read mail out of maildir direcory with correct encoding and use in bash

Question

I'm trying to implement some sort of spamreport for my Mail Server. Mails are sorted by sieve and are all in one folder called Spam. I loop through the folder with bash. With this i get the necessary informations from the mail:

cat $f | grep '^From' | head -n1 >> $TMPFILE
cat $f | grep '^Subject' | head -n1 >> $TMPFILE

but in some mails the subject is encoded like this

Subject: =?ISO-8859-1?Q?Test:_Jaguar_XKR-S:_Unter_dem_Blech_lauert_d?=

How i can get the subject in correct encoding? I tried to use mail, mailx, mutt... but no one was able to simply load a mail from file.

FYI, no need for a pipeline: `grep -m1 '^FROM' $f >> $TMPFILE` — chepner, Aug 20 '12 at 15:03
Or even `sed -n '/^$[Ff]rom\|[Ss]ubject$:/p;/^$/q' "$f" >$TMPFILE`; but if you are switching to Perl anyway, this is moot. — tripleee, Aug 21 '12 at 05:51
@tripleee i'm scared of that regexp, but thanks for it. maybe i can use it later :) — Max, Aug 21 '12 at 08:11

score 1 · Accepted Answer · answered Aug 20 '12 at 15:34

1

The encoding in the Subject line looks like MIME Words. One possible way to decode the data is write a perl script that uses the MIME::Words module. You could make the perl script a shell script and call it from your bash script.

convert_subject.sh:

 #!/bin/sh
 /usr/bin/perl -pe 'use MIME::Words(decode_mimewords); $_=decode_mimewords($_);'

Example of using the script:

$ echo "=?ISO-8859-1?Q?Test:_Jaguar_XKR-S:_Unter_dem_Blech_lauert_d?=" | sh convert_subject.sh

Which outputs:

Test: Jaguar XKR-S: Unter dem Blech lauert d

answered Aug 20 '12 at 15:34

j.w.r

4,136
2
27
29

works pretty! one small thing, the german umlauts are not converted correct. Any ideas? :) _Subject: Nur fÃŒr kurze Zeit: 15 Euro Gutschein fÃŒr_ – Max Aug 20 '12 at 15:59
working fine, failure of the umlauts is caused by my formatting of the mail as html and not fault of your little script. **THANKS** – Max Aug 20 '12 at 16:03
Actually it looks like you have UTF-8 but viewing it in an ISO-8859-1 terminal (or other legacy 8-bit Western, could be Windows code page 1251 or etc). – tripleee Aug 21 '12 at 03:07
i added `"Content-Type: text/html; charset='utf-8'"` to the mail script, now the most umlauts are fine. just some here and there are now question mark in a box. any suggestions? my code: `( echo "Subject: Spam" echo "MIME-Version: 1.0" echo "Content-Type: text/html; charset='utf-8'" echo "Content-Disposition: inline" echo "
```
"
  cat $TMPFILE
  echo "
```
" ) | sendmail $i`
– Max Aug 21 '12 at 08:23
i figured out that lines starting with `=?utf-8?Q?` are working, lines with `=?iso-8859-1?Q?` get ? instead of umlauts. – Max Aug 21 '12 at 12:35
working: `=?utf-8?Q?Top-_Reiseschn=C3=A4ppchen_+_Urlaubsgeld?=`. not working: `=?iso-8859-1?Q?K=F6rperliche_N=E4he?=`. executed over bash (echo "?..." | sh convert...) – Max Aug 21 '12 at 12:40

read mail out of maildir direcory with correct encoding and use in bash

1 Answers1