Weird subject wrapping

Weird subject wrapping

[Hannu Koivisto]

Greetings,

Gnus 5.8.8 from CVS, Emacs 20.7.  For some reason, subject line
"Subject: Auttakaahan DR:n koneen kanssa, Murphy iski jälleen"
ends up being
"Subject: Auttakaahan DR:n koneen kanssa, Murphy iski
  =?iso-8859-1?q?j=E4lleen?="
in the article that is actually sent to news server.  I've faced a
similar case before; also that time the line after split started
with =?iso... stuff.

Is this some sort of a feature (how to disable it?) or a bug?  That
is, what I'm worried about is the wrapping plus those two extra
spaces (so that Gnus displays the subject "...iski   jälleen", which
is ugly and I don't even dare to imagine how other news readers
display that), not the encoding which of course I know is a feature.

-- 
Hannu

Re: Weird subject wrapping

[Kai Großjohann]

On 08 Sep 2000, Hannu Koivisto wrote:

> Is this some sort of a feature (how to disable it?) or a bug?  That
> is, what I'm worried about is the wrapping plus those two extra
> spaces (so that Gnus displays the subject "...iski   jälleen", which
> is ugly and I don't even dare to imagine how other news readers
> display that), not the encoding which of course I know is a feature.

Gnus tries to avoid long lines, and in the example you showed the
Subject line would have been longer than 70 (or was it 72, or 78?)
characters without the breaking.

I think that MUAs are supposed to remove the leading whitespace from
continuation lines.  Or are they supposed to replace that with a
single space?  I'm not sure.

kai
-- 
I like BOTH kinds of music.

Re: Weird subject wrapping

[Paul Jarc]

Kai.Grossjohann@CS.Uni-Dortmund.DE (Kai =?iso-8859-1?q?Gro=DFjohann?=) writes:
> I think that MUAs are supposed to remove the leading whitespace from
> continuation lines.  Or are they supposed to replace that with a
> single space?  I'm not sure.

RFC822 (3.1.1) says: "Unfolding is accomplished by regarding CRLF
immediately followed by a LWSP-char as equivalent to the LWSP-char."

So the leading whitespace shouldn't be removed, which means extra
whitespace, other than the line break, shouldn't have been inserted.
(Assuming that this isn't another case of RFC822 being out of line
with reality.)


paul

Re: Weird subject wrapping

[Kai Großjohann]

On 08 Sep 2000, Paul Jarc wrote:

> Kai.Grossjohann@CS.Uni-Dortmund.DE (Kai
> =?iso-8859-1?q?Gro=DFjohann?=) writes:
>> I think that MUAs are supposed to remove the leading whitespace
>> from continuation lines.  Or are they supposed to replace that with
>> a single space?  I'm not sure.
> 
> RFC822 (3.1.1) says: "Unfolding is accomplished by regarding CRLF
> immediately followed by a LWSP-char as equivalent to the LWSP-char."

Ick!  Having the following in a message header looks really awful:

Subject: this is a long
 subject continued on the next line

I'd prefer something like

Subject: this is a long
         subject continued on the next line

But I can't use that with the RFC :-|

Darn.

kai
-- 
I like BOTH kinds of music.

Re: Weird subject wrapping

[Paul Jarc]

Kai.Grossjohann@CS.Uni-Dortmund.DE (Kai =?iso-8859-1?q?Gro=DFjohann?=) writes:
> On 08 Sep 2000, Paul Jarc wrote:
> > RFC822 (3.1.1) says: "Unfolding is accomplished by regarding CRLF
> > immediately followed by a LWSP-char as equivalent to the LWSP-char."
> 
> Ick!  Having the following in a message header looks really awful:
> 
> Subject: this is a long
>  subject continued on the next line
> 
> I'd prefer something like
> 
> Subject: this is a long
>          subject continued on the next line

You want it to *appear* that way.  You don't care how it's stored,
right?  So there's no problem.  Gnus could be made to display long
fields however you like; if it's a lossy conversion, then that's
something you take into consideration when deciding how you want Gnus
to display long fields.  (I'm not claiming that Gnus already supports
this configurability, just that it could be added if it's not there
already, and that (I think) it would satisfy your desire.)


paul

Re: Weird subject wrapping

[Kai Großjohann]

I was talking about the composition buffer.  Hm.  But maybe you're
right anyway.  When writing the message, one maybe shouldn't look at
the header section too closely...
kai
-- 
I like BOTH kinds of music.

Re: Weird subject wrapping

[Paul Jarc]

Kai.Grossjohann@CS.Uni-Dortmund.DE (Kai =?iso-8859-1?q?Gro=DFjohann?=) writes:
> I was talking about the composition buffer.

Ah.  In that case, then, arbitrary transformations could be applied to
the message before sending - before it enters the domain of RFC822.


paul

Re: Weird subject wrapping

[Hannu Koivisto]

prj@po.cwru.edu (Paul Jarc) writes:

| RFC822 (3.1.1) says: "Unfolding is accomplished by regarding CRLF
| immediately followed by a LWSP-char as equivalent to the LWSP-char."
| 
| So the leading whitespace shouldn't be removed, which means extra
| whitespace, other than the line break, shouldn't have been inserted.

You mean "other than the line break and one space"?  And, in
addition, Gnus shouldn't insert extra whitespace to the subject
display.  I.e. now, a subject that is ok by that RFC,

"""
Subject: COMMERCIAL: Extensys 1.0.1 Released. - Linux based platform for
 developing communications applications
"""

shows in the summary buffer like this:

"...for  developing..."

-- 
Hannu

Re: Weird subject wrapping

[Paul Jarc]

Hannu Koivisto <azure@iki.fi> writes:
> prj@po.cwru.edu (Paul Jarc) writes:
> 
> | RFC822 (3.1.1) says: "Unfolding is accomplished by regarding CRLF
> | immediately followed by a LWSP-char as equivalent to the LWSP-char."
> | 
> | So the leading whitespace shouldn't be removed, which means extra
> | whitespace, other than the line break, shouldn't have been inserted.
> 
> You mean "other than the line break and one space"?

No, I meant what I said.  If "foo bar" is encoded as "foo<CRLF> bar"
(only a line break is inserted, before the existing space), then it
will be decoded as "foo bar", according to RFC822.  If it is encoded
instead as "foo<CRLF>  bar" (a line break and a space are inserted),
then it will be decoded as "foo  bar".

> And, in addition, Gnus shouldn't insert extra whitespace to the
> subject display.  I.e. now, a subject that is ok by that RFC,
> 
> """
> Subject: COMMERCIAL: Extensys 1.0.1 Released. - Linux based platform for
>  developing communications applications
> """
> 
> shows in the summary buffer like this:
> 
> "...for  developing..."

Yes, there should be only one space there, according to the RFC.


paul