Browse Category

Web

Laravel: “Allowed memory size exhausted” error during unit tests

I’ve been using in-memory testing for my projects based on Laravel every since I found this great  tutorial: http://net.tutsplus.com/tutorials/php/testing-like-a-boss-in-laravel-models/

It was working very well as my test cases increases, to a point when I’m starting to get this error message:

Fatal error: Allowed memory size of 134217728 bytes exhausted (tried to allocate 16 bytes) in /Applications/AMPPS/www/project/vendor/symfony/console/Sym
fony/Component/Console/Command/Command.php on line 57

I couldn’t find the proper solution, but managed to get it working with a workaround here: http://forums.laravel.io/viewtopic.php?id=11723

We can temporarily increase the memory limit using this function:

// Temporarily increase memory limit to 256MB
ini_set('memory_limit','256M');

In my case, I needed more than 128MB of memory, so I conveniently increased it to 256MB. It depends on your usage and your hardware. Choose accordingly.

Since this is only required during my test (I follow the TDD approach, so testing comes regularly), I had added it in the TestCase class, inside method createApplication().

This is the full code in TestCase.php.

<?php

class TestCase extends Illuminate\Foundation\Testing\TestCase {

public function createApplication()
{
    // Temporarily increase memory limit to 256MB
    ini_set('memory_limit','256M');

    $unitTesting = true;
    $testEnvironment = 'testing';

    return require __DIR__.'/../../bootstrap/start.php';
}    

/**
 * Default preparation for each test
 */
public function setUp()
{
    parent::setUp();
    $this->prepareForTests();
}

/**
 * Migrates the database and set the mailer to 'pretend'.
 * This will cause the tests to run quickly.
 */
private function prepareForTests()
{
    Artisan::call('migrate');
    Mail::pretend(true);
} 

}

jQuery using $.on to call newly added elements

More than a year ago, I had written a blog about how to bind events to newly added elements using jQuery.

The scenario can be found here: http://blog.kongnir.com/2012/02/08/jquery-function-not-binding-to-newly-added-dom-elements/

In my previous solution, I had used $.live() in place of $.change(). However, as noted by a commenter in that blog, as of time of writing, $.live() is deprecated.

Instead, we should now use $.on() if you need to capture events of newly added elements.

For example, in normal circumstance, we would do this:

$(".existing-button").click(function() {
    alert("This will work if button already exist on load");
});

However, if you add new elements to the document using AJAX, then the above will not work. That button doesn’t exist.

In both circumstances, this will work:

$(document).on("click",".new-button", function() {
    alert("This will work even if element is newly added");
});

If you notice, the syntax is quite different. The element to call on is the containing element, not the element that the event is called. So let’s say “.new-button” is contained within “#main”, then the following should work too.

$("#main").on("click",".new-button", function() {
    alert("This will work even if element is newly added"); 
});

Reference: http://api.jquery.com/on/

Laravel 4: things to note when creating Package

NOTE: This tutorial was written for Laravel version 4.0. I’ve added some notes for version 4.1 but they’re not tested yet. Do let me know if you see any problem.

When I first started using Laravel 4, I had been creating all my models, controllers and views in the root app folder. While things can get too comfortable and convenient here, it may not be a good idea to dump everything at the top. That’s when Package (or Bundle) comes in.

According to Laravel’s documentation:

“Packages are the primary way of adding functionality to Laravel.”

I found 2 great tutorials on creating your own Laravel 4 Package:

  1. http://culttt.com/2013/06/24/creating-a-laravel-4-package/
  2. http://jasonlewis.me/article/laravel-4-develop-packages-using-the-workbench

The following are some of the problems I’ve encountered and it took me quite some time to look for the solutions. Hope this will help someone else too.

1. Define your dependencies

A Package is like a sandbox module where it should contain its own dependencies (or vendors Packages). So you need to add your dependencies to your Package’s composer.json file.

As an example, I needed to include Cartalyst Sentry and Twitter Bootstrap to my Package, this is what I did under “require”:

"require": {
        "php": ">=5.3.0",
        "illuminate/support": "4.0.x",
        "cartalyst/sentry": "2.0.*",
        "twbs/bootstrap": "3.0.*"
},

 NOTE on Sentry:

After writing this blog, I’ve decided to move Sentry out of my package. Firstly, it will be cleaner and allow people using my package to choose their own authentication vendor. Secondly, I was getting so much trouble trying to get it to work properly. So it’s advisable to just put it on the Laravel app root instead.

