To avoid time zone issues and catch duplicates with non-identical times, it looks like the program generates a hashes and compares them, meaning with the "ignore time" option, it'll delete every "hi" message you send a certain person, which obviously isn't the intent of message deduplication.
Instead, how about looking through each thread for identical messages less than n messages ago? Normally duplicates occur either back-to-back or immediately after a long message is broken into parts.
To avoid time zone issues and catch duplicates with non-identical times, it looks like the program generates a hashes and compares them, meaning with the "ignore time" option, it'll delete every "hi" message you send a certain person, which obviously isn't the intent of message deduplication.
Instead, how about looking through each thread for identical messages less than n messages ago? Normally duplicates occur either back-to-back or immediately after a long message is broken into parts.