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 */
14 #include "xbt/strbuff.h"
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=nullptr), 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)) nullptr.
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.
31 void xbt_str_rtrim(char *s, const char *char_list)
34 const char *__char_list = " \t\n\r\x0B";
35 char white_char[256] = { 1, 0 };
41 while (*__char_list) {
42 white_char[(unsigned char) *__char_list++] = 1;
46 white_char[(unsigned char) *char_list++] = 1;
53 while ((cur >= s) && white_char[(unsigned char) *cur])
59 /** @brief Strip whitespace (or other characters) from the beginning of a string.
61 * Strips the whitespaces from the begining of s.
62 * By default (when char_list=nullptr), these characters get stripped:
64 * - " " (ASCII 32 (0x20)) space.
65 * - "\t" (ASCII 9 (0x09)) tab.
66 * - "\n" (ASCII 10 (0x0A)) line feed.
67 * - "\r" (ASCII 13 (0x0D)) carriage return.
68 * - "\0" (ASCII 0 (0x00)) nullptr.
69 * - "\x0B" (ASCII 11 (0x0B)) vertical tab.
71 * @param s The string to strip. Modified in place.
72 * @param char_list A string which contains the characters you want to strip.
74 void xbt_str_ltrim(char *s, const char *char_list)
77 const char *__char_list = " \t\n\r\x0B";
78 char white_char[256] = { 1, 0 };
84 while (*__char_list) {
85 white_char[(unsigned char) *__char_list++] = 1;
89 white_char[(unsigned char) *char_list++] = 1;
93 while (*cur && white_char[(unsigned char) *cur])
96 memmove(s, cur, strlen(cur) + 1);
99 /** @brief Strip whitespace (or other characters) from the end and the begining of a string.
101 * Strips the whitespaces from both the beginning and the end of s.
102 * By default (when char_list=nullptr), these characters get stripped:
104 * - " " (ASCII 32 (0x20)) space.
105 * - "\t" (ASCII 9 (0x09)) tab.
106 * - "\n" (ASCII 10 (0x0A)) line feed.
107 * - "\r" (ASCII 13 (0x0D)) carriage return.
108 * - "\0" (ASCII 0 (0x00)) nullptr.
109 * - "\x0B" (ASCII 11 (0x0B)) vertical tab.
111 * @param s The string to strip.
112 * @param char_list A string which contains the characters you want to strip.
114 void xbt_str_trim(char *s, const char *char_list)
119 xbt_str_rtrim(s, char_list);
120 xbt_str_ltrim(s, char_list);
123 /** @brief Substitutes a char for another in a string
125 * @param str the string to modify
126 * @param from char to search
127 * @param to char to put instead
128 * @param occurence number of changes to do (=0 means all)
130 void xbt_str_subst(char *str, char from, char to, int occurence)
144 /** @brief Replaces a set of variables by their values
146 * @param str The input of the replacement process
147 * @param patterns The changes to apply
148 * @return The string modified
150 * Both '$toto' and '${toto}' are valid (and the two writing are equivalent).
152 * If the variable name contains spaces, use the brace version (ie, ${toto tutu})
154 * You can provide a default value to use if the variable is not set in the dict by using '${var:=default}' or
155 * '${var:-default}'. These two forms are equivalent, even if they shouldn't to respect the shell standard (:= form
156 * should set the value in the dict, but does not) (BUG).
158 char *xbt_str_varsubst(const char *str, xbt_dict_t patterns)
160 xbt_strbuff_t buff = xbt_strbuff_new_from(str);
162 xbt_strbuff_varsubst(buff, patterns);
164 xbt_strbuff_free_container(buff);
169 /** @brief Splits a string into a dynar of strings
171 * @param s: the string to split
172 * @param sep: a string of all chars to consider as separator.
174 * By default (with sep=nullptr), these characters are used as separator:
176 * - " " (ASCII 32 (0x20)) space.
177 * - "\t" (ASCII 9 (0x09)) tab.
178 * - "\n" (ASCII 10 (0x0A)) line feed.
179 * - "\r" (ASCII 13 (0x0D)) carriage return.
