1 /* xbt_str.cpp - various helping functions to deal with strings */
3 /* Copyright (c) 2007-2020. The SimGrid Team. All rights reserved. */
5 /* This program is free software; you can redistribute it and/or modify it
6 * under the terms of the license (GNU LGPL) which comes with this package. */
8 #include "simgrid/Exception.hpp"
10 #include "xbt/str.h" /* headers of these functions */
11 #include "xbt/string.hpp"
13 /** @brief Splits a string into a dynar of strings
15 * @param s: the string to split
16 * @param sep: a string of all chars to consider as separator.
18 * By default (with sep=nullptr), these characters are used as separator:
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 xbt_dynar_t xbt_str_split(const char *s, const char *sep)
29 xbt_dynar_t res = xbt_dynar_new(sizeof(char *), &xbt_free_ref);
30 const char *sep_dflt = " \t\n\r\x0B";
31 char is_sep[256] = { 1, 0 };
33 /* check what are the separators */
34 memset(is_sep, 0, sizeof(is_sep));
37 is_sep[(unsigned char) *sep_dflt++] = 1;
40 is_sep[(unsigned char) *sep++] = 1;
42 is_sep[0] = 1; /* End of string is also separator */
54 while (not is_sep[(unsigned char)*q]) {
60 topush = (char*) xbt_malloc(q - p + 1);
61 memcpy(topush, p, q - p);
63 xbt_dynar_push(res, &topush);
70 /** @brief Just like @ref xbt_str_split_quoted (Splits a string into a dynar of strings), but without memory allocation
72 * The string passed as argument must be writable (not const)
73 * The elements of the dynar are just parts of the string passed as argument.
74 * So if you don't store that argument elsewhere, you should free it in addition to freeing the dynar. This can be done
75 * by simply freeing the first argument of the dynar:
76 * free(xbt_dynar_get_ptr(dynar,0));
78 * Actually this function puts a bunch of \0 in the memory area you passed as argument to separate the elements, and
79 * pushes the address of each chunk in the resulting dynar. Yes, that's uneven. Yes, that's gory. But that's efficient.
81 xbt_dynar_t xbt_str_split_quoted_in_place(char *s) {
82 xbt_dynar_t res = xbt_dynar_new(sizeof(char *), nullptr);
84 char* end; /* pointers around the parsed chunk */
85 int in_simple_quote = 0;
86 int in_double_quote = 0;
88 int ctn = 0; /* Got something in this block */
95 /* do not trim leading spaces: caller responsibility to clean his cruft */
102 /* Protected char; move it closer */
103 memmove(end, end + 1, strlen(end));
105 throw std::invalid_argument("String ends with \\");
106 end++; /* Pass the protected char */
110 if (not in_double_quote) {
111 in_simple_quote = not in_simple_quote;
112 memmove(end, end + 1, strlen(end));
114 /* simple quote protected by double ones */
120 if (not in_simple_quote) {
121 in_double_quote = not in_double_quote;
122 memmove(end, end + 1, strlen(end));
124 /* double quote protected by simple ones */
132 if (*end == '\0' && (in_simple_quote || in_double_quote)) {
133 throw std::invalid_argument(simgrid::xbt::string_printf("End of string found while searching for %c in %s",
134 (in_simple_quote ? '\'' : '"'), s));
136 if (in_simple_quote || in_double_quote) {
144 /* Found a separator. Push the string if contains something */
145 xbt_dynar_push(res, &beg);
153 /* trim within the string, manually to speed things up */
167 /** @brief Splits a string into a dynar of strings, taking quotes into account
169 * It basically does the same argument separation than the shell, where white spaces can be escaped and where arguments
170 * are never split within a quote group.
171 * Several subsequent spaces are ignored (unless within quotes, of course).
172 * You may want to trim the input string, if you want to avoid empty entries
174 xbt_dynar_t xbt_str_split_quoted(const char *s)
176 xbt_dynar_t res = xbt_dynar_new(sizeof(char *), &xbt_free_ref);
178 char *str_to_free; /* we have to copy the string before, to handle backslashes */
184 str_to_free = xbt_strdup(s);
186 parsed = xbt_str_split_quoted_in_place(str_to_free);
187 xbt_dynar_foreach(parsed,cursor,p) {
188 char *q=xbt_strdup(p);
189 xbt_dynar_push(res,&q);
191 xbt_free(str_to_free);
192 xbt_dynar_shrink(res, 0);
193 xbt_dynar_free(&parsed);
197 /** @brief Parse an integer out of a string, or raise an error
199 * The @a str is passed as argument to your @a error_msg, as follows:
200 * @verbatim throw std::invalid_argument(simgrid::xbt::string_printf(error_msg, str)); @endverbatim
202 long int xbt_str_parse_int(const char* str, const char* error_msg)
205 if (str == nullptr || str[0] == '\0')
206 throw std::invalid_argument(simgrid::xbt::string_printf(error_msg, str));
208 long int res = strtol(str, &endptr, 10);
209 if (endptr[0] != '\0')
210 throw std::invalid_argument(simgrid::xbt::string_printf(error_msg, str));
215 /** @brief Parse a double out of a string, or raise an error
217 * The @a str is passed as argument to your @a error_msg, as follows:
218 * @verbatim throw std::invalid_argument(simgrid::xbt::string_printf(error_msg, str)); @endverbatim
220 double xbt_str_parse_double(const char* str, const char* error_msg)
223 if (str == nullptr || str[0] == '\0')
224 throw std::invalid_argument(simgrid::xbt::string_printf(error_msg, str));
226 double res = strtod(str, &endptr);
227 if (endptr[0] != '\0')
228 throw std::invalid_argument(simgrid::xbt::string_printf(error_msg, str));