broken "[Bb]oundary=" recognition?

broken "[Bb]oundary=" recognition?

[Karl Kleinpaste]

There is an article, news:9jrinq0dcf@enews4.newsguy.com, which
purports to be:

	 Mime-Version: 1.0
	 Content-Type: multipart/alternative;
		Boundary=----=DD12CF7A829611D599AF00104B9FF285

containing text/plain and text/html -- pretty typical.  Yet it displays
as though it were all text/plain, with all the internal MIME headers
visible.  I'm wondering whether Gnus is failing to notice a valid
"boundary=" due to the word being capitalized in the message header.

Is it legitimate for "Boundary=" to be capitalized?  I'd like to think
MIME keywords are supposed to be considered case-insensitively.

FTR, running Oort CVS-updated as of 5 minutes ago.

Re: broken "[Bb]oundary=" recognition?

[ShengHuo ZHU]

Karl Kleinpaste <karl@charcoal.com> writes:

> There is an article, news:9jrinq0dcf@enews4.newsguy.com, which
> purports to be:
> 
> 	 Mime-Version: 1.0
> 	 Content-Type: multipart/alternative;
> 		Boundary=----=DD12CF7A829611D599AF00104B9FF285
> 
> containing text/plain and text/html -- pretty typical.  Yet it displays
> as though it were all text/plain, with all the internal MIME headers
> visible.  I'm wondering whether Gnus is failing to notice a valid
> "boundary=" due to the word being capitalized in the message header.
> 
> Is it legitimate for "Boundary=" to be capitalized?  I'd like to think
> MIME keywords are supposed to be considered case-insensitively.

It is not legitimate, not because of the capitalized Boundary, but
because the boundary value contains special character =.  It should be
a quoted-string.

ShengHuo

Re: broken "[Bb]oundary=" recognition?

[Lars Magne Ingebrigtsen]

ShengHuo ZHU <zsh@cs.rochester.edu> writes:

>> 	 Content-Type: multipart/alternative;
>> 		Boundary=----=DD12CF7A829611D599AF00104B9FF285

[...]

> It is not legitimate, not because of the capitalized Boundary, but
> because the boundary value contains special character =.  It should be
> a quoted-string.

This is probably the most common way MIME handling "breaks" for Gnus
-- somebody uses a "=" character in a non-quoted separator.

Should we try to parse these invalid messages, or should we just let
the user intervene manually with `K m' to repair the message?

-- 
(domestic pets only, the antidote for overdose, milk.)
   larsi@gnus.org * Lars Magne Ingebrigtsen