JLoe.net

Zend Caveats

Jonathan — Thu, 24 Dec 2009 08:41:40 +0000

Ever gets frustrated during your coding? I’m pretty sure the answer is a big YES especially if you’re new with the programming language you’re trying to develop or you might have forgotten the old programming language you used to. Likewise, most of the time we get frustrated when there is a stumbling block during your development especially when the deadline is imminent. The expression “Argh!!” or banging on a table are common to convey the frustration. :)

Anyway, here I’m going to share with some of my encounters that I can still recall during my past development using Zend Framework.

Controller

$this->_forward() somehow reset protected variables upon two consecutive executions. Unfortunately I am unable to recall the exact situation during the encounter.
Avoid using Zend_Form_Element_Submit->setAttrib('value','submit') but rather to use Zend_Form_Element_Submit->setValue().
$this->url( array( 'controler' => 'control', 'action' => 'action' )) will only set or replace the specified parameters and retain other parameters of the current url. In this example, the function will only set the controller name and action name with ‘control’ and ‘action’ respectively. Thus if there’s additional parameter in the url, say id, then the id and its value will be retained in the new url.
If your code needs to access variable(s) within a Controller, then put your code within an override function preDispatch() instead of init(). Function init() does not allow your code to access the desired variable within the Controller.

Database

To order your query result, use $select->order() instead of $select->fetch($query, $order).

Use $db->select()->setIntegrityCheck(false) if you encounter the following error message.

Multiple SQL statements are not allowed (SELECT ...; DELETE..;)
$db->select()->join()
Exception information:
Message: Select query cannot join with another table

You may refer to this site for more information.

Hope this post helps settle down your unhealthy frustration. Enjoy! :)

Don't leave just yet! You may also be interested to take a quick look at my other posts.

Start ORM with LINQ in .Net Framework

Jonathan — Wed, 23 Dec 2009 15:19:16 +0000

In this post, I’m gonna show you how you can kick start your development with LINQ, new extension to .NET Framework that encompasses language-integrated query, set, and transform operations.

If you are unfamiliar with the terms ORM and LINQ then refer to the following articles to help you understand the idea and the purpose behind them.

I have heard of LINQ years back but did not have a chance to give it a test-drive since its first release. For your information LINQ was first introduced in .NET 3.0 and had been evangelized long before its public release.

Without going further into unnecessary explanation, I’m going to start guiding you on how to start LINQ in .Net.

To make things simple and easy to follow, let’s first create a console application project from your Visual Studio. Once the new project is created, then add a new Reference System.Data.Linq into the project from the Solution Explorer under .NET tab. If you are unable to find the specified reference then .NET 3.0 Framework is not yet installed on your PC. Please install the required Framework prior to proceeding to the next steps below. You may then verify the new reference under References tree node in your Solution Explorer.

Next is to create a simple class to represent a Person which, for the sake of simplicity, only consists of several attributes like Title, Name, Country. I have created a sample class for your reference as shown below.

using System;
using System.Collections.Generic;
using System.Data.Linq.Mapping;
using System.Linq;
using System.Text;

namespace JLoeConsole
{
   [Table(Name = "Person")]
   class Person
   {
      private string _title;
      [Column(Storage = "_title", Name = "Title")]
      public string Title
      {
         get { return _title; }
         set { _title = value; }
      }

      private string _name;
      [Column(Storage = "_name", Name = "Name")]
      public string Name
      {
         get { return _name; }
         set { _name = value; }
      }

      private string _country;
      [Column(Storage = "_country", Name = "Country")]
      public string Country
      {
         get { return _country; }
         set { _country = value; }
      }
   }
}

As you may notice, I have included a reference System.Data.Linq.Mapping into the class header. This is to enable the use of attribute [Table(Name = "Person")] as well as [Column(Storage = "_title", Name = "Title")].

[Table(Name = "Person")] is used to let LINQ know that the class Person maps to a database table named “Person” whereas [Column(Storage = "_title", Name = "Title")] tells LINQ to store the value of a column “Title” in the table “Person” into a variable named “_title”. Notice that _title property is also declared in the class as the storage. The same explanation goes to the rest of the properties.

At this juncture, we just completed the creation of a simple mapping class or ORM class.

