1 /* xbt_str.c - various helping functions to deal with strings */
3 /* Copyright (c) 2007-2014. The SimGrid Team.
4 * All rights reserved. */
6 /* This program is free software; you can redistribute it and/or modify it
7 * under the terms of the license (GNU LGPL) which comes with this package. */
11 #include "xbt/sysdep.h"
12 #include "xbt/str.h" /* headers of these functions */
13 #include "xbt/strbuff.h"
14 #include "xbt/matrix.h" /* for the diff */
16 /** @brief Strip whitespace (or other characters) from the end of a string.
18 * Strips the whitespaces from the end of s.
19 * By default (when char_list=NULL), these characters get stripped:
21 * - " " (ASCII 32 (0x20)) space.
22 * - "\t" (ASCII 9 (0x09)) tab.
23 * - "\n" (ASCII 10 (0x0A)) line feed.
24 * - "\r" (ASCII 13 (0x0D)) carriage return.
25 * - "\0" (ASCII 0 (0x00)) NULL.
26 * - "\x0B" (ASCII 11 (0x0B)) vertical tab.
28 * @param s The string to strip. Modified in place.
29 * @param char_list A string which contains the characters you want to strip.
32 void xbt_str_rtrim(char *s, const char *char_list)
35 const char *__char_list = " \t\n\r\x0B";
36 char white_char[256] = { 1, 0 };
42 while (*__char_list) {
43 white_char[(unsigned char) *__char_list++] = 1;
47 white_char[(unsigned char) *char_list++] = 1;
54 while ((cur >= s) && white_char[(unsigned char) *cur])
60 /** @brief Strip whitespace (or other characters) from the beginning of a string.
62 * Strips the whitespaces from the begining of s.
63 * By default (when char_list=NULL), these characters get stripped:
65 * - " " (ASCII 32 (0x20)) space.
66 * - "\t" (ASCII 9 (0x09)) tab.
67 * - "\n" (ASCII 10 (0x0A)) line feed.
68 * - "\r" (ASCII 13 (0x0D)) carriage return.
69 * - "\0" (ASCII 0 (0x00)) NULL.
70 * - "\x0B" (ASCII 11 (0x0B)) vertical tab.
72 * @param s The string to strip. Modified in place.
73 * @param char_list A string which contains the characters you want to strip.
76 void xbt_str_ltrim(char *s, const char *char_list)
79 const char *__char_list = " \t\n\r\x0B";
80 char white_char[256] = { 1, 0 };
86 while (*__char_list) {
87 white_char[(unsigned char) *__char_list++] = 1;
91 white_char[(unsigned char) *char_list++] = 1;
95 while (*cur && white_char[(unsigned char) *cur])
98 memmove(s, cur, strlen(cur) + 1);
101 /** @brief Strip whitespace (or other characters) from the end and the begining of a string.
103 * Strips the whitespaces from both the beginning and the end of s.
104 * By default (when char_list=NULL), these characters get stripped:
106 * - " " (ASCII 32 (0x20)) space.
107 * - "\t" (ASCII 9 (0x09)) tab.
108 * - "\n" (ASCII 10 (0x0A)) line feed.
109 * - "\r" (ASCII 13 (0x0D)) carriage return.
110 * - "\0" (ASCII 0 (0x00)) NULL.
111 * - "\x0B" (ASCII 11 (0x0B)) vertical tab.
113 * @param s The string to strip.
114 * @param char_list A string which contains the characters you want to strip.
117 void xbt_str_trim(char *s, const char *char_list)
123 xbt_str_rtrim(s, char_list);
124 xbt_str_ltrim(s, char_list);
127 /** @brief Replace double whitespaces (but no other characters) from the string.
129 * The function modifies the string so that each time that several spaces appear,
130 * they are replaced by a single space. It will only do so for spaces (ASCII 32, 0x20).
132 * @param s The string to strip. Modified in place.
