charset=unknown-8bit

charset=unknown-8bit

[Simon Josefsson]

I get lots of mail that has a C-T header looking like this:

Content-Type: text/plain; charset=unknown-8bit

In XEmacs the mail get rendered correctly since all mail with broken
C-T headers I get seem to be iso-8859-1.

In Emacs all 8bit characters in such mail are displayed as octal
sequences such as \345. Technically, that is probably the correct
thing to do since you can't know how to display them. However, I'd
like them to be rendered as being iso-8859-1 anyway.

So, how do I make Gnus do that?

Are the programs sending `charset=unknown-8bit' broken or is that a
valid thing to do?

(Emacs 20.3, all pGnus versions)

Re: charset=unknown-8bit

[Shenghuo ZHU]

>>>>> "Simon" == Simon Josefsson <jas@pdc.kth.se> writes:

Simon> I get lots of mail that has a C-T header looking like this:
Simon> Content-Type: text/plain; charset=unknown-8bit

Simon> In XEmacs the mail get rendered correctly since all mail with broken
Simon> C-T headers I get seem to be iso-8859-1.

Simon> In Emacs all 8bit characters in such mail are displayed as octal
Simon> sequences such as \345. Technically, that is probably the correct
Simon> thing to do since you can't know how to display them. However, I'd
Simon> like them to be rendered as being iso-8859-1 anyway.

Simon> So, how do I make Gnus do that?

Simon> Are the programs sending `charset=unknown-8bit' broken or is that a
Simon> valid thing to do?

Simon> (Emacs 20.3, all pGnus versions)

In my experience, some mailers send any 8-bit message with their
default charsets. If the charsets are not set, they could be
"unknown-8bit" or some other wired string, or "iso-8859-1". I'd like
them (even "iso-8859-1" in some groups that the messages are written
in other languages) to be rendered to gnus-newsgroup-charset.

-- 
Shenghuo

Re: charset=unknown-8bit

[Lars Magne Ingebrigtsen]

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

> In my experience, some mailers send any 8-bit message with their
> default charsets. If the charsets are not set, they could be
> "unknown-8bit" or some other wired string, or "iso-8859-1". I'd like
> them (even "iso-8859-1" in some groups that the messages are written
> in other languages) to be rendered to gnus-newsgroup-charset.

Right...  Hm.  I see two approaches:

1) Have a mapping from given charsets to what it really probably is
   (unknown-8bit -> iso-8859-1, or iso-8859-1 -> iso-2022-jp) in mm.

2) Have a Gnus regexp that says "if the given charset matches this,
   use the default charset instead".

Version 1) has the advantage that it will work anywhere, and version
2) has the advantage that it only works in some groups.

What would be most useful?  Hm.  I think number 2.  I'll add a new
variable, `gnus-newsgroup-ignored-charsets', that will default to
"unknown-8bit". 

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