Remember to run the following command line at your Package’s root folder:

php composer.phar update
php composer.phar dump-autoload

Note: On your first install, run “php composer.phar install”. Although using update will do the same thing too.

For dump-autoload, on the official documentation, they are using artisan instead of composer. I’m still not sure what’s the difference, but both seem to work.

php artisan dump-autoload

2. Extending Controller

When you download the Laravel 4 master, you will get a BaseController.php under the app\controllers folder. It looks something like this:

<?php

class BaseController extends Controller {

    /**
    * Setup the layout used by the controller.
    *
    * @return void
    */
    protected function setupLayout()
    {
        if ( ! is_null($this->layout))
        {
            $this->layout = View::make($this->layout);
        }
    }
}

And then all your custom controllers will extend this BaseController like this:

class CustomController extends BaseController {}

If you copy and paste this BaseController to your Package, it will not work. The Package will try to find “Controller” in your namespace Vendor\Package.

So to make it work, you must first define BaseController under your namespace, and then let Package knows that it needs to find the correct Controller.

In short, use this:

<?php namespace YourVendor\YourPackage;

use \Illuminate\Routing\Controllers\Controller;

class BaseController extends Controller {

     /**
     * Setup the layout used by the controller.
     *
     * @return void
     */
    protected function setupLayout()
    {
        if ( ! is_null($this->layout))
        {
            $this->layout = View::make($this->layout);
        }
    } 
}

Note for Laravel 4.1:

In Laravel 4.1, the Controller path has been moved. See http://laravel.com/docs/upgrade

Use this instead:

<?php namespace YourVendor\YourPackage;

use \Illuminate\Routing\Controller;

class BaseController extends Controller {

     // Your code
}

3. Using Sentry in your Package

Just to be clear, as some readers mistaken “Sentry” as my Package’s name. It’s a third party Package created by Cartalyst: “Sentry is a simple, powerful and easy to use authorisation and authentication package.” I wanted to include this third party Package in my custom Package to add authorisation and authentication capability.

But once Sentry is downloaded and installed to my Package, I still couldn’t get it to recognise the alias “Sentry”. So the following code produces an error:

Sentry::logout();

The FatalErrorException message was:

Class ‘Vendor\Package\Sentry’ not found

That’s right, the Package has mistakenly treated Sentry under my Package’s namespace.

If you remember from Laravel’s basic documentation, while in the app’s root folder, we can define aliases such as “Sentry” in the \app\config\app.php file. But there doesn’t seem to have any way to register that in my Package.

So what I did, as a workaround, is to include the namespace at the top of my file. Like this:

<?php namespace YourVendor\YourPackage;

use Cartalyst\Sentry\Facades\Laravel\Sentry;

class CustomController extends BaseController {
    public function logOut()
    {
        Sentry::logout();
    }
}

Works like charm! But if you know of a shorter and better way, please leave me a message!

Better solution (edited)

Thanks to hardik dangar, there’s actually a shorter way to do this. Simply add a backslash before Sentry so that php doesn’t search Sentry within my namespace.

<?php namespace YourVendor\YourPackage;

class CustomController extends BaseController {
    public function logOut()
    {
        \Sentry::logout();
    }
}

4. Using Illuminate Facades

Just as you thought everything is well and ready to code, you found yourself stuck at this line!

return View::make('vendorpackage::users/view');

And the FatalErrorException message is:

Class ‘Vendor\Package\View’ not found

What the?!

The same goes to Redirect, Input and Validator.

Base on my previous experience with Sentry, I got just the right idea.

Simply add the following to the top of your code:

<?php namespace YourVendor\YourPackage;

use Illuminate\Support\Facades\Validator;
use Illuminate\Support\Facades\Redirect;
use Illuminate\Support\Facades\Input;
use Illuminate\Support\Facades\View;

class CustomController extends BaseController {

Tada!!

Better solution (edited)

Thanks to hardik dangar, there’s actually a shorter way to do this. Simply add a backslash before View so that php doesn’t search View within my namespace.

return \View::make('vendorpackage::users/view');

5. Using Views in your Package

That right, you didn’t see it wrongly. This is how you call your Package’s view.