Next step is to open the default driver class Program.cs and complete the project with the following codes. The codes need to be inserted in the driver function static void Main(string[] args){ ... }.

// Set your connection string below.
string connectionString = @"";

// Initialize DataContext class for LINQ
// with the specified connection string
DataContext db = new DataContext(connectionString);

// Create a new instance of Table class
// to contain list of persons
Table Persons = db.GetTable();

// Below is the Lambda expressions for LINQ
// to search people who live in Singapore
var persons =
   from p in Persons
   where p.Country == "SINGAPORE"
   select p;

// Print names of all the person who stay in Singapore
foreach (Person person in persons)
{
   Console.WriteLine(person.Name);
}

I hope the comments inside the codes above are sufficient to help you understand. Just in case you’re not sure what to fill in for your connection string you may refer to this site for clue.

Well I can say that, with these codes inside your project, the application is now ready for your build or debug. Hopefully you don’t encounter any issue following these steps and this example gives you a good start into your exploration of LINQ.

Last but not least, since Christmas and New year is just a few days away from the time I’m writing this post, I would like to wish you a Wonderful Christmas and Joyful New Year. Happy Holiday! :)

I would also like to give my token of appreciation to the following sites.

Don't leave just yet! You may also be interested to take a quick look at my other posts.

Simple Steps to Start WCF with Visual Studio 2008

Jonathan — Sat, 19 Dec 2009 10:21:12 +0000

WCF, as you may have known, stands for Windows Communication Foundation. It is an interface provided in .NET framework for building connected and service-oriented application to support Service Oriented Architecture, in short SOA. However I won’t dig into detail about WCF and the relevant topics such as SOA. If you are interested to find out more about this technology, check out the WCF whitepaper or feel free to ask Uncle Google for help ;).

The objective of this post is to assist and demonstrate you on how you can start WCF using Microsoft Visual Studio 2008 in just few simple steps. FYI I just started experimenting this technology yesterday so forgive me should any of these steps are not achievable.

Before I start, please make sure you have the following installed on your computer.

Internet Information Services
Microsoft Visual Studio 2008

To start, first of course is to launch Microsoft Visual Studio 2008 which is then followed by creating a new project. Next is to select one of the available Web Templates, that is WCF Service Application. Refer to Figure 1 below.

Figure 1 - Create New Project

As depicted in figure 1, I chose C# as my preferred programming language for this project. If you prefer otherwise, then expand your desired language from the Project types panel and locate the same project WCF Service Application from the tree. Then type your preferred name. For this case JLoeService is used as the project name. Finally click Ok to confirm. Upon confirmation you should expect to see the following screenshot.

Figure 2 - Solution Explorer

As depicted on the left, a new project called JLoeService is created along with the necessary files: IService1.cs, Service1.svc, Service1.svc.cs, and Web.config. IService1.cs refers to an interface or the service contract provided by the application. Service1.svc is merely a service markup whereas Service1.svc.cs contains the code implementation of the contract defined in IService1.cs. Finally Web.config stores all your application settings or configuration values such as connection string, service declaration, etc. I will not explain here on some of advanced features you can define in Web.config like transforming the service envelop to JSON. JSON stands for Java Script Object Notation and is very popular for integration with AJAX (Asynchronous JavaScript and XML). The collaboration gives the most lightweight interaction for data interchange between web client and server.

Enough for that. Now lets try to build and run the new project in Debug mode by clicking Play button in the toolbar or press F5. If there is a prompt like shown in figure 3 and figure 4, click Ok and Yes respectively to ignore.

Figure 3 - Debugging Not Enabled

Figure 4 - Script Debugging Prompt

Figure 5 - ASP .NET Development Server

Upon successful build, please take note of the URL of ASP .NET Development Server is running as depicted in a tool tip located on top of your Windows task bar. In this example, the development server is running at http://localhost:2105/. A successful build could then be verified by a launch of new Internet Explorer or other default browser with a new page as illustrated in figure 6 and figure 7. Figure 7 is achieved by clicking Service1.svc from the figure 6.

