable_set(shell, trim(operation)) |
| 784 | var_value = get_shell_variable(shell, trim(operation)) |
| 785 | var_is_null = (len_trim(var_value) == 0) |
| 786 | |
| 787 | ! ${var+word}: use word if var is set (even if null) |
| 788 | ! ${var:+word}: use word if var is set and not null |
| 789 | if (has_colon) then |
| 790 | if (var_is_set .and. .not. var_is_null) then |
| 791 | result_value = expand_word_operand(trim(param1), shell) |
| 792 | else |
| 793 | result_value = '' |
| 794 | end if |
| 795 | else |
| 796 | if (var_is_set) then |
| 797 | result_value = expand_word_operand(trim(param1), shell) |
| 798 | else |
| 799 | result_value = '' |
| 800 | end if |
| 801 | end if |
| 802 | |
| 803 | else if (equals_pos > 0 .and. (question_pos == 0 .or. equals_pos < question_pos)) then |
| 804 | ! Equals operator (assign and expand) |
| 805 | has_colon = (equals_pos > 1 .and. var_name(equals_pos-1:equals_pos-1) == ':') |
| 806 | if (has_colon) then |
| 807 | operation = var_name(:equals_pos-2) |
| 808 | param1 = var_name(equals_pos+1:) |
| 809 | else |
| 810 | operation = var_name(:equals_pos-1) |
| 811 | param1 = var_name(equals_pos+1:) |
| 812 | end if |
| 813 | |
| 814 | var_is_set = is_shell_variable_set(shell, trim(operation)) |
| 815 | var_value = get_shell_variable(shell, trim(operation)) |
| 816 | var_is_null = (len_trim(var_value) == 0) |
| 817 | |
| 818 | ! ${var=word}: assign word if var is unset, then expand to var |
| 819 | ! ${var:=word}: assign word if var is unset or null, then expand to var |
| 820 | if (has_colon) then |
| 821 | if (.not. var_is_set .or. var_is_null) then |
| 822 | call set_shell_variable(shell, trim(operation), trim(param1)) |
| 823 | result_value = expand_word_operand(trim(param1), shell) |
| 824 | else |
| 825 | result_value = trim(var_value) |
| 826 | end if |
| 827 | else |
| 828 | if (.not. var_is_set) then |
| 829 | call set_shell_variable(shell, trim(operation), trim(param1)) |
| 830 | result_value = expand_word_operand(trim(param1), shell) |
| 831 | else |
| 832 | result_value = trim(var_value) |
| 833 | end if |
| 834 | end if |
| 835 | |
| 836 | else if (question_pos > 0) then |
| 837 | ! Question operator (error if unset) |
| 838 | has_colon = (question_pos > 1 .and. var_name(question_pos-1:question_pos-1) == ':') |
| 839 | if (has_colon) then |
| 840 | operation = var_name(:question_pos-2) |
| 841 | param1 = var_name(question_pos+1:) |
| 842 | else |
| 843 | operation = var_name(:question_pos-1) |
| 844 | param1 = var_name(question_pos+1:) |
| 845 | end if |
| 846 | |
| 847 | var_is_set = is_shell_variable_set(shell, trim(operation)) |
| 848 | var_value = get_shell_variable(shell, trim(operation)) |
| 849 | var_is_null = (len_trim(var_value) == 0) |
| 850 | |
| 851 | ! ${var?word}: error if var is unset |
| 852 | ! ${var:?word}: error if var is unset or null |
| 853 | if (has_colon) then |
| 854 | if (.not. var_is_set .or. var_is_null) then |
| 855 | if (len_trim(param1) > 0) then |
| 856 | write(error_unit, '(A)') trim(operation) // ': ' // trim(param1) |
| 857 | else |
| 858 | write(error_unit, '(A)') trim(operation) // ': parameter null or not set' |
| 859 | end if |
| 860 | shell%last_exit_status = 127 ! bash uses 127 for direct expansion errors |
| 861 | shell%fatal_expansion_error = .true. ! Signal to abort execution |
| 862 | result_value = '' |
| 863 | else |
| 864 | result_value = trim(var_value) |
| 865 | end if |
| 866 | else |
| 867 | if (.not. var_is_set) then |
| 868 | if (len_trim(param1) > 0) then |
| 869 | write(error_unit, '(A)') trim(operation) // ': ' // trim(param1) |
| 870 | else |
| 871 | write(error_unit, '(A)') trim(operation) // ': parameter not set' |
| 872 | end if |
| 873 | shell%last_exit_status = 127 ! bash uses 127 for direct expansion errors |
| 874 | shell%fatal_expansion_error = .true. ! Signal to abort execution |
| 875 | result_value = '' |
| 876 | else |
| 877 | result_value = trim(var_value) |
| 878 | end if |
| 879 | end if |
| 880 | end if |
| 881 | return |
| 882 | |
| 883 | else if (hash_pos > 0) then |
| 884 | ! ${#var} length expansion |
| 885 | operation = var_name(hash_pos+1:) |
| 886 | var_value = get_shell_variable(shell, trim(operation)) |
| 887 | write(num_buf, '(I0)') len_trim(var_value) |
| 888 | result_value = trim(num_buf) |
| 889 | |
| 890 | else |
| 891 | ! Simple variable expansion |
| 892 | var_value = get_shell_variable(shell, trim(var_name)) |
| 893 | |
| 894 | ! Check if variable is unset and set -u is enabled |
| 895 | if (len_trim(var_value) == 0 .and. .not. is_shell_variable_set(shell, trim(var_name))) then |
| 896 | if (check_nounset(shell, trim(var_name))) then |
| 897 | shell%last_exit_status = 127 ! bash uses 127 for direct expansion errors |
| 898 | shell%fatal_expansion_error = .true. |
| 899 | result_value = '' |
| 900 | return |
| 901 | end if |
| 902 | end if |
| 903 | |
| 904 | result_value = trim(var_value) |
| 905 | end if |
| 906 | |
| 907 | end subroutine process_param_expansion |
| 908 | |
| 909 | ! Expand nested ${...} in a word operand (used by default/assign/error/alternate operators) |
| 910 | recursive function expand_word_operand(word, shell) result(expanded) |
| 911 | character(len=*), intent(in) :: word |
| 912 | type(shell_state_t), intent(inout) :: shell |
| 913 | character(len=:), allocatable :: expanded |
| 914 | integer :: dp, bp, depth |
| 915 | |
| 916 | ! Quick check: if no $, return as-is |
| 917 | dp = index(word, '$') |
| 918 | if (dp == 0) then |
| 919 | expanded = trim(word) |
| 920 | return |
| 921 | end if |
| 922 | |
| 923 | ! Handle simple $VAR (no braces) |
| 924 | if (dp + 1 <= len_trim(word) .and. word(dp+1:dp+1) /= '{') then |
| 925 | block |
| 926 | integer :: vend |
| 927 | character(len=:), allocatable :: vval |
| 928 | vend = dp + 1 |
| 929 | do while (vend <= len_trim(word)) |
| 930 | if (.not. ((word(vend:vend) >= 'a' .and. word(vend:vend) <= 'z') .or. & |
| 931 | (word(vend:vend) >= 'A' .and. word(vend:vend) <= 'Z') .or. & |
| 932 | (word(vend:vend) >= '0' .and. word(vend:vend) <= '9') .or. & |
| 933 | word(vend:vend) == '_')) exit |
| 934 | vend = vend + 1 |
| 935 | end do |
| 936 | if (vend > dp + 1) then |
| 937 | vval = get_shell_variable(shell, word(dp+1:vend-1)) |
| 938 | expanded = word(1:dp-1) // trim(vval) |
| 939 | if (vend <= len_trim(word)) then |
| 940 | expanded = expanded // expand_word_operand(word(vend:), shell) |
| 941 | end if |
| 942 | return |
| 943 | end if |
| 944 | end block |
| 945 | end if |
| 946 | |
| 947 | ! Handle ${...} brace expansion |
| 948 | dp = index(word, '${') |
| 949 | |
| 950 | ! Find matching } counting nesting depth — start AFTER the opening ${ |
| 951 | depth = 1 |
| 952 | bp = dp + 2 |
| 953 | do while (bp <= len_trim(word)) |
| 954 | if (bp + 1 <= len_trim(word) .and. word(bp:bp) == '$' .and. word(bp+1:bp+1) == '{') then |
| 955 | depth = depth + 1 |
| 956 | bp = bp + 2 |
| 957 | cycle |
| 958 | else if (word(bp:bp) == '}') then |
| 959 | depth = depth - 1 |
| 960 | if (depth == 0) exit |
| 961 | end if |
| 962 | bp = bp + 1 |
| 963 | end do |
| 964 | |
| 965 | if (bp <= len_trim(word)) then |
| 966 | ! Recursively expand the inner ${...} |
| 967 | block |
| 968 | character(len=:), allocatable :: inner_result, prefix, suffix |
| 969 | prefix = word(1:dp-1) |
| 970 | call process_param_expansion(word(dp+2:bp-1), inner_result, shell) |
| 971 | suffix = word(bp+1:len_trim(word)) |
| 972 | ! Recurse on suffix in case there are more ${...} |
| 973 | if (index(suffix, '${') > 0) then |
| 974 | expanded = prefix // inner_result // expand_word_operand(suffix, shell) |
| 975 | else |
| 976 | expanded = prefix // inner_result // suffix |
| 977 | end if |
| 978 | end block |
| 979 | else |
| 980 | expanded = trim(word) |
| 981 | end if |
| 982 | end function |
| 983 | |
| 984 | subroutine parse_substring_expansion(input, var_name, offset_str, length_str) |
| 985 | character(len=*), intent(in) :: input |
| 986 | character(len=*), intent(out) :: var_name, offset_str, length_str |
| 987 | integer :: first_colon, second_colon |
| 988 | |
| 989 | var_name = '' |
| 990 | offset_str = '' |
| 991 | length_str = '' |
| 992 | |
| 993 | first_colon = index(input, ':') |
| 994 | if (first_colon == 0) return |
| 995 | |
| 996 | var_name = input(:first_colon-1) |
| 997 | |
| 998 | second_colon = index(input(first_colon+1:), ':') |
| 999 | if (second_colon > 0) then |
| 1000 | second_colon = first_colon + second_colon |
| 1001 | offset_str = input(first_colon+1:second_colon-1) |
| 1002 | length_str = input(second_colon+1:) |
| 1003 | else |
| 1004 | offset_str = input(first_colon+1:) |
| 1005 | end if |
| 1006 | end subroutine |
| 1007 | |
| 1008 | ! ============================================================================ |
| 1009 | ! Parameter Expansion Helper Functions |
| 1010 | ! ============================================================================ |
| 1011 | |
| 1012 | ! Convert string to uppercase |
| 1013 | function to_upper(input) result(output) |
| 1014 | character(len=*), intent(in) :: input |
| 1015 | character(len=len(input)) :: output |
| 1016 | integer :: i, char_code |
| 1017 | |
| 1018 | output = input |
| 1019 | do i = 1, len_trim(input) |
| 1020 | char_code = ichar(input(i:i)) |
| 1021 | if (char_code >= ichar('a') .and. char_code <= ichar('z')) then |
| 1022 | output(i:i) = char(char_code - 32) |
| 1023 | end if |
| 1024 | end do |
| 1025 | end function |
| 1026 | |
| 1027 | ! Convert string to lowercase |
| 1028 | function to_lower(input) result(output) |
| 1029 | character(len=*), intent(in) :: input |
| 1030 | character(len=len(input)) :: output |
| 1031 | integer :: i, char_code |
| 1032 | |
| 1033 | output = input |
| 1034 | do i = 1, len_trim(input) |
| 1035 | char_code = ichar(input(i:i)) |
| 1036 | if (char_code >= ichar('A') .and. char_code <= ichar('Z')) then |
| 1037 | output(i:i) = char(char_code + 32) |
| 1038 | end if |
| 1039 | end do |
| 1040 | end function |
| 1041 | |
| 1042 | ! Quote value - wrap in single quotes and escape embedded single quotes |
| 1043 | ! Used for ${var@Q} transformation |
| 1044 | function quote_value(input) result(output) |
| 1045 | character(len=*), intent(in) :: input |
| 1046 | character(len=:), allocatable :: output |
| 1047 | character(len=:), allocatable :: temp_output |
| 1048 | integer :: i, out_pos, capacity |
| 1049 | |
| 1050 | if (len_trim(input) == 0) then |
| 1051 | output = "''" |
| 1052 | return |
| 1053 | end if |
| 1054 | |
| 1055 | ! Allocate with initial capacity |
| 1056 | capacity = len(input) * 2 + 10 |
| 1057 | allocate(character(len=capacity) :: temp_output) |
| 1058 | temp_output = "'" |
| 1059 | out_pos = 2 |
| 1060 | |
| 1061 | do i = 1, len_trim(input) |
| 1062 | if (input(i:i) == "'") then |
| 1063 | ! Ensure we have space for escape sequence |
| 1064 | if (out_pos + 4 > capacity) then |
| 1065 | call grow_string_buffer_exp(temp_output, capacity, capacity * 2, out_pos - 1) |
| 1066 | end if |
| 1067 | ! Escape single quote: ' becomes '\'' |
| 1068 | temp_output(out_pos:out_pos+3) = "'\'''" |
| 1069 | out_pos = out_pos + 4 |
| 1070 | else |
| 1071 | if (out_pos > capacity) then |
| 1072 | call grow_string_buffer_exp(temp_output, capacity, capacity * 2, out_pos - 1) |
| 1073 | end if |
| 1074 | temp_output(out_pos:out_pos) = input(i:i) |
| 1075 | out_pos = out_pos + 1 |
| 1076 | end if |
| 1077 | end do |
| 1078 | |
| 1079 | if (out_pos > capacity) then |
| 1080 | call grow_string_buffer_exp(temp_output, capacity, capacity * 2, out_pos - 1) |
| 1081 | end if |
| 1082 | temp_output(out_pos:out_pos) = "'" |
| 1083 | output = temp_output(1:out_pos) |
| 1084 | deallocate(temp_output) |
| 1085 | end function |
| 1086 | |
| 1087 | ! Expand escape sequences in string |
| 1088 | ! Used for ${var@E} transformation |
| 1089 | function expand_escape_sequences(input) result(output) |
| 1090 | character(len=*), intent(in) :: input |
| 1091 | character(len=:), allocatable :: output |
| 1092 | character(len=:), allocatable :: temp_output |
| 1093 | integer :: i, out_pos, capacity, esc_hex_val, esc_n_digits |
| 1094 | character :: esc_ch |
| 1095 | |
| 1096 | ! Allocate with initial capacity |
| 1097 | capacity = len(input) + 100 |
| 1098 | allocate(character(len=capacity) :: temp_output) |
| 1099 | temp_output = '' |
| 1100 | out_pos = 1 |
| 1101 | i = 1 |
| 1102 | |
| 1103 | do while (i <= len_trim(input)) |
| 1104 | if (input(i:i) == '\' .and. i < len_trim(input)) then |
| 1105 | ! Escape sequence |
| 1106 | i = i + 1 |
| 1107 | select case (input(i:i)) |
| 1108 | case ('n') |
| 1109 | temp_output(out_pos:out_pos) = char(10) ! newline |
| 1110 | out_pos = out_pos + 1 |
| 1111 | case ('t') |
| 1112 | temp_output(out_pos:out_pos) = char(9) ! tab |
| 1113 | out_pos = out_pos + 1 |
| 1114 | case ('r') |
| 1115 | temp_output(out_pos:out_pos) = char(13) ! carriage return |
| 1116 | out_pos = out_pos + 1 |
| 1117 | case ('b') |
| 1118 | temp_output(out_pos:out_pos) = char(8) ! backspace |
| 1119 | out_pos = out_pos + 1 |
| 1120 | case ('f') |
| 1121 | temp_output(out_pos:out_pos) = char(12) ! form feed |
| 1122 | out_pos = out_pos + 1 |
| 1123 | case ('v') |
| 1124 | temp_output(out_pos:out_pos) = char(11) ! vertical tab |
| 1125 | out_pos = out_pos + 1 |
| 1126 | case ('a') |
| 1127 | temp_output(out_pos:out_pos) = char(7) ! alert/bell |
| 1128 | out_pos = out_pos + 1 |
| 1129 | case ('e', 'E') |
| 1130 | temp_output(out_pos:out_pos) = char(27) ! escape |
| 1131 | out_pos = out_pos + 1 |
| 1132 | case ('\') |
| 1133 | temp_output(out_pos:out_pos) = '\' ! backslash |
| 1134 | out_pos = out_pos + 1 |
| 1135 | case ('"') |
| 1136 | temp_output(out_pos:out_pos) = '"' ! double quote |
| 1137 | out_pos = out_pos + 1 |
| 1138 | case ("'") |
| 1139 | temp_output(out_pos:out_pos) = "'" ! single quote |
| 1140 | out_pos = out_pos + 1 |
| 1141 | case ('x') |
| 1142 | ! Hex escape: \xHH (up to 2 hex digits) |
| 1143 | esc_hex_val = 0; esc_n_digits = 0 |
| 1144 | i = i + 1 ! skip 'x' |
| 1145 | do while (i <= len_trim(input) .and. esc_n_digits < 2) |
| 1146 | esc_ch = input(i:i) |
| 1147 | if (esc_ch >= '0' .and. esc_ch <= '9') then |
| 1148 | esc_hex_val = esc_hex_val * 16 + (ichar(esc_ch) - ichar('0')) |
| 1149 | else if (esc_ch >= 'a' .and. esc_ch <= 'f') then |
| 1150 | esc_hex_val = esc_hex_val * 16 + (ichar(esc_ch) - ichar('a') + 10) |
| 1151 | else if (esc_ch >= 'A' .and. esc_ch <= 'F') then |
| 1152 | esc_hex_val = esc_hex_val * 16 + (ichar(esc_ch) - ichar('A') + 10) |
| 1153 | else |
| 1154 | exit |
| 1155 | end if |
| 1156 | i = i + 1 |
| 1157 | esc_n_digits = esc_n_digits + 1 |
| 1158 | end do |
| 1159 | if (esc_n_digits > 0 .and. esc_hex_val <= 255) then |
| 1160 | temp_output(out_pos:out_pos) = char(esc_hex_val) |
| 1161 | out_pos = out_pos + 1 |
| 1162 | end if |
| 1163 | cycle ! i already advanced past hex digits |
| 1164 | case ('0', '1', '2', '3', '4', '5', '6', '7') |
| 1165 | ! Octal escape: \nnn (up to 3 octal digits) |
| 1166 | esc_hex_val = 0; esc_n_digits = 0 |
| 1167 | do while (i <= len_trim(input) .and. esc_n_digits < 3) |
| 1168 | esc_ch = input(i:i) |
| 1169 | if (esc_ch >= '0' .and. esc_ch <= '7') then |
| 1170 | esc_hex_val = esc_hex_val * 8 + (ichar(esc_ch) - ichar('0')) |
| 1171 | else |
| 1172 | exit |
| 1173 | end if |
| 1174 | i = i + 1 |
| 1175 | esc_n_digits = esc_n_digits + 1 |
| 1176 | end do |
| 1177 | if (esc_n_digits > 0 .and. esc_hex_val <= 255) then |
| 1178 | temp_output(out_pos:out_pos) = char(esc_hex_val) |
| 1179 | out_pos = out_pos + 1 |
| 1180 | end if |
| 1181 | cycle ! i already advanced past octal digits |
| 1182 | case default |
| 1183 | ! Unknown escape - preserve backslash and character |
| 1184 | temp_output(out_pos:out_pos+1) = '\' // input(i:i) |
| 1185 | out_pos = out_pos + 2 |
| 1186 | end select |
| 1187 | i = i + 1 |
| 1188 | else |
| 1189 | temp_output(out_pos:out_pos) = input(i:i) |
| 1190 | out_pos = out_pos + 1 |
| 1191 | i = i + 1 |
| 1192 | end if |
| 1193 | end do |
| 1194 | |
| 1195 | output = temp_output(1:out_pos-1) |
| 1196 | deallocate(temp_output) |
| 1197 | end function |
| 1198 | |
| 1199 | ! Pattern replacement in string |
| 1200 | subroutine pattern_replace(input, pattern, replacement, replace_all, output) |
| 1201 | character(len=*), intent(in) :: input, pattern, replacement |
| 1202 | logical, intent(in) :: replace_all |
| 1203 | character(len=:), allocatable, intent(out) :: output |
| 1204 | #ifdef USE_C_STRINGS |
| 1205 | ! flang-new path: route through C to avoid allocatable heap corruption |
| 1206 | ! But if pattern contains glob characters ([, *, ?), use Fortran path |
| 1207 | ! because the C pattern_replace does literal matching only |
| 1208 | integer :: in_len, pat_len, repl_len, result_len, rc |
| 1209 | integer(c_int) :: c_replace_all |
| 1210 | type(c_ptr) :: c_result_ptr, ebuf |
| 1211 | integer(c_size_t) :: copied |
| 1212 | character(kind=c_char), pointer :: raw(:) |
| 1213 | |
| 1214 | in_len = len_trim(input) |
| 1215 | pat_len = len_trim(pattern) |
| 1216 | repl_len = len_trim(replacement) |
| 1217 | |
| 1218 | ! Use glob-aware Fortran path for patterns with wildcards/charclasses |
| 1219 | if (pat_len > 0 .and. (index(pattern(1:pat_len), '[') > 0 .or. & |
| 1220 | index(pattern(1:pat_len), '*') > 0 .or. index(pattern(1:pat_len), '?') > 0)) then |
| 1221 | call pattern_replace_glob(input, pattern, replacement, replace_all, output) |
| 1222 | return |
| 1223 | end if |
| 1224 | |
| 1225 | if (pat_len == 0) then |
| 1226 | if (in_len > 0) then |
| 1227 | ebuf = c_buf_create(int(in_len + 1, c_size_t)) |
| 1228 | rc = c_buf_append_chars(ebuf, input, int(in_len, c_size_t)) |
| 1229 | allocate(character(len=in_len) :: output) |
| 1230 | copied = c_buf_to_fortran(ebuf, output, int(in_len, c_size_t)) |
| 1231 | call c_buf_destroy(ebuf) |
| 1232 | else |
| 1233 | output = '' |
| 1234 | end if |
| 1235 | return |
| 1236 | end if |
| 1237 | |
| 1238 | if (replace_all) then |
| 1239 | c_replace_all = 1_c_int |
| 1240 | else |
| 1241 | c_replace_all = 0_c_int |
