Thomas Klausner wrote:

> Dillo and I have added base64 support to uu{de,en}code, per IEEE
> Std 1003.1-2004. The diff and a tarball with regression tests are
> attached. (The regression Makefile will need some more work.)
> While there, we also added support for -o to uudecode and made
> some minor style fixes.
> 
> Since uudecode is a host tool, we'd like to know if this breaks
> any hosts, we only tested on NetBSD/i386.
> 
> Any comments before we commit?

The removal of optimization from the decode_uu() loop isn't a style
fix, but the compiler should handle that.

There should be no magical treatment of the "/dev/stdout" pathname in
decode(), nor should that be part of the manual page. On a related
note, given the availability of the -p flag since 1999 you should
reconsider whether there's an obvious need to add the ambiguity of
a "-" option argument meaning stdout, which also isn't supported for
the input file operand.

There are several KNF problems with regard to the use of binary
operators in the changes.


- Klaus