>>>>> "EM" == Etienne Millon <etienne.mil...@gmail.com> writes:
EM> Can you try to put the following line in your config.py ?
EM> CHARSET_LIST='US-ASCII', 'UTF-8', 'BIG5', 'ISO-2022-JP', 'ISO-8859-1'
Ahh. It had been so long since I set my current config up, that either
I had forgotten about r2e's config.py or my setup predated CHARSET_LIST...
I used »CHARSET_LIST='US-ASCII', 'ISO-8859-1', 'UTF-8'« because those
will cover all of the feeds which I monitor.
There is little reason to have 8859-* after utf-8; it would never fall
through to it. But having it ahead of utf-8 can have benefit.
Something in my chain forces r2e's 8859-1 to cte:qp and its utf-8 to
cte:b64. Given that the former can be read w/o decoding, it is useful
to permit its use.
I don't know where the CJ encodings should fall in the default set.
Having them ahead of utf-8 causes harm for non-asian-language feeds,
but having utf-8 first will unify zh and jp text. That is, if the
reader's MUA is configured to prefer a jp font for ideographs, then
zh text in utf-8 will be rendered with that jp font. And visa-versa
if their MUA prefers a zh_{CN,TW,HK} font. I don't know how much
harm that would do. I usually can recognize zh_CN vs zh_TW vs jp
vs kn text, but cannot actually read any of them....
Were the character-set matching to limit big5 and 2022 to characters
which are not used outside Asia, though, they could remain before utf-8.
That means that neutral chars -- like the quotes -- should not match
the CJK character sets. Only matching characters which have width
property W in unicode's EastAsianWidth.txt should do the trick.
(The quotes have A, presumably for Ambiguous.)
Explicitly configuring it, though, does fix things for me.
Thanks.
-JimC
--
James Cloos <cl...@jhcloos.com> OpenPGP: 1024D/ED7DAEA6
--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
