Currently, any attachments of a mail message get extracted by ftw.mails own SearchableTextExtender.
This is error prone, requires hacks for broken MTAs / MUAs and non-trivial to debug. In setups that already use ftw.tika it's also unnecessary - Tika already extracts full text from mail attachments, and probably does so in a much more reliable way than we can.
I would therefore propose to make this behavior in ftw.mail configurable via a registry entry (defaulting to True (extract attachments)).
An alternative approach would be to check whether ftw.tika is installed, and simply skip the extraction of attachments if it is.
@jone @maethu thoughts?
Currently, any attachments of a mail message get extracted by
ftw.mails own SearchableTextExtender.This is error prone, requires hacks for broken MTAs / MUAs and non-trivial to debug. In setups that already use
ftw.tikait's also unnecessary - Tika already extracts full text from mail attachments, and probably does so in a much more reliable way than we can.I would therefore propose to make this behavior in
ftw.mailconfigurable via a registry entry (defaulting toTrue(extract attachments)).An alternative approach would be to check whether
ftw.tikais installed, and simply skip the extraction of attachments if it is.@jone @maethu thoughts?