Re: Additional img with every cronjob

[Earl Hood <ehood@hydra.acs.uci.edu>]

On February 5, 2002 at 14:12, Wilhelm Alm wrote:

> I have app. 150 messages in the archive and encounter that every cronjob
> is creating several gif/jpg's .
> These are inline images.
> <MIMEFilters>
> application/octet-stream;  m2h_external::filter;        mhexternal.pl
> text/plain;                m2h_text_plain::filter;      mhtxtplain.pl
> </MIMEFilters>
> <MIMEArgs>
> m2h_text_plain::filter; maxwidth=80; nourl
> <!--===  m2h_text_html::filter; allownoncidurls; maxwidth=80;  nourl -->
> </MIMEArgs>

(FYI: The "comment" in the MIMEArgs is actually not recognized as a
 comment by the resource file parser.  But what you have has the effect
 of "commenting out" the args line since there will definitely not
 be a filter called "<!--=== ...")

> 1.I had tried m2h_text_html::filter to allow noncidurls be shown and commente
> d
> that filter out (as above) as a first step, but without a result.

The "noncidurls" just tells the filter to preserver external URL
references in the HTML.

> 2. What is wrong? My archive is increasing with every cronjob by some 100 KBy
> te.

Are you using the -add option when invoking MHonArc?  If not, you are
recreating the archive each time, and if there are attachments, this
will cause the duplication of attachment files to be created since
mhonarc tries to avoid stomping on files.  If you are doing updates to
an archive, make sure to use -add.

If they are not updates, make dure to delete all files in the
directory before calling mhonarc.

> 3. How may I allow external images?

I'm unsure on what you are exactly asking here.

> 4. Is there a mode to parse html messages and extract only the plain text?

There is no built-in filter that converts HTML data into text.  If
your messages contain text/plain alternative parts, you can exclude
HTML data from the archives and the text/plain alternatives would
be used.  However, if all a message has is HTML data, then nothing
would show up for the message.

Note, you could write a MIMEFILTER that does HTML->text.  An option is
to leverage something like lynx or w3m to do the conversion for you via
a system() call.  BTW, w3m is better than lynx at converting HTML to
text.

--ewh