[mew-int 2830] Re: Encoding issues at Summary Mode

[Diogo F. S. Ramos]

Harald Hanche-Olsen <hanche at example.com> wrote:
> Ah, but those that don't, are usually multipart messages, as indicated
> by the Content-Type header. And then the Content-Transfer-Encoding is
> in each part. It appears that mew doesn't give you an easy way to read
> the headers for individual parts, so you'll have to check the raw
> message (accessible by typing C-TAB in summary mode).

Yes, you are right. The multipart messages also have the
Content-Transfer-Encoding and they are quoted-printable too.

>> This seems to happen especially with e-mails from gmail, which, as far
>> as I can tell, do not provide a Content-Transfer-Encoding.
> 
> They sure do, if they're single part. As far as I can tell from a few
> ones in my mailboxes, at least.

I've made some tests and I think I kind of gather some info about how
gmail encode the e-mails, but frankly, it's a little confusing. It
seems to change based on the e-mail's size or the accents used. I
tabled some samples of it.

Here is the legend:

  Title
    A - Accented
    N - Non Accented
  Body
    A - Accented
    N - Non Accented
  TOE - Type of e-mail
    R - Rich Text
    S - Simple Text
  CTEP - Content-Transfer-Encoding presence
    Yes
    No
  NCTE - Number of Content-Transfer-Encoding headers
  CTET - Content-Transfer-Encoding type
    b64 - base64
    qp - quoted-printable

|------+------+-----+------+------+---------|
| Tile | Body | TOE | CTEP | NCTE | CTET    |
|------+------+-----+------+------+---------|
| A    | A    | R   | Y    |    2 | b64     |
| A    | N    | R   | N    |      |         |
| N    | A    | R   | Y    |    2 | b64, qp |
| A    | A    | S   | Y    |    1 | b64     |
| A    | N    | S   | N    |      |         |
| N    | A    | S   | Y    |    1 | b64     |
|------+------+-----+------+------+---------|

I don't know if this can be useful for anything.

There were some e-mails that don't have their body excerpts displaying
at the (Summary) mode, but I was unable to see a pattern there.

> Anyway, it seems the trouble is that no decoding of the body is being
> done for the sample to go into the message summary. I certainly hadn't
> noticed, since I don't use body excerpts in my message summaries.

The problem is that I think this is causing me search errors. Maybe
the way Mew deal with displaying body excerpts is the same it deals
with searches.

-- 
Diogo F. S. Ramos