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. */
10 #include "src/internal_config.h"
12 #include "xbt/sysdep.h"
13 #include "xbt/str.h" /* headers of these functions */
15 /** @brief Strip whitespace (or other characters) from the end of a string.
17 * Strips the whitespaces from the end of s.
18 * By default (when char_list=nullptr), these characters get stripped:
20 * - " " (ASCII 32 (0x20)) space.
21 * - "\t" (ASCII 9 (0x09)) tab.
22 * - "\n" (ASCII 10 (0x0A)) line feed.
23 * - "\r" (ASCII 13 (0x0D)) carriage return.
24 * - "\0" (ASCII 0 (0x00)) nullptr.
25 * - "\x0B" (ASCII 11 (0x0B)) vertical tab.
27 * @param s The string to strip. Modified in place.
28 * @param char_list A string which contains the characters you want to strip.
30 void xbt_str_rtrim(char *s, const char *char_list)
33 const char *__char_list = " \t\n\r\x0B";
34 char white_char[256] = { 1, 0 };
40 while (*__char_list) {
41 white_char[(unsigned char) *__char_list++] = 1;
45 white_char[(unsigned char) *char_list++] = 1;
52 while ((cur >= s) && white_char[(unsigned char) *cur])
58 /** @brief Strip whitespace (or other characters) from the beginning of a string.
60 * Strips the whitespaces from the begining of s.
61 * By default (when char_list=nullptr), these characters get stripped:
63 * - " " (ASCII 32 (0x20)) space.
64 * - "\t" (ASCII 9 (0x09)) tab.
65 * - "\n" (ASCII 10 (0x0A)) line feed.
66 * - "\r" (ASCII 13 (0x0D)) carriage return.
67 * - "\0" (ASCII 0 (0x00)) nullptr.
68 * - "\x0B" (ASCII 11 (0x0B)) vertical tab.
70 * @param s The string to strip. Modified in place.
71 * @param char_list A string which contains the characters you want to strip.
73 void xbt_str_ltrim(char *s, const char *char_list)
76 const char *__char_list = " \t\n\r\x0B";
77 char white_char[256] = { 1, 0 };
83 while (*__char_list) {
84 white_char[(unsigned char) *__char_list++] = 1;
88 white_char[(unsigned char) *char_list++] = 1;
92 while (*cur && white_char[(unsigned char) *cur])
95 memmove(s, cur, strlen(cur) + 1);
98 /** @brief Strip whitespace (or other characters) from the end and the begining of a string.
100 * Strips the whitespaces from both the beginning and the end of s.
101 * By default (when char_list=nullptr), these characters get stripped:
103 * - " " (ASCII 32 (0x20)) space.
104 * - "\t" (ASCII 9 (0x09)) tab.
105 * - "\n" (ASCII 10 (0x0A)) line feed.
106 * - "\r" (ASCII 13 (0x0D)) carriage return.
107 * - "\0" (ASCII 0 (0x00)) nullptr.
108 * - "\x0B" (ASCII 11 (0x0B)) vertical tab.
110 * @param s The string to strip.
111 * @param char_list A string which contains the characters you want to strip.
113 void xbt_str_trim(char *s, const char *char_list)
118 xbt_str_rtrim(s, char_list);
119 xbt_str_ltrim(s, char_list);
122 /** @brief Substitutes a char for another in a string
124 * @param str the string to modify
125 * @param from char to search
126 * @param to char to put instead
127 * @param occurence number of changes to do (=0 means all)
129 void xbt_str_subst(char *str, char from, char to, int occurence)
143 /** @brief Splits a string into a dynar of strings
145 * @param s: the string to split
146 * @param sep: a string of all chars to consider as separator.
148 * By default (with sep=nullptr), these characters are used as separator:
150 * - " " (ASCII 32 (0x20)) space.
151 * - "\t" (ASCII 9 (0x09)) tab.
152 * - "\n" (ASCII 10 (0x0A)) line feed.
153 * - "\r" (ASCII 13 (0x0D)) carriage return.
154 * - "\0" (ASCII 0 (0x00)) nullptr.
