How do I filter MIME encoded 'Subject' or 'To' headers?

[dcollins]

Apparently, the message filter compares match strings against undecoded MIME in headers, rather than the decoded result. For example, the header:

Subject: New topic notification - "Board Forum"

appears as follows when the "blah,blah" (full headers) option is enabled:

Subject: =?UTF-8?B?TmV3IHRvcGljIG5vdGlmaWNhdGlvbiAtICJCb2FyZCBGb3 J1bSI=?=

I discovered this when debugging a straight forward filter that was not finding the string 'notification' in the subject header.

The filter doesn't work ... apparently because it's looking at the undecoded text.

This seems like a problem others would have, but I'll be damned if I can think of a search handle that reveals any discussion of same. I looked at regular expressions, but couldn't find anything sophisticated enough to decode the input before doing the match.

(If you are wondering, the subject header was produced by a phpBB mailing engine which MIME encodes headers (to make them more universally distributable ???) )

[jhmeyers]

Subject: =?UTF-8?B?TmV3IHRvcGljIG5vdGlmaWNhdGlvbiAtICJCb2FyZCBGb3 J1bSI=?=
[which decodes from "base64" as the following plain ascii text]
Subject: New topic notification - "Board Forum"

Hard way:

Encode your search string to "base64"
added as an extra ("OR" logic) condition.

If you don't know that your search string will be left-justified,
say, then you may actually need up to three extra tests,
because of how "base64" encoding works :(

Another way:

If "filtering" was missed, due to "base64" encoding,
then try a subsequent "search" on message "Summary,"
which is almost certainly by what's in the TOC, which may work,
if subject or "to" has, by then, been decoded properly.

I don't know whether searches on "subject" or "to"
go by the TOC version or by the message body version;
perhaps an experiment will reveal.

Third way:

Write your own plugin (or find someone else who's interested :)

I just realized -- there IS a "Mime Headers" plugin,
but I'm not sure whether it's for Windows or Macintosh ???

[dcollins]

Thanks. Your hard way was my easy way. In this special case, message subjects I'm testing for have one of two possible (undecoded) initial character strings ... so I just cut'n paste the two undecoded exemplaries, OR 'em up with regexp and escape the ?'s. Piece of cake.

It's a kludge but this problem is rare so doesn't warrant more time spent on a general solution.

I was just hoping that maybe there was something already out there. My initial google on the MIME headers plug in found something apparently written for Windows Eudora a while ago ... on first glance, it appears there was a time when Eudora didn't decode MIME headers by default when presenting in the non-expanded header mode ... hence the need for the plug in. Not sure this plug in would run prior to the filter.

My version (Eudora 6.2.4 for Mac) automatically presents decoded headers in non-expanded mode ... but apparently it runs the filter prior to decoding the headers.

Thanks, jh, for your timely response

[jhmeyers]

Yes, MimeHeaders plugin, by Alexander Lishansky, is for Windows
(and I've found all versions through 2.04!)

Glad to hear you have solved the problem, and best wishes.