/* This program is free software; you can redistribute it and/or modify it
* under the terms of the license (GNU LGPL) which comes with this package.
*/
-
-/* Returns the string without leading whitespaces (xbt_str_ltrim),
- * trailing whitespaces (xbt_str_rtrim),
- * or both leading and trailing whitespaces (xbt_str_trim).
- * (in-place modification of the string)
- */
#include "xbt/misc.h"
#include "xbt/sysdep.h"
#include "portable.h"
#include "xbt/matrix.h" /* for the diff */
-static void free_string(void *d){
- free(*(void**)d);
-}
-
/** @brief Strip whitespace (or other characters) from the end of a string.
*
* Strips the whitespaces from the end of s.
while(*cur && white_char[(unsigned char)*cur])
++cur;
- memmove(s,cur, strlen(cur));
+ memmove(s,cur, strlen(cur)+1);
}
/** @brief Strip whitespace (or other characters) from the end and the begining of a string.
* - "\x0B" (ASCII 11 (0x0B)) vertical tab.
*/
-xbt_dynar_t xbt_str_split(char *s, const char *sep) {
- char *str=xbt_strdup(s);
- xbt_dynar_t res = xbt_dynar_new(sizeof(char*), free_string);
- char *p, *q;
+xbt_dynar_t xbt_str_split(const char *s, const char *sep) {
+ xbt_dynar_t res = xbt_dynar_new(sizeof(char*), xbt_free_ref);
+ const char *p, *q;
int done;
const char* sep_dflt = " \t\n\r\x0B";
char is_sep[256] = {1,0};
is_sep[0] = 1; /* End of string is also separator */
/* Do the job */
- p=q=str;
+ p=q=s;
done=0;
+
+ if (s[0] == '\0')
+ return res;
+
while (!done) {
char *topush;
while (!is_sep[(unsigned char)*q]) {
q++;
}
- if (*q == '\0') {
-#ifdef UNDEF
- if (p==q) {
- /* do not push last empty line */
- free(str);
- return res;
- }
-#endif
+ if (*q == '\0')
done = 1;
- } else {
- *q='\0';
- }
- topush=xbt_strdup(p);
+
+ topush=xbt_malloc(q-p+1);
+ memcpy(topush,p,q-p);
+ topush[q - p] = '\0';
xbt_dynar_push(res,&topush);
p = ++q;
}
- free (str);
+
+ return res;
+}
+
+/**
+ * \brief This functions splits a string after using another string as separator
+ * For example A!!B!!C splitted after !! will return the dynar {A,B,C}
+ * \return An array of dynars containing the string tokens
+*/
+xbt_dynar_t xbt_str_split_str(const char *s, const char *sep) {
+ xbt_dynar_t res = xbt_dynar_new(sizeof(char*), xbt_free_ref);
+ int done;
+ const char *p, *q;
+
+ p = q = s;
+ done = 0;
+
+ if (s[0] == '\0')
+ return res;
+ if (sep[0] == '\0') {
+ s = xbt_strdup(s);
+ xbt_dynar_push(res, &s);
+ return res;
+ }
+
+ while (!done) {
+ char *to_push;
+ int v = 0;
+ //get the start of the first occurence of the substring
+ q = strstr(p, sep);
+ //if substring was not found add the entire string
+ if (NULL == q) {
+ v = strlen(p);
+ to_push = malloc(v + 1);
+ memcpy(to_push, p, v);
+ to_push[v] = '\0';
+ xbt_dynar_push(res, &to_push);
+ done = 1;
+ }
+ else {
+ //get the appearance
+ to_push = malloc(q - p + 1);
+ memcpy(to_push, p, q - p);
+ //add string terminator
+ to_push[q - p] = '\0';
+ xbt_dynar_push(res, &to_push);
+ p = q +strlen(sep);
+ }
+ }
+ return res;
+}
+
+/** @brief Splits a string into a dynar of strings, taking quotes into account
+ *
+ * It basically does the same argument separation than the shell, where white
+ * spaces can be escaped and where arguments are never splitted within a
+ * quote group.
+ * Several subsequent spaces are ignored (unless within quotes, of course).
