Re: [Lynx-dev] lynx newspost:// does not correctly encode non-7bit chars

[Thomas Dickey]

On Sat, May 11, 2013 at 07:42:32PM +0000, Thorsten Glaser wrote:
> Hi,
> 
> I discovered newsreply:// and newspost:// today and couldn’t resist
> testing it: news://news.gmane.org/address@hidden
> 
> The nōn-7bit-ASCII characters break, and I can trace the real cause
> of the breakage back to lynx. It really should encode the messages
> as quoted-printable if and only if they contain non-7bit chars (as
> MIME is otherwise still frowned upon in the Usenet).

I suppose that taking RFC 3977 and working back to see where lynx is
not up-to-date would help.
 
> Apparently, post_article() in W/L/I/HTNews.c is the culprit, but it
> doesn’t handle the input at all. I believe I can change it to do
> what I want, if you’d accept such a patch. One thing I’d need to
> know is how to get the “current charset” (i.e. the one Lynx believes
> the texteditor wrote).

src/LYNews.c refers to current_char_set
 
> Also, it doesn’t properly decode either:
> news://news.gmane.org/address@hidden
> 
> This might be read_article() in the same file. It doesn’t do MIME

It's unsurprising, since most of the attention on charsets has been
for the html pages.

> even for the bare Content-Transfer-Encoding minimum. I believe I
> could change it to parse at least nōn-multipart messages in either
> quoted-printable or base64 correctly, but then I’ve got a charset
> to indicate to Lynx; how would I pass that on?
> 
> Another thing I’m missing, compared with reading nntp via pine,
> is that lynx doesn’t offer to view the raw article, like this:
> 
yes... that sounds like more work.

At the moment I've been working down a small backlog on top of
fixes for Coverity, thinking to make that dev.16/pre.1

> --- snip ---
> # call this with one of:
> #     getarticle gmane.foo.bar 1234
> #     getarticle $message_id
> # ensure ~/mail/x can be a unix/mbox format mail folder we append to
> 
> ns=news.gmane.org
> if [[ $1 = -n ]]; then
>       ns=ncnntp
>       shift
> fi
> 
> if (( $# == 2 )); then
>       printf 'mode reader\r\ngroup %s\r\narticle %d\r\nquit\r\n' "$1" "$2"
> elif [[ $1 = */+([0-9]) ]]; then
>       printf 'mode reader\r\ngroup %s\r\narticle %d\r\nquit\r\n' "${1%/*}" 
> "${1##*/}"
> else
>       printf 'mode reader\r\narticle <%s>\r\nquit\r\n' "$1"
> fi | \
>     nc $ns 119 | sed -e $'s/\r$//' -e '1,/^[^0-9]/d' -e '/^\.$/,$d' | \
>     (date -u +'From MAILER-DAEMON %a %b %e %H:%M:%S %Y'; cat) >>~/mail/x
> --- snip ---
> 
> This can be useful in debugging, or manually decoding messages
> Lynx cannot parse. No idea how to add that best; the URL scheme
> news://server/msgid isn’t really extensible, but newsraw:// might
> be possible. I don’t currently offer an attempt at a patch.
> 
> bye,
> //mirabilos
> -- 
> Sorry,  I’m annoyed today and you came by as an Arch user. These are the
> perfect victims for any crime against humanity, like  systemd,  feminism
> or social democracy.
>               -- Christoph Lohmann on address@hidden
> 
> _______________________________________________
> Lynx-dev mailing list
> address@hidden
> https://lists.nongnu.org/mailman/listinfo/lynx-dev

-- 
Thomas E. Dickey <address@hidden>
http://invisible-island.net
ftp://invisible-island.net

signature.asc
Description: Digital signature