Figure 6 - Directory Listing

Figure 7 - Service1 Service

With this address, proxy class and contract could be generated by svcutil.exe which will be used by a client to access the service. Svcutil.exe could be located inside C:\Program Files\Microsoft SDKs\Windows\v6.0A\bin folder. To generate the files, use your command prompt and move the current directory to the specified folder and run the command “svcutil.exe http://localhost:2105/Service1.svc?wdsl“. The execution will generate two new files and they are Service1.cs and output.config.

Figure 8 - Svcutil.exe on Command Prompt

Next step is to create a simple client to access the service. To begin with, a new client project has to be created. Here I’m going to create a console application called JLoeConsole. To create, stop the running project in previous step by clicking the Stop button and then click File > Add > New Project from the menu bar to create a new Console Application and click Ok to confirm.

Figure 9 - Create Console Application

Upon creation, move the generated proxy files: Service1.svc and output.config to the folder where the console application was previously created. As an example, I’m going to move the files from C:\Program Files\Microsoft SDKs\Windows\v6.0A\bin to my project folder D:\Jonathan Documents\Visual Studio 2008\Projects\JLoeService\JLoeConsole.

Figure 10 - Include Files in Project

Remember to rename output.config to app.config as the default configuration file for any console application. And finally include these files into the project. To include, first select the Show All Files in the Solution Explorer, select the two hidden files, right-click, and select Include In Project from the right-click menu. After the inclusion we need to include new references as these references are required for the compilation. Those references are System.Runtime.Serialization and System.ServiceModel. To include these references, right click on the References under JLoeConsole tree, select the mentioned references in the .NET tab, and click Ok. Next we need to write some codes into the main function in Program.cs. Copy and paste the following codes into the function.

static void Main(string[] args)
{
   Service1Client client = new Service1Client();
   string output = client.GetData( 10 );
   Console.WriteLine( output );
}

Now the project is ready for compilation. Select Build > Build JLoeConsole from menu bar to compile. A successful build is indicated by a text “Build succeeded” located at the bottom left of the status bar.

Finally both the service and the client are now ready for testing. To test, the service needs to run before the execution of client. To run the service, simply press the Play button or press F5 on the JLoeService project and ignore any warning prompts. Let the Internet Explorer run indefinitely to simulate as the running service. Last is to execute the client JLoeConsole.exe located in bin/Debug folder using Command Prompt. Before the execution, make sure the current directory in your Command Prompt point to the folder where the compiled JLoeConsole.exe is resided. And finally, execute the command JLoeConsole.exe. By then you should expect an output from the execution. That is “You entered: 10″ as an output in your Command Prompt. This indicates that the console application JLoeConsole has successfully communicated with the service JLoeService.

Lastly I hope the above explanation is easy to follow and you’re able to accomplish what I’ve accomplished without any unexpected encounters. I also hope this sample can help you kick-start your development on WCF. Feel free to download this example from this link. Have fun! :)

Don't leave just yet! You may also be interested to take a quick look at my other posts.

Hey, I am still alive!

Jonathan — Wed, 11 Nov 2009 13:19:18 +0000

As the title says, yeah I’m still alive. :)

I just started working in a new company and have been really busy with the new stuff going on in my new company. Things have been going really fast and it’s coming to my three months at the time I’m writing this post.

All I can say more challenges are coming in my new company, be it the technology or people management skills. I’m sure everything happens for a purpose.

Anyway stay tuned for my new posts. I will try to put as much information as I can if time permits. For now let’s just enjoy and learn something for the last several days before the year of 2009 ends. And yeah time flies so better make use of these times thoughtfully and wisely.

SQL Queries Performance Findings on MySql

Jonathan — Tue, 11 Aug 2009 09:02:43 +0000

Recently I was given a chance to get my mind and fingers into optimizing some heavy SQL queries on MySQL server. These examples are taken from real queries. I must admit that I am not an expert on this SQL stuff but I would like to share my useful findings in optimizing these SQL queries. Details of the queries, mainly table names, have been removed for clarity and security reasons. So check them out.