135 void xbt_str_strip_spaces(char *s)
171 /** @brief Substitutes a char for another in a string
173 * @param str the string to modify
174 * @param from char to search
175 * @param to char to put instead
176 * @param occurence number of changes to do (=0 means all)
178 void xbt_str_subst(char *str, char from, char to, int occurence)
192 /** @brief Replaces a set of variables by their values
194 * @param str The input of the replacement process
195 * @param patterns The changes to apply
196 * @return The string modified
198 * Both '$toto' and '${toto}' are valid (and the two writing are equivalent).
200 * If the variable name contains spaces, use the brace version (ie, ${toto tutu})
202 * You can provide a default value to use if the variable is not set in the dict by using
203 * '${var:=default}' or '${var:-default}'. These two forms are equivalent, even if they
204 * shouldn't to respect the shell standard (:= form should set the value in the dict,
205 * but does not) (BUG).
208 char *xbt_str_varsubst(const char *str, xbt_dict_t patterns)
210 xbt_strbuff_t buff = xbt_strbuff_new_from(str);
212 xbt_strbuff_varsubst(buff, patterns);
214 xbt_strbuff_free_container(buff);
219 /** @brief Splits a string into a dynar of strings
221 * @param s: the string to split
222 * @param sep: a string of all chars to consider as separator.
224 * By default (with sep=NULL), these characters are used as separator:
226 * - " " (ASCII 32 (0x20)) space.
227 * - "\t" (ASCII 9 (0x09)) tab.
228 * - "\n" (ASCII 10 (0x0A)) line feed.
229 * - "\r" (ASCII 13 (0x0D)) carriage return.
230 * - "\0" (ASCII 0 (0x00)) NULL.
231 * - "\x0B" (ASCII 11 (0x0B)) vertical tab.
234 xbt_dynar_t xbt_str_split(const char *s, const char *sep)
236 xbt_dynar_t res = xbt_dynar_new(sizeof(char *), &xbt_free_ref);
239 const char *sep_dflt = " \t\n\r\x0B";
240 char is_sep[256] = { 1, 0 };
242 /* check what are the separators */
243 memset(is_sep, 0, sizeof(is_sep));
246 is_sep[(unsigned char) *sep_dflt++] = 1;
249 is_sep[(unsigned char) *sep++] = 1;
251 is_sep[0] = 1; /* End of string is also separator */
262 while (!is_sep[(unsigned char) *q]) {
268 topush = xbt_malloc(q - p + 1);
269 memcpy(topush, p, q - p);
270 topush[q - p] = '\0';
271 xbt_dynar_push(res, &topush);
279 * \brief This functions splits a string after using another string as separator
280 * For example A!!B!!C splitted after !! will return the dynar {A,B,C}
281 * \return An array of dynars containing the string tokens
283 xbt_dynar_t xbt_str_split_str(const char *s, const char *sep)
285 xbt_dynar_t res = xbt_dynar_new(sizeof(char *), &xbt_free_ref);
294 if (sep[0] == '\0') {
296 xbt_dynar_push(res, &s);
303 //get the start of the first occurence of the substring
305 //if substring was not found add the entire string
308 to_push = xbt_malloc(v + 1);
309 memcpy(to_push, p, v);
311 xbt_dynar_push(res, &to_push);
315 to_push = xbt_malloc(q - p + 1);
316 memcpy(to_push, p, q - p);
317 //add string terminator
318 to_push[q - p] = '\0';
319 xbt_dynar_push(res, &to_push);
326 /** @brief Just like @ref xbt_str_split_quoted (Splits a string into a dynar of strings), but without memory allocation
328 * The string passed as argument must be writable (not const)
329 * The elements of the dynar are just parts of the string passed as argument.