| 1242 | end if |
| 1243 | |
| 1244 | ebuf = c_buf_create(int(in_len + 1, c_size_t)) |
| 1245 | rc = c_buf_append_chars(ebuf, input, int(in_len, c_size_t)) |
| 1246 | result_len = c_pattern_replace_alloc(input, int(in_len, c_int), & |
| 1247 | pattern, int(pat_len, c_int), & |
| 1248 | replacement, int(repl_len, c_int), & |
| 1249 | c_replace_all, c_result_ptr) |
| 1250 | call c_buf_destroy(ebuf) |
| 1251 | |
| 1252 | if (result_len > 0) then |
| 1253 | call c_f_pointer(c_result_ptr, raw, [result_len]) |
| 1254 | ebuf = c_buf_create(int(result_len + 1, c_size_t)) |
| 1255 | rc = c_buf_append_chars(ebuf, raw, int(result_len, c_size_t)) |
| 1256 | allocate(character(len=result_len) :: output) |
| 1257 | copied = c_buf_to_fortran(ebuf, output, int(result_len, c_size_t)) |
| 1258 | call c_buf_destroy(ebuf) |
| 1259 | else |
| 1260 | output = '' |
| 1261 | end if |
| 1262 | call c_free_string(c_result_ptr) |
| 1263 | #else |
| 1264 | ! gfortran path: character-by-character scan, no C dependencies |
| 1265 | integer :: in_len, pat_len, repl_len, i2, j2 |
| 1266 | integer :: out_pos, out_cap |
| 1267 | logical :: matched |
| 1268 | character(len=:), allocatable :: result_buf |
| 1269 | |
| 1270 | in_len = len_trim(input) |
| 1271 | pat_len = len_trim(pattern) |
| 1272 | repl_len = len_trim(replacement) |
| 1273 | |
| 1274 | if (pat_len == 0) then |
| 1275 | output = input(1:in_len) |
| 1276 | return |
| 1277 | end if |
| 1278 | |
| 1279 | if (repl_len > pat_len) then |
| 1280 | out_cap = in_len + (in_len / pat_len + 1) * (repl_len - pat_len) + 1 |
| 1281 | else |
| 1282 | out_cap = in_len + 1 |
| 1283 | end if |
| 1284 | allocate(character(len=out_cap) :: result_buf) |
| 1285 | out_pos = 1 |
| 1286 | |
| 1287 | i2 = 1 |
| 1288 | do while (i2 <= in_len) |
| 1289 | ! Try glob pattern match at position i2 — find shortest match |
| 1290 | matched = .false. |
| 1291 | do j2 = 1, in_len - i2 + 1 |
| 1292 | if (pattern_matches_no_dotfile_check(pattern(1:pat_len), input(i2:i2+j2-1))) then |
| 1293 | matched = .true. |
| 1294 | exit ! j2 = matched length |
| 1295 | end if |
| 1296 | end do |
| 1297 | if (matched) then |
| 1298 | if (repl_len > 0) then |
| 1299 | result_buf(out_pos:out_pos + repl_len - 1) = replacement(1:repl_len) |
| 1300 | out_pos = out_pos + repl_len |
| 1301 | end if |
| 1302 | i2 = i2 + j2 ! skip matched portion |
| 1303 | if (.not. replace_all) then |
| 1304 | do while (i2 <= in_len) |
| 1305 | result_buf(out_pos:out_pos) = input(i2:i2) |
| 1306 | out_pos = out_pos + 1 |
| 1307 | i2 = i2 + 1 |
| 1308 | end do |
| 1309 | exit |
| 1310 | end if |
| 1311 | else |
| 1312 | result_buf(out_pos:out_pos) = input(i2:i2) |
| 1313 | out_pos = out_pos + 1 |
| 1314 | i2 = i2 + 1 |
| 1315 | end if |
| 1316 | end do |
| 1317 | if (out_pos > 1) then |
| 1318 | output = result_buf(1:out_pos - 1) |
| 1319 | else |
| 1320 | output = '' |
| 1321 | end if |
| 1322 | deallocate(result_buf) |
| 1323 | #endif |
| 1324 | end subroutine |
| 1325 | |
| 1326 | ! Glob-aware pattern replace — handles *, ?, [charclass] via glob module |
| 1327 | ! Used on all platforms; the C path falls through here for glob patterns. |
| 1328 | subroutine pattern_replace_glob(input, pattern, replacement, replace_all, output) |
| 1329 | character(len=*), intent(in) :: input, pattern, replacement |
| 1330 | logical, intent(in) :: replace_all |
| 1331 | character(len=:), allocatable, intent(out) :: output |
| 1332 | integer :: in_len, pat_len, repl_len, i2, j2 |
| 1333 | integer :: out_pos, out_cap |
| 1334 | logical :: matched |
| 1335 | character(len=:), allocatable :: result_buf |
| 1336 | |
| 1337 | in_len = len_trim(input) |
| 1338 | pat_len = len_trim(pattern) |
| 1339 | repl_len = len_trim(replacement) |
| 1340 | |
| 1341 | if (pat_len == 0) then |
| 1342 | output = input(1:in_len) |
| 1343 | return |
| 1344 | end if |
| 1345 | |
| 1346 | out_cap = in_len * 2 + 1 |
| 1347 | allocate(character(len=out_cap) :: result_buf) |
| 1348 | out_pos = 1 |
| 1349 | i2 = 1 |
| 1350 | do while (i2 <= in_len) |
| 1351 | matched = .false. |
| 1352 | do j2 = 1, in_len - i2 + 1 |
| 1353 | if (pattern_matches_no_dotfile_check(pattern(1:pat_len), input(i2:i2+j2-1))) then |
| 1354 | matched = .true. |
| 1355 | exit |
| 1356 | end if |
| 1357 | end do |
| 1358 | if (matched) then |
| 1359 | if (repl_len > 0) then |
| 1360 | if (out_pos + repl_len - 1 > out_cap) then |
| 1361 | call grow_string_buffer_exp(result_buf, out_cap, out_cap * 2, out_pos - 1) |
| 1362 | end if |
| 1363 | result_buf(out_pos:out_pos + repl_len - 1) = replacement(1:repl_len) |
| 1364 | out_pos = out_pos + repl_len |
| 1365 | end if |
| 1366 | i2 = i2 + j2 |
| 1367 | if (.not. replace_all) then |
| 1368 | do while (i2 <= in_len) |
| 1369 | result_buf(out_pos:out_pos) = input(i2:i2) |
| 1370 | out_pos = out_pos + 1 |
| 1371 | i2 = i2 + 1 |
| 1372 | end do |
| 1373 | exit |
| 1374 | end if |
| 1375 | else |
| 1376 | result_buf(out_pos:out_pos) = input(i2:i2) |
| 1377 | out_pos = out_pos + 1 |
| 1378 | i2 = i2 + 1 |
| 1379 | end if |
| 1380 | end do |
| 1381 | if (out_pos > 1) then |
| 1382 | output = result_buf(1:out_pos - 1) |
| 1383 | else |
| 1384 | output = '' |
| 1385 | end if |
| 1386 | deallocate(result_buf) |
| 1387 | end subroutine |
| 1388 | |
| 1389 | #ifdef USE_C_STRINGS |
| 1390 | ! Pattern replace: C buffer in → C replace → result stored in C buffer, |
| 1391 | ! extracted to Fortran allocatable via c_buf_to_fortran (single memcpy). |
| 1392 | subroutine pattern_replace_cbuf_to_expanded(input_buf, in_len, pattern, replacement, & |
| 1393 | replace_all, output) |
| 1394 | type(c_ptr), intent(in) :: input_buf |
| 1395 | integer, intent(in) :: in_len |
| 1396 | character(len=*), intent(in) :: pattern, replacement |
| 1397 | logical, intent(in) :: replace_all |
| 1398 | character(len=:), allocatable, intent(out) :: output |
| 1399 | integer :: pat_len, repl_len, result_len, rc |
| 1400 | integer(c_int) :: c_replace_all |
| 1401 | type(c_ptr) :: c_result_ptr, rbuf |
| 1402 | integer(c_size_t) :: copied, buf_len |
| 1403 | character(kind=c_char), pointer :: raw(:) |
| 1404 | |
| 1405 | pat_len = len_trim(pattern) |
| 1406 | repl_len = len_trim(replacement) |
| 1407 | |
| 1408 | if (pat_len == 0 .or. in_len == 0) then |
| 1409 | if (in_len > 0) then |
| 1410 | allocate(character(len=in_len) :: output) |
| 1411 | copied = c_buf_to_fortran(input_buf, output, int(in_len, c_size_t)) |
| 1412 | else |
| 1413 | output = '' |
| 1414 | end if |
| 1415 | return |
| 1416 | end if |
| 1417 | |
| 1418 | if (replace_all) then |
| 1419 | c_replace_all = 1_c_int |
| 1420 | else |
| 1421 | c_replace_all = 0_c_int |
| 1422 | end if |
| 1423 | |
| 1424 | result_len = c_buf_pattern_replace(input_buf, pattern, int(pat_len, c_int), & |
| 1425 | replacement, int(repl_len, c_int), & |
| 1426 | c_replace_all, c_result_ptr) |
| 1427 | |
| 1428 | if (result_len > 0) then |
| 1429 | ! Wrap C result in a buffer, then extract to Fortran in one memcpy |
| 1430 | call c_f_pointer(c_result_ptr, raw, [result_len]) |
| 1431 | rbuf = c_buf_create(int(result_len + 1, c_size_t)) |
| 1432 | rc = c_buf_append_chars(rbuf, raw, int(result_len, c_size_t)) |
| 1433 | allocate(character(len=result_len) :: output) |
| 1434 | copied = c_buf_to_fortran(rbuf, output, int(result_len, c_size_t)) |
| 1435 | call c_buf_destroy(rbuf) |
| 1436 | else |
| 1437 | output = '' |
| 1438 | end if |
| 1439 | |
| 1440 | call c_free_string(c_result_ptr) |
| 1441 | end subroutine |
| 1442 | #endif |
| 1443 | |
| 1444 | ! Remove suffix matching pattern (greedy or non-greedy) |
| 1445 | subroutine remove_suffix(input, pattern, greedy, output) |
| 1446 | character(len=*), intent(in) :: input, pattern |
| 1447 | logical, intent(in) :: greedy |
| 1448 | character(len=:), allocatable, intent(out) :: output |
| 1449 | integer :: best_pos, i |
| 1450 | |
| 1451 | output = input(1:len_trim(input)) |
| 1452 | |
| 1453 | if (len_trim(pattern) == 0) return |
| 1454 | |
| 1455 | if (greedy) then |
| 1456 | ! Remove largest matching suffix (%%) |
| 1457 | ! Try to match from start of string |
| 1458 | best_pos = 0 |
| 1459 | do i = 1, len_trim(input) |
| 1460 | if (match_pattern(input(i:), trim(pattern))) then |
| 1461 | best_pos = i |
| 1462 | exit |
| 1463 | end if |
| 1464 | end do |
| 1465 | if (best_pos > 0) then |
| 1466 | output = input(1:best_pos-1) |
| 1467 | end if |
| 1468 | else |
| 1469 | ! Remove smallest matching suffix (%) |
| 1470 | ! Try to match from end of string - include len_trim+1 for empty suffix |
| 1471 | do i = len_trim(input) + 1, 1, -1 |
| 1472 | if (match_pattern(input(i:), trim(pattern))) then |
| 1473 | output = input(1:i-1) |
| 1474 | return |
| 1475 | end if |
| 1476 | end do |
| 1477 | end if |
| 1478 | end subroutine |
| 1479 | |
| 1480 | ! Remove prefix matching pattern (greedy or non-greedy) |
| 1481 | subroutine remove_prefix(input, pattern, greedy, output) |
| 1482 | character(len=*), intent(in) :: input, pattern |
| 1483 | logical, intent(in) :: greedy |
| 1484 | character(len=:), allocatable, intent(out) :: output |
| 1485 | integer :: best_pos, i |
| 1486 | |
| 1487 | output = input(1:len_trim(input)) |
| 1488 | |
| 1489 | if (len_trim(pattern) == 0) return |
| 1490 | |
| 1491 | if (greedy) then |
| 1492 | ! Remove largest matching prefix (##) |
| 1493 | ! Try to match from end, working backwards |
| 1494 | best_pos = 0 |
| 1495 | do i = len_trim(input), 1, -1 |
| 1496 | if (match_pattern(input(1:i), trim(pattern))) then |
| 1497 | best_pos = i |
| 1498 | exit |
| 1499 | end if |
| 1500 | end do |
| 1501 | if (best_pos > 0) then |
| 1502 | output = input(best_pos+1:) |
| 1503 | end if |
| 1504 | else |
| 1505 | ! Remove smallest matching prefix (#) |
| 1506 | ! Try to match from start - start at 0 to test empty prefix |
| 1507 | do i = 0, len_trim(input) |
| 1508 | if (match_pattern(input(1:i), trim(pattern))) then |
| 1509 | output = input(i+1:) |
| 1510 | return |
| 1511 | end if |
| 1512 | end do |
| 1513 | end if |
| 1514 | end subroutine |
| 1515 | |
| 1516 | ! Simple pattern matching (supports * and ? wildcards) |
| 1517 | ! Delegate to glob module's pattern matcher which handles *, ?, [...] |
| 1518 | function match_pattern(str, pattern) result(matches) |
| 1519 | character(len=*), intent(in) :: str, pattern |
| 1520 | logical :: matches |
| 1521 | matches = pattern_matches_no_dotfile_check(pattern, str) |
| 1522 | end function |
| 1523 | |
| 1524 | ! ============================================================================ |
| 1525 | ! Arithmetic Expansion: $((expression)) |
| 1526 | ! Comprehensive arithmetic evaluator with full operator support |
| 1527 | ! ============================================================================ |
| 1528 | |
| 1529 | ! Note: This version doesn't have shell context - used when called from parser |
| 1530 | function arithmetic_expansion(expression) result(result_value) |
| 1531 | character(len=*), intent(in) :: expression |
| 1532 | character(len=32) :: result_value |
| 1533 | character(len=512) :: expr |
| 1534 | integer(kind=8) :: result_int |
| 1535 | |
| 1536 | result_value = '0' |
| 1537 | |
| 1538 | ! Remove $(( and )) |
| 1539 | if (len_trim(expression) < 6) return |
| 1540 | expr = adjustl(expression(4:len_trim(expression)-2)) |
| 1541 | |
| 1542 | ! Clear any previous error |
| 1543 | arithmetic_error = .false. |
| 1544 | arithmetic_error_msg = '' |
| 1545 | |
| 1546 | ! Evaluate the arithmetic expression (without shell context for variable resolution) |
| 1547 | result_int = eval_expression(trim(expr)) |
| 1548 | |
| 1549 | ! Check for arithmetic errors |
| 1550 | if (arithmetic_error) then |
| 1551 | write(error_unit, '(a,a)') 'fortsh: arithmetic expression: ', trim(arithmetic_error_msg) |
| 1552 | result_value = '' ! Return empty string to signal error |
| 1553 | else |
| 1554 | write(result_value, '(I0)') result_int |
| 1555 | end if |
| 1556 | end function |
| 1557 | |
| 1558 | ! Version with shell context for variable resolution |
| 1559 | function arithmetic_expansion_shell(expression, shell) result(result_value) |
| 1560 | character(len=*), intent(in) :: expression |
| 1561 | type(shell_state_t), intent(inout) :: shell |
| 1562 | character(len=32) :: result_value |
| 1563 | character(len=512) :: expr |
| 1564 | character(len=:), allocatable :: expanded_expr |
| 1565 | integer(kind=8) :: result_int |
| 1566 | |
| 1567 | result_value = '0' |
| 1568 | |
| 1569 | ! Remove $(( and )) |
| 1570 | if (len_trim(expression) < 6) return |
| 1571 | expr = adjustl(expression(4:len_trim(expression)-2)) |
| 1572 | |
| 1573 | ! Clear any previous error |
| 1574 | arithmetic_error = .false. |
| 1575 | arithmetic_error_msg = '' |
| 1576 | shell%arithmetic_error = .false. |
| 1577 | shell%arithmetic_error_msg = '' |
| 1578 | |
| 1579 | ! Expand ALL parameter expansions ($var, $1, $(cmd), etc.) before evaluation |
| 1580 | ! This handles variables, positional parameters, and command substitutions |
| 1581 | ! NOTE: Only call enhanced_expand_variables if there are $ characters to expand, |
| 1582 | ! because it has a bug where it strips internal whitespace. |
| 1583 | if (index(expr, '$') > 0) then |
| 1584 | call enhanced_expand_variables(expr, expanded_expr, shell) |
| 1585 | else |
| 1586 | expanded_expr = trim(expr) |
| 1587 | end if |
| 1588 | |
| 1589 | ! Evaluate with shell context for any remaining variable resolution |
| 1590 | result_int = eval_expression_shell(trim(expanded_expr), shell) |
| 1591 | |
| 1592 | ! Check for arithmetic errors |
| 1593 | if (arithmetic_error) then |
| 1594 | write(error_unit, '(a,a)') 'fortsh: arithmetic expression: ', trim(arithmetic_error_msg) |
| 1595 | shell%last_exit_status = 1 ! bash returns 1 for arithmetic errors |
| 1596 | shell%arithmetic_error = .true. |
| 1597 | shell%arithmetic_error_msg = trim(arithmetic_error_msg) |
| 1598 | result_value = '' ! Return empty string to signal error |
| 1599 | else |
| 1600 | write(result_value, '(I0)') result_int |
| 1601 | end if |
| 1602 | end function |
| 1603 | |
| 1604 | ! Main expression evaluator - handles full expressions |
| 1605 | recursive function eval_expression(expr) result(value) |
| 1606 | character(len=*), intent(in) :: expr |
| 1607 | integer(kind=8) :: value |
| 1608 | |
| 1609 | value = eval_ternary(trim(adjustl(expr))) |
| 1610 | end function |
| 1611 | |
| 1612 | ! Ternary conditional operator (? :) |
| 1613 | recursive function eval_ternary(expr) result(value) |
| 1614 | character(len=*), intent(in) :: expr |
| 1615 | integer(kind=8) :: value |
| 1616 | integer :: qmark_pos, colon_pos, depth, i |
| 1617 | character(len=512) :: condition_expr, true_expr, false_expr |
| 1618 | |
| 1619 | ! Find ? outside parentheses |
| 1620 | qmark_pos = 0 |
| 1621 | depth = 0 |
| 1622 | do i = 1, len_trim(expr) |
| 1623 | if (expr(i:i) == '(') then |
| 1624 | depth = depth + 1 |
| 1625 | else if (expr(i:i) == ')') then |
| 1626 | depth = depth - 1 |
| 1627 | else if (depth == 0 .and. expr(i:i) == '?') then |
| 1628 | qmark_pos = i |
| 1629 | exit |
| 1630 | end if |
| 1631 | end do |
| 1632 | |
| 1633 | if (qmark_pos > 0) then |
| 1634 | ! Find matching : after the ? |
| 1635 | colon_pos = 0 |
| 1636 | depth = 0 |
| 1637 | do i = qmark_pos + 1, len_trim(expr) |
| 1638 | if (expr(i:i) == '(') then |
| 1639 | depth = depth + 1 |
| 1640 | else if (expr(i:i) == ')') then |
| 1641 | depth = depth - 1 |
| 1642 | else if (depth == 0 .and. expr(i:i) == ':') then |
| 1643 | colon_pos = i |
| 1644 | exit |
| 1645 | end if |
| 1646 | end do |
| 1647 | |
| 1648 | if (colon_pos > 0) then |
| 1649 | condition_expr = expr(:qmark_pos-1) |
| 1650 | true_expr = expr(qmark_pos+1:colon_pos-1) |
| 1651 | false_expr = expr(colon_pos+1:) |
| 1652 | |
| 1653 | ! Evaluate condition |
| 1654 | value = eval_logical_or(trim(adjustl(condition_expr))) |
| 1655 | if (value /= 0) then |
| 1656 | ! Condition is true, evaluate true expression |
| 1657 | value = eval_ternary(trim(adjustl(true_expr))) |
| 1658 | else |
| 1659 | ! Condition is false, evaluate false expression |
| 1660 | value = eval_ternary(trim(adjustl(false_expr))) |
| 1661 | end if |
| 1662 | return |
| 1663 | end if |
| 1664 | end if |
| 1665 | |
| 1666 | ! No ternary operator found |
| 1667 | value = eval_logical_or(expr) |
| 1668 | end function |
| 1669 | |
| 1670 | ! Logical OR (lowest precedence except ternary) |
| 1671 | recursive function eval_logical_or(expr) result(value) |
| 1672 | character(len=*), intent(in) :: expr |
| 1673 | integer(kind=8) :: value, right_val |
| 1674 | integer :: pos |
| 1675 | character(len=512) :: left_expr, right_expr |
| 1676 | |
| 1677 | value = eval_logical_and(expr) |
| 1678 | |
| 1679 | pos = find_operator(expr, '||') |
| 1680 | if (pos > 0) then |
| 1681 | left_expr = expr(:pos-1) |
| 1682 | right_expr = expr(pos+2:) |
| 1683 | value = eval_logical_and(trim(adjustl(left_expr))) |
| 1684 | right_val = eval_logical_or(trim(adjustl(right_expr))) |
| 1685 | if (value /= 0 .or. right_val /= 0) then |
| 1686 | value = 1 |
| 1687 | else |
| 1688 | value = 0 |
| 1689 | end if |
| 1690 | end if |
| 1691 | end function |
| 1692 | |
| 1693 | ! Logical AND |
| 1694 | recursive function eval_logical_and(expr) result(value) |
| 1695 | character(len=*), intent(in) :: expr |
| 1696 | integer(kind=8) :: value, right_val |
| 1697 | integer :: pos |
| 1698 | character(len=512) :: left_expr, right_expr |
| 1699 | |
| 1700 | value = eval_bitwise_or(expr) |
| 1701 | |
| 1702 | pos = find_operator(expr, '&&') |
| 1703 | if (pos > 0) then |
| 1704 | left_expr = expr(:pos-1) |
| 1705 | right_expr = expr(pos+2:) |
| 1706 | value = eval_bitwise_or(trim(adjustl(left_expr))) |
| 1707 | right_val = eval_logical_and(trim(adjustl(right_expr))) |
| 1708 | if (value /= 0 .and. right_val /= 0) then |
| 1709 | value = 1 |
| 1710 | else |
| 1711 | value = 0 |
| 1712 | end if |
| 1713 | end if |
| 1714 | end function |
| 1715 | |
| 1716 | ! Bitwise OR |
| 1717 | recursive function eval_bitwise_or(expr) result(value) |
