1 /* xbt_str.cpp - 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 */
14 /** @brief Strip whitespace (or other characters) from the end of a string.
16 * Strips the whitespaces from the end of s.
17 * By default (when char_list=nullptr), these characters get stripped:
19 * - " " (ASCII 32 (0x20)) space.
20 * - "\t" (ASCII 9 (0x09)) tab.
21 * - "\n" (ASCII 10 (0x0A)) line feed.
22 * - "\r" (ASCII 13 (0x0D)) carriage return.
23 * - "\0" (ASCII 0 (0x00)) nullptr.
24 * - "\x0B" (ASCII 11 (0x0B)) vertical tab.
26 * @param s The string to strip. Modified in place.
27 * @param char_list A string which contains the characters you want to strip.
29 void xbt_str_rtrim(char *s, const char *char_list)
32 const char *__char_list = " \t\n\r\x0B";
33 char white_char[256] = { 1, 0 };
39 while (*__char_list) {
40 white_char[(unsigned char) *__char_list++] = 1;
44 white_char[(unsigned char) *char_list++] = 1;
51 while ((cur >= s) && white_char[(unsigned char) *cur])
57 /** @brief Strip whitespace (or other characters) from the beginning of a string.
59 * Strips the whitespaces from the beginning of s.
60 * By default (when char_list=nullptr), these characters get stripped:
62 * - " " (ASCII 32 (0x20)) space.
63 * - "\t" (ASCII 9 (0x09)) tab.
64 * - "\n" (ASCII 10 (0x0A)) line feed.
65 * - "\r" (ASCII 13 (0x0D)) carriage return.
66 * - "\0" (ASCII 0 (0x00)) nullptr.
67 * - "\x0B" (ASCII 11 (0x0B)) vertical tab.
69 * @param s The string to strip. Modified in place.
70 * @param char_list A string which contains the characters you want to strip.
72 void xbt_str_ltrim(char *s, const char *char_list)
75 const char *__char_list = " \t\n\r\x0B";
76 char white_char[256] = { 1, 0 };
82 while (*__char_list) {
83 white_char[(unsigned char) *__char_list++] = 1;
87 white_char[(unsigned char) *char_list++] = 1;
91 while (*cur && white_char[(unsigned char) *cur])
94 memmove(s, cur, strlen(cur) + 1);
97 /** @brief Strip whitespace (or other characters) from the end and the beginning of a string.
99 * Strips the whitespaces from both the beginning and the end of s.
100 * By default (when char_list=nullptr), these characters get stripped:
102 * - " " (ASCII 32 (0x20)) space.
103 * - "\t" (ASCII 9 (0x09)) tab.
104 * - "\n" (ASCII 10 (0x0A)) line feed.
105 * - "\r" (ASCII 13 (0x0D)) carriage return.
106 * - "\0" (ASCII 0 (0x00)) nullptr.
107 * - "\x0B" (ASCII 11 (0x0B)) vertical tab.
109 * @param s The string to strip.
110 * @param char_list A string which contains the characters you want to strip.
112 void xbt_str_trim(char *s, const char *char_list)
117 xbt_str_rtrim(s, char_list);
118 xbt_str_ltrim(s, char_list);
121 /** @brief Substitutes a char for another in a string
123 * @param str the string to modify
124 * @param from char to search
125 * @param to char to put instead
126 * @param occurrence number of changes to do (=0 means all)
128 void xbt_str_subst(char *str, char from, char to, int occurence)
142 /** @brief Splits a string into a dynar of strings
144 * @param s: the string to split
145 * @param sep: a string of all chars to consider as separator.
147 * By default (with sep=nullptr), these characters are used as separator:
149 * - " " (ASCII 32 (0x20)) space.
150 * - "\t" (ASCII 9 (0x09)) tab.
151 * - "\n" (ASCII 10 (0x0A)) line feed.
152 * - "\r" (ASCII 13 (0x0D)) carriage return.
153 * - "\0" (ASCII 0 (0x00)) nullptr.
154 * - "\x0B" (ASCII 11 (0x0B)) vertical tab.