330 * So if you don't store that argument elsewhere, you should free it in addition
331 * to freeing the dynar. This can be done by simply freeing the first argument
333 * free(xbt_dynar_get_ptr(dynar,0));
335 * Actually this function puts a bunch of \0 in the memory area you passed as
336 * argument to separate the elements, and pushes the address of each chunk
337 * in the resulting dynar. Yes, that's uneven. Yes, that's gory. But that's efficient.
339 xbt_dynar_t xbt_str_split_quoted_in_place(char *s) {
340 xbt_dynar_t res = xbt_dynar_new(sizeof(char *), NULL);
341 char *beg, *end; /* pointers around the parsed chunk */
342 int in_simple_quote = 0, in_double_quote = 0;
344 int ctn = 0; /* Got something in this block */
351 /* do not trim leading spaces: caller responsibility to clean his cruft */
360 /* Protected char; move it closer */
361 memmove(end, end + 1, strlen(end));
363 THROWF(arg_error, 0, "String ends with \\");
364 end++; /* Pass the protected char */
369 if (!in_double_quote) {
370 in_simple_quote = !in_simple_quote;
371 memmove(end, end + 1, strlen(end));
373 /* simple quote protected by double ones */
379 if (!in_simple_quote) {
380 in_double_quote = !in_double_quote;
381 memmove(end, end + 1, strlen(end));
383 /* double quote protected by simple ones */
392 if (*end == '\0' && (in_simple_quote || in_double_quote)) {
394 "End of string found while searching for %c in %s",
395 (in_simple_quote ? '\'' : '"'), s);
397 if (in_simple_quote || in_double_quote) {
405 /* Found a separator. Push the string if contains something */
406 xbt_dynar_push(res, &beg);
414 /* trim within the string, manually to speed things up */
429 /** @brief Splits a string into a dynar of strings, taking quotes into account
431 * It basically does the same argument separation than the shell, where white
432 * spaces can be escaped and where arguments are never split within a
434 * Several subsequent spaces are ignored (unless within quotes, of course).
435 * You may want to trim the input string, if you want to avoid empty entries
439 xbt_dynar_t xbt_str_split_quoted(const char *s)
441 xbt_dynar_t res = xbt_dynar_new(sizeof(char *), &xbt_free_ref);
443 char *str_to_free; /* we have to copy the string before, to handle backslashes */
449 str_to_free = xbt_strdup(s);
451 parsed = xbt_str_split_quoted_in_place(str_to_free);
452 xbt_dynar_foreach(parsed,cursor,p) {
453 char *q=xbt_strdup(p);
454 xbt_dynar_push(res,&q);
457 xbt_dynar_shrink(res, 0);
458 xbt_dynar_free(&parsed);
462 /** @brief Join a set of strings as a single string */
463 char *xbt_str_join(xbt_dynar_t dyn, const char *sep)
465 int len = 1, dyn_len = xbt_dynar_length(dyn);
471 return xbt_strdup("");
473 /* compute the length */
474 xbt_dynar_foreach(dyn, cpt, cursor) {
475 len += strlen(cursor);
477 len += strlen(sep) * dyn_len;
479 res = xbt_malloc(len);
481 xbt_dynar_foreach(dyn, cpt, cursor) {
482 if ((int) cpt < dyn_len - 1)
483 p += sprintf(p, "%s%s", cursor, sep);
485 p += sprintf(p, "%s", cursor);
489 /** @brief Join a set of strings as a single string
491 * The parameter must be a NULL-terminated array of chars,
492 * just like xbt_dynar_to_array() produces
494 char *xbt_str_join_array(const char *const *strs, const char *sep)
497 int amount_strings=0;
501 if ((!strs) || (!strs[0]))
502 return xbt_strdup("");
504 /* compute the length before malloc */
505 for (i=0;strs[i];i++) {
506 len += strlen(strs[i]);
509 len += strlen(sep) * amount_strings;
512 q = res = xbt_malloc(len);
513 for (i=0;strs[i];i++) {
514 if (i!=0) { // not first loop
515 q += sprintf(q, "%s%s", sep, strs[i]);
517 q += sprintf(q,"%s",strs[i]);
523 /** @brief Get a single line from the stream (reimplementation of the GNU getline)
525 * This is a reimplementation of the GNU getline function, so that our code don't depends on the GNU libc.