For example, we want view “users\view”, add your Package’s name to the front like this:

View::make('vendorpackage::users/view');

Then how about using master layout? Same logic, like this:

@extends('vendorpackage::layouts.master')

Where the directory is:

workbench\yourvendor\yourpackage\src\views\layouts\master.blade.php

6. Extending Eloquet in your own model

Laravel documentation taught us this:

class User extends Eloquent {}

This won’t work in a Package. Instead, you should do this:

<?php namespace YourVendor\YourPackage;

use Illuminate\Database\Eloquent\Model;

class User extends Model {}

7. Using Controller in your Package

This used to work:

{{ Form::open(array('action' => 'UserController@postStore')) }}

But to use your Package’s Controller, you need to add your Vendor\Package namespace to the Controller like this:

{{ Form::open(array('action' => 'YourVendor\YourPackage\UserController@postStore')) }}

8. Loading Package Config file

If you want to create a separate database for your package, you can define your database configuration within your package directory:

workspace\vendor\packagename\src\config\database.php

Then add this to the package service provider boot() method:

public function boot()
{
    $this->package('vendor/package');
    include __DIR__."/routes.php";

    // Add my database configurations to the default set of configurations                        
    $this->app['config']['database.connections'] = array_merge(
        $this->app['config']['database.connections']
       ,Config::get('package::database.connections')
    );
}

Source: http://stackoverflow.com/questions/15304722/eloquent-laravel-4-package-database-configuration-format

9. Loading external package within Providers and Aliases

In normal circumstances, when we add a new package to a Laravel installation, we would edit the config/app.php by adding the package namespace within the “Providers” and “Aliases” arrays.

However, when we need to add another external package that is a requirement within our own custom package, we wouldn’t want our users to edit a long lists of Providers and Aliases. Ideally, they should only include our custom package.

I’ve been searching for awhile and finally found a solution from this forum.

Solution:

In this example, we want to add an external package call “Markdown”.

In the file myLaravelProject/workbench/my-vendor/my-package/src/MyVendor/MyPackage/MyPackageServiceProvider,

put this in the boot() method:

$this->app->register('SomeExternalPackage\Markdown\MarkdownServiceProvider');

and this in the register() method:

$this->app->booting(function()
{
    $loader = \Illuminate\Foundation\AliasLoader::getInstance();
    $loader->alias('MyPackage', 'MyVendor\MyPackage\Facades\MyPackage');
    $loader->alias('Markdown', 'SomeExternalPackage\Markdown\Facades\Markdown');
});

10. Override config files of package

If you want to be able to allow the users who use your package to publish and modify the package’s config files, following this tutorial:

How to override config files of Laravel 4 package

Tricky WebDev: Fix missing list’s text in Mobile browsers

I had this very strange bug that didn’t show up in Desktop browsers (Chrome and Firefox) but in Mobile browsers, specifically iOS Safari and Chrome.

On Desktop:
tags_on_desktop

On Mobile:
tags_on_mobile

Since I couldn’t find a way to debug on the Mobile browsers, I had spent quite a while trying to figure out what went wrong. It’s also very strange to actually see the texts that have more than 1 word to appear, but yet it only shows the first word.

So I checked the HTML code from Chrome and saw this:

<li>
    "                             Face                             "
</li>

That’s because in my PHP code, I had actually echoed the text on a new line like this:

<li>
    <?php echo $tag; ?>
</li>

Out of curiosity and pure trial-and-error, I had decided to try this:

<li><?php echo $tag; ?></li>

To produce this:

<li>Face</li>

Guess what?
That solved the bloody issue I had been investigating for many hours.
I think I really need to re-learn the basics of HTML.

Missing notes for Laravel 4

Tutorial: Missing notes for Laravel 4

Note: This is an on-going blog. I’ll be updating it whenever I have new findings for Laravel 4.

I had been learning CodeIgniter and CakePHP for some time now when I chanced upon this new(er) framework called Laravel (http://laravel.com) via Twitter.

I’m not going to tell you how wonderful it is, but here’re some references:

There’s an official documentation and another one which is almost recommended everywhere I googled.

The thing is, the official documentation isn’t very complete. It covers most of the basics, but when you want to dig deeper, you’ll get lost. The other recommended tutorial site is almost outdated. I couldn’t find some of the things when I’m comparing it to the official documentation.

