From mboxrd@z Thu Jan  1 00:00:00 1970
X-Msuck: nntp://news.gmane.io/gmane.emacs.gnus.general/38947
Path: main.gmane.org!not-for-mail
From: Michael.Cook@cisco.com
Newsgroups: gmane.emacs.gnus.general
Subject: mm-qp-or-base64
Date: Wed, 26 Sep 2001 12:51:32 -0400
Sender: zzz@cisco.com
Message-ID: <g088zf1zyi3.fsf@zzz.cisco.com>
NNTP-Posting-Host: coloc-standby.netfonds.no
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
X-Trace: main.gmane.org 1035174733 25271 80.91.224.250 (21 Oct 2002 04:32:13 GMT)
X-Complaints-To: usenet@main.gmane.org
NNTP-Posting-Date: Mon, 21 Oct 2002 04:32:13 +0000 (UTC)
Return-Path: <zzz@cisco.com>
Return-Path: <zzz@cisco.com>
Original-Received: (qmail 12147 invoked from network); 26 Sep 2001 16:51:59 -0000
Original-Received: from frampton.cisco.com (161.44.253.15)
  by gnus.org with SMTP; 26 Sep 2001 16:51:59 -0000
Original-Received: from zzz.cisco.com (zzz.cisco.com [10.89.5.93])
	by frampton.cisco.com (8.8.8/2.6/Cisco List Logging/8.8.8) with ESMTP id MAA08220
	for <ding@gnus.org>; Wed, 26 Sep 2001 12:51:32 -0400 (EDT)
Original-Received: (from zzz@localhost)
	by zzz.cisco.com (8.11.6/8.11.6) id f8QGpWK16906;
	Wed, 26 Sep 2001 12:51:32 -0400
Original-To: ding@gnus.org
User-Agent: Gnus/5.090004 (Oort Gnus v0.04) Emacs/20.7
Original-Lines: 26
Xref: main.gmane.org gmane.emacs.gnus.general:38947
X-Report-Spam: http://spam.gmane.org/gmane.emacs.gnus.general:38947

pdf files often seem to look like plain text files because they tend
to contain a lot of "text" characters, especially at the beginning
of the file.  but pdf files also usually contain a lot of "binary"
characters.  in other words, mm-qp-or-base64 tends to return
'quoted-printable for pdf files, but it should probably return
'base64.  should we make mm-qp-or-base64 smarter?  maybe use
something like

 (goto-char (point-min))
 (if (looking-at "%PDF-")
    'base64
   ...)

suitably generalized, of course.  maybe with:

 (defvar mm-content-transfer-encoding-smart-defaults
   '(("\\`%PDF-" base64))
   "Alist of regexps to match against the beginning of a buffer
to determine which encoding to use.  If none of these regexps
matches, then a heuristic is applied to choose between
quoted-printable and base64.")

or maybe this idea could be folded into
mm-content-transfer-encoding-defaults somehow.

m.