[Mailman-Users] ISO-8859-1/Latin1 vs UTF-8

Mark Sapiro msapiro at value.net
Mon Oct 24 18:52:24 CEST 2005

Bernd Petrovitsch wrote:
>The problem is that I enter text in the web-interface on a default UTF-8
>system and it is apparently stored as UTF-8. The pages are delivered as
>ISO-8859-1 according to the HTTP  header and the header in the file.
>So the CGI scripts actually should convert correctly the ML-admins data
>from the given charset into ISO-8859-1 (which is not 100% possible but
>for German and in practice it will be good enough IMO).
>The other solution is to use ö and brothers and blame the ML admin
>if he doesn't do so (and put in somwhere into the docs or so).
>There is BTW a similar problem: If I enter " (quotes) in the web
>interface, they are converted to " (probably by the browser - I
>didn't check) and then sent out in a plain/text email.
>So in both cases I actually want to know what the way to go is.

As Brad points out in another reply, some of this problem is because
all text entered in the web interface (except for General
Options->info which is a special case) is HTML escaped to prevent XSS
attacks. Mailman arguably goes overboard on this, but the 4 characters
'&' '<' '>' and '"' are changed respectively to &amp;, &lt;, &gt; and
&quot; by Python's cgi.escape() method.

Thus, you can't even enter &ouml; and have it work in HTML or plain

You can convert Mailman to use utf-8 for German language, but this will
not solve the html escaping issue. If you are interested in converting
to utf-8, there is relevant information in the archives of this list.
See the thread beginning at
and continued at
and the thread beginning at
and continued at

Mark Sapiro <msapiro at value.net>       The highway is for gamblers,
San Francisco Bay Area, California    better use your sense - B. Dylan

More information about the Mailman-Users mailing list