527 * xbt_getline() reads an entire line from stream, storing the address of the
528 * buffer containing the text into *buf. The buffer is null-terminated and
529 * includes the newline character, if one was found.
531 * If *buf is NULL, then xbt_getline() will allocate a buffer for storing the
532 * line, which should be freed by the user program.
534 * Alternatively, before calling xbt_getline(), *buf can contain a pointer to a
535 * malloc()-allocated buffer *n bytes in size. If the buffer is not large
536 * enough to hold the line, xbt_getline() resizes it with realloc(), updating
537 * *buf and *n as necessary.
539 * In either case, on a successful call, *buf and *n will be updated to reflect
540 * the buffer address and allocated size respectively.
542 ssize_t xbt_getline(char **buf, size_t *n, FILE *stream)
548 if (ferror(stream) || feof(stream))
553 *buf = xbt_malloc(*n);
559 *buf = xbt_realloc(*buf, *n += 512);
561 } while (ch != '\n' && (ch = getc(stream)) != EOF);
564 *buf = xbt_realloc(*buf, *n += 1);
571 * Diff related functions
573 * Implementation of the algorithm described in "An O(NP) Sequence Comparison
574 * Algorithm", by Sun Wu, Udi Manber, Gene Myers, and Webb Miller (Information
575 * Processing Letters 35(6):317-323, 1990), with the linear-space
576 * divide-and-conquer strategy described in "An O(ND) Difference Algorithm and
577 * Its Variations", by Eugene W. Myers (Algorithmica 1:251-266, 1986).
581 int x, y; /* starting coordinates */
582 int len; /* length */
586 void diff_snake(const char *vec_a[], int a0, int len_a,
587 const char *vec_b[], int b0, int len_b,
588 struct subsequence *seqs, int *fp, int k, int limit)
592 int fp_left = fp[k - 1] + 1;
593 int fp_right = fp[k + 1];
594 if (fp_left > fp_right) {
602 if (x + y <= limit) {
607 seqs[k] = seqs[record_seq];
609 while (x < len_a && y < len_b && !strcmp(vec_a[a0 + x], vec_b[b0 + y]))
613 seqs[k].len = x - seqs[k].x;
616 /* Returns the length of a shortest edit script, and a common
617 * subsequence from the middle.
619 static int diff_middle_subsequence(const char *vec_a[], int a0, int len_a,
620 const char *vec_b[], int b0, int len_b,
621 struct subsequence *subseq,
622 struct subsequence *seqs, int *fp)
624 const int delta = len_a - len_b;
625 const int limit = (len_a + len_b) / 2;
638 for (k = kmin; k <= kmax; k++)
642 fp[kmin - p - 1] = fp[kmax + p + 1] = -1;
643 for (k = kmax + p; k > delta; k--)
644 diff_snake(vec_a, a0, len_a, vec_b, b0, len_b, seqs, fp, k, limit);
645 for (k = kmin - p; k <= delta; k++)
646 diff_snake(vec_a, a0, len_a, vec_b, b0, len_b, seqs, fp, k, limit);
647 } while (fp[delta] != len_a);
649 subseq->x = a0 + seqs[delta].x;
650 subseq->y = b0 + seqs[delta].y;
651 subseq->len = seqs[delta].len;
652 return abs(delta) + 2 * p;;
655 /* Finds a longest common subsequence.
656 * Returns its length.