180 * - "\0" (ASCII 0 (0x00)) nullptr.
181 * - "\x0B" (ASCII 11 (0x0B)) vertical tab.
183 xbt_dynar_t xbt_str_split(const char *s, const char *sep)
185 xbt_dynar_t res = xbt_dynar_new(sizeof(char *), &xbt_free_ref);
186 const char *sep_dflt = " \t\n\r\x0B";
187 char is_sep[256] = { 1, 0 };
189 /* check what are the separators */
190 memset(is_sep, 0, sizeof(is_sep));
193 is_sep[(unsigned char) *sep_dflt++] = 1;
196 is_sep[(unsigned char) *sep++] = 1;
198 is_sep[0] = 1; /* End of string is also separator */
210 while (!is_sep[(unsigned char) *q]) {
216 topush = (char*) xbt_malloc(q - p + 1);
217 memcpy(topush, p, q - p);
218 topush[q - p] = '\0';
219 xbt_dynar_push(res, &topush);
227 * \brief This functions splits a string after using another string as separator
228 * For example A!!B!!C split after !! will return the dynar {A,B,C}
229 * \return An array of dynars containing the string tokens
231 xbt_dynar_t xbt_str_split_str(const char *s, const char *sep)
233 xbt_dynar_t res = xbt_dynar_new(sizeof(char *), &xbt_free_ref);
241 if (sep[0] == '\0') {
243 xbt_dynar_push(res, &s);
250 // get the start of the first occurrence of the substring
252 //if substring was not found add the entire string
255 to_push = (char*) xbt_malloc(v + 1);
256 memcpy(to_push, p, v);
258 xbt_dynar_push(res, &to_push);
262 to_push = (char*) xbt_malloc(q - p + 1);
263 memcpy(to_push, p, q - p);
264 //add string terminator
265 to_push[q - p] = '\0';
266 xbt_dynar_push(res, &to_push);
273 /** @brief Just like @ref xbt_str_split_quoted (Splits a string into a dynar of strings), but without memory allocation
275 * The string passed as argument must be writable (not const)
276 * The elements of the dynar are just parts of the string passed as argument.
277 * So if you don't store that argument elsewhere, you should free it in addition to freeing the dynar. This can be done
278 * by simply freeing the first argument of the dynar:
279 * free(xbt_dynar_get_ptr(dynar,0));
281 * Actually this function puts a bunch of \0 in the memory area you passed as argument to separate the elements, and
282 * pushes the address of each chunk in the resulting dynar. Yes, that's uneven. Yes, that's gory. But that's efficient.
284 xbt_dynar_t xbt_str_split_quoted_in_place(char *s) {
285 xbt_dynar_t res = xbt_dynar_new(sizeof(char *), nullptr);
287 char* end; /* pointers around the parsed chunk */
288 int in_simple_quote = 0;
289 int in_double_quote = 0;
291 int ctn = 0; /* Got something in this block */
298 /* do not trim leading spaces: caller responsibility to clean his cruft */
305 /* Protected char; move it closer */
306 memmove(end, end + 1, strlen(end));
308 THROWF(arg_error, 0, "String ends with \\");
309 end++; /* Pass the protected char */
313 if (!in_double_quote) {
314 in_simple_quote = !in_simple_quote;
315 memmove(end, end + 1, strlen(end));
317 /* simple quote protected by double ones */
323 if (!in_simple_quote) {
324 in_double_quote = !in_double_quote;
325 memmove(end, end + 1, strlen(end));
327 /* double quote protected by simple ones */
335 if (*end == '\0' && (in_simple_quote || in_double_quote)) {
336 THROWF(arg_error, 0, "End of string found while searching for %c in %s", (in_simple_quote ? '\'' : '"'), s);
338 if (in_simple_quote || in_double_quote) {
346 /* Found a separator. Push the string if contains something */
347 xbt_dynar_push(res, &beg);
355 /* trim within the string, manually to speed things up */
369 /** @brief Splits a string into a dynar of strings, taking quotes into account
371 * It basically does the same argument separation than the shell, where white spaces can be escaped and where arguments
372 * are never split within a quote group.