So I’m taking notes on the things that I’ve found out (either through googling or the official API documentation). Here we go:

1. Form::open_for_files() is outdated

If you’re trying to create a form for uploading files, the old way was to open the form with Form::open_for_files().

It isn’t clear in the official documentation or the API that we should now use the same Form::open() but now with an additional argument ‘file’ => true.

{{ Form::open(array('action' => 'UploadController@postImage','files'=>true)) }}

2. Using Input::file()

You must always set ‘files’ => true in the Form::open() otherwise Input::file() will never work.

3. Getting filename from Input::file()

There were many tutorials saying that the following will return the filename of the upload file. But it didn’t work for me.

$filename_neverwork_forme = Input::file('image.name');

Instead, use this to get the filename of the upload file.

$filename_worked = Input::file('image')->getClientOriginalName();

Of course, using the filename directly without verification is dangerous. But that’s up to you to decide.

4. Redirect with validation error message and inputs

I realize most methods with underscores are outdated.

So this is wrong:

return Redirect::to('form/create')->with_errors($validation)->with_input();

This is the correct method:

return Redirect::to('form/create')->withErrors($validation)->withInput();

5. Getting URL of a location using URL::to()

To get the full URL of any location or route, simply use URL::to().

For example, if you need the URL of the route ‘about’

{{ URL::to('about') }}

Note: Remember to set your route.php correctly. This function depends on that.

6. Checking current location with Request::is()

I needed to track the current location so that I can mark the corresponding navigation button as active. In this case, you can use Request::is().

Example

@if ( Request::is('about') )
    <li class='active'>
@else
    <li>
@endif
        <a href='{{ URL::to("about") }}'>About</a>
    </li>

 7. Get distinct values from a table column

I had a list of string from a table column that can be repeated. So I needed to query the database to give me an array of string which are distinct.

This is how you can do it:

DB::table('tableName')->select('columnName')->distinct()->get()

 8. Using Mail::send()

This is quite frustrating because the official documentation didn’t describe how to use the $data parameter which is required when using Mail::send().

Found this very good tutorial:

http://maxoffsky.com/code-blog/sending-e-mail-with-laravel-4-using-mail/

 9. Deploying your Laravel project to a shared hosting

Reference: http://stackoverflow.com/questions/16683046/how-to-install-laravel-4-to-a-web-host-subfolder-without-publicly-exposing-app

9.1 Look for your root or home directory

If you’re not sure where is your root or home directory, login to your shared hosting’s cpanel (control panel) and look for the Home Directory. It’s usually something like “/home/domainname”.

9.2 Create a new folder to host the protected Laravel files

We need to create a new folder in the root directory so that they’re not accessible to public. Under the root directory (e.g. /home/domainname/ or example.com), create a folder “applications”.

Under “applications”, create a folder with your project name (e.g. project_name) and place all your Laravel files (except the “public” folder) into this folder.

9.3 Public folder

Put all your files inside Laravel’s public folder into /home/domainname/public_html/project_name.

9.4 Final folder structure

The end result of the folder structure will look like this:

- /home/domainname
    - applications
        - project_name
            - app
            - bootstrap
            - vendor
            - artisan
            - composer.json
            - phpunit.xml
            - server.php
    - public_html
        - project_name

9.5 Edit Laravel’s /bootstrap/paths.php

I was very lost how many “../” I need to add in order to get to the correct path. So we won’t use __DIR__ at all. We’re going to define the full path base on the root directory so that there’s no confusion.

So change from this:

return array(
    'app'     => __DIR__.'/../app', 
    'public'  => __DIR__.'/../public',
    'base'    => __DIR__.'/..',
    'storage' => __DIR__.'/../app/storage',
);

To this:

return array(
    'app'     => '/home/domainname/applications/project_name/app', 
    'public'  => '/home/domainname/public_html/project_name',
    'base'    => '/home/domainname/applications/project_name',
    'storage' => '/home/domainname/applications/project_name/app/storage',
);

9.6 Edit Laravel’s /public/index.php

Same as the point above, we’re not going to use __DIR__. Let’s change all the path to full path.