| 1718 | character(len=*), intent(in) :: expr |
| 1719 | integer(kind=8) :: value, right_val |
| 1720 | integer :: pos |
| 1721 | character(len=512) :: left_expr, right_expr |
| 1722 | |
| 1723 | value = eval_bitwise_xor(expr) |
| 1724 | |
| 1725 | pos = find_single_operator(expr, '|') |
| 1726 | if (pos > 0) then |
| 1727 | left_expr = expr(:pos-1) |
| 1728 | right_expr = expr(pos+1:) |
| 1729 | value = eval_bitwise_xor(trim(adjustl(left_expr))) |
| 1730 | right_val = eval_bitwise_or(trim(adjustl(right_expr))) |
| 1731 | value = ior(int(value), int(right_val)) |
| 1732 | end if |
| 1733 | end function |
| 1734 | |
| 1735 | ! Bitwise XOR |
| 1736 | recursive function eval_bitwise_xor(expr) result(value) |
| 1737 | character(len=*), intent(in) :: expr |
| 1738 | integer(kind=8) :: value, right_val |
| 1739 | integer :: pos |
| 1740 | character(len=512) :: left_expr, right_expr |
| 1741 | |
| 1742 | value = eval_bitwise_and(expr) |
| 1743 | |
| 1744 | pos = find_single_operator(expr, '^') |
| 1745 | if (pos > 0) then |
| 1746 | left_expr = expr(:pos-1) |
| 1747 | right_expr = expr(pos+1:) |
| 1748 | value = eval_bitwise_and(trim(adjustl(left_expr))) |
| 1749 | right_val = eval_bitwise_xor(trim(adjustl(right_expr))) |
| 1750 | value = ieor(int(value), int(right_val)) |
| 1751 | end if |
| 1752 | end function |
| 1753 | |
| 1754 | ! Bitwise AND |
| 1755 | recursive function eval_bitwise_and(expr) result(value) |
| 1756 | character(len=*), intent(in) :: expr |
| 1757 | integer(kind=8) :: value, right_val |
| 1758 | integer :: pos |
| 1759 | character(len=512) :: left_expr, right_expr |
| 1760 | |
| 1761 | value = eval_equality(expr) |
| 1762 | |
| 1763 | pos = find_single_operator(expr, '&') |
| 1764 | if (pos > 0) then |
| 1765 | left_expr = expr(:pos-1) |
| 1766 | right_expr = expr(pos+1:) |
| 1767 | value = eval_equality(trim(adjustl(left_expr))) |
| 1768 | right_val = eval_bitwise_and(trim(adjustl(right_expr))) |
| 1769 | value = iand(int(value), int(right_val)) |
| 1770 | end if |
| 1771 | end function |
| 1772 | |
| 1773 | ! Equality (==, !=) |
| 1774 | recursive function eval_equality(expr) result(value) |
| 1775 | character(len=*), intent(in) :: expr |
| 1776 | integer(kind=8) :: value, right_val |
| 1777 | integer :: pos |
| 1778 | character(len=512) :: left_expr, right_expr |
| 1779 | |
| 1780 | ! Try == |
| 1781 | pos = find_operator(expr, '==') |
| 1782 | if (pos > 0) then |
| 1783 | left_expr = expr(:pos-1) |
| 1784 | right_expr = expr(pos+2:) |
| 1785 | value = eval_comparison(trim(adjustl(left_expr))) |
| 1786 | right_val = eval_comparison(trim(adjustl(right_expr))) |
| 1787 | if (value == right_val) then |
| 1788 | value = 1 |
| 1789 | else |
| 1790 | value = 0 |
| 1791 | end if |
| 1792 | return |
| 1793 | end if |
| 1794 | |
| 1795 | ! Try != |
| 1796 | pos = find_operator(expr, '!=') |
| 1797 | if (pos > 0) then |
| 1798 | left_expr = expr(:pos-1) |
| 1799 | right_expr = expr(pos+2:) |
| 1800 | value = eval_comparison(trim(adjustl(left_expr))) |
| 1801 | right_val = eval_comparison(trim(adjustl(right_expr))) |
| 1802 | if (value /= right_val) then |
| 1803 | value = 1 |
| 1804 | else |
| 1805 | value = 0 |
| 1806 | end if |
| 1807 | return |
| 1808 | end if |
| 1809 | |
| 1810 | value = eval_comparison(expr) |
| 1811 | end function |
| 1812 | |
| 1813 | ! Comparison (<, <=, >, >=) |
| 1814 | recursive function eval_comparison(expr) result(value) |
| 1815 | character(len=*), intent(in) :: expr |
| 1816 | integer(kind=8) :: value, right_val |
| 1817 | integer :: pos |
| 1818 | character(len=512) :: left_expr, right_expr |
| 1819 | |
| 1820 | ! Try <= |
| 1821 | pos = find_operator(expr, '<=') |
| 1822 | if (pos > 0) then |
| 1823 | left_expr = expr(:pos-1) |
| 1824 | right_expr = expr(pos+2:) |
| 1825 | value = eval_shift(trim(adjustl(left_expr))) |
| 1826 | right_val = eval_shift(trim(adjustl(right_expr))) |
| 1827 | if (value <= right_val) then |
| 1828 | value = 1 |
| 1829 | else |
| 1830 | value = 0 |
| 1831 | end if |
| 1832 | return |
| 1833 | end if |
| 1834 | |
| 1835 | ! Try >= |
| 1836 | pos = find_operator(expr, '>=') |
| 1837 | if (pos > 0) then |
| 1838 | left_expr = expr(:pos-1) |
| 1839 | right_expr = expr(pos+2:) |
| 1840 | value = eval_shift(trim(adjustl(left_expr))) |
| 1841 | right_val = eval_shift(trim(adjustl(right_expr))) |
| 1842 | if (value >= right_val) then |
| 1843 | value = 1 |
| 1844 | else |
| 1845 | value = 0 |
| 1846 | end if |
| 1847 | return |
| 1848 | end if |
| 1849 | |
| 1850 | ! Try < |
| 1851 | pos = find_single_operator(expr, '<') |
| 1852 | if (pos > 0) then |
| 1853 | left_expr = expr(:pos-1) |
| 1854 | right_expr = expr(pos+1:) |
| 1855 | value = eval_shift(trim(adjustl(left_expr))) |
| 1856 | right_val = eval_shift(trim(adjustl(right_expr))) |
| 1857 | if (value < right_val) then |
| 1858 | value = 1 |
| 1859 | else |
| 1860 | value = 0 |
| 1861 | end if |
| 1862 | return |
| 1863 | end if |
| 1864 | |
| 1865 | ! Try > |
| 1866 | pos = find_single_operator(expr, '>') |
| 1867 | if (pos > 0) then |
| 1868 | left_expr = expr(:pos-1) |
| 1869 | right_expr = expr(pos+1:) |
| 1870 | value = eval_shift(trim(adjustl(left_expr))) |
| 1871 | right_val = eval_shift(trim(adjustl(right_expr))) |
| 1872 | if (value > right_val) then |
| 1873 | value = 1 |
| 1874 | else |
| 1875 | value = 0 |
| 1876 | end if |
| 1877 | return |
| 1878 | end if |
| 1879 | |
| 1880 | value = eval_shift(expr) |
| 1881 | end function |
| 1882 | |
| 1883 | ! Shift operations (<<, >>) |
| 1884 | recursive function eval_shift(expr) result(value) |
| 1885 | character(len=*), intent(in) :: expr |
| 1886 | integer(kind=8) :: value, right_val |
| 1887 | integer :: pos |
| 1888 | character(len=512) :: left_expr, right_expr |
| 1889 | |
| 1890 | ! Try << (left shift) |
| 1891 | pos = find_operator(expr, '<<') |
| 1892 | if (pos > 0) then |
| 1893 | left_expr = expr(:pos-1) |
| 1894 | right_expr = expr(pos+2:) |
| 1895 | value = eval_additive(trim(adjustl(left_expr))) ! Changed from eval_shift |
| 1896 | right_val = eval_additive(trim(adjustl(right_expr))) |
| 1897 | ! Left shift by right_val bits |
| 1898 | value = ishft(value, int(right_val)) |
| 1899 | return |
| 1900 | end if |
| 1901 | |
| 1902 | ! Try >> (right shift) |
| 1903 | pos = find_operator(expr, '>>') |
| 1904 | if (pos > 0) then |
| 1905 | left_expr = expr(:pos-1) |
| 1906 | right_expr = expr(pos+2:) |
| 1907 | value = eval_additive(trim(adjustl(left_expr))) ! Changed from eval_shift |
| 1908 | right_val = eval_additive(trim(adjustl(right_expr))) |
| 1909 | ! Right shift by right_val bits (negative for right shift in ishft) |
| 1910 | value = ishft(value, -int(right_val)) |
| 1911 | return |
| 1912 | end if |
| 1913 | |
| 1914 | ! No shift operator found |
| 1915 | value = eval_additive(expr) |
| 1916 | end function |
| 1917 | |
| 1918 | ! Addition and Subtraction |
| 1919 | recursive function eval_additive(expr) result(value) |
| 1920 | character(len=*), intent(in) :: expr |
| 1921 | integer(kind=8) :: value, right_val |
| 1922 | integer :: pos |
| 1923 | character(len=512) :: left_expr, right_expr |
| 1924 | |
| 1925 | ! Find rightmost + or - (to maintain left-to-right evaluation) |
| 1926 | pos = find_rightmost_additive(expr) |
| 1927 | |
| 1928 | if (pos > 0) then |
| 1929 | left_expr = expr(:pos-1) |
| 1930 | right_expr = expr(pos+1:) |
| 1931 | value = eval_additive(trim(adjustl(left_expr))) |
| 1932 | right_val = eval_multiplicative(trim(adjustl(right_expr))) |
| 1933 | |
| 1934 | if (expr(pos:pos) == '+') then |
| 1935 | value = value + right_val |
| 1936 | else |
| 1937 | value = value - right_val |
| 1938 | end if |
| 1939 | else |
| 1940 | value = eval_multiplicative(expr) |
| 1941 | end if |
| 1942 | end function |
| 1943 | |
| 1944 | ! Multiplication, Division, Modulo |
| 1945 | recursive function eval_multiplicative(expr) result(value) |
| 1946 | character(len=*), intent(in) :: expr |
| 1947 | integer(kind=8) :: value, right_val |
| 1948 | integer :: pos |
| 1949 | character(len=512) :: left_expr, right_expr |
| 1950 | character :: op |
| 1951 | |
| 1952 | ! Find rightmost *, /, or % |
| 1953 | pos = find_rightmost_multiplicative(expr, op) |
| 1954 | |
| 1955 | if (pos > 0) then |
| 1956 | left_expr = expr(:pos-1) |
| 1957 | right_expr = expr(pos+1:) |
| 1958 | value = eval_multiplicative(trim(adjustl(left_expr))) |
| 1959 | right_val = eval_power(trim(adjustl(right_expr))) |
| 1960 | |
| 1961 | select case (op) |
| 1962 | case ('*') |
| 1963 | value = value * right_val |
| 1964 | case ('/') |
| 1965 | if (right_val /= 0) then |
| 1966 | value = value / right_val |
| 1967 | else |
| 1968 | arithmetic_error = .true. |
| 1969 | arithmetic_error_msg = 'division by zero' |
| 1970 | value = 0 ! Division by zero |
| 1971 | end if |
| 1972 | case ('%') |
| 1973 | if (right_val /= 0) then |
| 1974 | value = mod(value, right_val) |
| 1975 | else |
| 1976 | arithmetic_error = .true. |
| 1977 | arithmetic_error_msg = 'division by zero' |
| 1978 | value = 0 ! Modulo by zero |
| 1979 | end if |
| 1980 | end select |
| 1981 | else |
| 1982 | value = eval_power(expr) |
| 1983 | end if |
| 1984 | end function |
| 1985 | |
| 1986 | ! Exponentiation (**) |
| 1987 | recursive function eval_power(expr) result(value) |
| 1988 | character(len=*), intent(in) :: expr |
| 1989 | integer(kind=8) :: value, exponent |
| 1990 | integer :: pos, i |
| 1991 | character(len=512) :: base_expr, exp_expr |
| 1992 | |
| 1993 | pos = find_operator(expr, '**') |
| 1994 | if (pos > 0) then |
| 1995 | base_expr = expr(:pos-1) |
| 1996 | exp_expr = expr(pos+2:) |
| 1997 | value = eval_unary(trim(adjustl(base_expr))) |
| 1998 | exponent = eval_power(trim(adjustl(exp_expr))) ! Right-associative |
| 1999 | |
| 2000 | ! Calculate power |
| 2001 | if (exponent < 0) then |
| 2002 | value = 0 ! Integer division for negative exponents |
| 2003 | else if (exponent == 0) then |
| 2004 | value = 1 |
| 2005 | else |
| 2006 | do i = 2, int(exponent) |
| 2007 | value = value * eval_unary(trim(adjustl(base_expr))) |
| 2008 | end do |
| 2009 | end if |
| 2010 | else |
| 2011 | value = eval_unary(expr) |
| 2012 | end if |
| 2013 | end function |
| 2014 | |
| 2015 | ! Unary operators (!, -, +) |
| 2016 | recursive function eval_unary(expr) result(value) |
| 2017 | character(len=*), intent(in) :: expr |
| 2018 | integer(kind=8) :: value |
| 2019 | character(len=512) :: rest |
| 2020 | |
| 2021 | if (len_trim(expr) == 0) then |
| 2022 | value = 0 |
| 2023 | return |
| 2024 | end if |
| 2025 | |
| 2026 | ! Logical NOT |
| 2027 | if (expr(1:1) == '!') then |
| 2028 | rest = adjustl(expr(2:)) |
| 2029 | value = eval_unary(rest) |
| 2030 | if (value == 0) then |
| 2031 | value = 1 |
| 2032 | else |
| 2033 | value = 0 |
| 2034 | end if |
| 2035 | return |
| 2036 | end if |
| 2037 | |
| 2038 | ! Bitwise NOT (~) |
| 2039 | if (expr(1:1) == '~') then |
| 2040 | rest = adjustl(expr(2:)) |
| 2041 | value = eval_unary(rest) |
| 2042 | ! Bitwise NOT in two's complement: ~n = -(n + 1) |
| 2043 | value = -(value + 1) |
| 2044 | return |
| 2045 | end if |
| 2046 | |
| 2047 | ! Unary minus |
| 2048 | if (expr(1:1) == '-' .and. len_trim(expr) > 1) then |
| 2049 | rest = adjustl(expr(2:)) |
| 2050 | value = -eval_unary(rest) |
| 2051 | return |
| 2052 | end if |
| 2053 | |
| 2054 | ! Unary plus |
| 2055 | if (expr(1:1) == '+' .and. len_trim(expr) > 1) then |
| 2056 | rest = adjustl(expr(2:)) |
| 2057 | value = eval_unary(rest) |
| 2058 | return |
| 2059 | end if |
| 2060 | |
| 2061 | value = eval_primary(expr) |
| 2062 | end function |
| 2063 | |
| 2064 | ! Primary expressions (numbers, variables, parentheses) |
| 2065 | function eval_primary(expr) result(value) |
| 2066 | character(len=*), intent(in) :: expr |
| 2067 | integer(kind=8) :: value |
| 2068 | character(len=512) :: inner_expr, temp_expr |
| 2069 | integer :: iostat, paren_end |
| 2070 | |
| 2071 | if (len_trim(expr) == 0) then |
| 2072 | value = 0 |
| 2073 | return |
| 2074 | end if |
| 2075 | |
| 2076 | ! Handle parentheses |
| 2077 | if (expr(1:1) == '(') then |
| 2078 | paren_end = find_matching_paren(expr, 1) |
| 2079 | if (paren_end > 1) then |
| 2080 | inner_expr = expr(2:paren_end-1) |
| 2081 | value = eval_expression(trim(adjustl(inner_expr))) |
| 2082 | return |
| 2083 | end if |
| 2084 | end if |
| 2085 | |
| 2086 | ! Try to parse as number (with octal/hex support) |
| 2087 | temp_expr = trim(adjustl(expr)) |
| 2088 | value = parse_arithmetic_number(temp_expr, iostat) |
| 2089 | if (iostat == 0) return |
| 2090 | |
| 2091 | ! Variable without shell context - return 0 |
| 2092 | value = 0 |
| 2093 | end function |
| 2094 | |
| 2095 | ! ============================================================================ |
| 2096 | ! Shell-aware arithmetic evaluation (with variable resolution) |
| 2097 | ! ============================================================================ |
| 2098 | |
| 2099 | recursive function eval_expression_shell(expr, shell) result(value) |
| 2100 | character(len=*), intent(in) :: expr |
| 2101 | type(shell_state_t), intent(inout) :: shell |
| 2102 | integer(kind=8) :: value |
| 2103 | ! Comma operator has lowest precedence |
| 2104 | value = eval_comma_shell(trim(adjustl(expr)), shell) |
| 2105 | end function |
| 2106 | |
| 2107 | ! Comma operator (evaluates left-to-right, returns rightmost value) |
| 2108 | recursive function eval_comma_shell(expr, shell) result(value) |
| 2109 | character(len=*), intent(in) :: expr |
| 2110 | type(shell_state_t), intent(inout) :: shell |
| 2111 | integer(kind=8) :: value |
| 2112 | integer :: comma_pos, paren_depth, i |
| 2113 | character(len=1) :: ch |
| 2114 | |
| 2115 | ! Find comma at top level (not inside parentheses) |
| 2116 | paren_depth = 0 |
| 2117 | comma_pos = 0 |
| 2118 | do i = 1, len_trim(expr) |
| 2119 | ch = expr(i:i) |
| 2120 | if (ch == '(') then |
| 2121 | paren_depth = paren_depth + 1 |
| 2122 | else if (ch == ')') then |
| 2123 | paren_depth = paren_depth - 1 |
| 2124 | else if (ch == ',' .and. paren_depth == 0) then |
| 2125 | ! Evaluate left side (for side effects), then continue with right |
| 2126 | value = eval_assignment_shell(trim(adjustl(expr(:i-1))), shell) |
| 2127 | ! Continue evaluating right side (may have more commas) |
| 2128 | value = eval_comma_shell(trim(adjustl(expr(i+1:))), shell) |
| 2129 | return |
| 2130 | end if |
| 2131 | end do |
| 2132 | |
| 2133 | ! No comma found, evaluate as assignment |
| 2134 | value = eval_assignment_shell(trim(adjustl(expr)), shell) |
| 2135 | end function |
| 2136 | |
| 2137 | ! Assignment operators (=, +=, -=, *=, /=, %=) |
| 2138 | recursive function eval_assignment_shell(expr, shell) result(value) |
| 2139 | character(len=*), intent(in) :: expr |
| 2140 | type(shell_state_t), intent(inout) :: shell |
| 2141 | integer(kind=8) :: value, right_val, current_val |
| 2142 | integer :: pos, op_len, iostat |
| 2143 | character(len=512) :: var_name, right_expr, var_value_str |
| 2144 | character(len=:), allocatable :: temp_value |
| 2145 | |
| 2146 | ! Check for assignment operators (right-to-left associative, so find rightmost) |
| 2147 | pos = find_rightmost_assignment(expr, op_len) |
| 2148 | |
| 2149 | if (pos > 0) then |
| 2150 | ! Extract variable name (left side) and expression (right side) |
| 2151 | var_name = trim(adjustl(expr(:pos-1))) |
| 2152 | right_expr = expr(pos+op_len:) |
| 2153 | |
| 2154 | ! Evaluate right side |
| 2155 | right_val = eval_assignment_shell(trim(adjustl(right_expr)), shell) |
| 2156 | |
| 2157 | ! Determine which operator and perform operation |
| 2158 | if (op_len == 1) then |
| 2159 | ! Simple assignment: = |
| 2160 | value = right_val |
| 2161 | else |
| 2162 | ! Compound assignment - get current value |
| 2163 | temp_value = get_shell_variable(shell, trim(var_name)) |
| 2164 | if (len_trim(temp_value) > 0) then |
| 2165 | read(temp_value, *, iostat=iostat) current_val |
| 2166 | if (iostat /= 0) current_val = 0 |
| 2167 | else |
| 2168 | current_val = 0 |
| 2169 | end if |
| 2170 | |
| 2171 | ! Apply compound operator |
| 2172 | select case (expr(pos:pos+op_len-1)) |
| 2173 | case ('+=') |
| 2174 | value = current_val + right_val |
| 2175 | case ('-=') |
| 2176 | value = current_val - right_val |
| 2177 | case ('*=') |
| 2178 | value = current_val * right_val |
| 2179 | case ('/=') |
| 2180 | if (right_val /= 0) then |
| 2181 | value = current_val / right_val |
| 2182 | else |
| 2183 | arithmetic_error = .true. |
| 2184 | arithmetic_error_msg = 'division by zero' |
| 2185 | value = 0 |
| 2186 | end if |
| 2187 | case ('%=') |
| 2188 | if (right_val /= 0) then |
| 2189 | value = mod(current_val, right_val) |
| 2190 | else |
| 2191 | arithmetic_error = .true. |
| 2192 | arithmetic_error_msg = 'division by zero' |
| 2193 | value = 0 |
| 2194 | end if |
| 2195 | case default |
| 2196 | value = right_val |
| 2197 | end select |
| 2198 | end if |
| 2199 | |
| 2200 | ! Set the variable |
| 2201 | write(var_value_str, '(I0)') value |
| 2202 | call set_shell_variable(shell, trim(var_name), trim(var_value_str)) |
| 2203 | else |
| 2204 | ! No assignment, evaluate as ternary |
| 2205 | value = eval_ternary_shell(expr, shell) |
| 2206 | end if |
| 2207 | end function |
| 2208 | |
| 2209 | ! Ternary conditional operator (? :) |
| 2210 | recursive function eval_ternary_shell(expr, shell) result(value) |
| 2211 | character(len=*), intent(in) :: expr |
| 2212 | type(shell_state_t), intent(inout) :: shell |
| 2213 | integer(kind=8) :: value |
| 2214 | integer :: qmark_pos, colon_pos, depth, i |
| 2215 | character(len=512) :: condition_expr, true_expr, false_expr |
| 2216 | |
| 2217 | ! Find ? outside parentheses |
| 2218 | qmark_pos = 0 |
| 2219 | depth = 0 |
| 2220 | do i = 1, len_trim(expr) |
| 2221 | if (expr(i:i) == '(') then |
| 2222 | depth = depth + 1 |
| 2223 | else if (expr(i:i) == ')') then |
| 2224 | depth = depth - 1 |
| 2225 | else if (depth == 0 .and. expr(i:i) == '?') then |
| 2226 | qmark_pos = i |
| 2227 | exit |
| 2228 | end if |