+ *
+ */
+
+xbt_dynar_t xbt_str_split_quoted(const char *s) {
+ xbt_dynar_t res = xbt_dynar_new(sizeof(char*), xbt_free_ref);
+ char *str; /* we have to copy the string before, to handle backslashes */
+ char *beg, *end; /* pointers around the parsed chunk */
+ int in_simple_quote=0, in_double_quote=0;
+ int done = 0;
+ int ctn = 0; /* Got something in this block */
+
+ if (s[0] == '\0')
+ return res;
+ beg = str = xbt_strdup(s);
+
+ /* trim leading spaces */
+ xbt_str_ltrim(beg," ");
+ end=beg;
+
+ while (!done) {
+
+
+ switch (*end) {
+ case '\\':
+ ctn = 1;
+ /* Protected char; move it closer */
+ memmove(end,end+1,strlen(end));
+ if (*end=='\0')
+ THROW0(arg_error,0,"String ends with \\");
+ end++; /* Pass the protected char */
+ break;
+
+ case '\'':
+ ctn = 1;
+ if (!in_double_quote) {
+ in_simple_quote = !in_simple_quote;
+ memmove(end,end+1,strlen(end));
+ } else {
+ /* simple quote protected by double ones */
+ end++;
+ }
+ break;
+ case '"':
+ ctn = 1;
+ if (!in_simple_quote) {
+ in_double_quote = !in_double_quote;
+ memmove(end,end+1,strlen(end));
+ } else {
+ /* double quote protected by simple ones */
+ end++;
+ }
+ break;
+
+ case ' ':
+ case '\t':
+ case '\n':
+ case '\0':
+ if (*end == '\0' && (in_simple_quote || in_double_quote)) {
+ THROW2(arg_error,0,
+ "End of string found while searching for %c in %s",
+ (in_simple_quote?'\'':'"'),
+ s);
+ }
+ if (in_simple_quote || in_double_quote) {
+ end++;
+ } else {
+ if (ctn) {
+ /* Found a separator. Push the string if contains something */
+ char *topush=xbt_malloc(end-beg+1);
+ memcpy(topush,beg,end-beg);
+ topush[end - beg] = '\0';
+ xbt_dynar_push(res,&topush);
+ }
+ ctn= 0;
+
+ if (*end == '\0') {
+ done = 1;
+ break;
+ }
+
+ beg=++end;
+ xbt_str_ltrim(beg," ");
+ end=beg;
+ }
+ break;
+
+ default:
+ ctn = 1;
+ end++;
+ }
+ }
+ free(str);
return res;
}
+#ifdef SIMGRID_TEST
+#include "xbt/str.h"
+
+#define mytest(name, input, expected) \
+ xbt_test_add0(name); \
+ d=xbt_str_split_quoted(input); \
+ s=xbt_str_join(d,"XXX"); \
+ xbt_test_assert3(!strcmp(s,expected),\
+ "Input (%s) leads to (%s) instead of (%s)", \
+ input,s,expected);\
+ free(s); \
+ xbt_dynar_free(&d);
+
+XBT_TEST_SUITE("xbt_str","String Handling");
+XBT_TEST_UNIT("xbt_str_split_quoted",test_split_quoted, "test the function xbt_str_split_quoted") {
+ xbt_dynar_t d;
+ char *s;
+
+ mytest("Empty", "", "");
+ mytest("Basic test", "toto tutu", "totoXXXtutu");
+ mytest("Useless backslashes", "\\t\\o\\t\\o \\t\\u\\t\\u", "totoXXXtutu");
+ mytest("Protected space", "toto\\ tutu", "toto tutu");
+ mytest("Several spaces", "toto tutu", "totoXXXtutu");
+ mytest("LTriming", " toto tatu", "totoXXXtatu");
+ mytest("Triming", " toto tutu ", "totoXXXtutu");
+ mytest("Single quotes", "'toto tutu' tata", "toto tutuXXXtata");
+ mytest("Double quotes", "\"toto tutu\" tata", "toto tutuXXXtata");
+ mytest("Mixed quotes", "\"toto' 'tutu\" tata", "toto' 'tutuXXXtata");
+ mytest("Backslashed quotes", "\\'toto tutu\\' tata", "'totoXXXtutu'XXXtata");
+ mytest("Backslashed quotes + quotes", "'toto \\'tutu' tata", "toto 'tutuXXXtata");
+
+}
+
+#define mytest_str(name, input, separator, expected) \
+ xbt_test_add0(name); \
+ d=xbt_str_split_str(input, separator); \
+ s=xbt_str_join(d,"XXX"); \
+ xbt_test_assert3(!strcmp(s,expected),\
+ "Input (%s) leads to (%s) instead of (%s)", \
+ input,s,expected);\
+ free(s); \
+ xbt_dynar_free(&d);
+
+XBT_TEST_UNIT("xbt_str_split_str",test_split_str, "test the function xbt_str_split_str") {
+ xbt_dynar_t d;
+ char *s;
+
+ mytest_str("Empty string and separator", "", "", "");
+ mytest_str("Empty string", "", "##", "");
+ mytest_str("Empty separator", "toto", "", "toto");
+ mytest_str("String with no separator in it", "toto", "##", "toto");
+ mytest_str("Basic test", "toto##tutu", "##", "totoXXXtutu");
+}
+#endif /* SIMGRID_TEST */
+
/** @brief Join a set of strings as a single string */
char *xbt_str_join(xbt_dynar_t dyn, const char*sep) {
- int len=2;
- int cpt;
+ int len=1,dyn_len=xbt_dynar_length(dyn);
+ unsigned int cpt;
char *cursor;
char *res,*p;
+
+ if (!dyn_len)
+ return xbt_strdup("");
+
/* compute the length */
xbt_dynar_foreach(dyn,cpt,cursor) {
len+=strlen(cursor);
}
- len+=strlen(sep)*(xbt_dynar_length(dyn)-1);
+ len+=strlen(sep)*dyn_len;
/* Do the job */
res = xbt_malloc(len);
p=res;
xbt_dynar_foreach(dyn,cpt,cursor) {
- p+=sprintf(p,"%s%s",cursor,sep);
+ if ((int)cpt<dyn_len-1)
+ p+=sprintf(p,"%s%s",cursor,sep);
+ else
+ p+=sprintf(p,"%s",cursor);
}
return res;
}
-#ifndef HAVE_GETLINE
+#if !defined(HAVE_GETLINE) || defined(DOXYGEN)
+/* prototype here, just in case */
+long getline(char **buf, size_t *n, FILE *stream);
+
+/** @brief Get a single line from the stream (reimplementation of the GNU getline)
+ *
+ * This is a redefinition of the GNU getline function, used on platforms where it does not exists.