155 * - "\x0B" (ASCII 11 (0x0B)) vertical tab.
157 xbt_dynar_t xbt_str_split(const char *s, const char *sep)
159 xbt_dynar_t res = xbt_dynar_new(sizeof(char *), &xbt_free_ref);
160 const char *sep_dflt = " \t\n\r\x0B";
161 char is_sep[256] = { 1, 0 };
163 /* check what are the separators */
164 memset(is_sep, 0, sizeof(is_sep));
167 is_sep[(unsigned char) *sep_dflt++] = 1;
170 is_sep[(unsigned char) *sep++] = 1;
172 is_sep[0] = 1; /* End of string is also separator */
184 while (!is_sep[(unsigned char) *q]) {
190 topush = (char*) xbt_malloc(q - p + 1);
191 memcpy(topush, p, q - p);
192 topush[q - p] = '\0';
193 xbt_dynar_push(res, &topush);
201 * \brief This functions splits a string after using another string as separator
202 * For example A!!B!!C split after !! will return the dynar {A,B,C}
203 * \return An array of dynars containing the string tokens
205 xbt_dynar_t xbt_str_split_str(const char *s, const char *sep)
207 xbt_dynar_t res = xbt_dynar_new(sizeof(char *), &xbt_free_ref);
215 if (sep[0] == '\0') {
217 xbt_dynar_push(res, &s);
224 // get the start of the first occurrence of the substring
226 //if substring was not found add the entire string
229 to_push = (char*) xbt_malloc(v + 1);
230 memcpy(to_push, p, v);
232 xbt_dynar_push(res, &to_push);
236 to_push = (char*) xbt_malloc(q - p + 1);
237 memcpy(to_push, p, q - p);
238 //add string terminator
239 to_push[q - p] = '\0';
240 xbt_dynar_push(res, &to_push);
247 /** @brief Just like @ref xbt_str_split_quoted (Splits a string into a dynar of strings), but without memory allocation
249 * The string passed as argument must be writable (not const)
250 * The elements of the dynar are just parts of the string passed as argument.
251 * So if you don't store that argument elsewhere, you should free it in addition to freeing the dynar. This can be done
252 * by simply freeing the first argument of the dynar:
253 * free(xbt_dynar_get_ptr(dynar,0));
255 * Actually this function puts a bunch of \0 in the memory area you passed as argument to separate the elements, and
256 * pushes the address of each chunk in the resulting dynar. Yes, that's uneven. Yes, that's gory. But that's efficient.
258 xbt_dynar_t xbt_str_split_quoted_in_place(char *s) {
259 xbt_dynar_t res = xbt_dynar_new(sizeof(char *), nullptr);
261 char* end; /* pointers around the parsed chunk */
262 int in_simple_quote = 0;
263 int in_double_quote = 0;
265 int ctn = 0; /* Got something in this block */
272 /* do not trim leading spaces: caller responsibility to clean his cruft */
279 /* Protected char; move it closer */
280 memmove(end, end + 1, strlen(end));
282 THROWF(arg_error, 0, "String ends with \\");
283 end++; /* Pass the protected char */
287 if (!in_double_quote) {
288 in_simple_quote = !in_simple_quote;
289 memmove(end, end + 1, strlen(end));
291 /* simple quote protected by double ones */
297 if (!in_simple_quote) {
298 in_double_quote = !in_double_quote;
299 memmove(end, end + 1, strlen(end));
301 /* double quote protected by simple ones */
309 if (*end == '\0' && (in_simple_quote || in_double_quote)) {
310 THROWF(arg_error, 0, "End of string found while searching for %c in %s", (in_simple_quote ? '\'' : '"'), s);
312 if (in_simple_quote || in_double_quote) {
320 /* Found a separator. Push the string if contains something */
321 xbt_dynar_push(res, &beg);
329 /* trim within the string, manually to speed things up */
343 /** @brief Splits a string into a dynar of strings, taking quotes into account
345 * It basically does the same argument separation than the shell, where white spaces can be escaped and where arguments
346 * are never split within a quote group.
