Android HTTP POST authentication error with Basic Auth

In one of my Android apps, a user should be able to push JSON data to the server. Of course this should only be allowed if the user is authorized by his credentials. For simplicity, I decided to use Basic Auth via HTTPS.

As I usually do not deal with Connectsions directly, I prefer to use the Apache Http classes to make life easier. So I used code like the following:

HttpParams httpParams = new BasicHttpParams();
HttpProtocolParams.setContentCharset(httpParams, HTTP.UTF_8);
HttpProtocolParams.setHttpElementCharset(httpParams, HTTP.UTF_8);
HttpConnectionParams.setConnectionTimeout(httpParams, 5000);
client = new DefaultHttpClient(httpParams);
    new AuthScope("localhost",80),
    new UsernamePasswordCredentials(username, password)

This was fine and worked in my local test environment and also in the live stage with GET requests. But when I started to POST data to the server I always got a “401 Unauthorized”. First I thought It’d be malformated JSON or encoding issues – but even with an empty string as payload I simply didn’t get through. Also the “same” POST request from the commandline using CURL worked like a charm! So it was very likely to be an issue with my handling of the Apache HTTP Lib. Stackoverflow also showed various people with the same problem. The proposed solution that came up several times was to manually set up the auth header. Not really the solution that I thought would be the greatest one.

After some more StackOverflow Posts and API crawling I found the very simple solution:

   new AuthScope(AuthScope.ANY),
   new UsernamePasswordCredentials(username, password)

And yes, it was that easy! Just set the Auth scope correctly.

I’m just very surprised that I only encountered issues with the POST request. The very same configuration worked with a GET without problems. This was actually the point that made me sure that this part of the code was correct.

RaspberryPi Weatherstation – The Sensor Station (Part II)

The second post of this series describes the RasperryPi that is connected to the sensors and used to record the data.


As described in the first part, I decided to use sensors from Tinkerforge as they are comparatively cheap, require no soldering, no electrical know how and they also provide a convenient API to get the data from the sensors. Ideal for people like me who want to spend more time with using the data instead of playing around with wires, resistors, etc.

The required hardware was just.

Raspberry Pi - Sensor Station


My first RaspberryPi project: TV-Controlled Weatherstation (Part I)


For quite some time I’ve been playing with the thought about bying a Raspberry Pi. Yet the biggest issue was: I had no use case for it!

I always liked information about the weather – especiallly I wanted to know in the morning how cold it was at night etc. So the use case was found: a weather station! Yet I was pretty distracted from all examples that I found on the web as I wanted a visually appealing UI so that it is fun to look at the data. Usually the samples just controlled a small LED display to display some values but – honestly – I don’t need a display at the weatherstation directly – I mean – shouldn’t it be somewhere outside and not in the room?


Tomcat MongoDB Authentication Realm

Tomcat MongoDB Authentication Realm

Tomcat provides a nice and powerful mechanism for authentication by the so called Tomcat Authentication Realms.
In the Tomcat Docs, Realms are explained as:

A Realm is a “database” of usernames and passwords that identify valid users of a web application (or set of web applications), plus an enumeration of the list of roles associated with each valid user. You can think of roles as similar to groups in Unix-like operating systems, because access to specific web application resources is granted to all users possessing a particular role (rather than enumerating the list of associated usernames). A particular user can have any number of roles associated with their username.


A basic Java WebApp example project supporting an xml free Jersey REST config

RESTful Webservices are an essential and convenient part in current web applications when Information should be exchanged. Additionally the main exchange formats in current webservices are either JSON or XML.

So when we want to implement the serverside of a web application using Java and especially REST services, we might end up serializing and deserializing objects to and from XML and JSON. Honestly not a very desirable task. But fortunately there is the Jersey project which can help us in this task.


“Easy going” vs “Taking care”

Freitag Abend. Heimfahrt in der BOB (BayerischeOberlandBahn). Wochenende here I come!

In Bad Tölz (eine Station vor meinem “Zielflughafen”) höre ich “Hausham? Da sind sind im falschen Zug.” (Hätte man vor einer halben Std in Holzkirchen umsteigen müssen). Ich denke mir “Arme Sau – jetzt fährst erst mal wieder zurück – oder du kennst hier jemanden der dich fährt – Ist mir aber wurscht – denn ICH bin gleich daheim. HA-HA!”.

Der Zug steht noch – ein junges asiatisches Mädchen das gerade ins Teenageralter eingetreten sein muss läuft etwas hektisch herum und spricht mich dann an “schuldi-gung – Hausham?”
Ohoh. “Hm tut mir leid, falscher Zug, steig schnell um in den Zug am anderen Gleis. Du musst zurückfahren.” Verzweifelter Blick ihrerseits. “ik ver-stehe nicht?”
Oh verdammt – kurzer Versuch in Englisch – nada. Noch maximum 2-3min bis der andere Zug fährt. Oh – scheiße – das bekomme ich jetzt nie erklärt. Die Hoffnung stirbt zuletzt: Noch ein Versuch es ihr zu erklären. Auch die Taxioption versteht sie natürlich nicht.

Option A) Schultern zucken, mich nicht weiter kümmern, das Mädel bis nach Lenggries fahren lassen (wo sie nicht weiter von Hausham weg sein könnte) und gleich daheim sein (also ich – sie sicher nicht). Ist ja schließlich nicht mein Kind.
Option B) Zusammenpacken, das Mädel in den anderen Zug setzen und hoffen, dass mich irgendwer fahren kann (oder dass diesmal wenigstens Taxis da sind) weil MEIN Zug dann ganz sicher weg ist.