658 static int diff_compute_lcs(const char *vec_a[], int a0, int len_a,
659 const char *vec_b[], int b0, int len_b,
660 xbt_dynar_t common_sequence,
661 struct subsequence *seqs, int *fp)
663 if (len_a > 0 && len_b > 0) {
664 struct subsequence subseq;
665 int ses_len = diff_middle_subsequence(vec_a, a0, len_a, vec_b, b0, len_b,
667 int lcs_len = (len_a + len_b - ses_len) / 2;
670 } else if (ses_len > 1) {
671 int lcs_len1 = subseq.len;
672 if (lcs_len1 < lcs_len)
673 lcs_len1 += diff_compute_lcs(vec_a, a0, subseq.x - a0,
674 vec_b, b0, subseq.y - b0,
675 common_sequence, seqs, fp);
677 xbt_dynar_push(common_sequence, &subseq);
678 if (lcs_len1 < lcs_len) {
679 int u = subseq.x + subseq.len;
680 int v = subseq.y + subseq.len;
681 diff_compute_lcs(vec_a, u, a0 + len_a - u, vec_b, v, b0 + len_b - v,
682 common_sequence, seqs, fp);
685 int len = MIN(len_a, len_b) - subseq.len;
686 if (subseq.x == a0 && subseq.y == b0) {
688 xbt_dynar_push(common_sequence, &subseq);
690 struct subsequence subseq0 = {a0 + len_a - len,
691 b0 + len_b - len, len};
692 xbt_dynar_push(common_sequence, &subseq0);
696 struct subsequence subseq0 = {a0, b0, len};
697 xbt_dynar_push(common_sequence, &subseq0);
700 xbt_dynar_push(common_sequence, &subseq);
709 static int diff_member(const char *s, const char *vec[], int from, int to)
711 for ( ; from < to ; from++)
712 if (!strcmp(s, vec[from]))
717 /* Extract common prefix.
719 static void diff_easy_prefix(const char *vec_a[], int *a0_p, int *len_a_p,
720 const char *vec_b[], int *b0_p, int *len_b_p,
721 xbt_dynar_t common_sequence)
725 int len_a = *len_a_p;
726 int len_b = *len_b_p;
728 while (len_a > 0 && len_b > 0) {
729 struct subsequence subseq = {a0, b0, 0};
730 while (len_a > 0 && len_b > 0 && !strcmp(vec_a[a0], vec_b[b0])) {
736 xbt_dynar_push(common_sequence, &subseq);
737 if (len_a > 0 && len_b > 0 &&
738 !diff_member(vec_a[a0], vec_b, b0 + 1, b0 + len_b)) {
751 /* Extract common suffix.
753 static void diff_easy_suffix(const char *vec_a[], int *a0_p, int *len_a_p,
754 const char *vec_b[], int *b0_p, int *len_b_p,
755 xbt_dynar_t common_suffix)
759 int len_a = *len_a_p;
760 int len_b = *len_b_p;
762 while (len_a > 0 && len_b > 0){
763 struct subsequence subseq;
765 while (len_a > 0 && len_b > 0 &&
766 !strcmp(vec_a[a0 + len_a - 1], vec_b[b0 + len_b - 1])) {
771 if (subseq.len > 0) {
772 subseq.x = a0 + len_a;
773 subseq.y = b0 + len_b;
774 xbt_dynar_push(common_suffix, &subseq);
776 if (len_a > 0 && len_b > 0 &&
777 !diff_member(vec_b[b0 + len_b - 1], vec_a, a0, a0 + len_a - 1)) {
790 /** @brief Compute the unified diff of two strings */
791 char *xbt_str_diff(const char *a, const char *b)
793 xbt_dynar_t da = xbt_str_split(a, "\n");
794 xbt_dynar_t db = xbt_str_split(b, "\n");
795 xbt_dynar_t common_sequence, common_suffix;
797 const char **vec_a, **vec_b;
802 struct subsequence *seqs_base, *seqs;
803 struct subsequence subseq;