347 * Several subsequent spaces are ignored (unless within quotes, of course).
348 * You may want to trim the input string, if you want to avoid empty entries
350 xbt_dynar_t xbt_str_split_quoted(const char *s)
352 xbt_dynar_t res = xbt_dynar_new(sizeof(char *), &xbt_free_ref);
354 char *str_to_free; /* we have to copy the string before, to handle backslashes */
360 str_to_free = xbt_strdup(s);
362 parsed = xbt_str_split_quoted_in_place(str_to_free);
363 xbt_dynar_foreach(parsed,cursor,p) {
364 char *q=xbt_strdup(p);
365 xbt_dynar_push(res,&q);
368 xbt_dynar_shrink(res, 0);
369 xbt_dynar_free(&parsed);
373 /** @brief Join a set of strings as a single string */
374 char *xbt_str_join(xbt_dynar_t dyn, const char *sep)
376 int len = 1, dyn_len = xbt_dynar_length(dyn);
381 return xbt_strdup("");
383 /* compute the length */
384 xbt_dynar_foreach(dyn, cpt, cursor) {
385 len += strlen(cursor);
387 len += strlen(sep) * dyn_len;
389 char* res = (char*)xbt_malloc(len);
391 xbt_dynar_foreach(dyn, cpt, cursor) {
392 if ((int) cpt < dyn_len - 1)
393 p += snprintf(p,len, "%s%s", cursor, sep);
395 p += snprintf(p,len, "%s", cursor);
400 /** @brief Join a set of strings as a single string
402 * The parameter must be a nullptr-terminated array of chars,
403 * just like xbt_dynar_to_array() produces
405 char *xbt_str_join_array(const char *const *strs, const char *sep)
407 int amount_strings=0;
410 if ((!strs) || (!strs[0]))
411 return xbt_strdup("");
413 /* compute the length before malloc */
414 for (int i = 0; strs[i]; i++) {
415 len += strlen(strs[i]);
418 len += strlen(sep) * amount_strings;
421 char* res = (char*)xbt_malloc(len);
423 for (int i = 0; strs[i]; i++) {
424 if (i != 0) { // not first loop
425 q += snprintf(q,len, "%s%s", sep, strs[i]);
427 q += snprintf(q,len, "%s",strs[i]);
433 /** @brief Parse an integer out of a string, or raise an error
435 * The @a str is passed as argument to your @a error_msg, as follows:
436 * @verbatim THROWF(arg_error, 0, error_msg, str); @endverbatim
438 long int xbt_str_parse_int(const char* str, const char* error_msg)
441 if (str == nullptr || str[0] == '\0')
442 THROWF(arg_error, 0, error_msg, str);
444 long int res = strtol(str, &endptr, 10);
445 if (endptr[0] != '\0')
446 THROWF(arg_error, 0, error_msg, str);
451 /** @brief Parse a double out of a string, or raise an error
453 * The @a str is passed as argument to your @a error_msg, as follows:
454 * @verbatim THROWF(arg_error, 0, error_msg, str); @endverbatim
456 double xbt_str_parse_double(const char* str, const char* error_msg)
459 if (str == nullptr || str[0] == '\0')
460 THROWF(arg_error, 0, error_msg, str);
462 double res = strtod(str, &endptr);
463 if (endptr[0] != '\0')
464 THROWF(arg_error, 0, error_msg, str);
470 #include <xbt/ex.hpp>
473 XBT_TEST_SUITE("xbt_str", "String Handling");
475 #define mytest(name, input, expected) \
476 xbt_test_add(name); \
477 d=xbt_str_split_quoted(input); \
478 s=xbt_str_join(d,"XXX"); \
479 xbt_test_assert(!strcmp(s,expected),\
480 "Input (%s) leads to (%s) instead of (%s)", \
484 XBT_TEST_UNIT("xbt_str_split_quoted", test_split_quoted, "test the function xbt_str_split_quoted")
489 mytest("Empty", "", "");
490 mytest("Basic test", "toto tutu", "totoXXXtutu");
