I wanted to dump the headers for a http request over ssl today. Pulled out socat and this command and it seemed to work quite nicely.

socat OPENSSL-LISTEN:4433,cert=server.crt,key=server.key,verify=0 -

You will need to create server.crt and server.key and if you don't want the other end to complain then they should be a valid keypair.

Socat is a really nice tool that exposes the power of Unix sockets and allows you to connect anything together. In this case a tcp socket with ssl support and the standard out. But it could be a UDP port and a serial cable for all all socat cares. A really nice tool.

I quite like unit tests. They give you that quick feedback that things are going to be okay. I can write a test watch it fail. Write some code and watch it fail also. Then learn very quickly what I have done wrong.

At some point last year I read the Clean Code book by Rober C Martin and it opened up to me a lot of the ideas about testing and code that I had not really got to grips with at the time. Martin takes you through the reasoning and process for producing code that, in a years time, you might want to actually look at and feel comfortable editing.

What has this go to do with bash and bats? Well we have been adding more tests to our puppet code and it came time to write a bit of bash for moving some files around. I thought there must be a simple way to check some of this and sure enought that is where bats comes in.

bats is a testing framework for bash scripts. I know it sounds crazy but it really does work. If you start applying some of the lessons from clean code and spliting your work into functions with local varaiables it soon becomes easier to test. You can, as ever, take things too far but this does at least allow you to test a lot of your logic before you roll it out. With the added benifit of makeing the code a little more readable.

I found this blog post by Spike cleared up some of the questions I had about testing functions as well. http://blog.spike.cx/post/60548255435/testing-bash-scripts-with-bats

Some other useful things are the google shell style guide and shUnit2 which I have yet to look at. The Defensive Shell Guide by Kfir Lavi and Better Bash Scripting in 15 Minutes by Robert Muth are also good.

If you are seeing some errors in the logs like the ones below:

pcscd[4489]: 58837251 ccid_usb.c:645:OpenUSBByName() Can't claim interface 001/110: Device or resource busy
pcscd[4489]: 00040743 ifdhandler.c:118:IFDHCreateChannelByName() failed
pcscd[4489]: 00000030 readerfactory.c:1043:RFInitializeReader() Open Port 0x200000 Failed (usb:072f/2200:libudev:0:/dev/bus/usb/001/110)
pcscd[4489]: 00000008 readerfactory.c:335:RFAddReader() ACS ACR122U init failed.

You need to prevent the kenel nfc drivers loading so add the following lines to /etc/modprobe.d/rfid-blacklist.conf

blacklist pn533
blacklist nfc

as suggested here: http://maciaszek.net/26522/rfid-reader-mit-debian/

Then remove those modules with modprobe -r pn533 nfc

Then install some packages

apt-get install libpcsclite1 pcsc-tools pcscd

Then run:

pcsc_scan

You should get some output similar to the stuff below.

PC/SC device scanner
V 1.4.23 (c) 2001-2011, Ludovic Rousseau <[email protected]>
Compiled with PC/SC lite version: 1.8.11
Using reader plug'n play mechanism
Scanning present readers...
0: ACS ACR122U 00 00

Thu Jan 22 23:00:58 2015
Reader 0: ACS ACR122U 00 00
  Card state: Card removed,

Thu Jan 22 23:01:06 2015
Reader 0: ACS ACR122U 00 00
  Card state: Card inserted,
  ATR: 3B 8F 81 01 80 4F 0C A0 00 00 04 06 03 00 01 00 00 00 00 6A

ATR: 3B 8F 81 01 80 4F 0C A0 00 00 04 06 03 00 01 00 00 00 00 6A
+ TS = 3B --> Direct Convention
+ T0 = 8F, Y(1): 1000, K: 15 (historical bytes)
  TD(1) = 80 --> Y(i+1) = 1000, Protocol T = 0
-----
  TD(2) = 01 --> Y(i+1) = 0000, Protocol T = 1
-----
+ Historical bytes: 80 4F 0C A0 00 00 04 06 03 00 01 00 00 00 00
  Category indicator byte: 80 (compact TLV data object)
    Tag: 4, len: F (initial access data)
      Initial access data: 0C A0 00 00 04 06 03 00 01 00 00 00 00
+ TCK = 6A (correct checksum)

