gnus sends corrupt attachments

gnus sends corrupt attachments

[kullstam]

i am using pgnus 0.96 and am having some trouble with attachments.
when i attach a binary (e.g., pdf or zip file) the file gets
corrupted.  the problem is with sending since 1) i can receive
attachments sent with other MUAs (mutt or netscape), and 2) other MUAs
(mutt) receive the same corrupted attachment.

i notice by looking at the received binary that i get extra line
breaks preceeded by a !.

here is an excerpt from the tail of a pdf file.  (please excuse long lines)

orig ending

R^M/Info 1 0 R^M/ID [<5c657a9ab424633503301772053391a4><5c657a9ab424633503301772053391a4>]^M>>^Mstartxref^M21150^M%%EOF

gnus mangled ending

R^M/Info 1 0 R^M/ID [<5c657a9ab42463350!
3301772053391a4><5c657a9ab424633503301772053391a4>]^M>>^Mstartxref^M21150^M%%EOF

as you can see, the !<lf> got somehow inserted.  is this a gnus
configuaration problem i can fix by setting the proper variable?  is
this a known gnus sending problem?  i am happy to send my config upon
request.

i'll delve into it, if no one already has a quick solution.

-- 
J o h a n  K u l l s t a m
[kullstam@ne.mediaone.net]
Don't Fear the Penguin!

Re: gnus sends corrupt attachments

[Rene H. Larsen]

kullstam@ne.mediaone.net writes:

> i am using pgnus 0.96 and am having some trouble with attachments.
> when i attach a binary (e.g., pdf or zip file) the file gets
> corrupted.  the problem is with sending since 1) i can receive
> attachments sent with other MUAs (mutt or netscape), and 2) other MUAs
> (mutt) receive the same corrupted attachment.

This is probably because pGnus by default attaches many binary files
using quoted printable encoding rather than the more robust base64
encoding.  Try putting this in your .gnus and see if it helps:

;;; Disable QP for unknown (binary) mime types.
(setq mm-content-transfer-encoding-defaults
  '(("text/x-patch" 8bit)
    ("text/.*" qp-or-base64)
    ("message/rfc822" 8bit)
    ("application/emacs-lisp" 8bit)
    ("application/x-patch" 8bit)
    (".*" base64)))

HTH.
-- 
#!/usr/bin/perl -w
for(0..12){$b[$_]="|".($_==6?"-":" ")x78}for(0..78){substr($b[-
(sin($_*3.1415/22)-1)*6.499],$_,1)="*"}print join "\n",@b,"";

Re: gnus sends corrupt attachments

[Johan Kullstam]

"Rene H. Larsen" <renehl@post1.tele.dk> writes:

> kullstam@ne.mediaone.net writes:
> 
> > i am using pgnus 0.96 and am having some trouble with attachments.
> > when i attach a binary (e.g., pdf or zip file) the file gets
> > corrupted.  the problem is with sending since 1) i can receive
> > attachments sent with other MUAs (mutt or netscape), and 2) other MUAs
> > (mutt) receive the same corrupted attachment.
> 
> This is probably because pGnus by default attaches many binary files
> using quoted printable encoding rather than the more robust base64
> encoding.  Try putting this in your .gnus and see if it helps:
> 
> ;;; Disable QP for unknown (binary) mime types.
> (setq mm-content-transfer-encoding-defaults
>   '(("text/x-patch" 8bit)
>     ("text/.*" qp-or-base64)
>     ("message/rfc822" 8bit)
>     ("application/emacs-lisp" 8bit)
>     ("application/x-patch" 8bit)
>     (".*" base64)))

thank you.  this seems to have worked beautifully.  

in light of the apparent unreliability of quoted-printable, should
this setup be made the default?

-- 
J o h a n  K u l l s t a m
[kullstam@ne.mediaone.net]
Don't Fear the Penguin!

Re: gnus sends corrupt attachments

[Kai Großjohann]

Johan Kullstam <kullstam@ne.mediaone.net> writes:

> in light of the apparent unreliability of quoted-printable, should
> this setup be made the default?

I don't think that QP is unreliable per se.  It's just that long lines
are a problem.  QP provides a way for dealing with long lines: if a
line ends with `=', a continuation line follows.

Maybe QP encoding should make use of this to restrict maximum line
lengths to 70 characters or so?

But maybe that PDF file was an all-ASCII file.  Gnus doesn't do any
encoding when that happens, I think.

kai
-- 
I like BOTH kinds of music.