491 mytest("Useless backslashes", "\\t\\o\\t\\o \\t\\u\\t\\u", "totoXXXtutu");
492 mytest("Protected space", "toto\\ tutu", "toto tutu");
493 mytest("Several spaces", "toto tutu", "totoXXXtutu");
494 mytest("LTriming", " toto tatu", "totoXXXtatu");
495 mytest("Triming", " toto tutu ", "totoXXXtutu");
496 mytest("Single quotes", "'toto tutu' tata", "toto tutuXXXtata");
497 mytest("Double quotes", "\"toto tutu\" tata", "toto tutuXXXtata");
498 mytest("Mixed quotes", "\"toto' 'tutu\" tata", "toto' 'tutuXXXtata");
499 mytest("Backslashed quotes", "\\'toto tutu\\' tata", "'totoXXXtutu'XXXtata");
500 mytest("Backslashed quotes + quotes", "'toto \\'tutu' tata", "toto 'tutuXXXtata");
503 #define mytest_str(name, input, separator, expected) \
504 xbt_test_add(name); \
505 d=xbt_str_split_str(input, separator); \
506 s=xbt_str_join(d,"XXX"); \
507 xbt_test_assert(!strcmp(s,expected),\
508 "Input (%s) leads to (%s) instead of (%s)", \
513 XBT_TEST_UNIT("xbt_str_split_str", test_split_str, "test the function xbt_str_split_str")
518 mytest_str("Empty string and separator", "", "", "");
519 mytest_str("Empty string", "", "##", "");
520 mytest_str("Empty separator", "toto", "", "toto");
521 mytest_str("String with no separator in it", "toto", "##", "toto");
522 mytest_str("Basic test", "toto##tutu", "##", "totoXXXtutu");
525 #define test_parse_error(function, name, variable, str) \
527 xbt_test_add(name); \
529 variable = function(str, "Parse error"); \
530 xbt_test_fail("The test '%s' did not detect the problem",name ); \
531 } catch(xbt_ex& e) { \
532 if (e.category != arg_error) { \
533 xbt_test_exception(e); \
537 #define test_parse_ok(function, name, variable, str, value) \
539 xbt_test_add(name); \
541 variable = function(str, "Parse error"); \
542 } catch(xbt_ex& e) { \
543 xbt_test_exception(e); \
545 xbt_test_assert(variable == value, "Fail to parse '%s'", str); \
548 XBT_TEST_UNIT("xbt_str_parse", test_parse, "Test the parsing functions")
551 test_parse_ok(xbt_str_parse_int, "Parse int", rint, "42", 42);
552 test_parse_ok(xbt_str_parse_int, "Parse 0 as an int", rint, "0", 0);
553 test_parse_ok(xbt_str_parse_int, "Parse -1 as an int", rint, "-1", -1);
555 test_parse_error(xbt_str_parse_int, "Parse int + noise", rint, "342 cruft");
556 test_parse_error(xbt_str_parse_int, "Parse nullptr as an int", rint, nullptr);
557 test_parse_error(xbt_str_parse_int, "Parse '' as an int", rint, "");
558 test_parse_error(xbt_str_parse_int, "Parse cruft as an int", rint, "cruft");
560 double rdouble = -9999;
561 test_parse_ok(xbt_str_parse_double, "Parse 42 as a double", rdouble, "42", 42);
562 test_parse_ok(xbt_str_parse_double, "Parse 42.5 as a double", rdouble, "42.5", 42.5);
563 test_parse_ok(xbt_str_parse_double, "Parse 0 as a double", rdouble, "0", 0);
564 test_parse_ok(xbt_str_parse_double, "Parse -1 as a double", rdouble, "-1", -1);
566 test_parse_error(xbt_str_parse_double, "Parse double + noise", rdouble, "342 cruft");
567 test_parse_error(xbt_str_parse_double, "Parse nullptr as a double", rdouble, nullptr);
568 test_parse_error(xbt_str_parse_double, "Parse '' as a double", rdouble, "");
569 test_parse_error(xbt_str_parse_double, "Parse cruft as a double", rdouble, "cruft");
571 #endif /* SIMGRID_TEST */