Here's how to change the rate and delay of your linux system, in direct console or in X, thanks to the command line.

Using ecryptfs outside of Ubuntu's very controlled environment can be tedious.

I wanted to have a fully automated script that could simply mount an ecryptfs partition given a single key (roughly along these lines):

ecryptfs-mount PATH KEY

Here are the main objective:

• non-interactive
• mount the encrypted partition on any host

An important point, is that we don't try to protect against compromission of the main host's root access. We just try to produce an encrypted filesystems that we can send externally (with rsync for example) safely. The host will have the key accessible by root anyway.

As I didn't have much time, I might have missed a simpler alternative so feel free to comment on this please. My knowledge on ecryptfs and security in general is very limited.

You'll find a breakdown of the process first and then a ecryptfs-mount script that support copy paste and will make all this easier.

$ tmp_wrapped_file=/tmp/wrapped-file
$ salt=ANYTHING
$ echo -en "$KEY\n$salt" | ecryptfs-wrap-passphrase "$tmp_wrapped_file"

$ echo -n "$salt" | ecryptfs-insert-wrapped-passphrase-into-keyring "$tmp_wrapped_file"

Inserted auth tok with sig [9f3193d42b4df62] into the user session keyring

mount -i -t ecryptfs $PATH_TO_CRYPT $PATH_TO_CRYPT \
      -o ecryptfs_enable_filename_crypto=y,ecryptfs_passthrough=n,ecryptfs_key_bytes=16,ecryptfs_cipher=aes,ecryptfs_sig=$SIG,ecryptfs_fnek_sig=$SIG,ecryptfs_unlink_sigs=y

ecryptfs-mount PATH KEY

I use org-mode and I have android mobile devices. I ought to try MobileOrg.

But configuration is quite rough, and these are information I found lacking:

• This won't be a synchronization of files as Dropbox handles usually.

  Instead you'll need to push/pull (thanks to M-x org-mobile-push or M-x org-mobile-pull) your normal off-line org-directory (which is outside of the dropbox) towards an on-line org-mobile-directory which is on the dropbox. The dropbox directory contents will be generated along with meta-information and new files (as the index.org) that allows the mobile application to find all your files.

  So don't try to synchronize directly your org files by putting them in the dropbox !

• Second, the MobileOrg Android app needs to know where to find the index.org file. You'll have to give the path of this file, and you'll have to start with a slash "/" for the root of the dropbox, then type the path including the index.org. For example:

  If your index.org file is in the dropbox root and then the MobileOrg directory, then you must inform your mobile devices that the path is /MobileOrg/index.org. Don't forget this first slash.

But if you want to install and use that application, I won't be giving you enough information: you can read OrgMode manual on that subject, then MobileOrg Android wiki... And that was enough for me to get it running.

Run this command in a more than 200+ column wide terminal:

for _c in $(seq 1 512); do
    for i in $(seq 1 $(stty size | cut -f 2 -d " ")); do
        echo -n x
    done
done

If your terminal is still responsive try to scroll up. If always responsive, try on widening window of gnome-terminal and relaunch the script.

This is a gnome-terminal bomb… If you want to test it without killing all your current gnome-terminal sessions you should launch a dummy instance with:

gnome-terminal --disable-factory

in any existing terminal. This will open a new windows that you can smash up safely.

The so-called "bomb" script will only display “x” all over the place on one big line (512 x column wide). Nothing big, it starts to collapse with much less chars on my computer.

Adding a line breaker will remove bomb effect:

for _c in $(seq 1 512); do
    for i in $(seq 1 $(stty size | cut -f 2 -d " ")); do
        echo -n x
    done
    echo
done

Both version of this code have the same visual output.
And are displayed with same blistering speed on urxvt (and probably a bunch of other terminal).

