Paul Eggert wrote: > > if (!mbi_avail (iter)) > > - abort (); > > + /* We can get here due to incomplete multibyte characters. > > */ > > + return false; > > mbi_advance (iter); > > If the string ends in an incomplete sequence and the pattern is the > empty string, won't this return false? Shouldn't the empty string be a > suffix of any string?
There's a comment /* Here, suffix is not empty. */ a few lines above. Also, feel free to add new test cases to the unit tests, if you think I missed to test some cases. Bruno