Base64 encoding applied to different unicode texts always yields different base64 texts ... true or false?

Steffen Nurpmeso via Unicode unicode at
Mon Oct 15 14:26:24 CDT 2018

Philippe Verdy via Unicode wrote in <CAGa7JC05XOjm87J1k92_moRyWWgfDfcbH5\
piMYTYcDLtbS4VZA at>:
 |Padding itself does not clearly indicate the length.
 |It's an artefact that **may** be infered only in some other layers \
 |of protocols which specify when and how padding is needed (and how \
 |many padding bytes 
 |are required or accepted), it works only if these upper layer protocols \
 |are using **octets** streams, but it is still not usable for more general 
 |bitstreams (with arbitrary bit lengths).
 |This RFC does not mandate/require these padding bytes and in fact many \
 |upper layer protocols do not ever need it (including UTF-7 for example), \
 |they are 
 |never necessary to infer a length in octets and insufficient for specify\
 |ing a length in bits.
 |As well the usage in MIME (where there's a requirement that lines of \
 |headers or in the content body is limited to 1000 bytes) requires free \
 |splitting of 
 |Base64 (there's no agreed maximum length, some sources insist it should \
 |not be more than 72 bytes, others use 80 bytes, but mail forwarding \
 |may add other 
 |characters at start of lines, forcing them to be shorter (leaving for \
 |example a line of 72 bytes+CRLF and another line of 8 bytes+CRLF): \
 |this means that 
 |padding may not be used where one would expect them, and padding can \
 |event occur in the middle of the encoded stream (not just at end) along \

That was actually a bug in my MUA.  Other MUAs were not capable of
decoding this correctly.
Sorry :-(!!

 |with other 
 |whitespaces or separators (like "> " at start of lines in cited messages).

In fact garbage bytes may be embedded explicitly says MIME.
Most handle that right, and skip (silently, maybe not right),
but some explicit base64 decoders fail miserably when such things
are seen (openssl base64, NetBSD base64 decoder (current)), others
do not (busybox base64, for example).

|Der Kragenbaer,                The moon bear,
|der holt sich munter           he cheerfully and one by one
|einen nach dem anderen runter  wa.ks himself off
|(By Robert Gernhardt)

More information about the Unicode mailing list