--- In [email protected], "swzoh" <sean...@...> wrote: > > --- In [email protected], "brucexs" <bswitzer@> wrote: > > > > I don't want to overcomplicate things. It only checks after finding a lead > > byte, and if the combination of that lead byte and next byte is not valid > > dbcs pair, then it replaces both by a blank. > > > > I didn't mean to make things complicated, on the contrary, I think it makes > the job simpler. With the current implementation, when mblen returns -1, > always have to check first if the second byte is NULL. But, if only first > byte is dropped/replaced, don't have to. Automatically we will get 0 at the > next test. So, the routine would be almost identical for both the DBCS & > SBCS, except when mblen returns 2. >
You are right that does save a couple of lines of code -- I uploaded a d6 which tries that approach. (Actually, it was checking for second byte zero before the mlen check, since I just kept what was working before).