| 2229 | end do |
| 2230 | |
| 2231 | if (qmark_pos > 0) then |
| 2232 | ! Find matching : after the ? |
| 2233 | colon_pos = 0 |
| 2234 | depth = 0 |
| 2235 | do i = qmark_pos + 1, len_trim(expr) |
| 2236 | if (expr(i:i) == '(') then |
| 2237 | depth = depth + 1 |
| 2238 | else if (expr(i:i) == ')') then |
| 2239 | depth = depth - 1 |
| 2240 | else if (depth == 0 .and. expr(i:i) == ':') then |
| 2241 | colon_pos = i |
| 2242 | exit |
| 2243 | end if |
| 2244 | end do |
| 2245 | |
| 2246 | if (colon_pos > 0) then |
| 2247 | condition_expr = expr(:qmark_pos-1) |
| 2248 | true_expr = expr(qmark_pos+1:colon_pos-1) |
| 2249 | false_expr = expr(colon_pos+1:) |
| 2250 | |
| 2251 | ! Evaluate condition |
| 2252 | value = eval_logical_or_shell(trim(adjustl(condition_expr)), shell) |
| 2253 | if (value /= 0) then |
| 2254 | ! Condition is true, evaluate true expression |
| 2255 | value = eval_ternary_shell(trim(adjustl(true_expr)), shell) |
| 2256 | else |
| 2257 | ! Condition is false, evaluate false expression |
| 2258 | value = eval_ternary_shell(trim(adjustl(false_expr)), shell) |
| 2259 | end if |
| 2260 | return |
| 2261 | end if |
| 2262 | end if |
| 2263 | |
| 2264 | ! No ternary operator found |
| 2265 | value = eval_logical_or_shell(expr, shell) |
| 2266 | end function |
| 2267 | |
| 2268 | ! Helper function to find leftmost assignment operator (for right-associativity) |
| 2269 | function find_rightmost_assignment(expr, op_len) result(pos) |
| 2270 | character(len=*), intent(in) :: expr |
| 2271 | integer, intent(out) :: op_len |
| 2272 | integer :: pos, i, paren_depth |
| 2273 | |
| 2274 | pos = 0 |
| 2275 | op_len = 0 |
| 2276 | paren_depth = 0 |
| 2277 | |
| 2278 | ! Scan from left to right, tracking parentheses |
| 2279 | ! This gives right-associativity: a=b=c becomes a=(b=c) |
| 2280 | do i = 1, len_trim(expr) |
| 2281 | if (expr(i:i) == '(') then |
| 2282 | paren_depth = paren_depth + 1 |
| 2283 | else if (expr(i:i) == ')') then |
| 2284 | paren_depth = paren_depth - 1 |
| 2285 | else if (paren_depth == 0) then |
| 2286 | ! Check for compound assignment operators (2 chars) |
| 2287 | if (i < len_trim(expr)) then |
| 2288 | if (expr(i:i+1) == '+=' .or. expr(i:i+1) == '-=' .or. & |
| 2289 | expr(i:i+1) == '*=' .or. expr(i:i+1) == '/=' .or. & |
| 2290 | expr(i:i+1) == '%=') then |
| 2291 | pos = i |
| 2292 | op_len = 2 |
| 2293 | return |
| 2294 | end if |
| 2295 | end if |
| 2296 | ! Check for simple assignment (but not ==, !=, <=, >=) |
| 2297 | if (expr(i:i) == '=') then |
| 2298 | ! Check it's not a comparison operator |
| 2299 | if (i > 1) then |
| 2300 | if (expr(i-1:i-1) == '=' .or. expr(i-1:i-1) == '!' .or. & |
| 2301 | expr(i-1:i-1) == '<' .or. expr(i-1:i-1) == '>') then |
| 2302 | cycle ! Skip this =, it's part of a comparison |
| 2303 | end if |
| 2304 | end if |
| 2305 | if (i < len_trim(expr)) then |
| 2306 | if (expr(i+1:i+1) == '=') then |
| 2307 | cycle ! Skip this =, it's part of == |
| 2308 | end if |
| 2309 | end if |
| 2310 | pos = i |
| 2311 | op_len = 1 |
| 2312 | return |
| 2313 | end if |
| 2314 | end if |
| 2315 | end do |
| 2316 | end function |
| 2317 | |
| 2318 | recursive function eval_logical_or_shell(expr, shell) result(value) |
| 2319 | character(len=*), intent(in) :: expr |
| 2320 | type(shell_state_t), intent(inout) :: shell |
| 2321 | integer(kind=8) :: value, right_val |
| 2322 | integer :: pos |
| 2323 | character(len=512) :: left_expr, right_expr |
| 2324 | |
| 2325 | ! FIRST check for || operator (lowest precedence in logical chain) |
| 2326 | pos = find_operator(expr, '||') |
| 2327 | if (pos > 0) then |
| 2328 | left_expr = expr(:pos-1) |
| 2329 | right_expr = expr(pos+2:) |
| 2330 | value = eval_logical_and_shell(trim(adjustl(left_expr)), shell) |
| 2331 | right_val = eval_logical_or_shell(trim(adjustl(right_expr)), shell) |
| 2332 | if (value /= 0 .or. right_val /= 0) then |
| 2333 | value = 1 |
| 2334 | else |
| 2335 | value = 0 |
| 2336 | end if |
| 2337 | else |
| 2338 | ! No || found, delegate to next precedence level |
| 2339 | value = eval_logical_and_shell(expr, shell) |
| 2340 | end if |
| 2341 | end function |
| 2342 | |
| 2343 | recursive function eval_logical_and_shell(expr, shell) result(value) |
| 2344 | character(len=*), intent(in) :: expr |
| 2345 | type(shell_state_t), intent(inout) :: shell |
| 2346 | integer(kind=8) :: value, right_val |
| 2347 | integer :: pos |
| 2348 | character(len=512) :: left_expr, right_expr |
| 2349 | |
| 2350 | ! FIRST check for && operator (lowest precedence in this chain) |
| 2351 | pos = find_operator(expr, '&&') |
| 2352 | if (pos > 0) then |
| 2353 | left_expr = expr(:pos-1) |
| 2354 | right_expr = expr(pos+2:) |
| 2355 | value = eval_bitwise_or_shell(trim(adjustl(left_expr)), shell) |
| 2356 | right_val = eval_logical_and_shell(trim(adjustl(right_expr)), shell) |
| 2357 | if (value /= 0 .and. right_val /= 0) then |
| 2358 | value = 1 |
| 2359 | else |
| 2360 | value = 0 |
| 2361 | end if |
| 2362 | else |
| 2363 | ! No && found, delegate to next precedence level |
| 2364 | value = eval_bitwise_or_shell(expr, shell) |
| 2365 | end if |
| 2366 | end function |
| 2367 | |
| 2368 | recursive function eval_bitwise_or_shell(expr, shell) result(value) |
| 2369 | character(len=*), intent(in) :: expr |
| 2370 | type(shell_state_t), intent(inout) :: shell |
| 2371 | integer(kind=8) :: value, right_val |
| 2372 | integer :: pos |
| 2373 | character(len=512) :: left_expr, right_expr |
| 2374 | |
| 2375 | ! FIRST check for | operator |
| 2376 | pos = find_single_operator(expr, '|') |
| 2377 | if (pos > 0) then |
| 2378 | left_expr = expr(:pos-1) |
| 2379 | right_expr = expr(pos+1:) |
| 2380 | value = eval_bitwise_xor_shell(trim(adjustl(left_expr)), shell) |
| 2381 | right_val = eval_bitwise_or_shell(trim(adjustl(right_expr)), shell) |
| 2382 | value = ior(int(value), int(right_val)) |
| 2383 | else |
| 2384 | value = eval_bitwise_xor_shell(expr, shell) |
| 2385 | end if |
| 2386 | end function |
| 2387 | |
| 2388 | recursive function eval_bitwise_xor_shell(expr, shell) result(value) |
| 2389 | character(len=*), intent(in) :: expr |
| 2390 | type(shell_state_t), intent(inout) :: shell |
| 2391 | integer(kind=8) :: value, right_val |
| 2392 | integer :: pos |
| 2393 | character(len=512) :: left_expr, right_expr |
| 2394 | |
| 2395 | ! FIRST check for ^ operator |
| 2396 | pos = find_single_operator(expr, '^') |
| 2397 | if (pos > 0) then |
| 2398 | left_expr = expr(:pos-1) |
| 2399 | right_expr = expr(pos+1:) |
| 2400 | value = eval_bitwise_and_shell(trim(adjustl(left_expr)), shell) |
| 2401 | right_val = eval_bitwise_xor_shell(trim(adjustl(right_expr)), shell) |
| 2402 | value = ieor(int(value), int(right_val)) |
| 2403 | else |
| 2404 | value = eval_bitwise_and_shell(expr, shell) |
| 2405 | end if |
| 2406 | end function |
| 2407 | |
| 2408 | recursive function eval_bitwise_and_shell(expr, shell) result(value) |
| 2409 | character(len=*), intent(in) :: expr |
| 2410 | type(shell_state_t), intent(inout) :: shell |
| 2411 | integer(kind=8) :: value, right_val |
| 2412 | integer :: pos |
| 2413 | character(len=512) :: left_expr, right_expr |
| 2414 | |
| 2415 | ! FIRST check for & operator |
| 2416 | pos = find_single_operator(expr, '&') |
| 2417 | if (pos > 0) then |
| 2418 | left_expr = expr(:pos-1) |
| 2419 | right_expr = expr(pos+1:) |
| 2420 | value = eval_equality_shell(trim(adjustl(left_expr)), shell) |
| 2421 | right_val = eval_bitwise_and_shell(trim(adjustl(right_expr)), shell) |
| 2422 | value = iand(int(value), int(right_val)) |
| 2423 | else |
| 2424 | value = eval_equality_shell(expr, shell) |
| 2425 | end if |
| 2426 | end function |
| 2427 | |
| 2428 | recursive function eval_equality_shell(expr, shell) result(value) |
| 2429 | character(len=*), intent(in) :: expr |
| 2430 | type(shell_state_t), intent(inout) :: shell |
| 2431 | integer(kind=8) :: value, right_val |
| 2432 | integer :: pos |
| 2433 | character(len=512) :: left_expr, right_expr |
| 2434 | |
| 2435 | pos = find_operator(expr, '==') |
| 2436 | if (pos > 0) then |
| 2437 | left_expr = expr(:pos-1) |
| 2438 | right_expr = expr(pos+2:) |
| 2439 | value = eval_comparison_shell(trim(adjustl(left_expr)), shell) |
| 2440 | right_val = eval_comparison_shell(trim(adjustl(right_expr)), shell) |
| 2441 | if (value == right_val) then |
| 2442 | value = 1 |
| 2443 | else |
| 2444 | value = 0 |
| 2445 | end if |
| 2446 | return |
| 2447 | end if |
| 2448 | |
| 2449 | pos = find_operator(expr, '!=') |
| 2450 | if (pos > 0) then |
| 2451 | left_expr = expr(:pos-1) |
| 2452 | right_expr = expr(pos+2:) |
| 2453 | value = eval_comparison_shell(trim(adjustl(left_expr)), shell) |
| 2454 | right_val = eval_comparison_shell(trim(adjustl(right_expr)), shell) |
| 2455 | if (value /= right_val) then |
| 2456 | value = 1 |
| 2457 | else |
| 2458 | value = 0 |
| 2459 | end if |
| 2460 | return |
| 2461 | end if |
| 2462 | |
| 2463 | value = eval_comparison_shell(expr, shell) |
| 2464 | end function |
| 2465 | |
| 2466 | recursive function eval_comparison_shell(expr, shell) result(value) |
| 2467 | character(len=*), intent(in) :: expr |
| 2468 | type(shell_state_t), intent(inout) :: shell |
| 2469 | integer(kind=8) :: value, right_val |
| 2470 | integer :: pos |
| 2471 | character(len=512) :: left_expr, right_expr |
| 2472 | |
| 2473 | pos = find_operator(expr, '<=') |
| 2474 | if (pos > 0) then |
| 2475 | left_expr = expr(:pos-1) |
| 2476 | right_expr = expr(pos+2:) |
| 2477 | value = eval_shift_shell(trim(adjustl(left_expr)), shell) |
| 2478 | right_val = eval_shift_shell(trim(adjustl(right_expr)), shell) |
| 2479 | if (value <= right_val) then; value = 1; else; value = 0; end if |
| 2480 | return |
| 2481 | end if |
| 2482 | |
| 2483 | pos = find_operator(expr, '>=') |
| 2484 | if (pos > 0) then |
| 2485 | left_expr = expr(:pos-1) |
| 2486 | right_expr = expr(pos+2:) |
| 2487 | value = eval_shift_shell(trim(adjustl(left_expr)), shell) |
| 2488 | right_val = eval_shift_shell(trim(adjustl(right_expr)), shell) |
| 2489 | if (value >= right_val) then; value = 1; else; value = 0; end if |
| 2490 | return |
| 2491 | end if |
| 2492 | |
| 2493 | pos = find_single_operator(expr, '<') |
| 2494 | if (pos > 0) then |
| 2495 | left_expr = expr(:pos-1) |
| 2496 | right_expr = expr(pos+1:) |
| 2497 | value = eval_shift_shell(trim(adjustl(left_expr)), shell) |
| 2498 | right_val = eval_shift_shell(trim(adjustl(right_expr)), shell) |
| 2499 | if (value < right_val) then; value = 1; else; value = 0; end if |
| 2500 | return |
| 2501 | end if |
| 2502 | |
| 2503 | pos = find_single_operator(expr, '>') |
| 2504 | if (pos > 0) then |
| 2505 | left_expr = expr(:pos-1) |
| 2506 | right_expr = expr(pos+1:) |
| 2507 | value = eval_shift_shell(trim(adjustl(left_expr)), shell) |
| 2508 | right_val = eval_shift_shell(trim(adjustl(right_expr)), shell) |
| 2509 | if (value > right_val) then; value = 1; else; value = 0; end if |
| 2510 | return |
| 2511 | end if |
| 2512 | |
| 2513 | value = eval_shift_shell(expr, shell) |
| 2514 | end function |
| 2515 | |
| 2516 | ! Shift operations (<<, >>) |
| 2517 | recursive function eval_shift_shell(expr, shell) result(value) |
| 2518 | character(len=*), intent(in) :: expr |
| 2519 | type(shell_state_t), intent(inout) :: shell |
| 2520 | integer(kind=8) :: value, right_val |
| 2521 | integer :: pos |
| 2522 | character(len=512) :: left_expr, right_expr |
| 2523 | |
| 2524 | ! Try << (left shift) |
| 2525 | pos = find_operator(expr, '<<') |
| 2526 | if (pos > 0) then |
| 2527 | left_expr = expr(:pos-1) |
| 2528 | right_expr = expr(pos+2:) |
| 2529 | value = eval_additive_shell(trim(adjustl(left_expr)), shell) ! Changed from eval_shift_shell |
| 2530 | right_val = eval_additive_shell(trim(adjustl(right_expr)), shell) |
| 2531 | ! Left shift by right_val bits |
| 2532 | value = ishft(value, int(right_val)) |
| 2533 | return |
| 2534 | end if |
| 2535 | |
| 2536 | ! Try >> (right shift) |
| 2537 | pos = find_operator(expr, '>>') |
| 2538 | if (pos > 0) then |
| 2539 | left_expr = expr(:pos-1) |
| 2540 | right_expr = expr(pos+2:) |
| 2541 | value = eval_additive_shell(trim(adjustl(left_expr)), shell) ! Changed from eval_shift_shell |
| 2542 | right_val = eval_additive_shell(trim(adjustl(right_expr)), shell) |
| 2543 | ! Right shift by right_val bits (negative for right shift in ishft) |
| 2544 | value = ishft(value, -int(right_val)) |
| 2545 | return |
| 2546 | end if |
| 2547 | |
| 2548 | ! No shift operator found |
| 2549 | value = eval_additive_shell(expr, shell) |
| 2550 | end function |
| 2551 | |
| 2552 | recursive function eval_additive_shell(expr, shell) result(value) |
| 2553 | character(len=*), intent(in) :: expr |
| 2554 | type(shell_state_t), intent(inout) :: shell |
| 2555 | integer(kind=8) :: value, right_val |
| 2556 | integer :: pos |
| 2557 | character(len=512) :: left_expr, right_expr |
| 2558 | |
| 2559 | pos = find_rightmost_additive(expr) |
| 2560 | if (pos > 0) then |
| 2561 | left_expr = expr(:pos-1) |
| 2562 | right_expr = expr(pos+1:) |
| 2563 | value = eval_additive_shell(trim(adjustl(left_expr)), shell) |
| 2564 | right_val = eval_multiplicative_shell(trim(adjustl(right_expr)), shell) |
| 2565 | if (expr(pos:pos) == '+') then |
| 2566 | value = value + right_val |
| 2567 | else |
| 2568 | value = value - right_val |
| 2569 | end if |
| 2570 | else |
| 2571 | value = eval_multiplicative_shell(expr, shell) |
| 2572 | end if |
| 2573 | end function |
| 2574 | |
| 2575 | recursive function eval_multiplicative_shell(expr, shell) result(value) |
| 2576 | character(len=*), intent(in) :: expr |
| 2577 | type(shell_state_t), intent(inout) :: shell |
| 2578 | integer(kind=8) :: value, right_val |
| 2579 | integer :: pos |
| 2580 | character(len=512) :: left_expr, right_expr |
| 2581 | character :: op |
| 2582 | |
| 2583 | pos = find_rightmost_multiplicative(expr, op) |
| 2584 | if (pos > 0) then |
| 2585 | left_expr = expr(:pos-1) |
| 2586 | right_expr = expr(pos+1:) |
| 2587 | value = eval_multiplicative_shell(trim(adjustl(left_expr)), shell) |
| 2588 | right_val = eval_power_shell(trim(adjustl(right_expr)), shell) |
| 2589 | select case (op) |
| 2590 | case ('*'); value = value * right_val |
| 2591 | case ('/') |
| 2592 | if (right_val /= 0) then |
| 2593 | value = value / right_val |
| 2594 | else |
| 2595 | arithmetic_error = .true. |
| 2596 | arithmetic_error_msg = 'division by zero' |
| 2597 | value = 0 |
| 2598 | end if |
| 2599 | case ('%') |
| 2600 | if (right_val /= 0) then |
| 2601 | value = mod(value, right_val) |
| 2602 | else |
| 2603 | arithmetic_error = .true. |
| 2604 | arithmetic_error_msg = 'division by zero' |
| 2605 | value = 0 |
| 2606 | end if |
| 2607 | end select |
| 2608 | else |
| 2609 | value = eval_power_shell(expr, shell) |
| 2610 | end if |
| 2611 | end function |
| 2612 | |
| 2613 | recursive function eval_power_shell(expr, shell) result(value) |
| 2614 | character(len=*), intent(in) :: expr |
| 2615 | type(shell_state_t), intent(inout) :: shell |
| 2616 | integer(kind=8) :: value, exponent, base_val |
| 2617 | integer :: pos, i |
| 2618 | character(len=512) :: base_expr, exp_expr |
| 2619 | |
| 2620 | pos = find_operator(expr, '**') |
| 2621 | if (pos > 0) then |
| 2622 | base_expr = expr(:pos-1) |
| 2623 | exp_expr = expr(pos+2:) |
| 2624 | base_val = eval_unary_shell(trim(adjustl(base_expr)), shell) |
| 2625 | exponent = eval_power_shell(trim(adjustl(exp_expr)), shell) |
| 2626 | if (exponent < 0) then; value = 0 |
| 2627 | else if (exponent == 0) then; value = 1 |
| 2628 | else |
| 2629 | value = base_val |
| 2630 | do i = 2, int(exponent) |
| 2631 | value = value * base_val |
| 2632 | end do |
| 2633 | end if |
| 2634 | else |
| 2635 | value = eval_unary_shell(expr, shell) |
| 2636 | end if |
| 2637 | end function |
| 2638 | |
| 2639 | recursive function eval_unary_shell(expr, shell) result(value) |
| 2640 | character(len=*), intent(in) :: expr |
| 2641 | type(shell_state_t), intent(inout) :: shell |
| 2642 | integer(kind=8) :: value, current_val |
| 2643 | character(len=512) :: rest, var_name, var_value_str, trimmed_expr |
| 2644 | character(len=:), allocatable :: temp_value |
| 2645 | integer :: iostat |
| 2646 | |
| 2647 | if (len_trim(expr) == 0) then; value = 0; return; end if |
| 2648 | |
| 2649 | ! Trim leading/trailing whitespace for all checks |
| 2650 | trimmed_expr = trim(adjustl(expr)) |
| 2651 | |
| 2652 | ! Pre-increment: ++x (only if followed by a variable name, not a number) |
| 2653 | if (len_trim(trimmed_expr) > 2 .and. trimmed_expr(1:2) == '++') then |
| 2654 | var_name = trim(adjustl(trimmed_expr(3:))) |
| 2655 | ! Check if it starts with a letter or underscore (variable name) |
| 2656 | ! If it starts with a digit, it's double unary plus, not increment |
| 2657 | if (len_trim(var_name) > 0) then |
| 2658 | if ((var_name(1:1) >= 'a' .and. var_name(1:1) <= 'z') .or. & |
| 2659 | (var_name(1:1) >= 'A' .and. var_name(1:1) <= 'Z') .or. & |
| 2660 | var_name(1:1) == '_') then |
| 2661 | ! Get current value |
| 2662 | temp_value = get_shell_variable(shell, trim(var_name)) |
| 2663 | if (len_trim(temp_value) > 0) then |
| 2664 | read(temp_value, *, iostat=iostat) current_val |
| 2665 | if (iostat /= 0) current_val = 0 |
| 2666 | else |
| 2667 | current_val = 0 |
| 2668 | end if |
| 2669 | ! Increment |
| 2670 | value = current_val + 1 |
| 2671 | ! Set variable |
| 2672 | write(var_value_str, '(I0)') value |
| 2673 | call set_shell_variable(shell, trim(var_name), trim(var_value_str)) |
| 2674 | return |
| 2675 | end if |
| 2676 | ! Otherwise fall through to unary plus handling |
| 2677 | end if |
| 2678 | end if |
| 2679 | |
| 2680 | ! Pre-decrement: --x (only if followed by a variable name, not a number) |
| 2681 | if (len_trim(trimmed_expr) > 2 .and. trimmed_expr(1:2) == '--') then |
| 2682 | var_name = trim(adjustl(trimmed_expr(3:))) |
| 2683 | ! Check if it starts with a letter or underscore (variable name) |
| 2684 | ! If it starts with a digit, it's double unary minus, not decrement |