SELECT pig.*, pdt.*, ph.*
FROM pig, pdt, ph
WHERE
	LOWER( pdt.Status ) = 'active' AND
	pig.pid = pdt.pid AND
	pdt.pid = ph.pid AND
	pdt.SearchFriendlyUrl = 'xxx-xxx-xxx'

Showing rows 0 – 5 (6 total, Query took 4.5421 sec)

SELECT pig.*, pdt.*, ph.*
FROM
	(
		SELECT *
		FROM pdt
		WHERE
			SearchFriendlyUrl = 'xxx-xxx-xxx' AND
			LOWER(Status) = 'active'
	) pdt
	INNER JOIN pig ON pig.pid = pdt.pid
	INNER JOIN ph ON ph.pid = pdt.pid

Showing rows 0 – 5 (6 total, Query took 0.0107 sec)

Here my optimization strategy was to reduce the number of rows returned by the joined tables as much as possible and then to join the tables using INNER JOIN clause instead of WHERE clause.

SELECT
	pdt.*,
	dtd.name as ProductType,
	IFNULL
	(
		(
			SELECT GROUP_CONCAT( dtd.name SEPARATOR ' ' )
			FROM pt
				INNER JOIN dtd on pt.tid = dtd.tid
				INNER JOIN dv on dv.vid = dtd.vid AND LOWER( dv.name ) = 'travel type'
			WHERE
				pt.pid = pdt.pid AND
				LOWER( dtd.name ) != 'classic'
			GROUP BY pt.pid
			LIMIT 1
		),
		'Travel'
	) TravelType,
	(
		SELECT country.Name countryName
		FROM pt
			INNER JOIN country ON country.TId = pt.TId
		WHERE pt.PId = pdt.PId and country.vid = 6
		ORDER BY field( countryName, 'Multi-Country' ) DESC
		LIMIT 1
	) ProductCountry
FROM pdt, ptx, dtd, dv
WHERE
	pdt.pid IN
	(
		SELECT pId
		FROM pdr
		WHERE
			pdr.pId = pdr.pId AND
			pdr.parent = '156'
	) AND
	ptx.pid = pdt.pId AND
	dtd.tid = ptx.tid AND
	LOWER( pdt.Status ) = 'active' AND
	dtd.vid = dv.vid AND
	LOWER( dv.name ) = 'product type';

Showing rows 0 – 1 (2 total, Query took 3.0657 sec)

SELECT
	pdt . * ,
	dtd.name AS ProductType,
	IFNULL
	(
		(
			SELECT GROUP_CONCAT( dtd.name SEPARATOR ' ' )
			FROM pt
				INNER JOIN dtd ON pt.tid = dtd.tid
				INNER JOIN dv ON dv.vid = dtd.vid
			WHERE
				pt.pid = pdt.pid AND
				LOWER( dtd.name ) != 'classic' AND
				LOWER( dv.name ) = 'travel type'
			GROUP BY pt.pid
			LIMIT 1
		),
		'Travel'
	) TravelType,
	(
		SELECT country.Name countryName
		FROM pt
			INNER JOIN country ON country.TId = pt.TId
		WHERE
			pt.PId = pdt.PId AND
			country.vid =6
		ORDER BY field( countryName, 'Multi-Country' ) DESC
		LIMIT 1
	) ProductCountry
FROM pdr
	INNER JOIN pdt ON pdr.pid = pdt.pid
	INNER JOIN ptx ON pdt.pid = ptx.pid
	INNER JOIN dtd ON ptx.tid = dtd.tid
	INNER JOIN dv ON dtd.vid = dv.vid
WHERE
	pdr.parent = '156' AND
	LOWER( pdt.Status ) = 'active' AND
	LOWER( dv.name ) = 'product type'
ORDER BY pId

Showing rows 0 – 1 (2 total, Query took 0.0047 sec)

If you look carefully at the query above, I have removed the use of IN clause from the previous query. This is proven to have improved the performance by 99.8% from 3.0657 sec to 0.0047 sec. Well it’s clear that the improvement gained was remarkably significant.

SELECT *
FROM
	(
		SELECT *
		FROM pd
		WHERE LOWER( status ) = 'active'
	) pd
	INNER JOIN
	(
		SELECT *
		FROM pe
		WHERE parent = '857'
	) pe ON pd.pid = pe.pId

Showing rows 0 – 1 (2 total, Query took 0.1062 sec)

SELECT *
FROM pd
	INNER JOIN pe ON pd.pid = pe.pId
WHERE
	pe.parent = '857' AND
	LOWER( pd.Status ) = 'active'

Showing rows 0 – 1 (2 total, Query took 0.0008 sec)

Again, looking at the numbers the improvement once again was quite significant. This is achieved by using WHERE clause in the main query instead of in each of the joined tables. This is apparently opposite to my discussion on the first example.

