Author Archive
I find it quite annoying that when ever I do a search related to ExtJS, two sites (codeweblog.com and defafe.com) are on the first page of results. These two sites simply aggregate ExtJS forum posts, and then present them in a SEO friendly page.
Given this, I took to finding a way that can block search results. Officially, there’s no way you can do that from within Google settings (which would have preserved that as you moved from machine to machine), but there are extensions for Firefox. Still looking for a Chrome one!
For Firefox, there are a few extensions. One is CustomizeGoogle, but it doesnt work in the latest version of Firefox.
The one that does work is SurfClarity. It allows you to list of a sites that should be filtered out. It doesn’t do it from the server-side which would save a 1kb of traffic, but nonetheless is good enough.
for (var key in object) {
alert('value of '+key+' is '+object[key]);
}
I’ve posted a tutorial on how to get wireshark up and running on Ubuntu. The one problem , though, is that sometimes (actually often) you get [truncated]. How can you untruncate or turn it off or increase its size.
This isn’t the answer, but a way of how to get access to the full response:
Right Click on the Packet in Line-based text data, then choose Copy > Bytes (Printable Text Only).
This will place the full response, and you can paste it in a text editor.
The one thing that horrified me in the latest version of Ubuntu was the ‘new’ toolbar for Nautilus or the file manager.

However, after reading Deconstructing Nautilus and rebuilding it better, I have to admit, I enjoy it. Why?
Even though I tend to browser elsewhere by typing in the path, the UP button is probably the most used button by myself. Switching to the location bar provides no alternative, but keeping the button bar allows one a form of up button.
Secondly, CTRL+L is probably something I can live with.
Given the whole debate around this, the lesson I believe is that don’t ignore your existing users. If you are going to change something, inform your users!
Given the opensource community spirit, it wont be long before someone does a quick script to bring back the old toolbar. If you are looking to do that today, have a look at: https://help.ubuntu.com/community/RestoreNautilusLocationBar
For the past couple of weeks, I’ve taken to reading Muhammad: His Life Based on the Earliest Sources by Martin Lings. One of the many touching incidents occurred in Madina.
The Prophet (PBUH) had to serve in multiple roles as the leader of the Muslims, governor of Madina, prophet, etc. At this time, he divided his day into three. 1/3 for family time, 1/3 for community affairs and a 1/3 to spent in worship and prayer.
Some of the companions felt as if they were losing touch with the Prophet, given his multiple roles, fixed time in them, and the large amount of people who were all vying for his attention.
In this regard, Allah revealed the verse:

