Happy Easter everyone, I don't have enough chocolate eggs for you all, so instead, here are five tips.

1. Give your screens names

I often have processes running in a GNU screen for long periods. For example, I almost always have an IRC client running, perhaps a screen for Gentoo Linux's emerge command, and so on.

You can resume a screen using the process number, for example:

screen -r 18203

However, it is more fun to give the screens human readable names. To do this, use the -S argument when you create the screen:

screen -S chat

Then you can resume the screen using:

screen -r chat

2. Go on the rampage with xargs rm

Sometimes I have a lot of files that I need to sort in a hurry, without going into specifics, usually the issue for me is that after I have gotten out the information that I really need; I then want to have a quick look at the rest before throwing it away.

So at this point, we need to separate almost useless data from completely useless data. In this I am not too concerned about the individual files themselves, but I just want to separate the noise out and look at the rest, normally before chucking those away too.

If the aim is throwing away as much unwanted data as possible, I will make a backup of the files first (in case I type the command wrong), then use "xargs rm" to decimate the working copy. If you feed a list of filenames to "xargs rm" then it will delete them.

So for example, lets throw away every .txt file in the working directory and below that is newer than two days old:

find * -iname "*.txt" -mtime -2 -print | xargs rm

To take another example, the comments in this blog are stored as XML files. If I get spams on this blog, then they usually are predictable, one recent spammer always used the same string in the author field. This command will delete all his comments:

grep -l *.xml "<author>Spammer" | xargs rm

3. Generate temporary filenames uniquely

Temporary files that have predictable filenames could (in theory) be used in race condition attacks or denial of service attacks, it is better to let the operating system assign a unique filename to use as a temporary file.

The way to do this in scripts is to use the mktemp command, this generates a unique name which you can then use in scripts. mktemp started in OpenBSD and appeared for a while in Linux distributions as an individual package. It is now included in coreutils by default.

Run the following command as many times as you like:

mktemp /tmp/example.XXXXXXXXXX

Each time it will substitute the Xs with random letters followed by the process number. This gives a unique filename.

The number of Xs is up to you. The manual page tells us that "ten 'Xs' will result in mktemp testing roughly 26 ** 10 combinations", that is 141.2 trillion plus change.

For example, our imaginary script records a program output to a file.

Lets start by making an environment variable called TMPFILE. The marks below are not single quotes, they are back-ticks (left of the 1 key on a British keyboard). This assigns the output of mktemp to the variable TMPFILE.

TMPFILE=`mktemp /tmp/datewatch.XXXXXXXXXX`

Now we can pipe output to $TMPFILE:

date >> $TMPFILE

Now lets print the output to the screen:

cat $TMPFILE

4. Include an RSS feed as a static component

Django has not taken over the world yet, and sometimes sites do not need to be that complicated. Some websites just use server sides includes to put together static content, for example it might have a header and a footer and the central content.

If you want to display items from another site's RSS feed, then you could just create another include for the feed; and then have that include recreated using cron.

Download this python script that I created. Do not forget to change the URL variable to contain the URL of the feed you want to include.

5. View your wireless history

Sometimes you might want to look at all the wireless networks that you have connected to in the past. If your laptop's Linux distribution uses NetworkManager, for example Ubuntu, then you can use my wireless network history script.

Download this file and rename it so the file ends with .py. Now run the script.

You can do this like so:

python netwomb.py

Or you can make the script executable:

chmod +x netwomb.py

Then just use

./netwomb.py

You can also import it as a module and do far more with it. For example, start the python interpreter:

python

and type the following commands:

import netwomb

wifihistory = netwomb.Network()

You now have an object called wifihistory that contains all the information about your past wireless connections.

To see the networks:

print wifihistory

Now take one of the results, for example, say one is called BTOpenzone. We can see when we set up the connection:

print wifihistory.BTOpenzone.time()

The other information available will depend on the type of connection. Imagine this connection used a WPA Certificate, we could see what certificate we used by going:

print wifihistory.BTOpenzone.wpa_eap_ca_cert_file

Each network is a regular python dictionary. You can see all the information about that connection by going:

print wifihistory.BTOpenzone.__dict__

Please let me know if you do anything cool with it.

That's all folks. Happy Easter!

Discuss this post - Leave a comment

Digg entry