809 /* Clean empty lines at the end of da and db */
811 if (len > 0 && a[len - 1] == '\n')
812 xbt_dynar_pop(da, NULL);
814 if (len > 0 && b[len - 1] == '\n')
815 xbt_dynar_pop(db, NULL);
817 /* Various initializations */
818 /* Assume that dynar's content is contiguous */
820 len_a = xbt_dynar_length(da);
821 vec_a = len_a ? xbt_dynar_get_ptr(da, 0) : NULL;
823 len_b = xbt_dynar_length(db);
824 vec_b = len_b ? xbt_dynar_get_ptr(db, 0) : NULL;
825 max = MAX(len_a, len_b) + 1;
826 fp_base = xbt_new(int, 2 * max + 1);
827 fp = fp_base + max; /* indexes in [-max..max] */
828 seqs_base = xbt_new(struct subsequence, 2 * max + 1);
829 seqs = seqs_base + max; /* indexes in [-max..max] */
830 common_sequence = xbt_dynar_new(sizeof(struct subsequence), NULL);
831 common_suffix = xbt_dynar_new(sizeof(struct subsequence), NULL);
833 /* Add a sentinel a the end of the sequence */
837 xbt_dynar_push(common_suffix, &subseq);
839 /* Compute the Longest Common Subsequence */
840 diff_easy_prefix(vec_a, &a0, &len_a, vec_b, &b0, &len_b, common_sequence);
841 diff_easy_suffix(vec_a, &a0, &len_a, vec_b, &b0, &len_b, common_suffix);
842 diff_compute_lcs(vec_a, a0, len_a, vec_b, b0, len_b, common_sequence, seqs, fp);
843 while (!xbt_dynar_is_empty(common_suffix)) {
844 xbt_dynar_pop(common_suffix, &subseq);
845 xbt_dynar_push(common_sequence, &subseq);
848 /* Build a Shortest Edit Script, and the final result */
849 diff = xbt_dynar_new(sizeof(char *), &xbt_free_ref);
852 xbt_dynar_foreach(common_sequence, s, subseq) {
853 while (x < subseq.x) {
854 char *topush = bprintf("- %s", vec_a[x++]);
855 xbt_dynar_push_as(diff, char*, topush);
857 while (y < subseq.y) {
858 char *topush = bprintf("+ %s", vec_b[y++]);
859 xbt_dynar_push_as(diff, char*, topush);
861 while (x < subseq.x + subseq.len) {
862 char *topush = bprintf(" %s", vec_a[x++]);
863 xbt_dynar_push_as(diff, char*, topush);
867 res = xbt_str_join(diff, "\n");
873 xbt_dynar_free(&common_sequence);
874 xbt_dynar_free(&common_suffix);
875 xbt_dynar_free(&diff);
881 /** @brief creates a new string containing what can be read on a fd
884 char *xbt_str_from_file(FILE * file)
886 xbt_strbuff_t buff = xbt_strbuff_new();
889 memset(bread, 0, 1024);
891 while (!feof(file)) {
892 int got = fread(bread, 1, 1023, file);
894 xbt_strbuff_append(buff, bread);
898 xbt_strbuff_free_container(buff);
902 /* @brief Retrun 1 if string 'str' starts with string 'start'
904 * \param str a string
905 * \param start the string to search in str
907 * \return 1 if 'str' starts with 'start'
909 int xbt_str_start_with(const char* str, const char* start)
912 size_t l_str = strlen(str);
913 size_t l_start = strlen(start);
915 if(l_start > l_str) return 0;
917 for(i = 0; i< l_start; i++){
918 if(str[i] != start[i]) return 0;
927 #define mytest(name, input, expected) \
928 xbt_test_add(name); \
929 d=xbt_str_split_quoted(input); \
930 s=xbt_str_join(d,"XXX"); \