Well, I won’t go into detail on why and how these figures come from as I don’t think I have the capability to explain the logic behind the numbers. I will probably leave the assumption to you now. I hope these examples may help you in producing more efficient SQL queries during your development. Have fun!

Don't leave just yet! You may also be interested to take a quick look at my other posts.

PHP ZipArchive corrupted

Jonathan — Mon, 06 Jul 2009 06:47:17 +0000

If you experienced a corrupted zip file produced by ZipArchive on WampServer you probably have come to the right place. Apparently there has been a discussion on this long time ago but looks like Wampserver still uses the old php_zip.dll.

Well, nothing much to write about as it’s quite straightforward. You might want to try downloading the latest binary file here to solve the issue. :)

Sources:

Zip becomes corrupt when using ZipArchive

Don't leave just yet! You may also be interested to take a quick look at my other posts.

Matt Mullenweg speaks about Improving Web Performance

Jonathan — Fri, 03 Jul 2009 06:15:56 +0000

Interesting and quite a coincidence I must say, to see another thought about web performance by Matt Mullenweg as recently I posted several posts about web performance: Web Performance Best Practices and One-Cent Tips to Increase Site Performance. For anyone who doesn’t know who is Matt Mullenweg, read more about him from Wikipedia. In short, he is the co-founder of a well-known software most bloggers used – Wordpress, and also a young entrepreneur living in San Fransisco. You may also want to check his blog.

There is a video of his recent talk in a technology conference where he spoke about how web performance is important and also shared his experience on handling these issues, mainly about his experiences in Wordpress and how the need for speed is really important.

Matt first talked about how Wordpress handles the 404 – aka Not Found page – causes a lot of performance overhead due to the fact that the 404 page generated by Wordpress is apparently database driven, unlike other static 404 pages. Matt also talked about how easy for developer to create a Wordpress plug-in and how these millions of plug-ins can contribute to the performance issue by introducing redundancies: multiple and identical javascript file inclusions in the HTML header. Therefore Wordpress created a new function for plug-in developers as a way to eliminate these redundancies.

He also spoke about how a millisecond can save time of people around the world collectively. He took a real example of his website WordPress.com where millions of bloggers have been using as their blog sites and explained how a small performance efficiency or improvement can indirectly promote a more eco-friendly environment: saving servers processing time thus less power consumption.

Interesting issue he raised is also on his experience when Wordpress made a total revamp of its user interface and apparently they faced a lot of issues when viewed on Internet Explorer which I believe most web developers have been having as a painful and frustrating task which we are forced to conform :P.

He also highlighted a few times on his favourite quote by Joe Clark that reads “We’ve upped our standards, up yours” which was coined at the beginning of the adoption of web standards. Anyway, check the video below for a full coverage of his talk.

Don't leave just yet! You may also be interested to take a quick look at my other posts.

Typogriph the Minimal Wordpress theme

Jonathan — Tue, 30 Jun 2009 08:48:33 +0000

Obvious enough, I just changed my theme from Fusion to a cleaner and lightweight theme called Typogriph as I had been looking around for a simple yet nice theme that suits my need. And finally came down to Typogriph as my best choice. There may be other popular theme called Thesis which most notable bloggers have been using but unfortunately this theme doesn’t come for free. It is priced at $87 for personal use and $164 for developer’s option. You can check the price in detail here.

The stuff that I like about this theme are:

Minimalist

Typogriph has a clean and minimal look and feel which hopefully results in more lightweight pages in order to give a faster loading and browsing experience – without relying so much on catchy images as well as heavy style sheet to appeal your readers.

Just a Switch

Changing to Typogriph from Fusion is just like pressing a switch. There was no hassle or other settings to make your pages to work correctly. With my activated plug-ins and just one or two tweaks to my AdSense widgets and there I have it, my new theme.

Fluid

Typogriph is a fluid theme which means that it makes full use of reader’s browser real estate and automatically adjusted when re-sized. This is definitely beneficial to meet various screen resolutions.

Browser Compatibility

Tested to render well across various browsers like Firefox, IE (6, 7, 8), Apple Safari, and Google Chrome.