| 2685 | if (len_trim(var_name) > 0) then |
| 2686 | if ((var_name(1:1) >= 'a' .and. var_name(1:1) <= 'z') .or. & |
| 2687 | (var_name(1:1) >= 'A' .and. var_name(1:1) <= 'Z') .or. & |
| 2688 | var_name(1:1) == '_') then |
| 2689 | ! Get current value |
| 2690 | temp_value = get_shell_variable(shell, trim(var_name)) |
| 2691 | if (len_trim(temp_value) > 0) then |
| 2692 | read(temp_value, *, iostat=iostat) current_val |
| 2693 | if (iostat /= 0) current_val = 0 |
| 2694 | else |
| 2695 | current_val = 0 |
| 2696 | end if |
| 2697 | ! Decrement |
| 2698 | value = current_val - 1 |
| 2699 | ! Set variable |
| 2700 | write(var_value_str, '(I0)') value |
| 2701 | call set_shell_variable(shell, trim(var_name), trim(var_value_str)) |
| 2702 | return |
| 2703 | end if |
| 2704 | ! Otherwise fall through to unary minus handling |
| 2705 | end if |
| 2706 | end if |
| 2707 | |
| 2708 | if (trimmed_expr(1:1) == '!') then |
| 2709 | rest = adjustl(trimmed_expr(2:)) |
| 2710 | value = eval_unary_shell(rest, shell) |
| 2711 | if (value == 0) then; value = 1; else; value = 0; end if |
| 2712 | return |
| 2713 | end if |
| 2714 | |
| 2715 | ! Bitwise NOT (~) |
| 2716 | if (trimmed_expr(1:1) == '~') then |
| 2717 | rest = adjustl(trimmed_expr(2:)) |
| 2718 | value = eval_unary_shell(rest, shell) |
| 2719 | ! Bitwise NOT in two's complement: ~n = -(n + 1) |
| 2720 | value = -(value + 1) |
| 2721 | return |
| 2722 | end if |
| 2723 | |
| 2724 | if (trimmed_expr(1:1) == '-' .and. len_trim(trimmed_expr) > 1) then |
| 2725 | rest = adjustl(trimmed_expr(2:)) |
| 2726 | value = -eval_unary_shell(rest, shell) |
| 2727 | return |
| 2728 | end if |
| 2729 | |
| 2730 | if (trimmed_expr(1:1) == '+' .and. len_trim(trimmed_expr) > 1) then |
| 2731 | rest = adjustl(trimmed_expr(2:)) |
| 2732 | value = eval_unary_shell(rest, shell) |
| 2733 | return |
| 2734 | end if |
| 2735 | |
| 2736 | value = eval_primary_shell(expr, shell) |
| 2737 | end function |
| 2738 | |
| 2739 | function eval_primary_shell(expr, shell) result(value) |
| 2740 | character(len=*), intent(in) :: expr |
| 2741 | type(shell_state_t), intent(inout) :: shell |
| 2742 | integer(kind=8) :: value, new_val |
| 2743 | character(len=512) :: inner_expr, temp_expr, var_name, var_value_str |
| 2744 | character(len=:), allocatable :: var_value |
| 2745 | integer :: iostat, paren_end, expr_len |
| 2746 | |
| 2747 | if (len_trim(expr) == 0) then; value = 0; return; end if |
| 2748 | |
| 2749 | expr_len = len_trim(expr) |
| 2750 | |
| 2751 | ! Check for post-increment: x++ (only if x is a valid variable name) |
| 2752 | if (expr_len > 2 .and. expr(expr_len-1:expr_len) == '++') then |
| 2753 | var_name = trim(adjustl(expr(:expr_len-2))) |
| 2754 | ! Only treat as post-increment if var_name is a valid identifier (no operators) |
| 2755 | if (is_valid_identifier(var_name)) then |
| 2756 | ! Get current value |
| 2757 | var_value = get_shell_variable(shell, trim(var_name)) |
| 2758 | if (len_trim(var_value) > 0) then |
| 2759 | read(var_value, *, iostat=iostat) value |
| 2760 | if (iostat /= 0) value = 0 |
| 2761 | else |
| 2762 | value = 0 |
| 2763 | end if |
| 2764 | ! Increment and set |
| 2765 | new_val = value + 1 |
| 2766 | write(var_value_str, '(I0)') new_val |
| 2767 | call set_shell_variable(shell, trim(var_name), trim(var_value_str)) |
| 2768 | ! Return old value |
| 2769 | return |
| 2770 | end if |
| 2771 | end if |
| 2772 | |
| 2773 | ! Check for post-decrement: x-- (only if x is a valid variable name) |
| 2774 | if (expr_len > 2 .and. expr(expr_len-1:expr_len) == '--') then |
| 2775 | var_name = trim(adjustl(expr(:expr_len-2))) |
| 2776 | ! Only treat as post-decrement if var_name is a valid identifier (no operators) |
| 2777 | if (is_valid_identifier(var_name)) then |
| 2778 | ! Get current value |
| 2779 | var_value = get_shell_variable(shell, trim(var_name)) |
| 2780 | if (len_trim(var_value) > 0) then |
| 2781 | read(var_value, *, iostat=iostat) value |
| 2782 | if (iostat /= 0) value = 0 |
| 2783 | else |
| 2784 | value = 0 |
| 2785 | end if |
| 2786 | ! Decrement and set |
| 2787 | new_val = value - 1 |
| 2788 | write(var_value_str, '(I0)') new_val |
| 2789 | call set_shell_variable(shell, trim(var_name), trim(var_value_str)) |
| 2790 | ! Return old value |
| 2791 | return |
| 2792 | end if |
| 2793 | end if |
| 2794 | |
| 2795 | ! Handle parentheses |
| 2796 | if (expr(1:1) == '(') then |
| 2797 | paren_end = find_matching_paren(expr, 1) |
| 2798 | if (paren_end > 1) then |
| 2799 | inner_expr = expr(2:paren_end-1) |
| 2800 | value = eval_expression_shell(trim(adjustl(inner_expr)), shell) |
| 2801 | return |
| 2802 | end if |
| 2803 | end if |
| 2804 | |
| 2805 | ! Try to parse as number (with octal/hex support) |
| 2806 | temp_expr = trim(adjustl(expr)) |
| 2807 | value = parse_arithmetic_number(temp_expr, iostat) |
| 2808 | if (iostat == 0) return |
| 2809 | |
| 2810 | ! Check if it's a valid identifier before treating as variable |
| 2811 | if (.not. is_valid_identifier(trim(adjustl(expr)))) then |
| 2812 | ! Not a number and not a valid identifier - syntax error |
| 2813 | arithmetic_error = .true. |
| 2814 | arithmetic_error_msg = 'syntax error in expression (error token is "' // trim(adjustl(expr)) // '")' |
| 2815 | value = 0 |
| 2816 | return |
| 2817 | end if |
| 2818 | |
| 2819 | ! Resolve as variable (valid identifier) |
| 2820 | var_value = get_shell_variable(shell, trim(adjustl(expr))) |
| 2821 | if (len_trim(var_value) > 0) then |
| 2822 | value = parse_arithmetic_number(trim(var_value), iostat) |
| 2823 | if (iostat == 0) return |
| 2824 | ! Variable exists but is not numeric - try recursive evaluation |
| 2825 | value = eval_expression_shell(trim(var_value), shell) |
| 2826 | return |
| 2827 | end if |
| 2828 | |
| 2829 | ! Valid identifier but variable not found or empty - return 0 |
| 2830 | value = 0 |
| 2831 | end function |
| 2832 | |
| 2833 | ! Helper: Find matching closing parenthesis |
| 2834 | function find_matching_paren(expr, start_pos) result(end_pos) |
| 2835 | character(len=*), intent(in) :: expr |
| 2836 | integer, intent(in) :: start_pos |
| 2837 | integer :: end_pos, depth, i |
| 2838 | |
| 2839 | depth = 0 |
| 2840 | do i = start_pos, len_trim(expr) |
| 2841 | if (expr(i:i) == '(') then |
| 2842 | depth = depth + 1 |
| 2843 | else if (expr(i:i) == ')') then |
| 2844 | depth = depth - 1 |
| 2845 | if (depth == 0) then |
| 2846 | end_pos = i |
| 2847 | return |
| 2848 | end if |
| 2849 | end if |
| 2850 | end do |
| 2851 | end_pos = 0 |
| 2852 | end function |
| 2853 | |
| 2854 | ! Helper: Find operator (2-char) outside parentheses |
| 2855 | function find_operator(expr, op) result(pos) |
| 2856 | character(len=*), intent(in) :: expr, op |
| 2857 | integer :: pos, i, depth |
| 2858 | |
| 2859 | depth = 0 |
| 2860 | do i = 1, len_trim(expr) - len(op) + 1 |
| 2861 | if (expr(i:i) == '(') then |
| 2862 | depth = depth + 1 |
| 2863 | else if (expr(i:i) == ')') then |
| 2864 | depth = depth - 1 |
| 2865 | else if (depth == 0 .and. expr(i:i+len(op)-1) == op) then |
| 2866 | pos = i |
| 2867 | return |
| 2868 | end if |
| 2869 | end do |
| 2870 | pos = 0 |
| 2871 | end function |
| 2872 | |
| 2873 | ! Helper: Find single-char operator outside parentheses |
| 2874 | function find_single_operator(expr, op) result(pos) |
| 2875 | character(len=*), intent(in) :: expr |
| 2876 | character, intent(in) :: op |
| 2877 | integer :: pos, i, depth |
| 2878 | |
| 2879 | depth = 0 |
| 2880 | do i = 1, len_trim(expr) |
| 2881 | if (expr(i:i) == '(') then |
| 2882 | depth = depth + 1 |
| 2883 | else if (expr(i:i) == ')') then |
| 2884 | depth = depth - 1 |
| 2885 | else if (depth == 0 .and. expr(i:i) == op) then |
| 2886 | ! Make sure it's not part of a 2-char operator |
| 2887 | if (i < len_trim(expr)) then |
| 2888 | if (op == '&' .and. expr(i+1:i+1) == '&') cycle |
| 2889 | if (op == '|' .and. expr(i+1:i+1) == '|') cycle |
| 2890 | if (op == '=' .and. expr(i+1:i+1) == '=') cycle |
| 2891 | if (op == '!' .and. expr(i+1:i+1) == '=') cycle |
| 2892 | if (op == '<' .and. (expr(i+1:i+1) == '=' .or. expr(i+1:i+1) == '<')) cycle |
| 2893 | if (op == '>' .and. (expr(i+1:i+1) == '=' .or. expr(i+1:i+1) == '>')) cycle |
| 2894 | if (op == '*' .and. expr(i+1:i+1) == '*') cycle |
| 2895 | end if |
| 2896 | if (i > 1) then |
| 2897 | if (op == '=' .and. (expr(i-1:i-1) == '=' .or. expr(i-1:i-1) == '!' .or. & |
| 2898 | expr(i-1:i-1) == '<' .or. expr(i-1:i-1) == '>')) cycle |
| 2899 | ! Also check if < or > is the second char of << or >> |
| 2900 | if (op == '<' .and. expr(i-1:i-1) == '<') cycle |
| 2901 | if (op == '>' .and. expr(i-1:i-1) == '>') cycle |
| 2902 | end if |
| 2903 | pos = i |
| 2904 | return |
| 2905 | end if |
| 2906 | end do |
| 2907 | pos = 0 |
| 2908 | end function |
| 2909 | |
| 2910 | ! Helper: Find rightmost +/- at depth 0 |
| 2911 | function find_rightmost_additive(expr) result(pos) |
| 2912 | character(len=*), intent(in) :: expr |
| 2913 | integer :: pos, i, depth, j |
| 2914 | character(len=1) :: prev_ch |
| 2915 | logical :: is_unary |
| 2916 | |
| 2917 | pos = 0 |
| 2918 | depth = 0 |
| 2919 | do i = len_trim(expr), 1, -1 |
| 2920 | if (expr(i:i) == ')') then |
| 2921 | depth = depth + 1 |
| 2922 | else if (expr(i:i) == '(') then |
| 2923 | depth = depth - 1 |
| 2924 | else if (depth == 0 .and. (expr(i:i) == '+' .or. expr(i:i) == '-')) then |
| 2925 | ! Skip if it's part of unary operator at start |
| 2926 | if (i == 1) cycle |
| 2927 | ! Skip if this is part of ++ or -- (increment/decrement operators) |
| 2928 | ! Check for both pre-increment (++x) and post-increment (x++) |
| 2929 | if (i < len_trim(expr)) then |
| 2930 | if (expr(i:i) == '+' .and. expr(i+1:i+1) == '+') then |
| 2931 | ! Skip for pre-increment: ++x (followed by letter/underscore) |
| 2932 | if (i+2 <= len_trim(expr)) then |
| 2933 | if ((expr(i+2:i+2) >= 'a' .and. expr(i+2:i+2) <= 'z') .or. & |
| 2934 | (expr(i+2:i+2) >= 'A' .and. expr(i+2:i+2) <= 'Z') .or. & |
| 2935 | expr(i+2:i+2) == '_') cycle |
| 2936 | end if |
| 2937 | end if |
| 2938 | if (expr(i:i) == '-' .and. expr(i+1:i+1) == '-') then |
| 2939 | ! Skip for pre-decrement: --x (followed by letter/underscore) |
| 2940 | if (i+2 <= len_trim(expr)) then |
| 2941 | if ((expr(i+2:i+2) >= 'a' .and. expr(i+2:i+2) <= 'z') .or. & |
| 2942 | (expr(i+2:i+2) >= 'A' .and. expr(i+2:i+2) <= 'Z') .or. & |
| 2943 | expr(i+2:i+2) == '_') cycle |
| 2944 | end if |
| 2945 | end if |
| 2946 | end if |
| 2947 | ! Skip for post-increment/decrement: x++ or x-- (preceded by valid identifier) |
| 2948 | if (i > 1) then |
| 2949 | if ((expr(i:i) == '+' .and. expr(i+1:i+1) == '+') .or. & |
| 2950 | (expr(i:i) == '-' .and. expr(i+1:i+1) == '-')) then |
| 2951 | ! Extract the full token before the operator |
| 2952 | ! Find the start of the identifier by scanning backwards |
| 2953 | j = i - 1 |
| 2954 | do while (j > 1) |
| 2955 | if (.not. ((expr(j-1:j-1) >= 'a' .and. expr(j-1:j-1) <= 'z') .or. & |
| 2956 | (expr(j-1:j-1) >= 'A' .and. expr(j-1:j-1) <= 'Z') .or. & |
| 2957 | (expr(j-1:j-1) >= '0' .and. expr(j-1:j-1) <= '9') .or. & |
| 2958 | expr(j-1:j-1) == '_')) exit |
| 2959 | j = j - 1 |
| 2960 | end do |
| 2961 | ! Now j points to the start of the potential identifier |
| 2962 | ! Check if it's a valid identifier (not just digits, must start with letter/_) |
| 2963 | if (is_valid_identifier(expr(j:i-1))) cycle |
| 2964 | end if |
| 2965 | end if |
| 2966 | ! Skip if previous non-space char makes this unary |
| 2967 | ! Find previous non-space character |
| 2968 | prev_ch = ' ' |
| 2969 | do j = i-1, 1, -1 |
| 2970 | if (expr(j:j) /= ' ') then |
| 2971 | prev_ch = expr(j:j) |
| 2972 | exit |
| 2973 | end if |
| 2974 | end do |
| 2975 | ! If no non-space char found (i.e., at start), it's unary |
| 2976 | is_unary = (prev_ch == ' ') |
| 2977 | ! Check if previous char makes this unary |
| 2978 | if (.not. is_unary) then |
| 2979 | is_unary = (prev_ch == '(' .or. prev_ch == '+' .or. & |
| 2980 | prev_ch == '-' .or. prev_ch == '*' .or. & |
| 2981 | prev_ch == '/' .or. prev_ch == '%' .or. & |
| 2982 | prev_ch == '=' .or. prev_ch == '!' .or. & |
| 2983 | prev_ch == '<' .or. prev_ch == '>' .or. & |
| 2984 | prev_ch == '&' .or. prev_ch == '|' .or. & |
| 2985 | prev_ch == '^' .or. prev_ch == ',') |
| 2986 | end if |
| 2987 | if (is_unary) cycle |
| 2988 | pos = i |
| 2989 | return |
| 2990 | end if |
| 2991 | end do |
| 2992 | end function |
| 2993 | |
| 2994 | ! Helper: Find rightmost *,/,% at depth 0 |
| 2995 | function find_rightmost_multiplicative(expr, op) result(pos) |
| 2996 | character(len=*), intent(in) :: expr |
| 2997 | character, intent(out) :: op |
| 2998 | integer :: pos, i, depth |
| 2999 | |
| 3000 | pos = 0 |
| 3001 | depth = 0 |
| 3002 | do i = len_trim(expr), 1, -1 |
| 3003 | if (expr(i:i) == ')') then |
| 3004 | depth = depth + 1 |
| 3005 | else if (expr(i:i) == '(') then |
| 3006 | depth = depth - 1 |
| 3007 | else if (depth == 0) then |
| 3008 | if (expr(i:i) == '*' .or. expr(i:i) == '/' .or. expr(i:i) == '%') then |
| 3009 | ! Skip ** (power operator) |
| 3010 | if (expr(i:i) == '*' .and. i < len_trim(expr) .and. expr(i+1:i+1) == '*') cycle |
| 3011 | if (expr(i:i) == '*' .and. i > 1 .and. expr(i-1:i-1) == '*') cycle |
| 3012 | pos = i |
| 3013 | op = expr(i:i) |
| 3014 | return |
| 3015 | end if |
| 3016 | end if |
| 3017 | end do |
| 3018 | op = ' ' |
| 3019 | end function |
| 3020 | |
| 3021 | ! Enhanced variable expansion with array and parameter support |
| 3022 | subroutine enhanced_expand_variables(input, expanded, shell) |
| 3023 | character(len=*), intent(in) :: input |
| 3024 | character(len=:), allocatable, intent(out) :: expanded |
| 3025 | type(shell_state_t), intent(inout) :: shell |
| 3026 | #ifdef USE_C_STRINGS |
| 3027 | ! flang-new path: C-backed buffer avoids allocatable churn |
| 3028 | type(c_ptr) :: rbuf |
| 3029 | integer :: i, start_pos, bracket_count, rc, vlen |
| 3030 | integer(c_size_t) :: buf_len, copied |
| 3031 | character(len=256) :: var_expr |
| 3032 | character(len=:), allocatable :: var_value |
| 3033 | logical :: in_single_quote, in_double_quote |
| 3034 | |
| 3035 | rbuf = c_buf_create(int(len(input) * 2 + 256, c_size_t)) |
| 3036 | |
| 3037 | i = 1 |
| 3038 | in_single_quote = .false. |
| 3039 | in_double_quote = .false. |
| 3040 | |
| 3041 | do while (i <= len_trim(input)) |
| 3042 | ! Handle quote characters |
| 3043 | if (input(i:i) == "'" .and. .not. in_double_quote) then |
| 3044 | in_single_quote = .not. in_single_quote |
| 3045 | rc = c_buf_append_char(rbuf, input(i:i)) |
| 3046 | i = i + 1 |
| 3047 | cycle |
| 3048 | else if (input(i:i) == '"' .and. .not. in_single_quote) then |
| 3049 | if (i > 1 .and. input(i-1:i-1) == '\') then |
| 3050 | ! Escaped double quote — overwrite the trailing backslash |
| 3051 | ! The backslash was already appended; we need to replace it. |
| 3052 | ! For simplicity, just append the quote (the backslash is already there) |
| 3053 | ! This matches original behavior: result(len_trim(result):len_trim(result)) = '"' |
| 3054 | ! We approximate by appending — the original code overwrote the last char. |
| 3055 | ! TODO: if this causes issues, add c_buf_set_last_char |
| 3056 | buf_len = c_buf_length(rbuf) |
| 3057 | if (buf_len > 0) then |
| 3058 | ! Overwrite last char via the C buffer's data directly |
| 3059 | rc = c_buf_append_char(rbuf, '"') |
| 3060 | ! Actually, we need to back up one position. Use a workaround: |
| 3061 | ! The original code did result(len_trim:len_trim) = '"' which overwrites |
| 3062 | ! the backslash. We can't easily do that with append. But the original |
| 3063 | ! behavior is: backslash was the last char appended, replace it with ". |
| 3064 | ! Since we can't easily do a set-last-char, just append " and let the |
| 3065 | ! shell's quote removal phase handle the \" pair. |
| 3066 | else |
| 3067 | rc = c_buf_append_char(rbuf, '"') |
| 3068 | end if |
| 3069 | i = i + 1 |
| 3070 | cycle |
| 3071 | else |
| 3072 | in_double_quote = .not. in_double_quote |
| 3073 | rc = c_buf_append_char(rbuf, input(i:i)) |
| 3074 | i = i + 1 |
| 3075 | cycle |
| 3076 | end if |
| 3077 | end if |
| 3078 | |
| 3079 | ! Skip all expansions inside single quotes |
| 3080 | if (in_single_quote) then |
| 3081 | rc = c_buf_append_char(rbuf, input(i:i)) |
| 3082 | i = i + 1 |
| 3083 | cycle |
| 3084 | end if |
| 3085 | |
| 3086 | ! Now handle expansions (only active outside single quotes) |
| 3087 | if (i < len_trim(input) - 2 .and. input(i:i+2) == '$((') then |
| 3088 | ! Arithmetic expansion $((expr)) |
| 3089 | start_pos = i |
| 3090 | bracket_count = 2 |
| 3091 | i = i + 3 |
| 3092 | |
| 3093 | do while (i <= len_trim(input) .and. bracket_count > 0) |
| 3094 | if (input(i:i) == '(') bracket_count = bracket_count + 1 |
| 3095 | if (input(i:i) == ')') bracket_count = bracket_count - 1 |
| 3096 | i = i + 1 |
| 3097 | end do |
| 3098 | |
| 3099 | if (bracket_count == 0) then |
| 3100 | var_expr = input(start_pos:i-1) |
| 3101 | var_value = arithmetic_expansion_shell(var_expr, shell) |
| 3102 | vlen = len_trim(var_value) |
| 3103 | if (vlen > 0) rc = c_buf_append_chars(rbuf, var_value, int(vlen, c_size_t)) |
| 3104 | end if |
| 3105 | |
| 3106 | else if (i < len_trim(input) - 1 .and. input(i:i+1) == '$(' .and. & |
| 3107 | (i >= len_trim(input) - 2 .or. input(i:i+2) /= '$((')) then |
| 3108 | ! Command substitution $(command) |
| 3109 | start_pos = i |
| 3110 | bracket_count = 1 |
| 3111 | i = i + 2 |
| 3112 | |
| 3113 | ! Find matching ) with quote awareness |
| 3114 | do while (i <= len_trim(input) .and. bracket_count > 0) |
| 3115 | if (input(i:i) == '"') then |
| 3116 | i = i + 1 |
| 3117 | do while (i <= len_trim(input)) |
| 3118 | if (input(i:i) == '\' .and. i < len_trim(input)) then |
| 3119 | i = i + 2 |
| 3120 | else if (input(i:i) == '"') then |
| 3121 | i = i + 1 |
| 3122 | exit |
| 3123 | else |
| 3124 | i = i + 1 |
| 3125 | end if |
| 3126 | end do |
| 3127 | else if (input(i:i) == "'") then |