From this:

require __DIR__.'/../bootstrap/autoload.php';
$app = require_once __DIR__.'/../bootstrap/start.php';

To this:

require '/home/domainname/applications/project_name/bootstrap/autoload.php'; 
$app = require_once '/home/domainname/applications/project_name/bootstrap/start.php';

9.7 Edit .htaccess

If your shared hosting’s default PHP version is < 5.3.7 but they support higher versions, you can edit the .htaccess file inside the “public” folder.

For Mac users, you can access it via the Terminal.

Use this command to go to the folder

cd your/laravel/path/public

Then list hidden files with this command

ls -a

You should see .htaccess there. By default, this file is protected. In order to change it, you need to open it using super user like this:

sudo open -e .htaccess

Note: -e means open with text editor.

Then append this line to the bottom and save it.

AddHandler application/x-httpd-php53 .php

Or if you want version 5.4

AddHandler application/x-httpd-php54 .php

Upload the updated .htaccess to the public_html/project_name folder and you’re done!

10. Returning view with Error message

When you need to return to a page with error messages (especially if you use validation), you can pass the error messages back to a view with this:

return Redirect::to('form')->withErrors($validation)->withInput();

However, if you need to return an error that is not captured in Laravel’s validation, for example, I needed to check if the reCAPTCHA returns a matching input, you can create a new MessageBag and return that error messages instead.

$errors = new IlluminateSupportMessageBag;
$errors->add('customError', "The reCAPTCHA wasn't entered correctly.");
return Redirect::to('form')->withErrors($errors)->withInput();

But what if you have both validation and custom message? This is how I’ll do it: first, check that validation has passed, if not, return with validation error messages. Then, check that, in my case, reCAPTCHA is matching, if not return that error message.

Example:

if( !$validation->passes() )
{
    return Redirect::to('form')->withErrors($validation)->withInput();
}

// CAPTCHA was entered incorrectly
if (!$reCaptcha->is_valid) {
    $errors = new IlluminateSupportMessageBag;
    $errors->add('reCaptcha', "The reCAPTCHA wasn't entered correctly.");
    return Redirect::to('form')->withErrors($errors)->withInput();
}

11.Creating a Package

Reference:

Important commands:

  1. To generate the class map of package:
    composer dump-autoload
  2. To move package’s assets into public/packages:
    php artisan asset:publish --bench="vendor/package"

12. Laravel requires Mcrypt PHP extension

If you’re using Mac like me, the default PHP installed on the Mac doesn’t insclude Mcrypt extension. I’m using MAMP for development and testing, so I followed this method and it worked for me.

Reference: http://stackoverflow.com/questions/16830405/laravel-requires-the-mcrypt-php-extension

Enter “which php” in terminal to see which PHP you are using.

which php

By default, Mac uses PHP from this path:

/usr/bin/php

If it’s not the PHP version from MAMP, you should edit or add .bash_profile under user root document (cd ~).

In .bash_profile, add following line (your MAMP version may be different, so does your PHP version. Check it up on the folder /Applications/MAMP/bin/php/):

export PATH=/Applications/MAMP/bin/php/php5.4.4/bin:$PATH

And restart Terminal to see which PHP you are using now. And it’ll be working by now.

13. Using lists for Form::select()

The shorter way to populate an array to be used in Form::select() is using query builder ‘lists’ to retrieve just the id column and a value column.

For example:

$select_array = Classname::lists('name', 'id');

Read the link below for more tips about this trick.

Laravel 4: Using query builder lists for Form::select

References to good BYO PHP MVC Framework tutorials

When I first came across the term MVC (Model-View-Controller) few years back, the whole concept sounded so complex, especially for people with (old) ASP and PHP background. In those days, we were so used to mixing business logics inside the presentation layer. It was so logical and convenient, until the code got big and things got out of hand. We forgot where the codes were, and we risked stability even for small changes to the UI, because they were all intertwined with business logics.

Typical PHP-HTML mixed code:

<?php
  include "someFileContainsFunction.php";
  public function someLocalFunction( $param )
  {
    return "Business logic in <b>$param</b>!";
  }
?>
<html> 
  <title>HTML with PHP</title>
  <body>
    <h1>My Example</h1>
    <?php
      print someLocalFunction( "Inside body" );
    ?>
    <b>Here is some more HTML</b>
    <?php
      print functionFromExternalPHP();
    ?>
  </body>
 </html>

Then I was introduced to MVC during a Microsoft .NET bootcamp in Singapore. I was quite fascinated with the idea. Separating the 3 components enable us to distribute the work to programmers and designers, allowable them to do their work without touching the fields they’re not familiar with. It is also an ideal solution to multinational collaboration.

