List Filtering (was Re: [Xapian-discuss] how to display the results in html which we r showing on konsole)

Olly Betts olly at survex.com
Thu Mar 30 17:13:05 BST 2006


On Thu, Mar 30, 2006 at 03:11:38PM +0100, James Aylett wrote:
> On Thu, Mar 30, 2006 at 01:51:44PM +0100, Olly Betts wrote:
> > The option is "convert_html_to_plaintext" in the "Content filtering"
> > section (and also make sure "filter_content" is enabled).
> 
> Any idea if this will just drop the HTML version if there's a text
> version, or if it will squash the text alternate with its converted
> version?

The explanations on mailman's "Content filtering" page are sadly rather
unclear about details, and about how the options interact.

I've used it in the past, and just assumed it did something sensible.
A bit of googling suggests that it just takes the first part (which
is usually text/plain but doesn't have to be).

But since multipart/alternative with HTML and text versions in reality
always has one auto-generated from the other, it doesn't really seem
too important which happens.
 
> (Also means that attachments will be junked. Hooray. What's the MIME
> type for patches again?)

I don't think there's an official type, but text/x-patch is common.
Attachments of text/x-c, text/x-perl, and several other text/x-* types
should probably not get stripped.  If you go that route, I'd probably
suggest allowing text/* with a few exceptions, and stripping anything
else.  

Incidentally, I don't think the recent rash of messages is an attempt to
spam.  Several of the posters seem to be trying to write a GIS system
which uses Xapian:

http://groups.google.com/group/pp_gis/

Cheers,
    Olly



More information about the Xapian-discuss mailing list