On Monday, August 20, 2007 at 12:34:49 -0600, Kyle Wheeler wrote:
> On Monday, August 20 at 06:49 PM, quoth Alain Bench:
>> a charset=none label. Does it really happen?
> I did actually see it in the wild. I think it was a misconfigured
> webmail
Thanks Kyle! Adopted "charset-hook ^none$ cp1252" in the hookset.
>> There is *no* charset auto-sensing for bodies.
> How come?
I can only guess that the initial author of the $assumed patch years
ago tried to make it the least invasive and the least cycles consuming
posible. I mean that given Mutt's design, $assuming one charset was dead
easy (a oneliner) and costless, while auto-sensing would have needed more
work and a partial redesign. And mutt-dev archives seem to show that EGE
didn't believe much in the (IMHO great) value of auto-sensing.
>> $assumed_charset=windows-1252 Appending anything is (practically)
>> useless.
> might benefit from matching all characters, in the guessed charset,
> against [:print:]
Available functions to determine what is printable and what is a
control char do work in the current locale, not in another arbitrary
charset. Convert to locale then check printability is not workable:
A poor locale would limit sensing richer charsets.
And if we had the magic universal isprint(), well... The practical
benefit to the auto-sensing guesswork would be much lower than expected.
"iso-8859-1:utf-8" would be possible, but unreliable (some UTF strings
happen to not contain 128-159 bytes, and would be wrongly sensed as L1).
"iso-8859-1:cp850" would probably work and have some value, especially
to BBS users. However that's not a very common case, and people would
anyway probably prefer "cp1252:cp850" which doesn't work well (nearly
everything would be sensed 1252, nothing 850). Right now I can't think
of any other example.
Japanese charsets would probably not benefit at all from such
printability check: Those charsets are already rather well
distinguishable by the simple validity check.
Bye! Alain.
--
How to Report Bugs Effectively
<URL:http://www.chiark.greenend.org.uk/~sgtatham/bugs.html>