373 * Several subsequent spaces are ignored (unless within quotes, of course).
374 * You may want to trim the input string, if you want to avoid empty entries
376 xbt_dynar_t xbt_str_split_quoted(const char *s)
378 xbt_dynar_t res = xbt_dynar_new(sizeof(char *), &xbt_free_ref);
380 char *str_to_free; /* we have to copy the string before, to handle backslashes */
386 str_to_free = xbt_strdup(s);
388 parsed = xbt_str_split_quoted_in_place(str_to_free);
389 xbt_dynar_foreach(parsed,cursor,p) {
390 char *q=xbt_strdup(p);
391 xbt_dynar_push(res,&q);
394 xbt_dynar_shrink(res, 0);
395 xbt_dynar_free(&parsed);
399 /** @brief Join a set of strings as a single string */
400 char *xbt_str_join(xbt_dynar_t dyn, const char *sep)
402 int len = 1, dyn_len = xbt_dynar_length(dyn);
407 return xbt_strdup("");
409 /* compute the length */
410 xbt_dynar_foreach(dyn, cpt, cursor) {
411 len += strlen(cursor);
413 len += strlen(sep) * dyn_len;
415 char* res = (char*)xbt_malloc(len);
417 xbt_dynar_foreach(dyn, cpt, cursor) {
418 if ((int) cpt < dyn_len - 1)
419 p += snprintf(p,len, "%s%s", cursor, sep);
421 p += snprintf(p,len, "%s", cursor);
426 /** @brief Join a set of strings as a single string
428 * The parameter must be a nullptr-terminated array of chars,
429 * just like xbt_dynar_to_array() produces
431 char *xbt_str_join_array(const char *const *strs, const char *sep)
433 int amount_strings=0;
436 if ((!strs) || (!strs[0]))
437 return xbt_strdup("");
439 /* compute the length before malloc */
440 for (int i = 0; strs[i]; i++) {
441 len += strlen(strs[i]);
444 len += strlen(sep) * amount_strings;
447 char* res = (char*)xbt_malloc(len);
449 for (int i = 0; strs[i]; i++) {
450 if (i != 0) { // not first loop
451 q += snprintf(q,len, "%s%s", sep, strs[i]);
453 q += snprintf(q,len, "%s",strs[i]);
459 /** @brief Parse an integer out of a string, or raise an error
461 * The @a str is passed as argument to your @a error_msg, as follows:
462 * @verbatim THROWF(arg_error, 0, error_msg, str); @endverbatim
464 long int xbt_str_parse_int(const char* str, const char* error_msg)
467 if (str == nullptr || str[0] == '\0')
468 THROWF(arg_error, 0, error_msg, str);
470 long int res = strtol(str, &endptr, 10);
471 if (endptr[0] != '\0')
472 THROWF(arg_error, 0, error_msg, str);
477 /** @brief Parse a double out of a string, or raise an error
479 * The @a str is passed as argument to your @a error_msg, as follows:
480 * @verbatim THROWF(arg_error, 0, error_msg, str); @endverbatim
482 double xbt_str_parse_double(const char* str, const char* error_msg)
485 if (str == nullptr || str[0] == '\0')
486 THROWF(arg_error, 0, error_msg, str);
488 double res = strtod(str, &endptr);
489 if (endptr[0] != '\0')
490 THROWF(arg_error, 0, error_msg, str);
496 #include <xbt/ex.hpp>
499 XBT_TEST_SUITE("xbt_str", "String Handling");
501 #define mytest(name, input, expected) \
502 xbt_test_add(name); \
503 d=xbt_str_split_quoted(input); \
504 s=xbt_str_join(d,"XXX"); \
505 xbt_test_assert(!strcmp(s,expected),\
506 "Input (%s) leads to (%s) instead of (%s)", \
510 XBT_TEST_UNIT("xbt_str_split_quoted", test_split_quoted, "test the function xbt_str_split_quoted")
515 mytest("Empty", "", "");
516 mytest("Basic test", "toto tutu", "totoXXXtutu");