Surely Allah and His Angels bless the Prophet. Oh you, who believe, call for blessings on him, and salute with (respectable) salutations. (Chapter 33: Verse 56)
This would be their (and our way) of staying connected with him!
The Prophet (PBUH) would also add: “Whoever invokes blessings on me once, Allah will invoke blessings upon him ten times”
One of the most annoying issues using Vodacom to connect to the internet is that sometimes the DNS server used is so badly rotten, internet use is practically useless.
DNS is the system that converts a domain example www.google.com to an IP Address, something ‘computers’ can interpret and understand. When this process fails, it may appear the internet is down. All that is happening, however, is that the process of converting domain names to IP Addresses is slow (or dead).
One way to fix this is to use alternative DNS servers/addresses. By switching DNS addresses, you are basically saying, use this server to convert domains to IP addresses.
First things to check:
- You have an internet connection. This will not help if you do have one
- Check that entering an IP address works, example: http://41.1.224.101
- http://41.1.224.101 (vodacom.co.za)
- http://41.203.21.136 (mybroadband.co.za)
- http:// 196.8.95.20 (standardbank.co.za)
- http://74.125.43.106 (google.co.za)
If this works, but entering www.vodacom.co.za doesn’t, then changing DNS entries will help. Edit your connection settings, and try one of the following:
196.43.46.190 – SAIX
196.207.40.165 – Vodacom
168.10.2.2 – Internet Solutions (IS)
208.67.222.222 – OpenDNS
208.67.220.220 – OpenDNS alt
It is possible to mix them up, rather than using two by the same company.
The National Energy Regulator of South Africa on Wednesday granted state-owned utility Eskom a 24,8% tariff increase for the 2010/11 financial year, falling short of the power firm’s request for a 35% hike. For the following two financial years, Eskom was granted increases of 25,8% and 25,9% respectively.
How much does that amount to for you? Try the calculator below:
Read the story on Mail and Guardian
Here’s a Google Map of the road closures for the opening of South African Parliament
View Opening of Parliament 2010 in a larger map
Would it not have been easier for them to stick to DeWaal Drive, Roeland Street?
Assume you have the following table called stories with the following data. And you have run the following query:
SELECT * FROM stories ORDER BY storydate DESC
| id |
source |
title |
storydate |
| 10 |
cnn |
Story 10 |
2010-01-21 |
| 9 |
bbcnews |
Story 9 |
2010-01-20 |
| 8 |
bbcnews |
Story 8 |
2010-01-19 |
| 7 |
skynews |
Story 7 |
2010-01-18 |
| 6 |
cnn |
Story 6 |
2010-01-17 |
| 5 |
bbcnews |
Story 5 |
2010-01-16 |
| 4 |
cnn |
Story 4 |
2010-01-15 |
| 3 |
skynews |
Story 3 |
2010-01-14 |
| 2 |
skynews |
Story 2 |
2010-01-13 |
| 1 |
cnn |
Story 1 |
2010-01-12 |
Now for the challenge. What happens if you are required to get the top two stories from each source? One option is to first get the order of the sources, and then loop through each one of them:
SELECT DISTINCT source FROM stories ORDER BY storydate DESC
<loop “source” as “sourceid”>
SELECT * FROM stories WHERE source = "{sourceid}" ORDER BY storydate DESC LIMIT 2
</loop>
Another option is to do the in MySQL itself! Unsure if there is a better way of doing this, but here’s mine using UNION and Sub Selects:
(SELECT * FROM stories WHERE source =
(SELECT DISTINCT source FROM stories ORDER BY storydate DESC LIMIT 0,1) ORDER BY storydate DESC LIMIT 2)
UNION
(SELECT * FROM stories WHERE source =
(SELECT DISTINCT source FROM stories ORDER BY storydate DESC LIMIT 1,1) ORDER BY storydate DESC LIMIT 2)
UNION
(SELECT * FROM stories WHERE source =
(SELECT DISTINCT source FROM stories ORDER BY storydate DESC LIMIT 2,1) ORDER BY storydate DESC LIMIT 2)
Explanation:
The bold section are subselects and form the WHERE clause of the query. Here we want the first, second and third as per source, but only one at a time. Once we have two results per item, we use UNION to join them together.
One thing to note though. Even though it returns the latest two stories per source, the final table is not sorted by storydate! This can be done using PHP sorting techniques. Is there a way of doing this one time in MySQL?
Fiddler is a useful tool for tracking http requests and responses. It’s similar to the Net tag in Firebug, except that it can be used to track all requests, not only browser-based ones.
For Ubuntu users, the recommended alternative is Wireshark. Wireshark is noted as being more powerful than Fiddler, but the focus will be on the features most commonly used by web developers for tracking.
Installation
Wireshark is found in the Ubuntu repositories, so simply search for wireshark in Synaptic, or enter sudo apt-get install wireshark
Usage
Under Applications > Internet, you will notice there are two options. Wireshark and Wireshark (as root). Run the second one (as root) as this gives you more data to track.

After wireshark has started, the process is straightforward:
- Choose which network interfaces to capture traffic from
- Start capturing
- Filter captured requests
1. Choosing which network interfaces to capture traffic from. From the menu, choose: Capture > Interfaces

This will bring up a list of network devices.
2. Since my work involves debugging work off my laptop (http://localhost etc), I only start device lo (127.0.0.1). This is useful as it immediately ignores all other traffic.
3. The one thing you will notice is that fiddler will log both the TCP and HTTP requests. For the purpose of debugging AJAX, etc., we are only interested in HTTP requests, TCP requests are not required. These can be hidden by adding a filter.

In the field next to Filter:, enter http and click on Apply.

The steps will now show all traffic coming on 127.0.0.1 via HTTP. The last step is picking up the AJAX parts for debugging. Notice that for each request, there is a response. Unlike Fiddler, Wireshark does not combine the request and the response. So click on the response line, and then expand the Line-based text data row. This will show the text of the AJAX response.

Conclusion
Wireshark takes more steps than Fiddler, and Fiddler is a more focussed program than Wireshark. Wireshark also captures traffic without having to install a plugin into firefox, etc., so it’s useful for debugging applications as well. The biggest improvement to make Wireshark truly replace Fiddler is to turn off truncation of the response text. Wish I knew how!