1*63d4abf0SagcForce PATTERN to match only whole words. A "whole word" is a 2*63d4abf0Sagcsubstring which either starts at the beginning or the record or is 3*63d4abf0Sagcpreceded by a non-word constituent character. Similarly, the 4*63d4abf0Sagcsubstring must either end at the end of the record or be followed by a 5*63d4abf0Sagcnon-word constituent character. Word-constituent characters are 6*63d4abf0Sagcalphanumerics (as defined by the current locale) and the underscore 7*63d4abf0Sagccharacter. Note that the non-word constituent characters must 8*63d4abf0Sagcsurround the match; they cannot be counted as errors. 9*63d4abf0Sagc 10*63d4abf0SagcSet cost of incorrect characters to NUM. Note that a deletion (a 11*63d4abf0Sagcmissing character) and an insertion (an extra character) together 12*63d4abf0Sagcconstitute a substituted character, but the cost will be the that of a 13*63d4abf0Sagcdeletion and an insertion added together. Thus, if the const of a 14*63d4abf0Sagcsubstitution is set to be larger than the sum of the costs of deletion 15*63d4abf0Sagcand insertion, direct substitutions will never be done. 16*63d4abf0Sagc 17*63d4abf0SagcSet the record delimiter regular expression to PATTERN. The text 18*63d4abf0Sagcbetween two delimiters, before the first delimiter, and after the last 19*63d4abf0Sagcdelimiter is considered to be a record. The default record delimiter 20*63d4abf0Sagcis the regexp "\n", so by default a record is a line. PATTERN can be 21*63d4abf0Sagcany regular expression that does not match the empty string. For 22*63d4abf0Sagcexample, using -d "^From " defines mail messages as records in a 23*63d4abf0SagcMailbox format file. 24