If you want a real boost you should consider another data-structure, instead of \0 as separator encode the length of the coming snippet. like this you can avoid tests which cross borders and you can take the value as offset to the next snippet.
In other words when checking if a 300 byte snippet is included within a 400 byte one you only need to check 100 possible slides.
Cheers Rolf
In reply to Re^16: list of unique strings, also eliminating matching substrings
by LanX
in thread list of unique strings, also eliminating matching substrings
by lindsay_grey
| For: | Use: | ||
| & | & | ||
| < | < | ||
| > | > | ||
| [ | [ | ||
| ] | ] |