Although the .NET MVC framework was great as a development tool, I couldn’t understand the fundamentals of building an MVC. We merely followed the coding style requirements by the framework, and work our way through to make the application work.

So I went in search of other languages, and found many other frameworks such as Ruby on Rails (for Ruby) and CakePHP (for PHP). They are also great MVC frameworks, but then again, they’re very established with quite stringent coding style requirements that I always got lost halfway down the development.

http://www.netmagazine.com/features/choose-right-php-framework

I thought the best way to learn about something is to start from the beginning, and keep on testing and failing until I understand it. The following 2 tutorials are great starting points. I managed to write my own MVC framework within a day (or maybe a few hours) by following the tutorial from Domagoj Salopek. I would recommend going through this tutorial first before going to the next one, which is slightly more complex but covers a little more for MVC.

http://www.domagojsalopek.com/Details/Create-a-simple-PHP-MVC-Framework/28

http://johnsquibb.com/tutorials

Of course, there’re people who think that using MVC on a small project is an overkill. Trust me, it’ll help you in the long run. By writing a core MVC framework, you’ll be able to use it in all other projects in the future, regardless how big or small it is. Unless the client only wanted a “simple” website. I know clients always think their requirements are simple because they don’t understand the simpler it seems, the more work it takes. I’m referring to those instances where the project can be done just with HTML and JavaScript. In those cases, MVC is really too much.

Now, going back to fiddling the simple MVC framework that I have written.

Edited: After writing my own simple MVC framework, I went on to study other major PHP framework such as Symfony, CakePHP and CodeIgniter. I noticed that CodeIgniter uses a very similar approach to the simple framework introduced by Domagoj. So if you’re advancing to a more complex framework, take a look at CodeIgniter. You’ll be glad he had written something so fundamental to get us started.

Setting Aptana to read another file type as PHP

So recently I’ve been following this simple tutorial to learn about creating a simple MVC PHP framework: http://www.domagojsalopek.com/Details/Create-a-simple-PHP-MVC-Framework/28

It’s so simple that I got a MVC template in less than an hour. But it definitely took me some time to fully understand the code. Luckily I have some experience with WordPress and CakePHP so somehow it wasn’t too long for me. *Trigger proud-mode* 😛

As usual I was using my favorite IDE Aptana but something struck me: I don’t see the PHP color code for the *.tpl files that I have created for the MVC framework. It’s quite annoying and inconvenient.

Turns out it’s actually quite simple to “fix” it.

  1. Just go to Preference (Mac users check the “Aptana Studio 3” menu at the top panel, WIndows and Linux users should check the File or Tool menu)
  2.  Under General -> Editors -> File Associations
  3. Under File Types -> Choose Add, and enter: *.tpl
  4. After added, click on that file type
  5. At the bottom, click Add and choose “PHP Source Editor”.

Done!

Now when you reopen all the *.tpl files in Aptana, it will be recognized as PHP files.

Hope this is useful to you. 🙂

C# .NET Entity Framework: How to insert record into database table with auto-increment primary ID?

Updated: 9 Sep 2016

Just got a comment asking me to take this post down because it’s outdated. Well, this was written in 2012 so it’s really VERY old. And I no longer use C# .NET Entity Framework. I’m leaving this post here for people who’re still using the old framework. All comments are closed now. I’ll no longer answer question relating to this topic. Please read it with caution.

——

I want to insert some records into the database table but it is giving an error which says “INDENTITY_INSERT is set to OFF”, something like that.

To resolve this, there’re 3 steps:

First, in Visual Studio 2010 and click on the .edmx file.
Look for the relevant and right click on the identity column (ID in most cases) and choose Properties.
In the properties window, look for StoreGeneratedPattern and change the value to “None”.
Save the changes.

In some cases (in my case), changing the value in step one is not enough. We still need to manually edit the .edmx file.
In Visual Studio 2010, right click the .edmx file -> “Open With…” -> Choose “XML (Text) Editor” to open the .edmx file
Look for the relevant table and change StoreGeneratedPattern=”Identity” to StoreGeneratedPattern=”None” for the identity column.
Save the changes.