156 xbt_dynar_t xbt_str_split(const char *s, const char *sep)
158 xbt_dynar_t res = xbt_dynar_new(sizeof(char *), &xbt_free_ref);
159 const char *sep_dflt = " \t\n\r\x0B";
160 char is_sep[256] = { 1, 0 };
162 /* check what are the separators */
163 memset(is_sep, 0, sizeof(is_sep));
166 is_sep[(unsigned char) *sep_dflt++] = 1;
169 is_sep[(unsigned char) *sep++] = 1;
171 is_sep[0] = 1; /* End of string is also separator */
183 while (not is_sep[(unsigned char)*q]) {
189 topush = (char*) xbt_malloc(q - p + 1);
190 memcpy(topush, p, q - p);
191 topush[q - p] = '\0';
192 xbt_dynar_push(res, &topush);
200 * \brief This functions splits a string after using another string as separator
201 * For example Anot not B!not C split after !! will return the dynar {A,B,C}
202 * \return An array of dynars containing the string tokens
204 xbt_dynar_t xbt_str_split_str(const char *s, const char *sep)
206 xbt_dynar_t res = xbt_dynar_new(sizeof(char *), &xbt_free_ref);
214 if (sep[0] == '\0') {
216 xbt_dynar_push(res, &s);
222 // get the start of the first occurrence of the substring
224 //if substring was not found add the entire string
227 to_push = (char*) xbt_malloc(v + 1);
228 memcpy(to_push, p, v);
230 xbt_dynar_push(res, &to_push);
234 to_push = (char*) xbt_malloc(q - p + 1);
235 memcpy(to_push, p, q - p);
236 //add string terminator
237 to_push[q - p] = '\0';
238 xbt_dynar_push(res, &to_push);
245 /** @brief Just like @ref xbt_str_split_quoted (Splits a string into a dynar of strings), but without memory allocation
247 * The string passed as argument must be writable (not const)
248 * The elements of the dynar are just parts of the string passed as argument.
249 * So if you don't store that argument elsewhere, you should free it in addition to freeing the dynar. This can be done
250 * by simply freeing the first argument of the dynar:
251 * free(xbt_dynar_get_ptr(dynar,0));
253 * Actually this function puts a bunch of \0 in the memory area you passed as argument to separate the elements, and
254 * pushes the address of each chunk in the resulting dynar. Yes, that's uneven. Yes, that's gory. But that's efficient.
256 xbt_dynar_t xbt_str_split_quoted_in_place(char *s) {
257 xbt_dynar_t res = xbt_dynar_new(sizeof(char *), nullptr);
259 char* end; /* pointers around the parsed chunk */
260 int in_simple_quote = 0;
261 int in_double_quote = 0;
263 int ctn = 0; /* Got something in this block */
270 /* do not trim leading spaces: caller responsibility to clean his cruft */
277 /* Protected char; move it closer */
278 memmove(end, end + 1, strlen(end));
280 THROWF(arg_error, 0, "String ends with \\");
281 end++; /* Pass the protected char */
285 if (not in_double_quote) {
286 in_simple_quote = not in_simple_quote;
287 memmove(end, end + 1, strlen(end));
289 /* simple quote protected by double ones */
295 if (not in_simple_quote) {
296 in_double_quote = not in_double_quote;
297 memmove(end, end + 1, strlen(end));
299 /* double quote protected by simple ones */
307 if (*end == '\0' && (in_simple_quote || in_double_quote)) {
308 THROWF(arg_error, 0, "End of string found while searching for %c in %s", (in_simple_quote ? '\'' : '"'), s);
310 if (in_simple_quote || in_double_quote) {
318 /* Found a separator. Push the string if contains something */
319 xbt_dynar_push(res, &beg);
327 /* trim within the string, manually to speed things up */
341 /** @brief Splits a string into a dynar of strings, taking quotes into account
343 * It basically does the same argument separation than the shell, where white spaces can be escaped and where arguments
344 * are never split within a quote group.
345 * Several subsequent spaces are ignored (unless within quotes, of course).
