From: Mike Nordell (tamlin@algonet.se)
Date: Fri Sep 26 2003 - 02:47:05 EDT
I just came to think of something. It's not a suggestion or something,
merely an Idea I got that perhaps someone would like to test.
Pseudoish code:
Prop* FindStringInSet(const char* szProp, PropSet& propset)
{
curr = propset.begin();
end = propset.end();
for (; curr != end; ++curr)
{
const char* pszVal = curr->val;
if (pszVal[0] == szProp[0] && !str(i)cmp(pszVal, szProp)) {
return curr;
}
}
return 0;
}
It doesn't change the complexity of the search, but it wouldn't surprise me
the least if the short-circuit check for first char matching could make
difference. I recall having made something similar in the past which proved
to have an *extreme* effect (2-3 orders of magnitude IIRC).
/Mike
This archive was generated by hypermail 2.1.4 : Fri Sep 26 2003 - 03:02:11 EDT