Once the EF settings are done, we can set the primary key at the EF side.
However, SQL Server does not allow the Identity column to be set unless IDENTITY_INSERT is ON. (http://msdn.microsoft.com/en-us/library/ms188059.aspx).
So we need to manually execute SQL command to turn on the IDENTITY_INSERT first before we insert new records.

using (TestDBEntities1 context = new TestDBEntities1())
{
	using (TransactionScope scope = new TransactionScope(TransactionScopeOption.RequiresNew))
	{
		context.Connection.Open();
		int i = context.ExecuteStoreCommand("SET IDENTITY_INSERT TestDB.dbo.Person ON");
		Person p = new Person();
		p.PersonID = 4;
		p.Name = "Michael";
		context.People.AddObject(p);
		context.SaveChanges();
		i = context.ExecuteStoreCommand("SET IDENTITY_INSERT dbo.Person OFF");
		scope.Complete();
	}
}

This is a workaround to let us manually set the identity column.
Warning: when IDENTITY_INSERT is set to ON, we need to manually assign the identity key (ID in most cases). You must ensure that it is unique and not null or empty. If you want to get the next available ID, you may add this:

Person p = context.Person.OrderByDescending( c => c.PersonID).FirstOrDefault();
int newId = (null == p ? 0 : p.Id) + 1;

Resource:
http://social.msdn.microsoft.com/Forums/eu/adodotnetentityframework/thread/48cfa132-8612-4353-b916-b29f32b4070e

http://social.msdn.microsoft.com/Forums/en-US/adonetefx/thread/84867836-2a52-4c8d-aa3d-e2d5f1845e66

Raphael.js IE7 issue with innerHTML on elements other than div

I kept having this unknown error on IE7 (IE9 compatibility mode) when I use Raphaël JavaScript Library. On deeper research, this error points me to line 5178’s “span.innerHTML”.

I found this very interesting article http://domscripting.com/blog/display/99 about usage of .innerHTML in IE7:

…innerHTML can cause some problems, particularly in Internet Explorer. If you use innerHTML to add or update form elements, all sorts of screwiness can occur. Sometimes the data from the newly added elements won’t be included when the form is submitted to the server.

and it provided a useful fix:

First of all, use the DOM to create a block level element—such as a div—using createElement. Then update the innerHTML property of this newly-created element. Finally, insert the updated element into the document using a DOM method like appendChild or insertBefore…

I thought it was a perfect fix, so I searched for the .innerHTML implementation in raphael.js (line 5178):

res.textpath.string && (span.innerHTML = Str(res.textpath.string).replace(/"));

and changed it to:

/*Resolve IE7 issue with innerHTML*/	
var newdiv = document.createElement("div");
res.textpath.string && (newdiv.innerHTML = Str(res.textpath.string).replace(/"));
span.appendChild(newdiv);
/*end of resolution*/

Somehow it worked very well. I had forked and submitted a fix but still waiting for the main master to commit this change. You can find the fix at my fork:
https://github.com/kongnir/raphael/compare/master…patch-1

Free Web development tools for Mac users

This is an on-going post which will be updated when I find or learn something better. Please leave a comment if you have any suggestion or if you find any error. Help me to help others! Thanks. 🙂 (I found many spams coming in, so please leave out the website if you don’t have any to share.)

Free Web development tools for Mac, especially for PHP developers. Of course, there’re many other free tools available out there. These are the tools that I have been using personally and I find them very useful. I hope you’ll find them useful too.

1. Local web server for testing

  • MAMP to manage local website (http://www.mamp.info/). This is only essential if you’re working with server side language PHP. Updated: If you’re too lazy to set up the Mac with its included Apache server like me, this is a very good app to start with. Just install, run the app and you’re good to go.
  • This tutorial shows how easy it is to set things up in Mac with MAMP for WordPress. I followed this to install a testing development for my work.
  • If you can’t get Aptana (mentioned below) to read the files from MAMP’s default Document root, scroll to the bottom to read more.
  • Mac’s Web Sharing for developers who are only developing using HTML, CSS and Javascript, Mac comes with a web server installed by default. You may want to google “How to enable web sharing on Mac Lion”. Updated: Thanks Glenn for pointing out. If you’re a little command line savvy, you may do away with MAMP and set up your Mac with PHP and MySQL. Just follow the tutorial here. I haven’t tested it though.
  • OS Mountain Lion: Unfortunately, Apple has removed the web sharing option in the sharing preference panel (see here). You can bing it back using command line by following this tutorial. But then again, they really look too much work. I prefer to use MAMP. And because the configuration is self-contained within MAMP, if I messed up the settings, I can simply delete the app and reinstall to reset to default.
  • AMPPS (http://www.ampps.com/): I was preparing for a web design class and was looking for a tool for Windows users similar to MAMP. And then I found this AMPPS. It supports both Windows and Mac, and is said to be portable. I’ve yet to give it a try, I probably will in the near future although MAMP is enough for me at the moment. Read this article about AMPPS vs MAMP: http://bit51.com/bye-bye-mamp-pro-hello-ampps/

