The function used here is from the leetcode. Details can be found in leetcode problem: Implement strStr()
The best explanation should be made in the comments, which can be understood by the leading of code.
// next[j]: the smallest valid position we need to check next when detect mismatch at jth character pattern[j]
// Here, valid position means "pattern[0, ..., next[j]-1]" are matched with "text".
void getNext(char *pattern, int next[]){
int i = 0, j = -1;
// If the "text[i]" fails to match "patter[0]", then we need to
// check "text[i+1]" and "patter[0]", which also means "text[i]"
// would check with "patter[-1]".
next[i] = j;
// while loop 1:
while(pattern[i] != '\0){
// while loop 2:
while(j >= 0 && pattern[i] != pattern[j]){
// First, j need to be valid index, so it needs to be not less than 0.
// Then, if "pattern[i]" fails to match "pattern[j]", we can also think as
// "text[i]" fails to match "pattern[j]".
// So we need to check if "text[i]" matches with "pattern[next[j]]", as next[j]
// is the position we need to check when we fail at position j.
j = next[j];
}
// After the above while loop, we can know that "text[0, ..., i]" matches
// "pattern[0, ..., j]", so we can move one more step for both "text" and "pattern".
++i; ++j;
// For the new i, marked as i_new, we can determine its "next value" now!!
// As we've known that "text[0, ..., i_new - 1]" matches "pattern[0, ..., j_new - 1]",
// if we fail to match at position "text[i_new]", we can move pattern to the j_new position to
// check if "text[i_new]" matches "pattern[j_new]".
// P.S:
// Also, we can know the j_new position is the optimized position. If we can get a valid position j' (valid
// means "pattern[0, ..., j'-1]" are matched with "text") smaller
// than j_new, then we'd get "(j' - 1)" (which is valid at position j'-1) is smaller than "next[j]",
// which is contradicted to the definition of "next" table.
if(pattern[i] == pattern[j])
next[i] = next[j];
else
next[i] = j;
}
}
char *strStr(char *text, char *pattern){
if(NULL == text || NULL == pattern)
return NULL;
if('\0' == pattern[0])
return text;
// i is the pointer of text, j is the pointer of pattern.
int i = 0, j = 0;
char *pos = NULL;
int *next = new int[strlen(pattern) + 1]; // include the '\0'
getNext(pattern, next);
while(text[i] != '\0'){
// Same optimization in getNext(), that is
// if we fail at one position, we may also fail at the
// next position, which means we can continue along the "next" table
// Also, we need the index to be valid first.
while(j >= 0 && text[i] != pattern[j])
j = next[j];
// After the while loop, we can know "text[0, ..., i]" matches "pattern[0, ..., j]"
// So we need to move one more step for both "text" and "pattern".
++i; ++j;
if(pattern[j] == '\0'){
pos = (text + i) - j; // The beginning position in text which corresponding to the matched pattern position.
return pos;
}
}
return pos;
}