boundary analysis test case on string match algorithm

boundary test is intuitive for numerical input, but when I try to do some boundary analysis on some string match algorithms (e.g boyer moore), I find it is hard to do the boundary analysis and derive some test cases for it.

for example, I have two equivalent classes: EC1: length(pattern) = length(text), pattern occurs in text. EC2: length(pattern) = length(text), pattern occurs not in text. So in my opinion, the boundary values are hard to determine cos it is not a numerical boundary, thus can I say boundary : pattern occurs in the text. on point: pattern occurs in the text. off point: pattern not occurs in the text