[Top] [All Lists]

Re: Forbidding 8bit text in Armor comments

2001-08-02 11:30:09

On 2001-08-02 17:25:49 +0100, Ian Bell wrote:

If utf-8 is allowed in Comments then RFC2015bis will need to be amended (if it hasn't got through last call yet) to tell clients what to do with it before sending.

It has gone through last call.

Double encoding is not a sensible option

No, of course not.

- either say RFC2015bis clients MUST NOT include comments lines with utf-8 characters, or say that RFC2015bis clients MAY drop utf-8 comments lines, but if they don't they MUST RFC2047 encode them.

This is plain silliness - just like double encoding. We are talking about a comment which is:

- hardly ever read by users, because it's not even on the application layer - mostly used to advertise the program which is used for generating it to those who wonder what kind of "data junk" they are receiving
- is not cryptographically protected
- and, when using utf-8 text, breaks assumptions which have been made for years now, and, in fact, breaks the whole point ASCII armor was designed for

The discussion we are having here is basically equivalent to discussing internationalization of SMTP greetings or User-Agent headers.

If you ask me, I'd just suggest to drop the Comment armor header. Since that's probably not an option, castrate it to 7bit characters.

Thomas Roessler              

<Prev in Thread] Current Thread [Next in Thread>