931 xbt_test_assert(!strcmp(s,expected),\
932 "Input (%s) leads to (%s) instead of (%s)", \
937 XBT_TEST_SUITE("xbt_str", "String Handling");
938 XBT_TEST_UNIT("xbt_str_split_quoted", test_split_quoted, "test the function xbt_str_split_quoted")
943 mytest("Empty", "", "");
944 mytest("Basic test", "toto tutu", "totoXXXtutu");
945 mytest("Useless backslashes", "\\t\\o\\t\\o \\t\\u\\t\\u",
947 mytest("Protected space", "toto\\ tutu", "toto tutu");
948 mytest("Several spaces", "toto tutu", "totoXXXtutu");
949 mytest("LTriming", " toto tatu", "totoXXXtatu");
950 mytest("Triming", " toto tutu ", "totoXXXtutu");
951 mytest("Single quotes", "'toto tutu' tata", "toto tutuXXXtata");
952 mytest("Double quotes", "\"toto tutu\" tata", "toto tutuXXXtata");
953 mytest("Mixed quotes", "\"toto' 'tutu\" tata", "toto' 'tutuXXXtata");
954 mytest("Backslashed quotes", "\\'toto tutu\\' tata",
955 "'totoXXXtutu'XXXtata");
956 mytest("Backslashed quotes + quotes", "'toto \\'tutu' tata",
957 "toto 'tutuXXXtata");
961 #define mytest_str(name, input, separator, expected) \
962 xbt_test_add(name); \
963 d=xbt_str_split_str(input, separator); \
964 s=xbt_str_join(d,"XXX"); \
965 xbt_test_assert(!strcmp(s,expected),\
966 "Input (%s) leads to (%s) instead of (%s)", \
971 XBT_TEST_UNIT("xbt_str_split_str", test_split_str, "test the function xbt_str_split_str")
976 mytest_str("Empty string and separator", "", "", "");
977 mytest_str("Empty string", "", "##", "");
978 mytest_str("Empty separator", "toto", "", "toto");
979 mytest_str("String with no separator in it", "toto", "##", "toto");
980 mytest_str("Basic test", "toto##tutu", "##", "totoXXXtutu");
983 /* Last args are format string and parameters for xbt_test_add */
984 #define mytest_diff(s1, s2, diff, ...) \
986 char *mytest_diff_res; \
987 xbt_test_add(__VA_ARGS__); \
988 mytest_diff_res = xbt_str_diff(s1, s2); \
989 xbt_test_assert(!strcmp(mytest_diff_res, diff), \
990 "Wrong output:\n--- got:\n%s\n--- expected:\n%s\n---", \
991 mytest_diff_res, diff); \
992 free(mytest_diff_res); \
995 XBT_TEST_UNIT("xbt_str_diff", test_diff, "test the function xbt_str_diff")
1000 mytest_diff("a", "a", " a", "1 word, no difference");
1001 mytest_diff("a", "A", "- a\n+ A", "1 word, different");
1002 mytest_diff("a\n", "a\n", " a", "1 line, no difference");
1003 mytest_diff("a\n", "A\n", "- a\n+ A", "1 line, different");
1006 mytest_diff("", "", "", "empty strings");
1007 mytest_diff("", "a", "+ a", "1 word, added");
1008 mytest_diff("a", "", "- a", "1 word, removed");
1009 mytest_diff("", "a\n", "+ a", "1 line, added");
1010 mytest_diff("a\n", "", "- a", "1 line, removed");
1011 mytest_diff("", "a\nb\nc\n", "+ a\n+ b\n+ c", "4 lines, all added");
1012 mytest_diff("a\nb\nc\n", "", "- a\n- b\n- c", "4 lines, all removed");
1015 mytest_diff("\n", "\n", " ", "empty lines");
1016 mytest_diff("", "\n", "+ ", "empty line, added");
1017 mytest_diff("\n", "", "- ", "empty line, removed");