On the contrary Typogriph does not render your page links well if you have pages as sub of other pages. The links were a mess so I had to fix that by moving sub pages as parent pages to quickly solve it. And everything else was acceptable and I love the arrangement of the widgets on the front page that showcases all my desired widgets, recent comments, and most popular posts.

In case you’re interested to explore other minimalist themes. Here are the sites I stumbled to.

So how you feel about this theme? Let me know ;)

Don't leave just yet! You may also be interested to take a quick look at my other posts.

Image Manipulation in PHP

Jonathan — Mon, 22 Jun 2009 08:49:17 +0000

Sometimes there is a need to do simple image manipulation for certain web projects. Take for example a popular site called Flickr. Flickr is an image hosting site that allows online community to store pictures for others to view. And for some cases some uploaders may have big image files and hence a site like Flickr definitely needs some image processing to reduce the actual image into a smaller dimension, hence smaller file size, to allow quick view before actually downloading the actual size.

I’ve been working with Zend Framework quite sometime for web projects. I found it’s quite interesting to find that Zend Framework does not have any class to perform image manipulation even though a proposal for Zend_Image had been submitted to Zend community as a wrapper of existing PHP interfaces: GD and ImageMagick but unfortunately the proposal was not considered.

And for your information, GD which stands for ‘Gif Draw’ is a standard PHP installation, unlike ImageMagick. If you wish to utilize these libraries, you need to have php extensions called php_gd2 and php_imagick respectively installed and enabled on the web server.

The noticeable difference between GD and ImageMagick is that GD simply does not support image format like TIFF (Tagged Image File Format). If TIFF is not your issue then go for GD. And if you wish to make your development easier, you may be interested to use Thumbnail class by our fellow developer. But if you’re looking into ImageMagick I’m going to show you the steps (from another site) on how to install ImageMagick on your Windows web server as Windows installation is a bit tricky since “pecl install imagick” does NOT work properly.

Download and install ImageMagick software from http://www.imagemagick.org/script/binary-releases.php#windows.
Download pecl-5.2-dev.zip (choose the version relevant to your PHP) from http://snaps.php.net/win32/
Copy php_imagick.dll from the archive you’ve downloaded to your PHP extension folder.
Add the following line to php.ini (in the extensions section):
extension=php_imagick.dll

or on WampServer click WampServer icon on taskbar, go to PHP > php.ini and add the above entry.
Restart your server.

Try this example script below to verify your installation.

header('Content-type: image/tif');

$image = new Imagick('example.tif');
// Create thumbnail for the specified image file.
$image->thumbnailImage(100, 0);

echo $image;

The script should display the image thumbnail with a maximum width of 100 pixels.

If you find the above steps are difficult to follow you may want to download php_imagick.dll here and start from step 3. And if you need help with these libraries, check the manual pages for GD and ImageMagick.

Don't leave just yet! You may also be interested to take a quick look at my other posts.

Web Performance Best Practices

Jonathan — Mon, 15 Jun 2009 07:29:56 +0000

Following-up my previous post on how to increase site performance. I received an email from my lead about a Firefox add-on named “Page Speed” that helps analyze a site performance.

In brief, Page Speed is an open-source Firefox/Firebug Add-on. Webmasters and web developers can use Page Speed to evaluate the performance of their web pages and to get suggestions on how to improve them.

Page Speed works by performing several tests on a site’s web server configuration and front-end code. These tests are based on a set of best practices known to enhance web page performance. Webmasters who run Page Speed on their pages get a set of scores for each page, as well as helpful suggestions on how to improve its performance.

But that’s not all, the good stuff is its article “Performance Best Practices” which explains in detail all aspects on what needs to be done to improve your site: from network (DNS, server settings) to browser layer (CSS, JavaScript, image, compression) which extends my previous post in greater detail.

However these best practices may not be the complete solutions, in my opinion. If your site deals with dynamic content, there should be programming logic behind the content. This means whoever creates the logic has to utilize the best logic for highest efficiency in terms of computation. And if your site deals with a database, optimizing the queries to the designated database is also required.

Ultimately basic rules for programming are to avoid redundancy and to distribute computation load whenever possible.

Don't leave just yet! You may also be interested to take a quick look at my other posts.