view README @ 8:a92d0d59b669 default tip

f-demime: indicate X-backslash-escapes encoding in output
author Mychaela Falconia <falcon@freecalypso.org>
date Sat, 06 May 2023 17:00:23 +0000
parents 36a95de13d30
children
line wrap: on
line source

The present code repository contains email software which Mother Mychaela
developed for her own personal use; it is unlikely to be of interest to other
people as my tastes in personal technology tend to be very different from the
rest of the world, but I publish almost all of my software nonetheless, as a
matter of general principle.

The main program in the present collection is f-demime - it is a de-MIME-ing
tool.  f-demime processes one or more email messages in UNIX mailbox format
(each message preceded and delimited by a '^From ' line) and applies the
following transformations to each message:

1] base64 blobs that aren't text parts are excised from the message; each such
   blob is decoded and saved into its own file.  These saved attachment files
   are written into a special directory that must be indicated with
   FDEMIME_ATT_DIR= environment variable.

2] text/plain parts that are encoded in either base64 or quoted-printable are
   decoded and written out in a form that (a) losslessly preserves the content
   of the text part that originally contained non-ASCII characters and/or
   insanely long lines, but also (b) is more suitable for direct consumption by
   my wetware, rather than MIME software.  The output form uses backslash
   escapes to represent non-ASCII characters and machine-inserted line breaks.

3] text/html and other non-plain text parts that are encoded in base64 are
   losslessly transcoded from base64 to quoted-printable.

The transformations applied by f-demime aren't fully lossless at level of
overall message lines and octets (one cannot take f-demime output and
reconstruct a message byte-identical to the original), but they are lossless at
the level of content which MIME itself promises to preserve.

One important demiming function which f-demime does NOT do is collapsing
multipart/alternative structures, keeping text/plain and discarding HTML.  The
rationale for this functional omission is that I will be inserting f-demime
into my incoming mail handling path, such that every incoming message will
automatically pass through f-demime before it hits my eyeballs.  Given the
current state of badness on the Internet, I occasionally receive mails in which
the HTML part (which would be thrown away in a multipart/alternative collapse)
contains some valuable or interesting information content that is missing in the
text/plain version - thus fully automatic collapsing is not a workable approach
currently.  I plan on implementing another program (resurrecting maltcollapse
from my previous cdemime attempt circa 2006) to do this multipart/alternative
collapsing, and then devising a practical way to invoke it once I have
determined that the HTML part is junk that can be safely thrown away.