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 Splits a string into a dynar of strings
146 * @param s: the string to split
147 * @param sep: a string of all chars to consider as separator.
149 * By default (with sep=nullptr), these characters are used as separator:
151 * - " " (ASCII 32 (0x20)) space.
152 * - "\t" (ASCII 9 (0x09)) tab.
153 * - "\n" (ASCII 10 (0x0A)) line feed.
154 * - "\r" (ASCII 13 (0x0D)) carriage return.
155 * - "\0" (ASCII 0 (0x00)) nullptr.
156 * - "\x0B" (ASCII 11 (0x0B)) vertical tab.
158 xbt_dynar_t xbt_str_split(const char *s, const char *sep)
160 xbt_dynar_t res = xbt_dynar_new(sizeof(char *), &xbt_free_ref);
161 const char *sep_dflt = " \t\n\r\x0B";
162 char is_sep[256] = { 1, 0 };
164 /* check what are the separators */
165 memset(is_sep, 0, sizeof(is_sep));
168 is_sep[(unsigned char) *sep_dflt++] = 1;
171 is_sep[(unsigned char) *sep++] = 1;
173 is_sep[0] = 1; /* End of string is also separator */
185 while (!is_sep[(unsigned char) *q]) {
191 topush = (char*) xbt_malloc(q - p + 1);
192 memcpy(topush, p, q - p);
193 topush[q - p] = '\0';
194 xbt_dynar_push(res, &topush);
202 * \brief This functions splits a string after using another string as separator
203 * For example A!!B!!C split after !! will return the dynar {A,B,C}
204 * \return An array of dynars containing the string tokens
206 xbt_dynar_t xbt_str_split_str(const char *s, const char *sep)
208 xbt_dynar_t res = xbt_dynar_new(sizeof(char *), &xbt_free_ref);
216 if (sep[0] == '\0') {
218 xbt_dynar_push(res, &s);
225 // get the start of the first occurrence of the substring
227 //if substring was not found add the entire string
230 to_push = (char*) xbt_malloc(v + 1);
231 memcpy(to_push, p, v);
233 xbt_dynar_push(res, &to_push);
237 to_push = (char*) xbt_malloc(q - p + 1);
238 memcpy(to_push, p, q - p);
239 //add string terminator
240 to_push[q - p] = '\0';
241 xbt_dynar_push(res, &to_push);
248 /** @brief Just like @ref xbt_str_split_quoted (Splits a string into a dynar of strings), but without memory allocation
250 * The string passed as argument must be writable (not const)
251 * The elements of the dynar are just parts of the string passed as argument.
252 * So if you don't store that argument elsewhere, you should free it in addition to freeing the dynar. This can be done
253 * by simply freeing the first argument of the dynar:
254 * free(xbt_dynar_get_ptr(dynar,0));
256 * Actually this function puts a bunch of \0 in the memory area you passed as argument to separate the elements, and
257 * pushes the address of each chunk in the resulting dynar. Yes, that's uneven. Yes, that's gory. But that's efficient.
259 xbt_dynar_t xbt_str_split_quoted_in_place(char *s) {
260 xbt_dynar_t res = xbt_dynar_new(sizeof(char *), nullptr);
262 char* end; /* pointers around the parsed chunk */
263 int in_simple_quote = 0;
264 int in_double_quote = 0;
266 int ctn = 0; /* Got something in this block */
273 /* do not trim leading spaces: caller responsibility to clean his cruft */
280 /* Protected char; move it closer */
281 memmove(end, end + 1, strlen(end));
283 THROWF(arg_error, 0, "String ends with \\");
284 end++; /* Pass the protected char */
288 if (!in_double_quote) {
289 in_simple_quote = !in_simple_quote;
290 memmove(end, end + 1, strlen(end));
292 /* simple quote protected by double ones */
298 if (!in_simple_quote) {
299 in_double_quote = !in_double_quote;
300 memmove(end, end + 1, strlen(end));
302 /* double quote protected by simple ones */
310 if (*end == '\0' && (in_simple_quote || in_double_quote)) {
311 THROWF(arg_error, 0, "End of string found while searching for %c in %s", (in_simple_quote ? '\'' : '"'), s);
313 if (in_simple_quote || in_double_quote) {
321 /* Found a separator. Push the string if contains something */
322 xbt_dynar_push(res, &beg);
330 /* trim within the string, manually to speed things up */
344 /** @brief Splits a string into a dynar of strings, taking quotes into account
346 * It basically does the same argument separation than the shell, where white spaces can be escaped and where arguments
347 * are never split within a quote group.