1019 mytest_diff("a", "\na", "+ \n a", "empty line added before word");
1020 mytest_diff("a", "a\n\n", " a\n+ ", "empty line added after word");
1021 mytest_diff("\na", "a", "- \n a", "empty line removed before word");
1022 mytest_diff("a\n\n", "a", " a\n- ", "empty line removed after word");
1024 mytest_diff("a\n", "\na\n", "+ \n a", "empty line added before line");
1025 mytest_diff("a\n", "a\n\n", " a\n+ ", "empty line added after line");
1026 mytest_diff("\na\n", "a\n", "- \n a", "empty line removed before line");
1027 mytest_diff("a\n\n", "a\n", " a\n- ", "empty line removed after line");
1029 mytest_diff("a\nb\nc\nd\n", "\na\nb\nc\nd\n", "+ \n a\n b\n c\n d",
1030 "empty line added before 4 lines");
1031 mytest_diff("a\nb\nc\nd\n", "a\nb\nc\nd\n\n", " a\n b\n c\n d\n+ ",
1032 "empty line added after 4 lines");
1033 mytest_diff("\na\nb\nc\nd\n", "a\nb\nc\nd\n", "- \n a\n b\n c\n d",
1034 "empty line removed before 4 lines");
1035 mytest_diff("a\nb\nc\nd\n\n", "a\nb\nc\nd\n", " a\n b\n c\n d\n- ",
1036 "empty line removed after 4 lines");
1038 /* Missing newline at the end of one of the strings */
1039 mytest_diff("a\n", "a", " a", "1 line, 1 word, no difference");
1040 mytest_diff("a", "a\n", " a", "1 word, 1 line, no difference");
1041 mytest_diff("a\n", "A", "- a\n+ A", "1 line, 1 word, different");
1042 mytest_diff("a", "A\n", "- a\n+ A", "1 word, 1 line, different");
1044 mytest_diff("a\nb\nc\nd", "a\nb\nc\nd\n", " a\n b\n c\n d",
1045 "4 lines, no newline on first");
1046 mytest_diff("a\nb\nc\nd\n", "a\nb\nc\nd", " a\n b\n c\n d",
1047 "4 lines, no newline on second");
1049 /* Four lines, all combinations of differences */
1050 for (i = 0 ; i < (1U << 4) ; i++) {
1053 char res[4 * 8 + 1];
1060 for (/* j */ ; j < 4 && !(i & (1U << j)) ; j++) {
1062 ps += sprintf(ps, "%c\n", "abcd"[j]);
1063 pr += sprintf(pr, " %c\n", "abcd"[j]);
1065 for (k = j ; k < 4 && (i & (1U << k)) ; k++) {
1067 ps += sprintf(ps, "%c\n", "ABCD"[k]);
1068 pr += sprintf(pr, "- %c\n", "abcd"[k]);
1070 for (/* j */ ; j < k ; j++) {
1071 pr += sprintf(pr, "+ %c\n", "ABCD"[j]);
1075 *--pr = '\0'; /* strip last '\n' from expected result */
1076 mytest_diff("a\nb\nc\nd\n", s2, res,
1077 "compare (abcd) with changed (%s)", d2);
1080 /* Subsets of four lines, do not test for empty subset */
1081 for (i = 1 ; i < (1U << 4) ; i++) {
1084 char res[4 * 8 + 1];
1090 for (/* j */ ; j < 4 && (i & (1U << j)) ; j++) {
1092 ps += sprintf(ps, "%c\n", "abcd"[j]);
1093 pr += sprintf(pr, " %c\n", "abcd"[j]);
1095 for (/* j */; j < 4 && !(i & (1U << j)) ; j++) {
1096 pr += sprintf(pr, "- %c\n", "abcd"[j]);
1100 *--pr = '\0'; /* strip last '\n' from expected result */
1101 mytest_diff("a\nb\nc\nd\n", s2, res,
1102 "compare (abcd) with subset (%s)", d2);
1104 for (pr = res ; *pr != '\0' ; pr++)
1107 mytest_diff(s2, "a\nb\nc\nd\n", res,
1108 "compare subset (%s) with (abcd)", d2);
1112 #endif /* SIMGRID_TEST */