Thanks for paying attention to this.
I want to know why will we meet the following situation during discretizng the left-most-suffix-substrings(I'll call them lms substrings):the two substrings are the same,one begin from the $$$i$$$ th character and the other begin from the $$$j$$$ th,for the $$$i+k$$$ and the the $$$j+k$$$ th place($$$k$$$ is smaller than the length of the substrings),one leads a smaller suffix than the next character(I'll call this S type) ans the other is opposite (L type).
I think this situation won't exist,which is different from the most articles I read online.Here is my reason: if there is a different character in the substrings after the $$$i+k$$$ and the $$$j+k$$$ th place,the type should be the same.For example,if the substrings are all aaaab...,the character b will make the first four characters in the two substrings S type,which is different from the assertion (from one of them is S and from the other is L).Otherwise,if the characters in the substrings after these two places are the same,for example,...aaaaaaaa,then the as should all be S or L type,but according to the defination, the lms substrings should have an L type character on the second last place and an S one on the last place(excepct for the last character we put by ourselves,which is smaller than other characters,obviously conflicts the assertion).In concluesion,if the two substrings are the same,every character on the same place must have the same type.
I'm not sure about whether what I said is true or not, but if I was right,can we discretize the substrings without considering whether the types are the same or not?Otherwise,can you please help me proof that this situation may exist?Thank you.
A few days ago, I submitted a code from shadowice1984's blog without the part of considering the following situation, and it was Accepted.
(I apologize for my poor English.If you know Chinese please view the Chinese discuss I posted on luogu )