348 * Several subsequent spaces are ignored (unless within quotes, of course).
349 * You may want to trim the input string, if you want to avoid empty entries
351 xbt_dynar_t xbt_str_split_quoted(const char *s)
353 xbt_dynar_t res = xbt_dynar_new(sizeof(char *), &xbt_free_ref);
355 char *str_to_free; /* we have to copy the string before, to handle backslashes */
361 str_to_free = xbt_strdup(s);
363 parsed = xbt_str_split_quoted_in_place(str_to_free);
364 xbt_dynar_foreach(parsed,cursor,p) {
365 char *q=xbt_strdup(p);
366 xbt_dynar_push(res,&q);
369 xbt_dynar_shrink(res, 0);
370 xbt_dynar_free(&parsed);
374 /** @brief Join a set of strings as a single string */
375 char *xbt_str_join(xbt_dynar_t dyn, const char *sep)
377 int len = 1, dyn_len = xbt_dynar_length(dyn);
382 return xbt_strdup("");
384 /* compute the length */
385 xbt_dynar_foreach(dyn, cpt, cursor) {
386 len += strlen(cursor);
388 len += strlen(sep) * dyn_len;
390 char* res = (char*)xbt_malloc(len);
392 xbt_dynar_foreach(dyn, cpt, cursor) {
393 if ((int) cpt < dyn_len - 1)
394 p += snprintf(p,len, "%s%s", cursor, sep);
396 p += snprintf(p,len, "%s", cursor);
401 /** @brief Join a set of strings as a single string
403 * The parameter must be a nullptr-terminated array of chars,
404 * just like xbt_dynar_to_array() produces
406 char *xbt_str_join_array(const char *const *strs, const char *sep)
408 int amount_strings=0;
411 if ((!strs) || (!strs[0]))
412 return xbt_strdup("");
414 /* compute the length before malloc */
415 for (int i = 0; strs[i]; i++) {
416 len += strlen(strs[i]);
419 len += strlen(sep) * amount_strings;
422 char* res = (char*)xbt_malloc(len);
424 for (int i = 0; strs[i]; i++) {
425 if (i != 0) { // not first loop
426 q += snprintf(q,len, "%s%s", sep, strs[i]);
428 q += snprintf(q,len, "%s",strs[i]);
434 /** @brief Parse an integer out of a string, or raise an error
436 * The @a str is passed as argument to your @a error_msg, as follows:
437 * @verbatim THROWF(arg_error, 0, error_msg, str); @endverbatim
439 long int xbt_str_parse_int(const char* str, const char* error_msg)
442 if (str == nullptr || str[0] == '\0')
443 THROWF(arg_error, 0, error_msg, str);
445 long int res = strtol(str, &endptr, 10);
446 if (endptr[0] != '\0')
447 THROWF(arg_error, 0, error_msg, str);
452 /** @brief Parse a double out of a string, or raise an error
454 * The @a str is passed as argument to your @a error_msg, as follows:
455 * @verbatim THROWF(arg_error, 0, error_msg, str); @endverbatim
457 double xbt_str_parse_double(const char* str, const char* error_msg)
460 if (str == nullptr || str[0] == '\0')
461 THROWF(arg_error, 0, error_msg, str);
463 double res = strtod(str, &endptr);
464 if (endptr[0] != '\0')
465 THROWF(arg_error, 0, error_msg, str);
471 #include <xbt/ex.hpp>
474 XBT_TEST_SUITE("xbt_str", "String Handling");
476 #define mytest(name, input, expected) \
477 xbt_test_add(name); \
478 d=xbt_str_split_quoted(input); \
479 s=xbt_str_join(d,"XXX"); \
480 xbt_test_assert(!strcmp(s,expected),\
481 "Input (%s) leads to (%s) instead of (%s)", \
485 XBT_TEST_UNIT("xbt_str_split_quoted", test_split_quoted, "test the function xbt_str_split_quoted")
490 mytest("Empty", "", "");
491 mytest("Basic test", "toto tutu", "totoXXXtutu");
