Bruno Haible <br...@clisp.org> writes: > Jim Meyering wrote: >> > Could also be done, but since the code to remove newlines was there, I >> > thought it is simpler to augment it to remove common garbage. >> >> I am reluctant to make this low-level decoding function ignore anything >> else. > > I agree. If you remove the garbage unconditionally, you have lost the > functionality of doing strict checking of the input. If you add a 'bool' > parameter that determines whether the remove it, you have added complexity > in the interface (.h file) for nearly nothing.
I agree too -- there are situations where you really don't want to ignore non-alphabetic characters. Pre-filtering seems like the right thing here. Possibly, the base64 module could provide a separate function for doing this pre-filtering though? That doesn't add significant complexity to the code, but it makes it easy for callers to use. /Simon