346 * You may want to trim the input string, if you want to avoid empty entries
348 xbt_dynar_t xbt_str_split_quoted(const char *s)
350 xbt_dynar_t res = xbt_dynar_new(sizeof(char *), &xbt_free_ref);
352 char *str_to_free; /* we have to copy the string before, to handle backslashes */
358 str_to_free = xbt_strdup(s);
360 parsed = xbt_str_split_quoted_in_place(str_to_free);
361 xbt_dynar_foreach(parsed,cursor,p) {
362 char *q=xbt_strdup(p);
363 xbt_dynar_push(res,&q);
366 xbt_dynar_shrink(res, 0);
367 xbt_dynar_free(&parsed);
371 /** @brief Join a set of strings as a single string */
372 char *xbt_str_join(xbt_dynar_t dyn, const char *sep)
375 int dyn_len = xbt_dynar_length(dyn);
380 return xbt_strdup("");
382 /* compute the length */
383 xbt_dynar_foreach(dyn, cpt, cursor) {
384 len += strlen(cursor);
386 len += strlen(sep) * dyn_len;
388 char* res = (char*)xbt_malloc(len);
390 xbt_dynar_foreach(dyn, cpt, cursor) {
391 if ((int) cpt < dyn_len - 1)
392 p += snprintf(p,len, "%s%s", cursor, sep);
394 p += snprintf(p,len, "%s", cursor);
399 /** @brief Join a set of strings as a single string
401 * The parameter must be a nullptr-terminated array of chars,
402 * just like xbt_dynar_to_array() produces
404 char *xbt_str_join_array(const char *const *strs, const char *sep)
406 int amount_strings=0;
409 if ((not strs) || (not strs[0]))
410 return xbt_strdup("");
412 /* compute the length before malloc */
413 for (int i = 0; strs[i]; i++) {
414 len += strlen(strs[i]);
417 len += strlen(sep) * amount_strings;
420 char* res = (char*)xbt_malloc(len);
422 for (int i = 0; strs[i]; i++) {
423 if (i != 0) { // not first loop
424 q += snprintf(q,len, "%s%s", sep, strs[i]);
426 q += snprintf(q,len, "%s",strs[i]);
432 /** @brief Parse an integer out of a string, or raise an error
434 * The @a str is passed as argument to your @a error_msg, as follows:
435 * @verbatim THROWF(arg_error, 0, error_msg, str); @endverbatim
437 long int xbt_str_parse_int(const char* str, const char* error_msg)
440 if (str == nullptr || str[0] == '\0')
441 THROWF(arg_error, 0, error_msg, str);
443 long int res = strtol(str, &endptr, 10);
444 if (endptr[0] != '\0')
445 THROWF(arg_error, 0, error_msg, str);
450 /** @brief Parse a double out of a string, or raise an error
452 * The @a str is passed as argument to your @a error_msg, as follows:
453 * @verbatim THROWF(arg_error, 0, error_msg, str); @endverbatim
455 double xbt_str_parse_double(const char* str, const char* error_msg)
458 if (str == nullptr || str[0] == '\0')
459 THROWF(arg_error, 0, error_msg, str);
461 double res = strtod(str, &endptr);
462 if (endptr[0] != '\0')
463 THROWF(arg_error, 0, error_msg, str);
469 #include <xbt/ex.hpp>
472 XBT_TEST_SUITE("xbt_str", "String Handling");
474 #define mytest(name, input, expected) \
475 xbt_test_add(name); \
476 d = xbt_str_split_quoted(input); \
477 s = xbt_str_join(d, "XXX"); \
478 xbt_test_assert(not strcmp(s, expected), "Input (%s) leads to (%s) instead of (%s)", input, s, expected); \
481 XBT_TEST_UNIT("xbt_str_split_quoted", test_split_quoted, "test the function xbt_str_split_quoted")
486 mytest("Empty", "", "");
487 mytest("Basic test", "toto tutu", "totoXXXtutu");
488 mytest("Useless backslashes", "\\t\\o\\t\\o \\t\\u\\t\\u", "totoXXXtutu");