“bad subifd directory” Replacing/removing bad SubIFD

Handling bad SubIFD entries in photos you want to geo tag / the problem

I just converted some of my RAW photos in RawTherapee and polished them in Photoshop CS. Of course I also wanted to GeoTag them (using GeoSetter) before putting the images in my archive. Fortunately, a GPS point was found in my trace for all photos. Unfortunately, I couldn’t write the Exif-GPS position due to a “bad subifd directory”. That’s what exiftool (which geosetter uses) tells me.

So – how to get the geo information into exif – or how to get the bad SubIFD directory out?

The solution:

In the end the issue was rather easy to solve:

  1. copy your images to a linux machine (or vm or whatever)
  2. copy all images to a directory “conv” AND in parallel to “conv2
  3. enter “conv2
  4. remove all exif information from the conv2 images by
    find ./*.jpg -exec exiftool -exif:all= {} ;
  5. enter “conv1” & copy the exif information (except SubIFD) to the cleaned images in “conv2” by
    find ./*.jpg -exec exiftool -tagsfromfile {} -exif:all --subifd:all ../conv2/{};
  6. you’re done.
  7. (fire up geo setter and geo tag your photos)

Be happy!

Convert Mercurial repository to Git

I have converted yet another googlecode Mercurial repository to Git – and as it took me (again) a bit too much time, here is my recipe:


Log into a Linux shell (Windows will hardly work – at least it didn’t work for me). If you don’t have some remote shell, download an Ubuntu VM and fire it up (you might need to install VirtualBox if you haven’t installed it already).

Check out your HG repo:

hg clone

Check if you have to remap some author information:
hg log MyHgProject

If you want/have to remap, simple create an author map file Lastname


Get fast-export and convert your repo:
git clone git://
mkdir new_git_repo
cd new_git_repo
git init
/path/to/ -A ../authormap -r /path/to/MyHgProject
git checkout HEAD


Now change the source setting in your google code repository to git and push the local codebase:
git remote add origin
git push --all

At this point, your upload might fail with something like “error: RPC failed; result=35, HTTP code = 0″.
This can happen if your upload takes too long and this is a documented bug.
I simply solved this issue by pushig the git repo from a shell with a fast enough upload speed as my local connection obviously was too slow.


Now you might realize that you have “lost” your complete wiki – don’t worry, it’s still there!
Switch back your repo setting to Mercurial and repeat the process for your wiki which you can usually find at

That’s it. You should now be ready to use Git!

Thanks to following sites hedonismbot, scrambled tofu

No more windows backup after uninstalling Acronis True Image Home 2011

I gave Acronis True Image Home 2010 and 2011 really quite several chances but it seems that either the product cannot reliably perform regular backups for my computer or I am just incomatible with it. Anyways – I finally wanted to draw a final stroke, get rid of Acronis True Image Home (ATIH), uninstall it and try Windows’ native Backup and Restore.

After unsintalling TIH I wanted to configure Windows’ backup and restore tool.
But – the according window just didn’t open!? Is TIH still bothering me AFTER I uninstalled it?

After quite some googling, I found one article in the Acronis Knowledge Base “14741: Acronis True Image Home 2011: Disabling “Integrate True Image into Windows” Does Not Work“. Unfortunately this only applies to an installed Version of TIH where Dis-Integrating TIH from windows backup does not work. – It’s quite a while since I tried it – but if I remember correctly, this didn’t work out for me. So another round at google was waiting for me…


How to write useful error messages

Each programmer has to deal with error mesages in some way. Either we have to check our data (and possibly raise an error/exception) or we are using libraries and have to deal with error messages or exceptions that are raised in these libraries. The purpose of this post is to point out how meaningfull error messages can save you time. Either your own time (by avoiding answering support questions) or your own debugging time.

So how can an error message save or waste time? Imagine the situation where you try to open a file and all you get is an error. What do you do next? I guess you’ll probably check if you’re trying to open the correct path. Maybe you did, then you possibly check if the permissions are okay, if you’re trying to open a directory or a file, etc until you found the problem that is causing the error. Actually this can be quite some wasted time – why? Becasue you’re doing all the stuff that the system already checked. The system decided that one of the condidtions failed and exited. So – shouldn’t the system just TELL the exact reason? And even better: shouldn’t it also just tell you a possible solution? Well – actually I do not blame “the system” but the guy who write the line that threw the error without any further information. Everytime I get an absolutely uninformative error I think “was it really so damn hard to add one more line of code that just mentions WHAT went wrong?!”.

Assume the situation where you try to open a file with the path being configured in an external file.
I think the quality of an error message can be categorized in one of the three categories:

  1. Reporting: The system just reports that there was an error.
    Example: IO Exception occured.
  2. Informing: Same as 1, just with more information.
    Example:  IO Exception occured while opening file: <filename>, null=false, exists=true, is file=false, is directory=false, readable=false, writable=false
  3. Supporting: same as 2 with additional information of how to solve the problem.
    Example: IO Exception occured while opening file: <filename>, null=false, exists=true, is file=false, is directory=true, readable=false, writable=false; Was the pat set correctly in

In the past weeks I have seen very useful error messages in Google’s Android (“… Have you declared this activity in your AndroidManifest.xml?”) or in twitter4j (Displaying a shortURL to an FAQ page explaingin a very common error).
Bad examples are plain NullPointerExceptions when 3 parameters were checked, one of them was null and the according is just followed by a throw new NullPointerException so that you even did not recognize from the message WHICH parameter was null.

Conclusion: Please write robust code and please provide helpful and supportive error messages to the folks using your code!