Re: v.98 pdf decoding broken?

["Johan Kullstam" <kullstam@ne.mediaone.net>]

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

> uliw@erdw.ethz.ch writes:
> 
> > I am unable to decode a properly encoded pdf attachment. The file is
> > written to disk and loaded, but remains empty.
> 
> Does this work better in Pterodactyl 0.99?

while looking at the qp-or-uuencode choser back in 0.96 (or was it
0.97...), i noticed some bad behavior with .pdf files.

it seems pdf are ascii but with a CR instead of LF or CF/LF to end the
lines.  (i think this is macintosh format as opposed to unix or ansi.)
anyhow, the choser doesn't see enough non-printable stuff and uses
quoted-printable format.

the problem is that CR doesn't signify end-of-line to many of the
tools and the QP can, therefore, have some extremely long lines.  i've
had the lines get broken and wrapped by my mail processing chain.  i
am not sure if it's gnus/emacs sending, my isp's sendmail, my
fetchmail, my qmail, or gnus/emacs receiving the mail where it gets
clobbered.  somewhere there be dragons.

has anyone else run across this?  how should we handle ascii with long
lines (very few LF chars)?  should this be made uuencoded or perhaps
there is some CR and LF stuff in QP to break the lines and reform them
upon unquoting?

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