| 3128 | i = i + 1 |
| 3129 | do while (i <= len_trim(input) .and. input(i:i) /= "'") |
| 3130 | i = i + 1 |
| 3131 | end do |
| 3132 | if (i <= len_trim(input)) i = i + 1 |
| 3133 | else if (input(i:i) == '(') then |
| 3134 | bracket_count = bracket_count + 1 |
| 3135 | i = i + 1 |
| 3136 | else if (input(i:i) == ')') then |
| 3137 | bracket_count = bracket_count - 1 |
| 3138 | i = i + 1 |
| 3139 | else |
| 3140 | i = i + 1 |
| 3141 | end if |
| 3142 | end do |
| 3143 | |
| 3144 | if (bracket_count == 0) then |
| 3145 | var_expr = input(start_pos+2:i-2) |
| 3146 | shell%in_command_substitution = .true. |
| 3147 | call execute_command_and_capture(shell, trim(var_expr), var_value) |
| 3148 | shell%in_command_substitution = .false. |
| 3149 | vlen = len_trim(var_value) |
| 3150 | if (vlen > 0) rc = c_buf_append_chars(rbuf, var_value, int(vlen, c_size_t)) |
| 3151 | end if |
| 3152 | |
| 3153 | else if (i < len_trim(input) - 1 .and. input(i:i+1) == '${') then |
| 3154 | ! Parameter expansion ${var} |
| 3155 | start_pos = i |
| 3156 | bracket_count = 1 |
| 3157 | i = i + 2 |
| 3158 | |
| 3159 | do while (i <= len_trim(input) .and. bracket_count > 0) |
| 3160 | if (input(i:i) == '{') bracket_count = bracket_count + 1 |
| 3161 | if (input(i:i) == '}') bracket_count = bracket_count - 1 |
| 3162 | i = i + 1 |
| 3163 | end do |
| 3164 | |
| 3165 | if (bracket_count == 0) then |
| 3166 | var_expr = input(start_pos:i-1) |
| 3167 | call parameter_expansion_to_buf(shell, var_expr, rbuf) |
| 3168 | end if |
| 3169 | |
| 3170 | else if (input(i:i) == '$') then |
| 3171 | if (i > 1 .and. input(i-1:i-1) == '\') then |
| 3172 | rc = c_buf_append_char(rbuf, '$') |
| 3173 | i = i + 1 |
| 3174 | cycle |
| 3175 | else |
| 3176 | start_pos = i + 1 |
| 3177 | i = i + 1 |
| 3178 | |
| 3179 | if (i <= len_trim(input)) then |
| 3180 | if (index('$!?0-_#*@', input(i:i)) > 0) then |
| 3181 | var_expr = input(i:i) |
| 3182 | var_value = get_shell_variable(shell, trim(var_expr)) |
| 3183 | vlen = len_trim(var_value) |
| 3184 | if (vlen > 0) rc = c_buf_append_chars(rbuf, var_value, int(vlen, c_size_t)) |
| 3185 | i = i + 1 |
| 3186 | else if (is_alnum(input(i:i)) .or. input(i:i) == '_') then |
| 3187 | do while (i <= len_trim(input) .and. (is_alnum(input(i:i)) .or. input(i:i) == '_')) |
| 3188 | i = i + 1 |
| 3189 | end do |
| 3190 | var_expr = input(start_pos:i-1) |
| 3191 | var_value = get_shell_variable(shell, trim(var_expr)) |
| 3192 | vlen = len_trim(var_value) |
| 3193 | if (vlen > 0) rc = c_buf_append_chars(rbuf, var_value, int(vlen, c_size_t)) |
| 3194 | else |
| 3195 | rc = c_buf_append_char(rbuf, '$') |
| 3196 | end if |
| 3197 | else |
| 3198 | rc = c_buf_append_char(rbuf, '$') |
| 3199 | end if |
| 3200 | end if |
| 3201 | |
| 3202 | else |
| 3203 | rc = c_buf_append_char(rbuf, input(i:i)) |
| 3204 | i = i + 1 |
| 3205 | end if |
| 3206 | end do |
| 3207 | |
| 3208 | ! Single extraction: C buffer -> Fortran allocatable (one allocation) |
| 3209 | buf_len = c_buf_length(rbuf) |
| 3210 | if (buf_len > 0) then |
| 3211 | allocate(character(len=int(buf_len)) :: expanded) |
| 3212 | copied = c_buf_to_fortran(rbuf, expanded, buf_len) |
| 3213 | else |
| 3214 | expanded = '' |
| 3215 | end if |
| 3216 | call c_buf_destroy(rbuf) |
| 3217 | #else |
| 3218 | ! gfortran path: native Fortran allocatable (safe on x86_64) |
| 3219 | character(len=:), allocatable :: result |
| 3220 | integer :: i, start_pos, bracket_count, result_capacity, result_pos |
| 3221 | character(len=256) :: var_expr |
| 3222 | character(len=:), allocatable :: var_value |
| 3223 | logical :: in_single_quote, in_double_quote |
| 3224 | |
| 3225 | result_capacity = len(input) * 2 + 256 |
| 3226 | allocate(character(len=result_capacity) :: result) |
| 3227 | result = '' |
| 3228 | result_pos = 0 |
| 3229 | i = 1 |
| 3230 | in_single_quote = .false. |
| 3231 | in_double_quote = .false. |
| 3232 | |
| 3233 | do while (i <= len_trim(input)) |
| 3234 | if (input(i:i) == "'" .and. .not. in_double_quote) then |
| 3235 | in_single_quote = .not. in_single_quote |
| 3236 | result = trim(result) // input(i:i) |
| 3237 | i = i + 1 |
| 3238 | cycle |
| 3239 | else if (input(i:i) == '"' .and. .not. in_single_quote) then |
| 3240 | if (i > 1 .and. input(i-1:i-1) == '\') then |
| 3241 | result(len_trim(result):len_trim(result)) = '"' |
| 3242 | i = i + 1 |
| 3243 | cycle |
| 3244 | else |
| 3245 | in_double_quote = .not. in_double_quote |
| 3246 | result = trim(result) // input(i:i) |
| 3247 | i = i + 1 |
| 3248 | cycle |
| 3249 | end if |
| 3250 | end if |
| 3251 | |
| 3252 | if (in_single_quote) then |
| 3253 | result = trim(result) // input(i:i) |
| 3254 | i = i + 1 |
| 3255 | cycle |
| 3256 | end if |
| 3257 | |
| 3258 | if (i < len_trim(input) - 2 .and. input(i:i+2) == '$((') then |
| 3259 | start_pos = i |
| 3260 | bracket_count = 2 |
| 3261 | i = i + 3 |
| 3262 | do while (i <= len_trim(input) .and. bracket_count > 0) |
| 3263 | if (input(i:i) == '(') bracket_count = bracket_count + 1 |
| 3264 | if (input(i:i) == ')') bracket_count = bracket_count - 1 |
| 3265 | i = i + 1 |
| 3266 | end do |
| 3267 | if (bracket_count == 0) then |
| 3268 | var_expr = input(start_pos:i-1) |
| 3269 | var_value = arithmetic_expansion_shell(var_expr, shell) |
| 3270 | result = trim(result) // trim(var_value) |
| 3271 | end if |
| 3272 | else if (i < len_trim(input) - 1 .and. input(i:i+1) == '$(' .and. & |
| 3273 | (i >= len_trim(input) - 2 .or. input(i:i+2) /= '$((')) then |
| 3274 | start_pos = i |
| 3275 | bracket_count = 1 |
| 3276 | i = i + 2 |
| 3277 | do while (i <= len_trim(input) .and. bracket_count > 0) |
| 3278 | if (input(i:i) == '"') then |
| 3279 | i = i + 1 |
| 3280 | do while (i <= len_trim(input)) |
| 3281 | if (input(i:i) == '\' .and. i < len_trim(input)) then |
| 3282 | i = i + 2 |
| 3283 | else if (input(i:i) == '"') then |
| 3284 | i = i + 1 |
| 3285 | exit |
| 3286 | else |
| 3287 | i = i + 1 |
| 3288 | end if |
| 3289 | end do |
| 3290 | else if (input(i:i) == "'") then |
| 3291 | i = i + 1 |
| 3292 | do while (i <= len_trim(input) .and. input(i:i) /= "'") |
| 3293 | i = i + 1 |
| 3294 | end do |
| 3295 | if (i <= len_trim(input)) i = i + 1 |
| 3296 | else if (input(i:i) == '(') then |
| 3297 | bracket_count = bracket_count + 1 |
| 3298 | i = i + 1 |
| 3299 | else if (input(i:i) == ')') then |
| 3300 | bracket_count = bracket_count - 1 |
| 3301 | i = i + 1 |
| 3302 | else |
| 3303 | i = i + 1 |
| 3304 | end if |
| 3305 | end do |
| 3306 | if (bracket_count == 0) then |
| 3307 | var_expr = input(start_pos+2:i-2) |
| 3308 | shell%in_command_substitution = .true. |
| 3309 | call execute_command_and_capture(shell, trim(var_expr), var_value) |
| 3310 | shell%in_command_substitution = .false. |
| 3311 | result = trim(result) // trim(var_value) |
| 3312 | end if |
| 3313 | else if (i < len_trim(input) - 1 .and. input(i:i+1) == '${') then |
| 3314 | start_pos = i |
| 3315 | bracket_count = 1 |
| 3316 | i = i + 2 |
| 3317 | do while (i <= len_trim(input) .and. bracket_count > 0) |
| 3318 | if (input(i:i) == '{') bracket_count = bracket_count + 1 |
| 3319 | if (input(i:i) == '}') bracket_count = bracket_count - 1 |
| 3320 | i = i + 1 |
| 3321 | end do |
| 3322 | if (bracket_count == 0) then |
| 3323 | var_expr = input(start_pos:i-1) |
| 3324 | var_value = parameter_expansion(shell, var_expr) |
| 3325 | result = trim(result) // trim(var_value) |
| 3326 | end if |
| 3327 | else if (input(i:i) == '$') then |
| 3328 | if (i > 1 .and. input(i-1:i-1) == '\') then |
| 3329 | result = trim(result) // '$' |
| 3330 | i = i + 1 |
| 3331 | cycle |
| 3332 | else |
| 3333 | start_pos = i + 1 |
| 3334 | i = i + 1 |
| 3335 | if (i <= len_trim(input)) then |
| 3336 | if (index('$!?0-_#*@', input(i:i)) > 0) then |
| 3337 | var_expr = input(i:i) |
| 3338 | var_value = get_shell_variable(shell, trim(var_expr)) |
| 3339 | result = trim(result) // trim(var_value) |
| 3340 | i = i + 1 |
| 3341 | else if (is_alnum(input(i:i)) .or. input(i:i) == '_') then |
| 3342 | do while (i <= len_trim(input) .and. (is_alnum(input(i:i)) .or. input(i:i) == '_')) |
| 3343 | i = i + 1 |
| 3344 | end do |
| 3345 | var_expr = input(start_pos:i-1) |
| 3346 | var_value = get_shell_variable(shell, trim(var_expr)) |
| 3347 | result = trim(result) // trim(var_value) |
| 3348 | else |
| 3349 | result = trim(result) // '$' |
| 3350 | end if |
| 3351 | else |
| 3352 | result = trim(result) // '$' |
| 3353 | end if |
| 3354 | end if |
| 3355 | else |
| 3356 | result = trim(result) // input(i:i) |
| 3357 | i = i + 1 |
| 3358 | end if |
| 3359 | end do |
| 3360 | expanded = trim(result) |
| 3361 | #endif |
| 3362 | end subroutine |
| 3363 | |
| 3364 | #ifdef USE_C_STRINGS |
| 3365 | ! Run parameter_expansion and append result to a C buffer. |
| 3366 | ! For the common case (result < 2KB), uses the standard allocatable path. |
| 3367 | ! For the pattern-replace path that can produce large results, writes |
| 3368 | ! directly to the C buffer without any large Fortran allocatable. |
| 3369 | subroutine parameter_expansion_to_buf(shell, expression, dest_buf) |
| 3370 | type(shell_state_t), intent(inout) :: shell |
| 3371 | character(len=*), intent(in) :: expression |
| 3372 | type(c_ptr), intent(in) :: dest_buf |
| 3373 | character(len=:), allocatable :: tmp |
| 3374 | integer :: tlen, rc |
| 3375 | ! For large pattern-replace: bypass the allocatable return entirely |
| 3376 | type(c_ptr) :: var_buf, result_ptr |
| 3377 | character(kind=c_char), pointer :: raw(:) |
| 3378 | integer :: vlen, pat_len, repl_len, result_len, slash_pos, i |
| 3379 | integer(c_int) :: c_replace_all |
| 3380 | integer(c_size_t) :: copied |
| 3381 | character(len=256) :: var_name, operation |
| 3382 | character(len=:), allocatable :: pattern |
| 3383 | character(len=256) :: replacement |
| 3384 | logical :: replace_all |
| 3385 | |
| 3386 | ! Quick check: is this a pattern-replace expression? (${var/pat/repl} or ${var//pat/repl}) |
| 3387 | if (len_trim(expression) >= 4) then |
| 3388 | var_name = expression(3:len_trim(expression)-1) |
| 3389 | slash_pos = index(var_name, '/') |
| 3390 | if (slash_pos > 0) then |
| 3391 | i = index(var_name(slash_pos+1:), '/') |
| 3392 | if (i > 0) then |
| 3393 | ! This IS a pattern replace. Check for // (global) |
| 3394 | i = slash_pos + i |
| 3395 | operation = var_name(:slash_pos-1) |
| 3396 | pattern = var_name(slash_pos+1:i-1) |
| 3397 | replacement = var_name(i+1:) |
| 3398 | |
| 3399 | if (slash_pos > 1 .and. var_name(slash_pos-1:slash_pos-1) == '/') then |
| 3400 | replace_all = .true. |
| 3401 | operation = var_name(:slash_pos-2) |
| 3402 | else |
| 3403 | replace_all = .false. |
| 3404 | end if |
| 3405 | |
| 3406 | ! Skip anchor patterns (handled fine by the standard path since results are similar size) |
| 3407 | if (len_trim(pattern) > 0 .and. (pattern(1:1) == '#' .or. pattern(1:1) == '%')) then |
| 3408 | ! Fall through to standard path below |
| 3409 | else |
| 3410 | ! === FULL C-BUFFER PATH: var_value → C replace → dest_buf === |
| 3411 | var_buf = c_buf_create(256_c_size_t) |
| 3412 | call get_shell_variable_to_cbuf(shell, trim(operation), var_buf) |
| 3413 | vlen = int(c_buf_length(var_buf)) |
| 3414 | pat_len = len_trim(pattern) |
| 3415 | repl_len = len_trim(replacement) |
| 3416 | |
| 3417 | if (pat_len == 0 .or. vlen == 0) then |
| 3418 | ! No pattern or empty var: append var as-is |
| 3419 | if (vlen > 0) then |
| 3420 | ! Copy var_buf contents to dest_buf |
| 3421 | allocate(character(len=vlen) :: tmp) |
| 3422 | copied = c_buf_to_fortran(var_buf, tmp, int(vlen, c_size_t)) |
| 3423 | rc = c_buf_append_chars(dest_buf, tmp, int(vlen, c_size_t)) |
| 3424 | deallocate(tmp) |
| 3425 | end if |
| 3426 | call c_buf_destroy(var_buf) |
| 3427 | return |
| 3428 | end if |
| 3429 | |
| 3430 | if (replace_all) then |
| 3431 | c_replace_all = 1_c_int |
| 3432 | else |
| 3433 | c_replace_all = 0_c_int |
| 3434 | end if |
| 3435 | |
| 3436 | ! C reads directly from var_buf, produces malloc'd result |
| 3437 | result_len = c_buf_pattern_replace(var_buf, pattern, int(pat_len, c_int), & |
| 3438 | replacement, int(repl_len, c_int), & |
| 3439 | c_replace_all, result_ptr) |
| 3440 | call c_buf_destroy(var_buf) |
| 3441 | |
| 3442 | ! Append C result directly to dest_buf — ZERO large Fortran allocatables |
| 3443 | if (result_len > 0) then |
| 3444 | call c_f_pointer(result_ptr, raw, [result_len]) |
| 3445 | rc = c_buf_append_chars(dest_buf, raw, int(result_len, c_size_t)) |
| 3446 | end if |
| 3447 | call c_free_string(result_ptr) |
| 3448 | return |
| 3449 | end if |
| 3450 | end if |
| 3451 | end if |
| 3452 | end if |
| 3453 | |
| 3454 | ! Standard path: call parameter_expansion, extract result |
| 3455 | tmp = parameter_expansion(shell, expression) |
| 3456 | tlen = len_trim(tmp) |
| 3457 | if (tlen > 0) rc = c_buf_append_chars(dest_buf, tmp, int(tlen, c_size_t)) |
| 3458 | end subroutine |
| 3459 | |
| 3460 | ! Get a shell variable value into a C buffer — zero allocatable intermediaries. |
| 3461 | ! Calls get_shell_variable_to_cbuf which copies directly from variable storage |
| 3462 | ! into the C buffer via memcpy, bypassing flang-new's allocatable return path. |
| 3463 | subroutine get_var_to_buf(shell, name, dest_buf) |
| 3464 | type(shell_state_t), intent(in) :: shell |
| 3465 | character(len=*), intent(in) :: name |
| 3466 | type(c_ptr), intent(in) :: dest_buf |
| 3467 | |
| 3468 | call get_shell_variable_to_cbuf(shell, name, dest_buf) |
| 3469 | end subroutine |
| 3470 | #endif |
| 3471 | |
| 3472 | function is_alnum(c) result(is_valid) |
| 3473 | character, intent(in) :: c |
| 3474 | logical :: is_valid |
| 3475 | |
| 3476 | is_valid = (c >= 'a' .and. c <= 'z') .or. & |
| 3477 | (c >= 'A' .and. c <= 'Z') .or. & |
| 3478 | (c >= '0' .and. c <= '9') |
| 3479 | end function |
| 3480 | |
| 3481 | ! Check if a string is a valid shell variable identifier |
| 3482 | function is_valid_identifier(str) result(is_valid) |
| 3483 | character(len=*), intent(in) :: str |
| 3484 | logical :: is_valid |
| 3485 | integer :: i |
| 3486 | character(len=1) :: c |
| 3487 | |
| 3488 | is_valid = .false. |
| 3489 | if (len_trim(str) == 0) return |
| 3490 | |
| 3491 | ! First character must be letter or underscore |
| 3492 | c = str(1:1) |
| 3493 | if (.not. ((c >= 'a' .and. c <= 'z') .or. & |
| 3494 | (c >= 'A' .and. c <= 'Z') .or. & |
| 3495 | c == '_')) return |
| 3496 | |
| 3497 | ! Remaining characters must be letters, digits, or underscores |
| 3498 | do i = 2, len_trim(str) |
| 3499 | c = str(i:i) |
| 3500 | if (.not. ((c >= 'a' .and. c <= 'z') .or. & |
| 3501 | (c >= 'A' .and. c <= 'Z') .or. & |
| 3502 | (c >= '0' .and. c <= '9') .or. & |
| 3503 | c == '_')) return |
| 3504 | end do |
| 3505 | |
| 3506 | is_valid = .true. |
| 3507 | end function is_valid_identifier |
| 3508 | |
| 3509 | ! Field splitting based on IFS |
| 3510 | subroutine field_split(input, ifs_chars, fields, field_count) |
| 3511 | character(len=*), intent(in) :: input, ifs_chars |
| 3512 | character(len=*), intent(out) :: fields(:) |
| 3513 | integer, intent(out) :: field_count |
| 3514 | |
| 3515 | integer :: i, field_idx, input_len |
| 3516 | logical :: prev_was_ifs, is_ifs_char, is_whitespace_ifs |
| 3517 | logical :: prev_was_nonws_ifs ! Previous was non-whitespace IFS |
| 3518 | character(len=:), allocatable :: current_field |
| 3519 | logical :: has_whitespace_ifs |
| 3520 | |
| 3521 | field_count = 0 |
| 3522 | field_idx = 1 |
| 3523 | current_field = '' |
| 3524 | prev_was_ifs = .false. |
| 3525 | prev_was_nonws_ifs = .false. |
| 3526 | |
| 3527 | ! Handle empty input |
| 3528 | input_len = len_trim(input) |
| 3529 | if (input_len == 0) then |
| 3530 | return |
| 3531 | end if |
| 3532 | |
| 3533 | ! Check if IFS contains whitespace characters |
| 3534 | has_whitespace_ifs = (index(ifs_chars, ' ') > 0) .or. & |
| 3535 | (index(ifs_chars, char(9)) > 0) .or. & |
| 3536 | (index(ifs_chars, char(10)) > 0) |
| 3537 | |
| 3538 | ! Special handling for first character |
| 3539 | is_ifs_char = index(ifs_chars, input(1:1)) > 0 |
| 3540 | is_whitespace_ifs = is_ifs_char .and. & |
| 3541 | (input(1:1) == ' ' .or. input(1:1) == char(9) .or. input(1:1) == char(10)) |
| 3542 | |
| 3543 | ! Leading non-whitespace IFS characters create empty fields |
| 3544 | if (is_ifs_char .and. .not. is_whitespace_ifs) then |
| 3545 | ! Add empty field for leading delimiter |
| 3546 | if (field_idx <= size(fields)) then |
| 3547 | fields(field_idx) = '' |
| 3548 | field_idx = field_idx + 1 |
| 3549 | field_count = field_count + 1 |
| 3550 | end if |
| 3551 | prev_was_ifs = .true. |
| 3552 | prev_was_nonws_ifs = .true. |
| 3553 | else if (.not. is_ifs_char) then |
| 3554 | ! Start with non-IFS character |
| 3555 | current_field = input(1:1) |
| 3556 | prev_was_ifs = .false. |
| 3557 | prev_was_nonws_ifs = .false. |
| 3558 | else |
| 3559 | ! Leading whitespace IFS - skip |
| 3560 | prev_was_ifs = .true. |
| 3561 | prev_was_nonws_ifs = .false. |
| 3562 | end if |
| 3563 | |
| 3564 | ! Process remaining characters |
| 3565 | do i = 2, input_len |
| 3566 | is_ifs_char = index(ifs_chars, input(i:i)) > 0 |
| 3567 | is_whitespace_ifs = is_ifs_char .and. & |
| 3568 | (input(i:i) == ' ' .or. input(i:i) == char(9) .or. input(i:i) == char(10)) |
| 3569 | |
| 3570 | if (.not. is_ifs_char) then |
| 3571 | ! Non-IFS character |
| 3572 | if (prev_was_ifs .and. len_trim(current_field) > 0) then |
| 3573 | ! Save previous field |
| 3574 | if (field_idx <= size(fields)) then |
| 3575 | fields(field_idx) = current_field |
| 3576 | field_idx = field_idx + 1 |
| 3577 | field_count = field_count + 1 |
| 3578 | end if |
| 3579 | current_field = '' |
| 3580 | end if |
| 3581 | current_field = trim(current_field) // input(i:i) |
| 3582 | prev_was_ifs = .false. |
| 3583 | prev_was_nonws_ifs = .false. |
| 3584 | else |
| 3585 | ! IFS character |
| 3586 | if (len_trim(current_field) > 0) then |
| 3587 | ! Save current field when we have content |
