short assumed (was: alternates (was: What should go into 1.5.7?))
On Sunday, January 30, 2005 at 8:56:18 AM +0900, Tamotsu Takahashi wrote:
>>> They are on http://www.momonga-linux.org/~tamo/ , too.
>>> (short_assumed_charset [...] patch)
Much thanks. Still two legacy and masking problems. Context: Three
broken mail cases:
-A) MIME without charset label body
-B) Non-MIME body (no "Content-Type:")
-C) Non-MIME raw 8bit header
Legacy Mutt treats A as implicit US-Ascii, B and C as pass-thru: No
conversion, no masking, just raw bytes sent to terminal hoping it will
know how to deal with them. A is subject to charset-hook, B and C are
not controlable.
Takashi's assumed_charset gives control over the 3 cases. One can
get as default Mutt exactly, or assume, convert, and mask as any given
charset. There is a bug of masking of unconvertables in case C, spotted
by Vincent in point 3 of msgid <20040521132932.GA16033@xxxxxxxxxxxxxx>
in bug #1876 here.
Tamo's short_assumed_charset addresses this bug, drops $strict_mime
bolean, drops WS around 2047 words treatment, and changes some details.
The fix for Vincent *seems* nearly OK: Whatever setting, C case is
converted at best, and unconvertables are correctly ?-masked. But... Not
in $assumed_charset=us-ascii default setting: A and B are correctly
masked, C is pass-thru. Ideally C should be masked as A and B.
And there probably should be a legacy mode where A is masked, and B
and C are pass-thru. Reintroduce $strict_mime bolean? No: Better when
$assumed_charset is unset or has an empty component. And probably legacy
mode should be default.
The meaning of "us-ascii" should be "us-ascii", not a special case
trigger. Evil in-band signalling and such.
Bye! Alain.
--
When you want to reply to a mailing list, please avoid doing so with
Microsoft-Entourage/11. This lacks necessary references and breaks threads.