517 mytest("Useless backslashes", "\\t\\o\\t\\o \\t\\u\\t\\u", "totoXXXtutu");
518 mytest("Protected space", "toto\\ tutu", "toto tutu");
519 mytest("Several spaces", "toto tutu", "totoXXXtutu");
520 mytest("LTriming", " toto tatu", "totoXXXtatu");
521 mytest("Triming", " toto tutu ", "totoXXXtutu");
522 mytest("Single quotes", "'toto tutu' tata", "toto tutuXXXtata");
523 mytest("Double quotes", "\"toto tutu\" tata", "toto tutuXXXtata");
524 mytest("Mixed quotes", "\"toto' 'tutu\" tata", "toto' 'tutuXXXtata");
525 mytest("Backslashed quotes", "\\'toto tutu\\' tata", "'totoXXXtutu'XXXtata");
526 mytest("Backslashed quotes + quotes", "'toto \\'tutu' tata", "toto 'tutuXXXtata");
529 #define mytest_str(name, input, separator, expected) \
530 xbt_test_add(name); \
531 d=xbt_str_split_str(input, separator); \
532 s=xbt_str_join(d,"XXX"); \
533 xbt_test_assert(!strcmp(s,expected),\
534 "Input (%s) leads to (%s) instead of (%s)", \
539 XBT_TEST_UNIT("xbt_str_split_str", test_split_str, "test the function xbt_str_split_str")
544 mytest_str("Empty string and separator", "", "", "");
545 mytest_str("Empty string", "", "##", "");
546 mytest_str("Empty separator", "toto", "", "toto");
547 mytest_str("String with no separator in it", "toto", "##", "toto");
548 mytest_str("Basic test", "toto##tutu", "##", "totoXXXtutu");
551 #define test_parse_error(function, name, variable, str) \
553 xbt_test_add(name); \
555 variable = function(str, "Parse error"); \
556 xbt_test_fail("The test '%s' did not detect the problem",name ); \
557 } catch(xbt_ex& e) { \
558 if (e.category != arg_error) { \
559 xbt_test_exception(e); \
563 #define test_parse_ok(function, name, variable, str, value) \
565 xbt_test_add(name); \
567 variable = function(str, "Parse error"); \
568 } catch(xbt_ex& e) { \
569 xbt_test_exception(e); \
571 xbt_test_assert(variable == value, "Fail to parse '%s'", str); \
574 XBT_TEST_UNIT("xbt_str_parse", test_parse, "Test the parsing functions")
577 test_parse_ok(xbt_str_parse_int, "Parse int", rint, "42", 42);
578 test_parse_ok(xbt_str_parse_int, "Parse 0 as an int", rint, "0", 0);
579 test_parse_ok(xbt_str_parse_int, "Parse -1 as an int", rint, "-1", -1);
581 test_parse_error(xbt_str_parse_int, "Parse int + noise", rint, "342 cruft");
582 test_parse_error(xbt_str_parse_int, "Parse nullptr as an int", rint, nullptr);
583 test_parse_error(xbt_str_parse_int, "Parse '' as an int", rint, "");
584 test_parse_error(xbt_str_parse_int, "Parse cruft as an int", rint, "cruft");
586 double rdouble = -9999;
587 test_parse_ok(xbt_str_parse_double, "Parse 42 as a double", rdouble, "42", 42);
588 test_parse_ok(xbt_str_parse_double, "Parse 42.5 as a double", rdouble, "42.5", 42.5);
589 test_parse_ok(xbt_str_parse_double, "Parse 0 as a double", rdouble, "0", 0);
590 test_parse_ok(xbt_str_parse_double, "Parse -1 as a double", rdouble, "-1", -1);
592 test_parse_error(xbt_str_parse_double, "Parse double + noise", rdouble, "342 cruft");
593 test_parse_error(xbt_str_parse_double, "Parse nullptr as a double", rdouble, nullptr);
594 test_parse_error(xbt_str_parse_double, "Parse '' as a double", rdouble, "");
595 test_parse_error(xbt_str_parse_double, "Parse cruft as a double", rdouble, "cruft");
597 #endif /* SIMGRID_TEST */