We will be using postconf to set values directly in the /etc/postfix/main.cf

Mail delivery

We will need now to give postfix a clue to which mail we'll want him to deliver localy. This is quite simple:

# postconf -e "mydestination = wiz.orgx, home.foo.comx"
# postfix reload

Now lets see how postfix reacts to the sending of mail towards wiz.orgx:

# nc localhost 25
220 myhostname.localdomain ESMTP Postfix
MAIL FROM:<gbush@us.comx>
250 2.1.0 Ok
RCPT TO:<bugsbunny@wiz.orgx>
550 5.1.1 <bugsbunny@wiz.orgx>: Recipient address rejected: User unknown in local recipient table
^C
#

As user "bugsbunny" does not exist, postfix refuses to send the mail. To continue this, let's suppose you know how to create a new user on your system called 'john' in this example. Now lets try to send a mail to john:

# nc localhost 25
220 myhostname.localdomain ESMTP Postfix
MAIL FROM:<gbush@us.comx>
250 2.1.0 Ok
RCPT TO:<john@wiz.orgx>
DATA
354 End data with <CR><LF>.<CR><LF>
Subject: hello friend

how are you ?
.
250 2.0.0 Ok: queued as 94B822AC246
QUIT
221 2.0.0 Bye
#

Note that the domain "wiz.orgx" is not a valid domain name currently. But this will work with exactly this domain. Postfix does not do any check (and this is normal), neither of your real network domain nor of ICANN legislation. No check at all are performed on the sender's mail address neither.

But where's our mail ?

Let's check where has been put our mail: /var/mail or /var/spool/mail (which should be the same as the second is a link towards the first):

# cat /var/spool/mail/john
From gbush@us.comx  Wed Mar 24 08:04:04 2010
Return-Path: <gbush@us.comx>
X-Original-To: john@wiz.orgx
Delivered-To: john@wiz.orgx
Received: from myhostname.mylocaldomain (myhostname.mylocaldomain [127.0.0.1])
    by myhostname.localdomain (Postfix) with SMTP id CD0CC2AC246
    for <john@wiz.orgx>; Wed, 24 Mar 2010 08:03:16 +0100 (CET)
Subject: my friend
Message-Id: <20100324070339.CD0CC2AC246@myhostname.localdomain>
Date: Wed, 24 Mar 2010 08:03:16 +0100 (CET)
From: gbush@us.comx
To: undisclosed-recipients:;

how are you ?

You should notice that several information have been set in the header of the mail and that our header as been kept intact (it contained only a Subject field).

You can choose to put this file in the home directory of the user by configuring 'home_mailbox':

# postconf -e home_mailbox=MyMailBox

Mail will then be created directly in ~john/MyMailBox as a file in mbox format, or you can choose to put:

# postconf -e home_mailbox=MyMailBox/

Notice the ending slash, this will tell postfix to store mail in ~john/MyMailBox as a directory in Maildir format.

Let's choose the maildir format:

# postconf -e home_mailbox=.mail/

In further blog entry, I'll describe how to plugin procmail as a mail delivery agent.

Before we start

I'll be illustrating what is directly possible through the use of netcat on SMTP/25 port. So you might concider installing it.

# sudo apt-get install netcat

And we'll check often the content of /var/log/mail.{err,log} which are full of information on what is happening in postfix. As a reminder, you can see this content with:

# sudo tail -f /var/log/mail.{err,log}

Postfix basic installation

Installation of postfix base repertories will be done through apt mecanism:

# sudo apt-get install postfix

As of Ubuntu Jaunty, this doesn't install the main configuration file of postfix which is /etc/postfix/main.cf . Without it, /etc/init.d/postfix won't do anything.