Possibly identified card (using /usr/share/pcsc/smartcard_list.txt):
3B 8F 81 01 80 4F 0C A0 00 00 03 06 04 00 01 00 00 00 00 6A
3B 8F 81 01 80 4F 0C A0 00 00 03 06 .. 00 01 00 00 00 00 ..
    Mifare Standard 1K (as per PCSC std part3)
3B 8F 81 01 80 4F 0C A0 00 00 03 06 04 00 01 00 00 00 00 6A
3B 8F 81 01 80 4F 0C A0 00 00 03 06 04 .. .. 00 00 00 00 ..
    RFID - ISO 14443 Type A Part 3 (as per PCSC std part3)
3B 8F 81 01 80 4F 0C A0 00 00 03 06 04 00 01 00 00 00 00 6A
    Philips MIFARE Standard (1 Kbytes EEPROM)
    http://www.nxp.com/#/pip/pip=[pfp=41863]|pp=[t=pfp,i=41863]
    RFID - ISO 14443 Type A - Transport for London Oyster
    ACOS5/1k Mirfare
    RFID - ISO 14443 Type A - NXP Mifare card with 1k EEPROM
    vivotech ViVOcard Contactless Test Card
    Bangkok BTS Sky SmartPass

More next time.

Spent a little time last night at Hackspace writing some basic code to scroll some text across a rainbowduino.

I use arduino-mk ( pacakge available on debian ) so here is the basic makefile to build it.

BOARD_TAG    = atmega328
ARDUINO_LIBS = Rainbowduino
ARDUINO_DIR   = /usr/share/arduino
ARDMK_DIR     = /usr/share/arduino
AVR_TOOLS_DIR = /usr

include /usr/share/arduino/Arduino.mk

/*
 Rainbowduino v3.0

 Scrolling text

*/

#include <Rainbowduino.h>

String message = "John is ace!";

void setup()
{
  Rb.init();
}

unsigned char x,y,z;

void scroll_text(String msg)
{
    int len = msg.length();
    const int kern = 6;
    byte msg_b[len+1];
    msg.getBytes(msg_b, len+1);

    for (int i = 0; i <= len * kern ; i++) {
        for (int b = 0 ; b < len ; b++) {
            if ((b * kern) > ( i - kern) && (b * kern) < (i + kern) ) {
                Rb.drawChar(msg_b[b],- i + (b * kern),1,random(0xFFFFFF));
            }
        }
        delay(200);
        Rb.blankDisplay();
    }
}

void loop()
{

    scroll_text(message);

}

A new year and a new blog post....

Trying to get back into the habit of sharing some interesting stuff. Again writing some comments etc.

Email can be very distracting and so we all try different techniques to limit those distractions. I am fairly settled now on an Inbox Zero type of approach. I don't always have me inbox open, I don't have any alerts or notifications and I read and file my mail at intervals during the day. It works for me for the most part and I don't find the nervous tick of checking my email every time my mind wanders to be too distracting.

All good, you may think. Lots of unbroken time for working on the real work but this does have one major fault. Less cake! Every so often we get an email that alerts us to the presense of cake in the building. These need an instant response. And this is where the cake flag comes in. Every time I get an email that is related to cake a Heath Robinson setup of filters, process and electronics kicks in a the cake flag is raised. We can then check our email and run to where the cake is. No more missed cake.

How does it work? There is a filter in gmail that identifies the important messages. (I know what you are thinking and unless you bring me cake an intentional false alarm means your email goes to the bottom of the long grass pile). This sends the email to another gmail account that I have registered for this porpose so that I don't have to store my main account details. I then have a small Python script that registers checks that gmail account for new mail. This then tells the Arduino on my desk, that is there to monitor if my chilli plant needs watering, that there is cake. The Arduino then prints a message on the LCD: ***CAKE*** and raised the cake flag by means of a servo. And everybody runs for cake.

There have been some improvements along the way. I have added a web api so people can reset the cake flag if I am not there. Later I noticed that we missed a call for Champagne this bug has now been fixed.

I use Vim as my everyday editor mostly because it's available on all the systems that I have ever administored. Once you get used to a modal editor though it gets very annoying trying to use anything else.

I have been trying to ignore the buzz around the Atom editor from GitHub for a while now but had to give in recently. So I downloaded, compiled and installed it. That was pretty painless and it makes a nice .deb by default so I can install it cleanly. Good so far.

Next up was to launch it and have a play. Wow! it's quick. I assumed it would be a fairly beefy startup and while nothing could be as slow as Eclipse I thought it might be of that order but no it's pretty lean.

Editing in it. Well it's pretty slick. There are a million new keyboard shortcuts and I think a lot of them are mac based so it's all a bit weird.

Making it easier with the vim plugin. I am not a great fan of plugins that make something behave like something else as it often makes things not work how they expect. In this case I don't think I can go back to a non modal editor.

It looks nicer. I can't deny that the smoothness of the interface and the cleanness of the style looks nicer than the harsh terminal interface. Things like having a builtin preview for Markdown or RST is nice.

In conclusion I might give this a bit of a go.

Some post about a lot of stuff, is plain text the way forward?