Re: apostrophes showing as ?
On 06Apr2006 11:34, David Woodfall <dave@xxxxxxxxxxxxxxx> wrote:
| On (12:17 06/04/06), Cameron Simpson <cs@xxxxxxxxxx> put forth the
proposition:
| > On 06Apr2006 01:53, David Woodfall <dave@xxxxxxxxxxxxxxx> wrote:
| > | When I view some emails I find ' replaced by qestion marks, but only some
| > | - some emails show correctly. I've tried using various ansi/non-ansi fonts
| > | but I can't seem to correct this. Is this perhaps due to the mailers
| > | setup rather than mine?
| >
| > You see similar stuff on the web sometimes too. A bunch of Microsoft
| > tools use other codes for the quote marks instead of just an ASCII
| > apostrophe. Do these messages correspond to particular mailer types
| > (as frequently shown in the X-Mailer: header)?
|
| I don't see an X-Mailer header in the affected mails.
Well one recent example is Kyle Wheeler, who is using mutt but
explicitly using an unusual codepage (well, "different, and therefore
wrong"). His email is PGP-signed and uses
Content-Type: text/plain; charset=windows-1252; format=flowed
for the text portion. Valid, but how much does the receiving system
need to know to transcode the curly-quotes into a working glyph in the
receiver's charset?
| > Meanwhile, amuse yourself here:
| > http://www.fourmilab.ch/webtools/demoroniser/
| > which addresses this kind of issue.
|
| I'm going to experiment putting this in my procmailrc to see if it fixes
| the issue:
|
| :0fw
| | perl /home/dave/scripts/demoroniser.pl
I fear that a blanket filter like that may do damage. It really wants to
be applied to text post-decode, inside any MIME part markers. Otherwise
it may do damage to other stuff.
Probably, instead of in your procmailrc, it should go in the mailcap
entry for particular types so the pager uses it when presenting text.
Kyle also says this, elsewhere (in the display-toggle-weed thread):
One other thing? not everybody uses a UTF-8-capable terminal (like uxterm
(which comes with xterm)). If your terminal is only capable of ASCII or
something like that, it?s probably smart to set your mutt?s $charset to
?us-ascii//TRANSLIT? ? that //TRANSLIT at the end will tell the iconv
library to transform characters that don?t *quite* map to your charset
(us-ascii) into something similar, so curly quotes: ?? will become
un-curled: ""
which seems far more precise and apt. If it works:-) I didn't even
know about "TRANSLIT".
Cheers,
--
Cameron Simpson <cs@xxxxxxxxxx> DoD#743
http://www.cskk.ezoshosting.com/cs/
Ya see, nothwithstanding the exchange of net.custard-pies, I enjoy discussing
stuff with Mike, because I know that the positions he adopts are ones he's
arrived at after careful thought and honest reflection. - Jeremy Henderson