Front page | perl.perl5.porters |
Postings from January 2017
On 12/31/2016 10:53 PM, David Nicol wrote: > iven the above, if the regex compiler knows that for a match on c[0] > to result in a match on the whole pattern, c[1] ... c[7] are going to > have to match too, it could make sense to deal with the alignment > issues somehow and then compare 64 bits at once. Outside of that > situation -- long sequences where they all must match to continue on > that path in the FSA -- I don't see doing vector comparisons helping > any, especially when the engine would have to backtrack and recheck > matches when ambiguity might get encoded as a loose test. If we have a sequence that must all match, it seems to me we should be using memcmp() regardless of alignment.Thread Previous | Thread Next