--- cut.c 2012-12-28 15:19:57.345985225 -0500 +++ src/cut.c 2012-12-28 16:55:51.906062821 -0500 @@ -45,6 +45,22 @@ proper_name ("David MacKenzie"), \ proper_name ("Jim Meyering") +/* When initializing an array to hold each line when using --preserve-order, + there must be a sentinel value that signifies that the next character added + to the array will be the first character in that field. */ +#define CELL_SENTINEL '\0' + +/* This should be analogous to the problem of searching for an object by + starting at the origin and walking either toward infinity or toward negative + infinity. No matter how far you walk the first time, when you turn around, + you should walk to the other side of the origin twice as far as you went + in the first direction. */ +#define ARRAY_SIZE_MULTIPLIER 2 + +/* FIXME It's not clear what value to use for the initial array size when using + --preserve-order. */ +#define INITIAL_ARRAY_SIZE 10 + #define FATAL_ERROR(Message) \ do \ { \ @@ -148,6 +164,10 @@ /* True if we have ever read standard input. */ static bool have_read_stdin; +/* Keep track of of the order of fields when using --preserve-order. */ +static size_t *field_order; +static bool preserve_field_order; + #define HT_RANGE_START_INDEX_INITIAL_CAPACITY 31 /* The set of range-start indices. For example, given a range-spec list like @@ -213,6 +233,10 @@ or fields\n\ "), stdout); fputs (_("\ + -p, --preserve-order preserve the given field order such that,\n\ + e.g., -fi,j and -fj,i, give different results\n\ +"), stdout); + fputs (_("\ -s, --only-delimited do not print lines not containing delimiters\n\ --output-delimiter=STRING use STRING as the output delimiter\n\ the default is to use the input delimiter\n\ @@ -504,6 +528,33 @@ printable_field = xzalloc (max_range_endpoint / CHAR_BIT + 1); + /* Allocate an array to keep track of the fields necessary when using + --preserve-order. */ + field_order = (size_t *) xmalloc ((max_range_endpoint + 1) * sizeof (size_t)); + memset (field_order, 0, (max_range_endpoint + 1) * sizeof (size_t)); + + /* pos will help key into field_order when using --preserve-order. */ + size_t pos = 0; + + /* Set the field_order when using --preserve-order. This must happen before + calling qsort on the ranges. */ + if (preserve_field_order) { + for (i = 0; i < n_rp; i++) { + size_t j = rp[i].lo; + + if (j <= rp[i].hi && !is_printable_field(j)) { + /* Key into field_order. This is the start of the range. */ + ++pos; + field_order[pos] = j; + } + for (++j; j <= rp[i].hi; j++) { + /* Since this is a range, we look at the previous cell and increment. */ + ++pos; + field_order[pos] = field_order[j-1] + 1; + } + } + } + qsort (rp, n_rp, sizeof (rp[0]), compare_ranges); /* Set the array entries corresponding to integers in the ranges of RP. */ @@ -516,11 +567,12 @@ index that is not part of any other (lo..hi] range. */ rsi_candidate = complement ? rp[i].hi + 1 : rp[i].lo; if (output_delimiter_specified - && !is_printable_field (rsi_candidate)) + && !is_printable_field (rsi_candidate)) { mark_range_start (rsi_candidate); - - for (j = rp[i].lo; j <= rp[i].hi; j++) + } + for (j = rp[i].lo; j <= rp[i].hi; j++) { mark_printable_field (j); + } } if (output_delimiter_specified @@ -533,6 +585,30 @@ return field_found; } +/* Initialize an array that will hold the printable fields when using + --preserve-order with CELL_SENTINEL. */ +static void +initialize_array(char *** array, size_t prev_arraysize, size_t new_arraysize) +{ + size_t print_i; + for (print_i = prev_arraysize; print_i < new_arraysize; print_i++) { + char ** a = 0; + a = xmalloc( INITIAL_ARRAY_SIZE * sizeof(char **) ); + a[0] = (char *) CELL_SENTINEL; + array[print_i] = a; + } +} + +/* Initialize an array that will hold the printable fields when using + --preserve-order. */ +static char *** +alloc_array(char *** array, size_t size) +{ + array = xmalloc( size * sizeof(char **) ); + initialize_array(array, 0, size); + return array; +} + /* Read from stream STREAM, printing to standard output any selected bytes. */ static void @@ -605,6 +681,14 @@ That is because a non-delimited line has exactly one field. */ buffer_first_field = (suppress_non_delimited ^ !print_kth (1, NULL)); + /* Keep track of the size of the array that we need and initialize the array + for --preserve-order. */ + size_t arraysize = INITIAL_ARRAY_SIZE; + char *** array = 0; + if (preserve_field_order) { + array = alloc_array(array, arraysize); + } + while (1) { if (field_idx == 1 && buffer_first_field) @@ -659,14 +743,46 @@ { if (found_any_selected_field) { + /* Write out the delimiter for --preserve-order. */ + if (!preserve_field_order) { fwrite (output_delimiter_string, sizeof (char), output_delimiter_length, stdout); + } } found_any_selected_field = true; while ((c = getc (stream)) != delim && c != '\n' && c != EOF) { - putchar (c); + if (preserve_field_order) { + /* Store the printable fields when using --preserve-order. */ + size_t output_field_idx = 0; + while (1) { + if (field_order[output_field_idx] == field_idx) { + break; + } + ++output_field_idx; + } + if ((output_field_idx + 1) >= arraysize) { + size_t prev_arraysize = arraysize; + while (arraysize < output_field_idx + 1) { + arraysize *= ARRAY_SIZE_MULTIPLIER; + } + array = realloc(array, arraysize * sizeof(char **)); + initialize_array(array, prev_arraysize, arraysize); + } + if (array[output_field_idx][0] == CELL_SENTINEL) { + array[output_field_idx][0] = (char *) c; + } else { + char tmp[2]; + tmp[0] = c; + tmp[1] = '\0'; + //strcat(&array[output_field_idx][0], &tmp); + strcat(array[output_field_idx], &tmp); + } + } else { + /* Print the current char when not using --preserve-order. */ + putchar (c); + } } } else @@ -680,6 +796,24 @@ if (c == '\n') { + /* Print the printable fields when using --preserve-order. */ + if (preserve_field_order) { + bool did_first = false; + size_t print_j = 0; + while (print_j < arraysize) { + if (print_kth(field_order[print_j], NULL)) { + if (did_first) { + fwrite (output_delimiter_string, sizeof (char), + output_delimiter_length, stdout); + } + printf (_("%s"), array[print_j]); + did_first = true; + } + ++print_j; + } + free (array); + array = alloc_array(array, arraysize); + } c = getc (stream); if (c != EOF) { @@ -778,7 +912,7 @@ delim = '\0'; have_read_stdin = false; - while ((optc = getopt_long (argc, argv, "b:c:d:f:ns", longopts, NULL)) != -1) + while ((optc = getopt_long (argc, argv, "b:c:d:f:nps", longopts, NULL)) != -1) { switch (optc) { @@ -799,6 +933,11 @@ spec_list_string = optarg; break; + case 'p': + /* Preserve field order. */ + preserve_field_order = true; + break; + case 'd': /* New delimiter. */ /* Interpret -d '' to mean 'use the NUL byte as the delimiter.' */ @@ -866,6 +1005,10 @@ FATAL_ERROR (_("missing list of positions")); } + if (complement && preserve_field_order) { + FATAL_ERROR (_("--complement is incompatible with --preserve-order")); + } + if (!delim_specified) delim = '\t';