So let's create an empty configuration file:

# touch /etc/postfix/main.cf

For this first configuration, to avoid giving a chance to some nasty guys to use an open smtp relay, let's protect us:

# cat <<EOF >> /etc/postfix/main.cf
## temporary protection against unwanted external connection
inet_interface = localhost
EOF

We are done, well not completely done yet, as after

# /etc/init.d/postfix start

we will try to send a mail through

# nc localhost 25

And this will stall. You'll notice in the logs that:

smtpd[30428]:  fatal:   open database /etc/aliases.db: No such file or directory
master[30332]: warning: process /usr/lib/postfix/smtpd pid 30428 exit status 1
master[30332]: warning: /usr/lib/postfix/smtpd: bad command startup -- throttling

To create the /etc/aliases.db, simply type:

# newaliases

It'll take content of /etc/aliases and create the /etc/aliases.db. Don't forget to restart postfix:

# /etc/init.d/postfix restart

First test of our open relay

Let's try to send a mail:

# nc localhost 25
220 myhostname.localdomain ESMTP Postfix
MAIL FROM:<foo@bar.comx>
250 2.1.0 Ok
RCPT TO:<wiz@bee.orgx>
250 2.1.5 Ok
DATA
354 End data with <CR><LF>.<CR><LF>
Subject: test mail

test content
.
250 2.0.0 Ok: queued as 94B822AC246
QUIT
221 2.0.0 Bye
#

Note that neither sender email foo@bar.comx nor destination email wiz@bee.orgx are correct mail addresses. This doesn't seem to bother postfix which seems to take care of my content. But in the log:

smtpd[30451]:   connect from myhostname.mydomain[127.0.0.1]
smtpd[30451]:   A7FB62AC246: client=myhostname.mydomain[127.0.0.1]
cleanup[30546]: A7FB62AC246: message-id=<20100317214750.A7FB62AC246@myhostname.mydomain>
smtp[30547]:    A7FB62AC246: to=<wiz@bee.orgx>, relay=none, delay=31, delays=31/0.02/0.11/0, dsn=5.4.4, status=bounced (Host or domain name not found. Name service error for name=bee.orgx type=A: Host not found)
cleanup[30546]: 4EBCA2AC249: message-id=<20100317214811.4EBCA2AC249@myhostname.mydomain>
qmgr[30446]:    4EBCA2AC249: from=<>, size=2282, nrcpt=1 (queue active)
bounce[30548]:  A7FB62AC246: sender non-delivery notification: 4EBCA2AC249
qmgr[30446]:    A7FB62AC246: removed
smtp[30547]:    4EBCA2AC249: to=<foo@bar.comx>, relay=none, delay=0.04, delays=0.02/0/0.03/0, dsn=5.4.4, status=bounced (Host or domain name not found. Name service error for name=bar.comx type=A: Host not found)
qmgr[30446]:    4EBCA2AC249: removed

You can see that postfix tried to send the mail, and received a Host not found error for host bee.orgx. An email is generated to warn the sender that his mail has not been sent correctly, but bar.comx ends also in a Host not found. Finally, the original message and the warning message are deleted from their queue as they can't be sent to their respective domain.

In this basic configuration, there are no check at all. This is an open relay. This is why we limited access to this postfix to localhost only.

Next stage we will setup postfix to deliver mail on our host...

This was tested on Ubuntu (karmic), and instructions will target this system.

Principle

We'll use a application to get vCards Contact list out of the phone through the bluetooth connection. Then importing this to thunderbird will be trivial.

Extraction of Contacts from Phone

I use gammu (you might need to check the installation pages to see how to install it):

# gammu backup save.vcf

This will create the "save.vcf" file containing all your contact. You should not use "save.ldif" format as it will mess up fields names and won't correctly be imported by Thunderbird.

Convert VCF file to LDIF file (correctly)

The trick now is to use this Online VCF to Ldif converter (which don't mess with Thunderbird fields) to convert your "save.vcf" file.

Import file in Thunderbird

You can now safely import contacts as Text/Ldif contact list in Thunderbird's Contact Manager. They will be put in a separate contact book. You'll have to move (drag & drop) all your contact to your synchronized contact.

You should check carefully that fields are correct. If you have several duplicates, you should check my next blog entry.

Et voilà !