Re: gnus sends corrupt attachments

[Johan Kullstam]

Kai.Grossjohann@CS.Uni-Dortmund.DE (Kai Großjohann) writes:

> Johan Kullstam <kullstam@ne.mediaone.net> writes:
> 
> > in light of the apparent unreliability of quoted-printable, should
> > this setup be made the default?
> 
> I don't think that QP is unreliable per se.  It's just that long lines
> are a problem.  QP provides a way for dealing with long lines: if a
> line ends with `=', a continuation line follows.
> 
> Maybe QP encoding should make use of this to restrict maximum line
> lengths to 70 characters or so?
> 
> But maybe that PDF file was an all-ASCII file.  Gnus doesn't do any
> encoding when that happens, I think.

the pdf was all ascii but had CR (^M) instead of LF (^J) or CR/LF
(^M^J) to break lines.

the handling of flat text files usually comes with some assumption
about maximum line length.  exceeding that length can cause trouble.
it will almost certainly be bigger than 70.  perhaps 140 would be a
good value?  (i.e., long enough so that regular text sprinkled with
the odd 8 bit value wouldn't normally wrap but short enough such that
tools wouldn't mangle it.)

-- 
J o h a n  K u l l s t a m
[kullstam@ne.mediaone.net]
Don't Fear the Penguin!

Re: gnus sends corrupt attachments

[Robert Bihlmeyer]

Hi,

>>>>> On 19 Sep 1999 22:37:56 +0200
>>>>> Kai.Grossjohann@CS.Uni-Dortmund.DE (Kai Großjohann) said:

 Kai> Maybe QP encoding should make use of this to restrict maximum
 Kai> line lengths to 70 characters or so?

I think patchmaster Shenghou has already provided a fix. We'll just
have to wait (with the help of the workaround) for Larsi to get back
here from Real Life.

        Robbe

-- 
Robert Bihlmeyer	reads: Deutsch, English, MIME, Latin-1, NO SPAM!
<robbe@orcus.priv.at>	<http://stud2.tuwien.ac.at/~e9426626/sig.html>

Re: gnus sends corrupt attachments

[Raja R Harinath]

Robert Bihlmeyer <e9426626@stud4.tuwien.ac.at> writes:
> >>>>> On 19 Sep 1999 22:37:56 +0200
> >>>>> Kai.Grossjohann@CS.Uni-Dortmund.DE (Kai Großjohann) said:
> 
>  Kai> Maybe QP encoding should make use of this to restrict maximum
>  Kai> line lengths to 70 characters or so?
> 
> I think patchmaster Shenghou has already provided a fix. We'll just
> have to wait (with the help of the workaround) for Larsi to get back
> here from Real Life.

I think Shenghou should release a 0.96.1 patch set to tide us over
till larsi returns :-) 

- Hari
-- 
Raja R Harinath ------------------------------ harinath@cs.umn.edu
"When all else fails, read the instructions."      -- Cahn's Axiom
"Our policy is, when in doubt, do the right thing."   -- Roy L Ash

Re: gnus sends corrupt attachments

[Shenghuo ZHU]

>>>>> "Raja" == Raja R Harinath <harinath@cs.umn.edu> writes:

Raja> I think Shenghou should release a 0.96.1 patch set to tide us
Raja> over till larsi returns :-)

I can't do that. On reason is not only I who sent patches. And I don't
want to get the version control confusing.

-- 
Shenghuo ZHU

Re: gnus sends corrupt attachments

[Lars Magne Ingebrigtsen]

Raja R Harinath <harinath@cs.umn.edu> writes:

> I think Shenghou should release a 0.96.1 patch set to tide us over
> till larsi returns :-) 

Kinda like Alan Cox does with linux-2.3.18ac-19 and stuff?  Yeah!
pgnus-0.96sz-14.  :-)

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

Re: gnus sends corrupt attachments

[Shenghuo ZHU]

>>>>> "LMI" == Lars Magne Ingebrigtsen <larsi@gnus.org> writes:

LMI> Raja R Harinath <harinath@cs.umn.edu> writes:
>> I think Shenghou should release a 0.96.1 patch set to tide us over
>> till larsi returns :-) 

LMI> Kinda like Alan Cox does with linux-2.3.18ac-19 and stuff?  Yeah!
LMI> pgnus-0.96sz-14.  :-)

It is not necessary, since 0.97 is coming soon. Right, Lars? :-)

-- 
Shenghuo ZHU

Re: gnus sends corrupt attachments

[Lars Magne Ingebrigtsen]

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

> It is not necessary, since 0.97 is coming soon. Right, Lars? :-)

Yup.  :-)

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

Re: gnus sends corrupt attachments

[Lars Magne Ingebrigtsen]

Raja R Harinath <harinath@cs.umn.edu> writes:

> I think Shenghou should release a 0.96.1 patch set to tide us over
> till larsi returns :-) 

Now that I think of it, I should set up a CVS server that's writable
by Shenghou (and others), and then he could just check stuff in.  And
have snapshots generated automatically, and stuff.

Well, for the next development cycle, at least, I think that would be
a good idea.

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