492 mytest("Useless backslashes", "\\t\\o\\t\\o \\t\\u\\t\\u", "totoXXXtutu");
493 mytest("Protected space", "toto\\ tutu", "toto tutu");
494 mytest("Several spaces", "toto tutu", "totoXXXtutu");
495 mytest("LTriming", " toto tatu", "totoXXXtatu");
496 mytest("Triming", " toto tutu ", "totoXXXtutu");
497 mytest("Single quotes", "'toto tutu' tata", "toto tutuXXXtata");
498 mytest("Double quotes", "\"toto tutu\" tata", "toto tutuXXXtata");
499 mytest("Mixed quotes", "\"toto' 'tutu\" tata", "toto' 'tutuXXXtata");
500 mytest("Backslashed quotes", "\\'toto tutu\\' tata", "'totoXXXtutu'XXXtata");
501 mytest("Backslashed quotes + quotes", "'toto \\'tutu' tata", "toto 'tutuXXXtata");
504 #define mytest_str(name, input, separator, expected) \
505 xbt_test_add(name); \
506 d=xbt_str_split_str(input, separator); \
507 s=xbt_str_join(d,"XXX"); \
508 xbt_test_assert(!strcmp(s,expected),\
509 "Input (%s) leads to (%s) instead of (%s)", \
514 XBT_TEST_UNIT("xbt_str_split_str", test_split_str, "test the function xbt_str_split_str")
519 mytest_str("Empty string and separator", "", "", "");
520 mytest_str("Empty string", "", "##", "");
521 mytest_str("Empty separator", "toto", "", "toto");
522 mytest_str("String with no separator in it", "toto", "##", "toto");
523 mytest_str("Basic test", "toto##tutu", "##", "totoXXXtutu");
526 #define test_parse_error(function, name, variable, str) \
528 xbt_test_add(name); \
530 variable = function(str, "Parse error"); \
531 xbt_test_fail("The test '%s' did not detect the problem",name ); \
532 } catch(xbt_ex& e) { \
533 if (e.category != arg_error) { \
534 xbt_test_exception(e); \
538 #define test_parse_ok(function, name, variable, str, value) \
540 xbt_test_add(name); \
542 variable = function(str, "Parse error"); \
543 } catch(xbt_ex& e) { \
544 xbt_test_exception(e); \
546 xbt_test_assert(variable == value, "Fail to parse '%s'", str); \
549 XBT_TEST_UNIT("xbt_str_parse", test_parse, "Test the parsing functions")
552 test_parse_ok(xbt_str_parse_int, "Parse int", rint, "42", 42);
553 test_parse_ok(xbt_str_parse_int, "Parse 0 as an int", rint, "0", 0);
554 test_parse_ok(xbt_str_parse_int, "Parse -1 as an int", rint, "-1", -1);
556 test_parse_error(xbt_str_parse_int, "Parse int + noise", rint, "342 cruft");
557 test_parse_error(xbt_str_parse_int, "Parse nullptr as an int", rint, nullptr);
558 test_parse_error(xbt_str_parse_int, "Parse '' as an int", rint, "");
559 test_parse_error(xbt_str_parse_int, "Parse cruft as an int", rint, "cruft");
561 double rdouble = -9999;
562 test_parse_ok(xbt_str_parse_double, "Parse 42 as a double", rdouble, "42", 42);
563 test_parse_ok(xbt_str_parse_double, "Parse 42.5 as a double", rdouble, "42.5", 42.5);
564 test_parse_ok(xbt_str_parse_double, "Parse 0 as a double", rdouble, "0", 0);
565 test_parse_ok(xbt_str_parse_double, "Parse -1 as a double", rdouble, "-1", -1);
567 test_parse_error(xbt_str_parse_double, "Parse double + noise", rdouble, "342 cruft");
568 test_parse_error(xbt_str_parse_double, "Parse nullptr as a double", rdouble, nullptr);
569 test_parse_error(xbt_str_parse_double, "Parse '' as a double", rdouble, "");
570 test_parse_error(xbt_str_parse_double, "Parse cruft as a double", rdouble, "cruft");
572 #endif /* SIMGRID_TEST */