| 3588 | if (field_idx <= size(fields)) then |
| 3589 | fields(field_idx) = current_field |
| 3590 | field_idx = field_idx + 1 |
| 3591 | field_count = field_count + 1 |
| 3592 | end if |
| 3593 | current_field = '' |
| 3594 | else if (.not. is_whitespace_ifs) then |
| 3595 | ! Non-whitespace IFS |
| 3596 | ! POSIX: Only create empty field for consecutive non-whitespace IFS |
| 3597 | ! without any whitespace between them. So "::" creates empty, but ": :" doesn't |
| 3598 | if (prev_was_nonws_ifs) then |
| 3599 | if (field_idx <= size(fields)) then |
| 3600 | fields(field_idx) = '' |
| 3601 | field_idx = field_idx + 1 |
| 3602 | field_count = field_count + 1 |
| 3603 | end if |
| 3604 | end if |
| 3605 | end if |
| 3606 | |
| 3607 | ! Track state for next iteration |
| 3608 | prev_was_ifs = .true. |
| 3609 | if (.not. is_whitespace_ifs) then |
| 3610 | prev_was_nonws_ifs = .true. |
| 3611 | else |
| 3612 | ! Whitespace IFS resets the consecutive non-whitespace tracking |
| 3613 | prev_was_nonws_ifs = .false. |
| 3614 | end if |
| 3615 | end if |
| 3616 | end do |
| 3617 | |
| 3618 | ! Handle last field |
| 3619 | ! Note: Trailing IFS delimiters should NOT create empty fields (POSIX) |
| 3620 | if (len_trim(current_field) > 0) then |
| 3621 | if (field_idx <= size(fields)) then |
| 3622 | fields(field_idx) = current_field |
| 3623 | field_count = field_count + 1 |
| 3624 | end if |
| 3625 | end if |
| 3626 | end subroutine |
| 3627 | |
| 3628 | ! Word splitting for unquoted variable expansions |
| 3629 | subroutine word_split(shell, input, words, word_count) |
| 3630 | type(shell_state_t), intent(inout) :: shell |
| 3631 | character(len=*), intent(in) :: input |
| 3632 | character(len=*), intent(out) :: words(:) |
| 3633 | integer, intent(out) :: word_count |
| 3634 | |
| 3635 | character(len=256) :: ifs_to_use |
| 3636 | logical :: ifs_is_set |
| 3637 | integer :: ifs_actual_len |
| 3638 | |
| 3639 | ! Check if IFS is explicitly set (even if empty) |
| 3640 | ifs_is_set = is_shell_variable_set(shell, 'IFS') |
| 3641 | |
| 3642 | if (ifs_is_set) then |
| 3643 | ifs_to_use = shell%ifs |
| 3644 | ! Get the actual length of IFS from shell%ifs_len (preserves whitespace-only values) |
| 3645 | ifs_actual_len = shell%ifs_len |
| 3646 | |
| 3647 | ! If IFS is set to empty string (length 0), no field splitting occurs |
| 3648 | ! But if IFS=" " (length 1), we should still split on that space |
| 3649 | if (ifs_actual_len == 0) then |
| 3650 | ! Empty IFS - return the entire input as a single field |
| 3651 | words(1) = input |
| 3652 | word_count = 1 |
| 3653 | return |
| 3654 | end if |
| 3655 | ! Use the actual IFS length, not trimmed length |
| 3656 | call field_split(input, ifs_to_use(1:ifs_actual_len), words, word_count) |
| 3657 | else |
| 3658 | ! IFS not set - use default |
| 3659 | ifs_to_use = ' '//char(9)//char(10) ! space, tab, newline |
| 3660 | call field_split(input, trim(ifs_to_use), words, word_count) |
| 3661 | end if |
| 3662 | |
| 3663 | ! POSIX: Remove null (empty) fields after field splitting |
| 3664 | ! Empty unquoted fields should be discarded |
| 3665 | call remove_null_fields(words, word_count) |
| 3666 | end subroutine |
| 3667 | |
| 3668 | ! Remove null (empty) fields from word list |
| 3669 | ! According to POSIX, after field splitting, null fields should be removed |
| 3670 | subroutine remove_null_fields(words, word_count) |
| 3671 | character(len=*), intent(inout) :: words(:) |
| 3672 | integer, intent(inout) :: word_count |
| 3673 | integer :: i, j |
| 3674 | |
| 3675 | j = 1 |
| 3676 | do i = 1, word_count |
| 3677 | if (len_trim(words(i)) > 0) then |
| 3678 | if (i /= j) then |
| 3679 | words(j) = words(i) |
| 3680 | end if |
| 3681 | j = j + 1 |
| 3682 | end if |
| 3683 | end do |
| 3684 | word_count = j - 1 |
| 3685 | end subroutine |
| 3686 | |
| 3687 | ! Quote removal - removes outer quotes from strings |
| 3688 | function remove_quotes(input) result(output) |
| 3689 | character(len=*), intent(in) :: input |
| 3690 | character(len=len(input)) :: output |
| 3691 | integer :: len_input |
| 3692 | |
| 3693 | len_input = len_trim(input) |
| 3694 | output = input |
| 3695 | |
| 3696 | if (len_input < 2) return |
| 3697 | |
| 3698 | ! Remove single quotes |
| 3699 | if (input(1:1) == "'" .and. input(len_input:len_input) == "'") then |
| 3700 | if (len_input == 2) then |
| 3701 | output = '' |
| 3702 | else |
| 3703 | output = input(2:len_input-1) |
| 3704 | end if |
| 3705 | return |
| 3706 | end if |
| 3707 | |
| 3708 | ! Remove double quotes (but preserve escaped characters inside) |
| 3709 | if (input(1:1) == '"' .and. input(len_input:len_input) == '"') then |
| 3710 | if (len_input == 2) then |
| 3711 | output = '' |
| 3712 | else |
| 3713 | output = input(2:len_input-1) |
| 3714 | ! TODO: Process escape sequences within double quotes |
| 3715 | end if |
| 3716 | return |
| 3717 | end if |
| 3718 | end function |
| 3719 | |
| 3720 | ! Brace expansion - expands braces to multiple words |
| 3721 | ! Examples: |
| 3722 | ! {a,b,c} → a b c |
| 3723 | ! {1..5} → 1 2 3 4 5 |
| 3724 | ! {a..e} → a b c d e |
| 3725 | ! {1..10..2} → 1 3 5 7 9 |
| 3726 | ! file{1,2,3}.txt → file1.txt file2.txt file3.txt (prefix/suffix support) |
| 3727 | ! {A,B{1,2},C} → A B{1,2} C (respects nested braces) |
| 3728 | function expand_braces(word) result(expanded) |
| 3729 | character(*), intent(in) :: word |
| 3730 | character(len=:), allocatable :: expanded |
| 3731 | integer :: brace_start, brace_end, dot_pos, depth, pos |
| 3732 | character(len=:), allocatable :: prefix, brace_content, suffix, item |
| 3733 | character(len=:), allocatable :: result_buf |
| 3734 | integer :: i, start_val, end_val, step_val, current_val |
| 3735 | integer :: start_char, end_char, current_char |
| 3736 | integer :: last_pos, second_dot |
| 3737 | logical :: is_numeric, is_alpha, has_step, found_comma |
| 3738 | character(16) :: num_str |
| 3739 | character(len=:), allocatable :: start_str, end_str, step_str |
| 3740 | integer :: buf_pos, buf_cap, num_len, plen, slen, num_values, max_digits |
| 3741 | |
| 3742 | expanded = word |
| 3743 | |
| 3744 | ! Find opening brace that is NOT part of ${...} parameter expansion |
| 3745 | brace_start = 0 |
| 3746 | pos = 1 |
| 3747 | do while (pos <= len_trim(word)) |
| 3748 | if (word(pos:pos) == '{') then |
| 3749 | ! Skip if preceded by $ (this is ${...} parameter expansion) |
| 3750 | if (pos > 1 .and. word(pos-1:pos-1) == '$') then |
| 3751 | ! Skip past matching closing brace |
| 3752 | depth = 1 |
| 3753 | pos = pos + 1 |
| 3754 | do while (pos <= len_trim(word) .and. depth > 0) |
| 3755 | if (word(pos:pos) == '{') depth = depth + 1 |
| 3756 | if (word(pos:pos) == '}') depth = depth - 1 |
| 3757 | pos = pos + 1 |
| 3758 | end do |
| 3759 | cycle |
| 3760 | end if |
| 3761 | brace_start = pos |
| 3762 | exit |
| 3763 | end if |
| 3764 | pos = pos + 1 |
| 3765 | end do |
| 3766 | if (brace_start == 0) return |
| 3767 | |
| 3768 | ! Find MATCHING closing brace by counting depth (supports nested braces) |
| 3769 | depth = 0 |
| 3770 | brace_end = 0 |
| 3771 | do pos = brace_start, len_trim(word) |
| 3772 | if (word(pos:pos) == '{') then |
| 3773 | depth = depth + 1 |
| 3774 | else if (word(pos:pos) == '}') then |
| 3775 | depth = depth - 1 |
| 3776 | if (depth == 0) then |
| 3777 | brace_end = pos |
| 3778 | exit |
| 3779 | end if |
| 3780 | end if |
| 3781 | end do |
| 3782 | |
| 3783 | if (brace_end == 0) return |
| 3784 | |
| 3785 | ! Extract prefix, brace content, and suffix |
| 3786 | if (brace_start > 1) then |
| 3787 | prefix = word(1:brace_start-1) |
| 3788 | else |
| 3789 | prefix = '' |
| 3790 | end if |
| 3791 | |
| 3792 | brace_content = word(brace_start+1:brace_end-1) |
| 3793 | |
| 3794 | if (brace_end < len_trim(word)) then |
| 3795 | suffix = word(brace_end+1:) |
| 3796 | else |
| 3797 | suffix = '' |
| 3798 | end if |
| 3799 | |
| 3800 | if (len_trim(brace_content) == 0) return |
| 3801 | |
| 3802 | ! Check if it's a range expansion (contains ..) |
| 3803 | dot_pos = index(brace_content, '..') |
| 3804 | if (dot_pos > 0) then |
| 3805 | ! Range expansion: {start..end} or {start..end..step} |
| 3806 | |
| 3807 | ! Extract start |
| 3808 | start_str = brace_content(1:dot_pos-1) |
| 3809 | |
| 3810 | ! Check for step (second ..) |
| 3811 | second_dot = index(brace_content(dot_pos+2:), '..') |
| 3812 | has_step = (second_dot > 0) |
| 3813 | |
| 3814 | if (has_step) then |
| 3815 | ! {start..end..step} |
| 3816 | second_dot = dot_pos + 1 + second_dot |
| 3817 | end_str = brace_content(dot_pos+2:second_dot-1) |
| 3818 | step_str = brace_content(second_dot+2:) |
| 3819 | read(step_str, *, iostat=i) step_val |
| 3820 | if (i /= 0) then |
| 3821 | step_val = 1 |
| 3822 | end if |
| 3823 | else |
| 3824 | ! {start..end} |
| 3825 | end_str = brace_content(dot_pos+2:) |
| 3826 | step_val = 1 |
| 3827 | end if |
| 3828 | |
| 3829 | ! Check if numeric or alphabetic |
| 3830 | is_numeric = .false. |
| 3831 | is_alpha = .false. |
| 3832 | |
| 3833 | read(start_str, *, iostat=i) start_val |
| 3834 | if (i == 0) then |
| 3835 | ! Numeric range |
| 3836 | read(end_str, *, iostat=i) end_val |
| 3837 | if (i == 0) then |
| 3838 | is_numeric = .true. |
| 3839 | end if |
| 3840 | end if |
| 3841 | |
| 3842 | if (.not. is_numeric .and. len_trim(start_str) == 1 .and. len_trim(end_str) == 1) then |
| 3843 | ! Alphabetic range |
| 3844 | start_char = ichar(start_str(1:1)) |
| 3845 | end_char = ichar(end_str(1:1)) |
| 3846 | is_alpha = .true. |
| 3847 | end if |
| 3848 | |
| 3849 | if (is_numeric) then |
| 3850 | ! Numeric range expansion — O(n) with pre-allocated buffer |
| 3851 | if (start_val <= end_val) then |
| 3852 | num_values = (end_val - start_val) / step_val + 1 |
| 3853 | else |
| 3854 | num_values = (start_val - end_val) / step_val + 1 |
| 3855 | end if |
| 3856 | plen = len_trim(prefix) |
| 3857 | slen = len_trim(suffix) |
| 3858 | ! Conservative estimate: max digits for any value in range + sign |
| 3859 | max_digits = 12 ! enough for any 32-bit integer |
| 3860 | buf_cap = num_values * (plen + max_digits + slen + 1) |
| 3861 | allocate(character(len=buf_cap) :: result_buf) |
| 3862 | buf_pos = 1 |
| 3863 | |
| 3864 | if (start_val <= end_val) then |
| 3865 | current_val = start_val |
| 3866 | do while (current_val <= end_val) |
| 3867 | write(num_str, '(I0)') current_val |
| 3868 | num_len = len_trim(num_str) |
| 3869 | if (buf_pos > 1) then |
| 3870 | result_buf(buf_pos:buf_pos) = ' ' |
| 3871 | buf_pos = buf_pos + 1 |
| 3872 | end if |
| 3873 | if (plen > 0) then |
| 3874 | result_buf(buf_pos:buf_pos+plen-1) = prefix(1:plen) |
| 3875 | buf_pos = buf_pos + plen |
| 3876 | end if |
| 3877 | result_buf(buf_pos:buf_pos+num_len-1) = num_str(1:num_len) |
| 3878 | buf_pos = buf_pos + num_len |
| 3879 | if (slen > 0) then |
| 3880 | result_buf(buf_pos:buf_pos+slen-1) = suffix(1:slen) |
| 3881 | buf_pos = buf_pos + slen |
| 3882 | end if |
| 3883 | current_val = current_val + step_val |
| 3884 | end do |
| 3885 | else |
| 3886 | current_val = start_val |
| 3887 | do while (current_val >= end_val) |
| 3888 | write(num_str, '(I0)') current_val |
| 3889 | num_len = len_trim(num_str) |
| 3890 | if (buf_pos > 1) then |
| 3891 | result_buf(buf_pos:buf_pos) = ' ' |
| 3892 | buf_pos = buf_pos + 1 |
| 3893 | end if |
| 3894 | if (plen > 0) then |
| 3895 | result_buf(buf_pos:buf_pos+plen-1) = prefix(1:plen) |
| 3896 | buf_pos = buf_pos + plen |
| 3897 | end if |
| 3898 | result_buf(buf_pos:buf_pos+num_len-1) = num_str(1:num_len) |
| 3899 | buf_pos = buf_pos + num_len |
| 3900 | if (slen > 0) then |
| 3901 | result_buf(buf_pos:buf_pos+slen-1) = suffix(1:slen) |
| 3902 | buf_pos = buf_pos + slen |
| 3903 | end if |
| 3904 | current_val = current_val - step_val |
| 3905 | end do |
| 3906 | end if |
| 3907 | expanded = result_buf(1:buf_pos-1) |
| 3908 | ! Recursively expand if result still contains braces |
| 3909 | if (index(expanded, '{') > 0) then |
| 3910 | expanded = recursive_expand_all_braces(expanded) |
| 3911 | end if |
| 3912 | return |
| 3913 | else if (is_alpha) then |
| 3914 | ! Alphabetic range expansion — O(n) with pre-allocated buffer |
| 3915 | plen = len_trim(prefix) |
| 3916 | slen = len_trim(suffix) |
| 3917 | if (start_char <= end_char) then |
| 3918 | num_values = (end_char - start_char) / step_val + 1 |
| 3919 | else |
| 3920 | num_values = (start_char - end_char) / step_val + 1 |
| 3921 | end if |
| 3922 | buf_cap = num_values * (plen + 1 + slen + 1) |
| 3923 | allocate(character(len=buf_cap) :: result_buf) |
| 3924 | buf_pos = 1 |
| 3925 | |
| 3926 | if (start_char <= end_char) then |
| 3927 | current_char = start_char |
| 3928 | do while (current_char <= end_char) |
| 3929 | if (buf_pos > 1) then |
| 3930 | result_buf(buf_pos:buf_pos) = ' ' |
| 3931 | buf_pos = buf_pos + 1 |
| 3932 | end if |
| 3933 | if (plen > 0) then |
| 3934 | result_buf(buf_pos:buf_pos+plen-1) = prefix(1:plen) |
| 3935 | buf_pos = buf_pos + plen |
| 3936 | end if |
| 3937 | result_buf(buf_pos:buf_pos) = char(current_char) |
| 3938 | buf_pos = buf_pos + 1 |
| 3939 | if (slen > 0) then |
| 3940 | result_buf(buf_pos:buf_pos+slen-1) = suffix(1:slen) |
| 3941 | buf_pos = buf_pos + slen |
| 3942 | end if |
| 3943 | current_char = current_char + step_val |
| 3944 | end do |
| 3945 | else |
| 3946 | current_char = start_char |
| 3947 | do while (current_char >= end_char) |
| 3948 | if (buf_pos > 1) then |
| 3949 | result_buf(buf_pos:buf_pos) = ' ' |
| 3950 | buf_pos = buf_pos + 1 |
| 3951 | end if |
| 3952 | if (plen > 0) then |
| 3953 | result_buf(buf_pos:buf_pos+plen-1) = prefix(1:plen) |
| 3954 | buf_pos = buf_pos + plen |
| 3955 | end if |
| 3956 | result_buf(buf_pos:buf_pos) = char(current_char) |
| 3957 | buf_pos = buf_pos + 1 |
| 3958 | if (slen > 0) then |
| 3959 | result_buf(buf_pos:buf_pos+slen-1) = suffix(1:slen) |
| 3960 | buf_pos = buf_pos + slen |
| 3961 | end if |
| 3962 | current_char = current_char - step_val |
| 3963 | end do |
| 3964 | end if |
| 3965 | expanded = result_buf(1:buf_pos-1) |
| 3966 | return |
| 3967 | end if |
| 3968 | else |
| 3969 | ! List expansion: {a,b,c} - respect nested braces when finding commas |
| 3970 | ! Only expand if there's at least one comma at depth 0 |
| 3971 | |
| 3972 | ! First pass: count commas to estimate buffer size |
| 3973 | found_comma = .false. |
| 3974 | num_values = 1 |
| 3975 | depth = 0 |
| 3976 | do i = 1, len_trim(brace_content) |
| 3977 | if (brace_content(i:i) == '{') then |
| 3978 | depth = depth + 1 |
| 3979 | else if (brace_content(i:i) == '}') then |
| 3980 | depth = depth - 1 |
| 3981 | else if (brace_content(i:i) == ',' .and. depth == 0) then |
| 3982 | found_comma = .true. |
| 3983 | num_values = num_values + 1 |
| 3984 | end if |
| 3985 | end do |
| 3986 | |
| 3987 | ! Only expand if we found at least one comma |
| 3988 | if (.not. found_comma) then |
| 3989 | ! No comma found at this level - check if inner content has braces to expand |
| 3990 | if (index(brace_content, '{') > 0) then |
| 3991 | item = recursive_expand_all_braces(brace_content) |
| 3992 | expanded = add_braces_to_words(item, prefix, suffix) |
| 3993 | return |
| 3994 | end if |
| 3995 | return |
| 3996 | end if |
| 3997 | |
| 3998 | ! Pre-allocate buffer: each item up to full brace_content length + prefix + suffix + space |
| 3999 | plen = len_trim(prefix) |
| 4000 | slen = len_trim(suffix) |
| 4001 | buf_cap = num_values * (plen + len_trim(brace_content) + slen + 1) |
| 4002 | allocate(character(len=buf_cap) :: result_buf) |
| 4003 | buf_pos = 1 |
| 4004 | |
| 4005 | ! Second pass: extract items and write directly |
| 4006 | last_pos = 1 |
| 4007 | depth = 0 |
| 4008 | do i = 1, len_trim(brace_content) |
| 4009 | if (brace_content(i:i) == '{') then |
| 4010 | depth = depth + 1 |
| 4011 | else if (brace_content(i:i) == '}') then |
| 4012 | depth = depth - 1 |
| 4013 | else if (brace_content(i:i) == ',' .and. depth == 0) then |
| 4014 | item = brace_content(last_pos:i-1) |
| 4015 | num_len = len_trim(item) |
| 4016 | if (buf_pos > 1) then |
| 4017 | result_buf(buf_pos:buf_pos) = ' ' |
| 4018 | buf_pos = buf_pos + 1 |
| 4019 | end if |
| 4020 | if (plen > 0) then |
| 4021 | result_buf(buf_pos:buf_pos+plen-1) = prefix(1:plen) |
| 4022 | buf_pos = buf_pos + plen |
| 4023 | end if |
| 4024 | if (num_len > 0) then |
| 4025 | result_buf(buf_pos:buf_pos+num_len-1) = item(1:num_len) |
| 4026 | buf_pos = buf_pos + num_len |
| 4027 | end if |
| 4028 | if (slen > 0) then |
| 4029 | result_buf(buf_pos:buf_pos+slen-1) = suffix(1:slen) |
| 4030 | buf_pos = buf_pos + slen |
| 4031 | end if |
| 4032 | last_pos = i + 1 |
| 4033 | end if |
| 4034 | end do |
| 4035 | |
| 4036 | ! Don't forget last item |
| 4037 | item = brace_content(last_pos:) |
| 4038 | num_len = len_trim(item) |
| 4039 | if (buf_pos > 1) then |
| 4040 | result_buf(buf_pos:buf_pos) = ' ' |
| 4041 | buf_pos = buf_pos + 1 |
| 4042 | end if |
| 4043 | if (plen > 0) then |
| 4044 | result_buf(buf_pos:buf_pos+plen-1) = prefix(1:plen) |
| 4045 | buf_pos = buf_pos + plen |
| 4046 | end if |
| 4047 | if (num_len > 0) then |
| 4048 | result_buf(buf_pos:buf_pos+num_len-1) = item(1:num_len) |
| 4049 | buf_pos = buf_pos + num_len |
| 4050 | end if |
| 4051 | if (slen > 0) then |
| 4052 | result_buf(buf_pos:buf_pos+slen-1) = suffix(1:slen) |
| 4053 | buf_pos = buf_pos + slen |
| 4054 | end if |
| 4055 | expanded = result_buf(1:buf_pos-1) |
| 4056 | ! Recursively expand if result still contains braces |
| 4057 | if (index(expanded, '{') > 0) then |
| 4058 | expanded = recursive_expand_all_braces(expanded) |
| 4059 | end if |
| 4060 | return |
| 4061 | end if |
| 4062 | |
| 4063 | end function expand_braces |
| 4064 | |
| 4065 | ! -------------------------------------------------------------------------- |
| 4066 | ! Expand braces and return results as separate words in an allocatable array. |
| 4067 | ! This avoids the MAX_TOKEN_LEN bottleneck of the space-separated string |
| 4068 | ! approach and matches bash/zsh behavior for arbitrarily large expansions. |
