Re: header un-folding & encoded headers

it's simple:

1. white space between two adjacent encoded-words is not displayed.
2. white space between an encoded-word and an ordinary rfc822 atom is 
displayed.

Keith


Paul Smith wrote:
> 
> I'm just struggling slightly to work out how you're meant to handle the 
> following situation. It seems a bit inconsistent, so I want to check I'm 
> interpreting it right.
> 
> If I have a header
> 
> Subject: hello
> there
> (no space after 'hello' and one space before 'there')
> 
> I should unwrap that to
> 
> Subject: hello there
> 
> (one space between 'hello' and 'there' - ie just remove the CRLF as 
> stated in RFC 2822 2.2.3)
> Yes?
> 
> ----
> 
> If I have a header
> 
> Subject: =?us-ascii?Q?hello?= =?us-ascii?Q?there?=
> 
> (two encoded words with no spaces in them, and a space between them) I 
> should decode that to
> 
> Subject: hello there
> 
> Yes?
> (to be honest, I'm not entirely sure about this because 'Subject: 
> =?us-ascii?Q?hello?=there' (without a space) is illegal, so if you 
> wanted that would you use 'Subject: =?us-ascii?Q?hello?= there' instead 
> and expect the space to be stripped out or can you just not do it - 
> section 5(1) of RFC 2047 isn't clear whether the separating LWS is 
> retained or not)
> 
> ----
> If I have a header
> 
> Subject: =?us-ascii?Q?hello?=
> =?us-ascii?Q?there?=
> (no space after the encoded 'hello' and one space before the encoded 
> 'there')
> 
> As I read it, I should decode it to
> Subject: hellothere
> i.e. in this case I should discard the CRLF *and* the space, so I can't 
> just unwrap it, and then decode as the space would be kept then.
> 
> Yes?
> 
> ----
> If so, what about
> 
> Subject: hello
> =?us-ascii?Q?there?=
> 
> or
> 
> Subject: =?us-ascii?Q?hello?=
> there
> 
> Do I keep the space after the CRLF in these situations or not?
> 
> RFC 2047 seems to ignore this little side effect of header folding, 
> although it says that wrapping should be used to make sure that no 
> encoded-word is longer than 76 characters.
> 
>