<<< Date Index >>>     <<< Thread Index >>>

charsets (was: feature freeze?)



Hello Brendan,

 On Tuesday, August 15, 2006 at 10:25:40 -0700, Brendan Cully wrote:

> assumed_charset and iconv-hook might be worth a review

    Short summary listing:

 -1) $assumed_charset: Feature patch, assigns a charset to non-MIME
messages (instead of pass-thru). Usefull to anybody, and even
indispensable in UTF-8 locales. Solves numerous user cases that can't
be cleanly solved by other means. Large user base since years (Debian,
Gentoo, Mutt-ng), very well tested, no known malfunction since 2 years.
Thomas dislikes it, but once said he'd accept to consider committing.
I'm 107% for inclusion.

TLR> Even though I don't really like the assumed_charset feature, I'm
TLR> somewhat tempted to adding it to mutt; I'd welcome comments on
TLR> that.

 -2) $attach_charset (formerly $file_charset): Feature patch, assigns an
origin charset to text files one attaches (instead of $charset), with
limited but quite handy optional auto-sensing ("utf-8:iso-8859-1"). Same
large user and tester base as $assumed. Very short and unintrusive
patch, because it in fact just makes use of functions already present in
Mutt. No known bug. Usefull and handy, I'm all for inclusion.

 -3) msyk.iconv-hook: Essentially bugfixes, that make iconv-hooks work.
Principle seems to me undiscussable, so once agreed on implementation,
this should go in.

 -4) M_ICONV_HOOK_sanitize: Bugfixes, clearing bad effects of some
charset-hook setups. Principle already approved. Implementation works
for me, but needs a review by other eyes. Then should go in.

 -5) $create_rfc2047_parameters: Feature, sends non-Ascii filenames to
Outlook users (and to some other 2231-unaware mailers and webmailers).
Rejected for obvious RFC conformance reasons. Still usefull as
independant feature patch. No inclusion.

 -6) pgp_charsethack: Feature, sends inline PGP clearsigned in
$send_charset (instead of Ascii or UTF-8 only). Enhances
interoperability in some cases, reduces it in others. I believe the
balance to be positive in practice, but this is discussable. And this
would go against some RFC. Independant feature patch. No inclusion.

 -7) $unknown_charset: Feature, assigns a replacement for any unknown
charset label (instead of pass-thru). Too young: Should not yet be
included. Quick more testers, or perhaps later for 1.7?

 -8) multiple_charsets: Not reviewed yet, so IIUC: Generalization to
Chinese and Korean charsets of features already existing for Japanese.
Seems "natural" inclusions, worth a serious review.

 -9) create_alias_charset: Not ready.


    So I think that 1-4 should go in ASAP, and 7-8 thoroughly considered
for next lap.


    Latest versions are:

 -1) <URL:http://www10.plala.or.jp/sanrinsha/tamo/patch-1.5.9.tt+tamo.assumed.3>

 -2) On <URL:http://www.emaillab.org/mutt/download1510.html>
patch-1.5.10.tt.attach_charset.1 and additionally
patch-1.5.10.tt.file_charset_compat.1 (readds old $file_charset as a
synonym).

 -3) Article <20060722161136.GA10407@xxxxxxx> here.

 -4) <20060802100503.GA1337@xxxxxxx>

 -7) <20060812142235.GA11118@xxxxxxx>


Bye!    Alain.
-- 
Give your computer's unused idle processor cycles to a scientific goal:
The Folding@home project at <URL:http://folding.stanford.edu/>.