Benchmarking with a user defined pattern gives an ERROR result if the user defined pattern is not in the text being searched.
This is because smart assumes that there must be a match in the text (as the normal process is to extract the pattern from the text). A count of zero is interpreted as ERROR, even if that is actually correct.
Fix is probably (when using a user-defined pattern) to perform a brute force search and compare the count results of the algorithms with the actual result. This could be done once per text , as the pattern does not change.
Benchmarking with a user defined pattern gives an ERROR result if the user defined pattern is not in the text being searched.
This is because smart assumes that there must be a match in the text (as the normal process is to extract the pattern from the text). A count of zero is interpreted as ERROR, even if that is actually correct.
Fix is probably (when using a user-defined pattern) to perform a brute force search and compare the count results of the algorithms with the actual result. This could be done once per text , as the pattern does not change.