+ *
+ * getline() reads an entire line from stream, storing the address of the buffer
+ * containing the text into *buf. The buffer is null-terminated and includes
+ * the newline character, if one was found.
+ *
+ * If *buf is NULL, then getline() will allocate a buffer for storing the line,
+ * which should be freed by the user program. Alternatively, before calling getline(),
+ * *buf can contain a pointer to a malloc()-allocated buffer *n bytes in size. If the buffer
+ * is not large enough to hold the line, getline() resizes it with realloc(), updating *buf and *n
+ * as necessary. In either case, on a successful call, *buf and *n will be updated to
+ * reflect the buffer address and allocated size respectively.
+ */
long getline(char **buf, size_t *n, FILE *stream) {
- int i, ch;
+ size_t i;
+ int ch;
if (!*buf) {
*buf = xbt_malloc(512);
static xbt_matrix_t diff_build_LCS(xbt_dynar_t da, xbt_dynar_t db) {
xbt_matrix_t C = xbt_matrix_new(xbt_dynar_length(da),xbt_dynar_length(db),
sizeof(int),NULL);
- int i,j;
+ unsigned long i,j;
+
/* Compute the LCS */
/*
C = array(0..m, 0..n)
C[i,j] := max(C[i,j-1], C[i-1,j])
return C[m,n]
*/
- for (i=0; i<xbt_dynar_length(da); i++)
- *((int*) xbt_matrix_get_ptr(C,i,0) ) = 0;
+ if (xbt_dynar_length(db) != 0)
+ for (i=0; i<xbt_dynar_length(da); i++)
+ *((int*) xbt_matrix_get_ptr(C,i,0) ) = 0;
- for (j=0; j<xbt_dynar_length(db); j++)
- *((int*) xbt_matrix_get_ptr(C,0,j) ) = 0;
+ if (xbt_dynar_length(da) != 0)
+ for (j=0; j<xbt_dynar_length(db); j++)
+ *((int*) xbt_matrix_get_ptr(C,0,j) ) = 0;
for (i=1; i<xbt_dynar_length(da); i++)
for (j=1; j<xbt_dynar_length(db); j++) {
diff_build_diff(res,C,da,db,i-1,j-1);
topush = bprintf(" %s",xbt_dynar_get_as(da,i,char*));
xbt_dynar_push(res, &topush);
- } else if (j>0 &&
- (i<=0 || xbt_matrix_get_as(C,i,j-1,int) >= xbt_matrix_get_as(C,i-1,j,int))) {
+ } else if (j>=0 &&
+ (i<=0 ||j==0|| xbt_matrix_get_as(C,i,j-1,int) >= xbt_matrix_get_as(C,i-1,j,int))) {
diff_build_diff(res,C,da,db,i,j-1);
topush = bprintf("+ %s",xbt_dynar_get_as(db,j,char*));
xbt_dynar_push(res,&topush);
- } else if (i>0 &&
+ } else if (i>=0 &&
(j<=0 || xbt_matrix_get_as(C,i,j-1,int) < xbt_matrix_get_as(C,i-1,j,int))) {
diff_build_diff(res,C,da,db,i-1,j);
topush = bprintf("- %s",xbt_dynar_get_as(da,i,char*));
} else {
THROW2(arg_error,0,"Invalid values: i=%d, j=%d",i,j);
}
+
}
/** @brief Compute the unified diff of two strings */
xbt_dynar_t db = xbt_str_split(b, "\n");
xbt_matrix_t C = diff_build_LCS(da,db);
- xbt_dynar_t diff = xbt_dynar_new(sizeof(char*),free_string);
+ xbt_dynar_t diff = xbt_dynar_new(sizeof(char*),xbt_free_ref);
char *res=NULL;
diff_build_diff(diff, C, da,db, xbt_dynar_length(da)-1, xbt_dynar_length(db)-1);
+ /* Clean empty lines at the end */
+ while (xbt_dynar_length(diff) > 0) {
+ char *str;
+ xbt_dynar_pop(diff,&str);
+ if (str[0]=='\0' || !strcmp(str," ")) {
+ free(str);
+ } else {
+ xbt_dynar_push(diff,&str);
+ break;
+ }
+ }
res = xbt_str_join(diff, "\n");
xbt_dynar_free(&da);