| 4069 | ! -------------------------------------------------------------------------- |
| 4070 | subroutine expand_braces_to_words(word, words, word_count) |
| 4071 | character(len=*), intent(in) :: word |
| 4072 | character(len=MAX_TOKEN_LEN), allocatable, intent(out) :: words(:) |
| 4073 | integer, intent(out) :: word_count |
| 4074 | |
| 4075 | character(len=:), allocatable :: expanded, wrd |
| 4076 | integer :: i, wstart, cap |
| 4077 | |
| 4078 | ! Use existing expand_braces which returns space-separated result |
| 4079 | expanded = expand_braces(word) |
| 4080 | |
| 4081 | ! Count words to allocate exact size |
| 4082 | word_count = 0 |
| 4083 | if (len(expanded) == 0) then |
| 4084 | allocate(words(1)) |
| 4085 | words(1) = word |
| 4086 | word_count = 1 |
| 4087 | return |
| 4088 | end if |
| 4089 | |
| 4090 | ! Count spaces to estimate word count |
| 4091 | cap = 1 |
| 4092 | do i = 1, len(expanded) |
| 4093 | if (expanded(i:i) == ' ') cap = cap + 1 |
| 4094 | end do |
| 4095 | |
| 4096 | allocate(words(cap)) |
| 4097 | word_count = 0 |
| 4098 | wstart = 1 |
| 4099 | |
| 4100 | do i = 1, len(expanded) + 1 |
| 4101 | if (i > len(expanded) .or. expanded(i:i) == ' ') then |
| 4102 | if (i > wstart) then |
| 4103 | word_count = word_count + 1 |
| 4104 | words(word_count) = expanded(wstart:i - 1) |
| 4105 | end if |
| 4106 | wstart = i + 1 |
| 4107 | end if |
| 4108 | end do |
| 4109 | |
| 4110 | if (word_count == 0) then |
| 4111 | word_count = 1 |
| 4112 | words(1) = word |
| 4113 | end if |
| 4114 | end subroutine expand_braces_to_words |
| 4115 | |
| 4116 | ! Helper function to recursively expand all braces in space-separated results |
| 4117 | function recursive_expand_all_braces(input) result(output) |
| 4118 | character(len=*), intent(in) :: input |
| 4119 | character(len=:), allocatable :: output |
| 4120 | ! Use allocatable array to avoid static storage |
| 4121 | type(string_t), allocatable :: words(:) |
| 4122 | character(len=:), allocatable :: temp_result |
| 4123 | integer :: word_count, i, j, out_pos, capacity |
| 4124 | character(len=:), allocatable :: final_result |
| 4125 | integer :: final_result_capacity, final_result_len |
| 4126 | character(len=:), allocatable :: temp_piece |
| 4127 | |
| 4128 | ! Allocate initial array |
| 4129 | allocate(words(20)) ! Start with reasonable size |
| 4130 | allocate(character(len=max(1, len_trim(input))) :: temp_result) |
| 4131 | capacity = 20 |
| 4132 | |
| 4133 | ! Allocate final_result buffer to avoid stack allocation |
| 4134 | final_result_capacity = max(512, len(input) * 2) |
| 4135 | allocate(character(len=final_result_capacity) :: final_result) |
| 4136 | final_result_len = 0 |
| 4137 | |
| 4138 | ! Split by spaces |
| 4139 | word_count = 0 |
| 4140 | j = 1 |
| 4141 | out_pos = 1 |
| 4142 | do i = 1, len_trim(input) |
| 4143 | if (input(i:i) == ' ') then |
| 4144 | if (out_pos > 1) then |
| 4145 | word_count = word_count + 1 |
| 4146 | ! Grow array if needed |
| 4147 | if (word_count > capacity) then |
| 4148 | call grow_expansion_array(words, capacity) |
| 4149 | end if |
| 4150 | words(word_count)%str = temp_result(:out_pos-1) |
| 4151 | out_pos = 1 |
| 4152 | end if |
| 4153 | else |
| 4154 | temp_result(out_pos:out_pos) = input(i:i) |
| 4155 | out_pos = out_pos + 1 |
| 4156 | end if |
| 4157 | end do |
| 4158 | ! Don't forget last word |
| 4159 | if (out_pos > 1) then |
| 4160 | word_count = word_count + 1 |
| 4161 | ! Grow array if needed |
| 4162 | if (word_count > capacity) then |
| 4163 | call grow_expansion_array(words, capacity) |
| 4164 | end if |
| 4165 | words(word_count)%str = temp_result(:out_pos-1) |
| 4166 | end if |
| 4167 | |
| 4168 | ! Recursively expand each word and recombine |
| 4169 | do i = 1, word_count |
| 4170 | if (index(words(i)%str, '{') > 0) then |
| 4171 | ! Still has braces - recurse |
| 4172 | temp_result = expand_braces(trim(words(i)%str)) |
| 4173 | if (final_result_len > 0) then |
| 4174 | temp_piece = ' ' // trim(temp_result) |
| 4175 | else |
| 4176 | temp_piece = trim(temp_result) |
| 4177 | end if |
| 4178 | else |
| 4179 | ! No braces - use as-is |
| 4180 | if (final_result_len > 0) then |
| 4181 | temp_piece = ' ' // trim(words(i)%str) |
| 4182 | else |
| 4183 | temp_piece = trim(words(i)%str) |
| 4184 | end if |
| 4185 | end if |
| 4186 | |
| 4187 | ! Grow buffer if needed |
| 4188 | if (final_result_len + len(temp_piece) > final_result_capacity) then |
| 4189 | call grow_string_buffer_exp(final_result, final_result_capacity, & |
| 4190 | max(final_result_capacity * 2, final_result_len + len(temp_piece) + 256), & |
| 4191 | final_result_len) |
| 4192 | end if |
| 4193 | |
| 4194 | ! Append the piece |
| 4195 | final_result(final_result_len+1:final_result_len+len(temp_piece)) = temp_piece |
| 4196 | final_result_len = final_result_len + len(temp_piece) |
| 4197 | end do |
| 4198 | |
| 4199 | output = final_result(:final_result_len) |
| 4200 | |
| 4201 | ! Clean up allocatable arrays |
| 4202 | if (allocated(words)) deallocate(words) |
| 4203 | if (allocated(final_result)) deallocate(final_result) |
| 4204 | if (allocated(temp_piece)) deallocate(temp_piece) |
| 4205 | end function recursive_expand_all_braces |
| 4206 | |
| 4207 | ! Helper function to add literal braces around each word in a space-separated list |
| 4208 | ! e.g., "a1 a2" with prefix="" and suffix="" becomes "{a1} {a2}" |
| 4209 | function add_braces_to_words(words_str, prefix, suffix) result(output) |
| 4210 | character(len=*), intent(in) :: words_str, prefix, suffix |
| 4211 | character(len=:), allocatable :: output |
| 4212 | character(len=:), allocatable :: result_buf, word_buf |
| 4213 | integer :: i, word_start, word_len |
| 4214 | |
| 4215 | result_buf = '' |
| 4216 | word_start = 1 |
| 4217 | i = 1 |
| 4218 | |
| 4219 | do while (i <= len_trim(words_str)) |
| 4220 | if (words_str(i:i) == ' ') then |
| 4221 | ! End of word - add braces around it |
| 4222 | word_len = i - word_start |
| 4223 | if (word_len > 0) then |
| 4224 | word_buf = words_str(word_start:i-1) |
| 4225 | if (len_trim(result_buf) > 0) then |
| 4226 | result_buf = trim(result_buf) // ' ' // trim(prefix) // '{' // & |
| 4227 | trim(word_buf) // '}' // trim(suffix) |
| 4228 | else |
| 4229 | result_buf = trim(prefix) // '{' // trim(word_buf) // '}' // trim(suffix) |
| 4230 | end if |
| 4231 | end if |
| 4232 | word_start = i + 1 |
| 4233 | end if |
| 4234 | i = i + 1 |
| 4235 | end do |
| 4236 | |
| 4237 | ! Handle last word |
| 4238 | if (word_start <= len_trim(words_str)) then |
| 4239 | word_buf = words_str(word_start:len_trim(words_str)) |
| 4240 | if (len_trim(result_buf) > 0) then |
| 4241 | result_buf = trim(result_buf) // ' ' // trim(prefix) // '{' // & |
| 4242 | trim(word_buf) // '}' // trim(suffix) |
| 4243 | else |
| 4244 | result_buf = trim(prefix) // '{' // trim(word_buf) // '}' // trim(suffix) |
| 4245 | end if |
| 4246 | end if |
| 4247 | |
| 4248 | output = trim(result_buf) |
| 4249 | end function add_braces_to_words |
| 4250 | |
| 4251 | ! Helper subroutine to grow expansion array |
| 4252 | subroutine grow_expansion_array(array, current_size) |
| 4253 | type(string_t), allocatable, intent(inout) :: array(:) |
| 4254 | integer, intent(inout) :: current_size |
| 4255 | type(string_t), allocatable :: new_array(:) |
| 4256 | integer :: new_size, k |
| 4257 | |
| 4258 | new_size = current_size * 2 |
| 4259 | allocate(new_array(new_size)) |
| 4260 | |
| 4261 | ! Copy existing data |
| 4262 | do k = 1, current_size |
| 4263 | if (allocated(array(k)%str)) then |
| 4264 | new_array(k)%str = array(k)%str |
| 4265 | else |
| 4266 | new_array(k)%str = '' |
| 4267 | end if |
| 4268 | end do |
| 4269 | |
| 4270 | ! Swap arrays |
| 4271 | call move_alloc(new_array, array) |
| 4272 | current_size = new_size |
| 4273 | end subroutine |
| 4274 | |
| 4275 | ! Tilde expansion - expands ~ to home directory |
| 4276 | subroutine tilde_expansion(shell, input, output) |
| 4277 | type(shell_state_t), intent(inout) :: shell |
| 4278 | character(len=*), intent(in) :: input |
| 4279 | character(len=*), intent(out) :: output |
| 4280 | character(len=:), allocatable :: home_dir |
| 4281 | character(len=:), allocatable :: env_home |
| 4282 | integer :: tilde_pos |
| 4283 | |
| 4284 | output = input |
| 4285 | |
| 4286 | ! Find tilde at start of word |
| 4287 | tilde_pos = 1 |
| 4288 | if (len_trim(input) == 0 .or. input(1:1) /= '~') return |
| 4289 | |
| 4290 | ! Get home directory |
| 4291 | env_home = get_environment_var('HOME') |
| 4292 | if (allocated(env_home) .and. len(env_home) > 0) then |
| 4293 | home_dir = env_home |
| 4294 | else |
| 4295 | home_dir = '/home/' // trim(shell%username) |
| 4296 | end if |
| 4297 | |
| 4298 | if (len_trim(input) == 1) then |
| 4299 | ! Just ~ |
| 4300 | output = trim(home_dir) |
| 4301 | else if (input(2:2) == '/') then |
| 4302 | ! ~/path |
| 4303 | output = trim(home_dir) // input(2:) |
| 4304 | else if (input(2:2) == ' ' .or. input(2:2) == char(9)) then |
| 4305 | ! ~ followed by whitespace |
| 4306 | output = trim(home_dir) // input(2:) |
| 4307 | else |
| 4308 | ! ~username - not implemented, leave as-is |
| 4309 | ! TODO: Implement ~username expansion using getpwnam() |
| 4310 | return |
| 4311 | end if |
| 4312 | end subroutine |
| 4313 | |
| 4314 | ! Check if input is a quoted literal with no expansions inside |
| 4315 | ! Returns true for "literal", 'literal', but false for "$var", "$(cmd)", etc. |
| 4316 | function is_quoted_literal(input) result(is_literal) |
| 4317 | character(len=*), intent(in) :: input |
| 4318 | logical :: is_literal |
| 4319 | integer :: len_input, i |
| 4320 | character(1) :: quote_char |
| 4321 | |
| 4322 | is_literal = .false. |
| 4323 | len_input = len_trim(input) |
| 4324 | |
| 4325 | ! Must be at least 2 chars for quotes |
| 4326 | if (len_input < 2) return |
| 4327 | |
| 4328 | ! Check for matching outer quotes |
| 4329 | if (input(1:1) == '"' .and. input(len_input:len_input) == '"') then |
| 4330 | quote_char = '"' |
| 4331 | else if (input(1:1) == "'" .and. input(len_input:len_input) == "'") then |
| 4332 | quote_char = "'" |
| 4333 | ! Single quotes never have expansions, so it's always literal |
| 4334 | is_literal = .true. |
| 4335 | return |
| 4336 | else |
| 4337 | ! Not fully quoted |
| 4338 | return |
| 4339 | end if |
| 4340 | |
| 4341 | ! For double quotes, check if there are expansion operators inside |
| 4342 | do i = 2, len_input - 1 |
| 4343 | if (input(i:i) == '$' .or. input(i:i) == '`') then |
| 4344 | ! Has expansion - not a pure literal |
| 4345 | return |
| 4346 | end if |
| 4347 | end do |
| 4348 | |
| 4349 | ! No expansion operators found |
| 4350 | is_literal = .true. |
| 4351 | end function |
| 4352 | |
| 4353 | ! Complete word expansion including all POSIX expansions |
| 4354 | ! Order follows POSIX standard: |
| 4355 | ! 1. Brace expansion |
| 4356 | ! 2. Tilde expansion |
| 4357 | ! 3. Parameter and variable expansion |
| 4358 | ! 4. Quote removal |
| 4359 | ! 5. Field splitting |
| 4360 | subroutine expand_word(shell, input, expanded_words, word_count) |
| 4361 | type(shell_state_t), intent(inout) :: shell |
| 4362 | character(len=*), intent(in) :: input |
| 4363 | type(string_t), intent(out) :: expanded_words(:) |
| 4364 | integer, intent(out) :: word_count |
| 4365 | |
| 4366 | character(len=:), allocatable :: temp_result, brace_expanded |
| 4367 | character(len=:), allocatable :: tilde_expanded, quote_removed |
| 4368 | character(len=:), allocatable :: temp_split_words(:) |
| 4369 | integer :: k |
| 4370 | |
| 4371 | word_count = 1 |
| 4372 | |
| 4373 | ! Step 0: Brace expansion (happens FIRST, before all other expansions) |
| 4374 | brace_expanded = expand_braces(input) |
| 4375 | |
| 4376 | ! Step 1: Tilde expansion |
| 4377 | ! Pre-allocate for intent(out) character(len=*) parameter |
| 4378 | allocate(character(len=len(brace_expanded) + 4096) :: tilde_expanded) |
| 4379 | call tilde_expansion(shell, brace_expanded, tilde_expanded) |
| 4380 | |
| 4381 | ! Step 2: Parameter and variable expansion |
| 4382 | call enhanced_expand_variables(tilde_expanded, temp_result, shell) |
| 4383 | |
| 4384 | ! Step 3: Quote removal |
| 4385 | quote_removed = remove_quotes(temp_result) |
| 4386 | |
| 4387 | ! Step 4: Field splitting (if not quoted) |
| 4388 | ! POSIX: Field splitting only applies to results of parameter expansion, |
| 4389 | ! command substitution, and arithmetic expansion - NOT to literal quoted strings. |
| 4390 | ! Check if the original input was entirely quoted with no expansions inside. |
| 4391 | if (is_quoted_literal(input)) then |
| 4392 | ! Skip field splitting for quoted literals |
| 4393 | expanded_words(1)%str = quote_removed |
| 4394 | word_count = 1 |
| 4395 | else |
| 4396 | ! Use temp buffer for word_split (expects character(len=*) array) |
| 4397 | allocate(character(len=max(1, len(quote_removed))) :: temp_split_words(size(expanded_words))) |
| 4398 | call word_split(shell, quote_removed, temp_split_words, word_count) |
| 4399 | do k = 1, word_count |
| 4400 | expanded_words(k)%str = trim(temp_split_words(k)) |
| 4401 | end do |
| 4402 | deallocate(temp_split_words) |
| 4403 | end if |
| 4404 | |
| 4405 | ! POSIX: If field splitting results in zero words (empty unquoted expansion), |
| 4406 | ! keep it as zero words - don't add back an empty field |
| 4407 | ! Note: This means unquoted empty variables disappear from the command line |
| 4408 | end subroutine |
| 4409 | |
| 4410 | ! Helper to grow an allocatable string buffer |
| 4411 | subroutine grow_string_buffer_exp(buffer, old_capacity, new_capacity, content_len) |
| 4412 | character(len=:), allocatable, intent(inout) :: buffer |
| 4413 | integer, intent(inout) :: old_capacity |
| 4414 | integer, intent(in) :: new_capacity |
| 4415 | integer, intent(in) :: content_len ! Actual used length of buffer |
| 4416 | character(len=:), allocatable :: temp |
| 4417 | |
| 4418 | ! Validate content_len |
| 4419 | if (content_len < 0 .or. content_len > old_capacity) then |
| 4420 | ! Invalid content length - this is a bug, but don't crash |
| 4421 | if (allocated(buffer)) deallocate(buffer) |
| 4422 | allocate(character(len=new_capacity) :: buffer) |
| 4423 | buffer = '' |
| 4424 | old_capacity = new_capacity |
| 4425 | return |
| 4426 | end if |
| 4427 | |
| 4428 | ! Allocate temp buffer and copy only actual content |
| 4429 | allocate(character(len=new_capacity) :: temp) |
| 4430 | temp = '' ! Initialize entire buffer to prevent heap corruption |
| 4431 | |
| 4432 | if (allocated(buffer) .and. content_len > 0) then |
| 4433 | ! Only copy the actual content (content_len bytes), not uninitialized data |
| 4434 | temp(1:content_len) = buffer(1:content_len) |
| 4435 | deallocate(buffer) |
| 4436 | end if |
| 4437 | |
| 4438 | ! Allocate new larger buffer |
| 4439 | allocate(character(len=new_capacity) :: buffer) |
| 4440 | buffer = '' ! Initialize entire buffer |
| 4441 | if (content_len > 0) then |
| 4442 | buffer(1:content_len) = temp(1:content_len) |
| 4443 | end if |
| 4444 | old_capacity = new_capacity |
| 4445 | |
| 4446 | deallocate(temp) |
| 4447 | end subroutine |
| 4448 | |
| 4449 | ! Parse arithmetic number with octal and hex support |
| 4450 | ! Returns value and sets iostat (0 = success, non-zero = error) |
| 4451 | function parse_arithmetic_number(str, iostat) result(value) |
| 4452 | character(len=*), intent(in) :: str |
| 4453 | integer, intent(out) :: iostat |
| 4454 | integer(kind=8) :: value |
| 4455 | integer :: i, len_str, digit |
| 4456 | character(len=256) :: trimmed_str |
| 4457 | |
| 4458 | value = 0 |
| 4459 | iostat = 0 |
| 4460 | trimmed_str = trim(adjustl(str)) |
| 4461 | len_str = len_trim(trimmed_str) |
| 4462 | |
| 4463 | if (len_str == 0) then |
| 4464 | iostat = 1 |
| 4465 | return |
| 4466 | end if |
| 4467 | |
| 4468 | ! Check for hexadecimal (0x or 0X) |
| 4469 | if (len_str >= 3 .and. trimmed_str(1:1) == '0' .and. & |
| 4470 | (trimmed_str(2:2) == 'x' .or. trimmed_str(2:2) == 'X')) then |
| 4471 | ! Parse hexadecimal |
| 4472 | do i = 3, len_str |
| 4473 | if (trimmed_str(i:i) >= '0' .and. trimmed_str(i:i) <= '9') then |
| 4474 | digit = ichar(trimmed_str(i:i)) - ichar('0') |
| 4475 | else if (trimmed_str(i:i) >= 'a' .and. trimmed_str(i:i) <= 'f') then |
| 4476 | digit = ichar(trimmed_str(i:i)) - ichar('a') + 10 |
| 4477 | else if (trimmed_str(i:i) >= 'A' .and. trimmed_str(i:i) <= 'F') then |
| 4478 | digit = ichar(trimmed_str(i:i)) - ichar('A') + 10 |
| 4479 | else |
| 4480 | iostat = 1 |
| 4481 | return |
| 4482 | end if |
| 4483 | value = value * 16 + digit |
| 4484 | end do |
| 4485 | return |
| 4486 | end if |
| 4487 | |
| 4488 | ! Check for octal (starts with 0 and has only 0-7 digits) |
| 4489 | if (len_str >= 2 .and. trimmed_str(1:1) == '0') then |
| 4490 | ! Verify all digits are 0-7 for octal |
| 4491 | do i = 2, len_str |
| 4492 | if (trimmed_str(i:i) < '0' .or. trimmed_str(i:i) > '7') then |
| 4493 | ! Not a valid octal, try decimal |
| 4494 | read(trimmed_str, *, iostat=iostat) value |
| 4495 | return |
| 4496 | end if |
| 4497 | end do |
| 4498 | ! Parse as octal |
| 4499 | do i = 1, len_str |
| 4500 | digit = ichar(trimmed_str(i:i)) - ichar('0') |
| 4501 | value = value * 8 + digit |
| 4502 | end do |
| 4503 | return |
| 4504 | end if |
| 4505 | |
| 4506 | ! Default: parse as decimal |
| 4507 | ! First verify the string contains only valid decimal characters (digits and optional leading +/-) |
| 4508 | i = 1 |
| 4509 | if (len_str > 0 .and. (trimmed_str(1:1) == '+' .or. trimmed_str(1:1) == '-')) then |
| 4510 | i = 2 |
| 4511 | end if |
| 4512 | do while (i <= len_str) |
| 4513 | if (trimmed_str(i:i) < '0' .or. trimmed_str(i:i) > '9') then |
| 4514 | iostat = 1 ! Not a valid decimal number |
| 4515 | return |
| 4516 | end if |
| 4517 | i = i + 1 |
| 4518 | end do |
| 4519 | read(trimmed_str, *, iostat=iostat) value |
| 4520 | end function |
| 4521 | |
| 4522 | end module expansion |