489 mytest("Protected space", "toto\\ tutu", "toto tutu");
490 mytest("Several spaces", "toto tutu", "totoXXXtutu");
491 mytest("LTriming", " toto tatu", "totoXXXtatu");
492 mytest("Triming", " toto tutu ", "totoXXXtutu");
493 mytest("Single quotes", "'toto tutu' tata", "toto tutuXXXtata");
494 mytest("Double quotes", "\"toto tutu\" tata", "toto tutuXXXtata");
495 mytest("Mixed quotes", "\"toto' 'tutu\" tata", "toto' 'tutuXXXtata");
496 mytest("Backslashed quotes", "\\'toto tutu\\' tata", "'totoXXXtutu'XXXtata");
497 mytest("Backslashed quotes + quotes", "'toto \\'tutu' tata", "toto 'tutuXXXtata");
500 #define mytest_str(name, input, separator, expected) \
501 xbt_test_add(name); \
502 d = xbt_str_split_str(input, separator); \
503 s = xbt_str_join(d, "XXX"); \
504 xbt_test_assert(not strcmp(s, expected), "Input (%s) leads to (%s) instead of (%s)", input, s, expected); \
508 XBT_TEST_UNIT("xbt_str_split_str", test_split_str, "test the function xbt_str_split_str")
513 mytest_str("Empty string and separator", "", "", "");
514 mytest_str("Empty string", "", "##", "");
515 mytest_str("Empty separator", "toto", "", "toto");
516 mytest_str("String with no separator in it", "toto", "##", "toto");
517 mytest_str("Basic test", "toto##tutu", "##", "totoXXXtutu");
520 #define test_parse_error(function, name, variable, str) \
522 xbt_test_add(name); \
524 variable = function(str, "Parse error"); \
525 xbt_test_fail("The test '%s' did not detect the problem",name ); \
526 } catch(xbt_ex& e) { \
527 if (e.category != arg_error) { \
528 xbt_test_exception(e); \
532 #define test_parse_ok(function, name, variable, str, value) \
534 xbt_test_add(name); \
536 variable = function(str, "Parse error"); \
537 } catch(xbt_ex& e) { \
538 xbt_test_exception(e); \
540 xbt_test_assert(variable == value, "Fail to parse '%s'", str); \
543 XBT_TEST_UNIT("xbt_str_parse", test_parse, "Test the parsing functions")
546 test_parse_ok(xbt_str_parse_int, "Parse int", rint, "42", 42);
547 test_parse_ok(xbt_str_parse_int, "Parse 0 as an int", rint, "0", 0);
548 test_parse_ok(xbt_str_parse_int, "Parse -1 as an int", rint, "-1", -1);
550 test_parse_error(xbt_str_parse_int, "Parse int + noise", rint, "342 cruft");
551 test_parse_error(xbt_str_parse_int, "Parse nullptr as an int", rint, nullptr);
552 test_parse_error(xbt_str_parse_int, "Parse '' as an int", rint, "");
553 test_parse_error(xbt_str_parse_int, "Parse cruft as an int", rint, "cruft");
555 double rdouble = -9999;
556 test_parse_ok(xbt_str_parse_double, "Parse 42 as a double", rdouble, "42", 42);
557 test_parse_ok(xbt_str_parse_double, "Parse 42.5 as a double", rdouble, "42.5", 42.5);
558 test_parse_ok(xbt_str_parse_double, "Parse 0 as a double", rdouble, "0", 0);
559 test_parse_ok(xbt_str_parse_double, "Parse -1 as a double", rdouble, "-1", -1);
561 test_parse_error(xbt_str_parse_double, "Parse double + noise", rdouble, "342 cruft");
562 test_parse_error(xbt_str_parse_double, "Parse nullptr as a double", rdouble, nullptr);
563 test_parse_error(xbt_str_parse_double, "Parse '' as a double", rdouble, "");
564 test_parse_error(xbt_str_parse_double, "Parse cruft as a double", rdouble, "cruft");
566 #endif /* SIMGRID_TEST */