2. Text Editor

Both tools provide syntax-highlighting, but Aptana provides auto-complete and syntax error alert features which are slightly more useful for me.

  • Aptana as web development IDE http://aptana.com/
  • I had been using TextWrangler before I finally found this tool. The syntax error alert feature really saves me a lot of trouble trying to debug. And the auto-complete saves me some typing too.
  • One thing I don’t really like is the default white text on black background theme. To change the white text on black background to black text on white background, click the color wheel icon on the top of the Aptana window and select “Aptana Studio 2.x”, “Eclipse” or “Dreamweaver” theme.
  • I didn’t spend time trying to set up Aptana for debugging web environment because I could use MAMP to run the web and see my development while I’m doing my work.
  • Aptana works even better if you have Git installed. It can track branches and changes by indicating them on the left panel. This is a simple yet extremely useful feature. Apple’s Xcode comes with Git. My iMac (with Xcode and SourceTree installed) worked pretty well for Aptana, but my MacBook (no Xcode, only SourceTree installed) didn’t work. So it’s probably advisable to install Xcode along with Git so you don’t have to figure out how to configure Git to work with Aptana.
  • TextWrangler as text editor http://www.barebones.com/products/textwrangler/
  • TextWrangler doesn’t support auto-complete and doesn’t prompt you if there’s any syntax error. You need to know exactly what you’re doing. I have had hard time debugging when I was using this tool. It is still a very good free tool though.

3. Source repository and management

Source control and management is crucial in a development work. Because it is the core of the business. It is important for us to find a good storage and to keep track of all changes done to the source.

  • BitBucket https://bitbucket.org/
  • You can create unlimited private repositories.
  • Each repository is limited to 5 users for free account.
  • I have been using this for several months. It is very easy to use and configure. The SourceTree client works out of the box with BitBucket. And recently they had refreshed and improved their web UI. Very impressed. Not to mention they provide unlimited private repositories.
  • GitHub https://github.com/
  • You can create unlimited public repositories.
  • Private repositories come with a price.

4. Source repository client

You can always use the web page to upload changes. This is just a convenient way to commit the changes to the repository from the Mac.

  • SourceTree app http://www.sourcetreeapp.com/
  • Any Git or Mercurial repositories.
  • I’ve been using this tool and find it very easy to use.
  • The client allows us to add any git repository, not limited to BitBucket. I had tried adding GitHub and “it just works”.
  • For Git beginners the concept of Fetch, Pull, Push, Commit and Checkout may seem overwhelming… but they are the fundamentals of Git that I think it’s good to know. It becomes very useful if you ever want to use Git with command line.
  • GitHub for Machttp://mac.github.com/
  • Limited to GitHub account.
  • I have not personally tried this but the UI looks very intuitive.
  • The UI is so simple that some concepts of Git have been hidden from users. It may be difficult for you to switch to other clients or command line if you don’t have the concept of Fetch, Pull, Push, etc. Nevertheless, it is a simple client for newbie to start using Git within minutes.

While the 2 clients are extremely useful for synchronizing your work on the cloud (their servers), you can always choose to leave the work in your local machine. Simply use commit and not use the Pull and Push commands. But repositories are best for collaboration, so usually developers will sync them to a server for sharing (and backup too).

Using Git from the Terminal

Sometimes you may need Git from the Terminal. In my case, I wanted to use Composer.org to install some dependencies in my Laravel project. So I need Git in my Terminal. It’s not that difficult to install, just go to this link, download the latest package and install it.

https://code.google.com/p/git-osx-installer/downloads/list

5. Aptana & MAMP troubleshoot (Document Root issue)

I didn’t use the default root in MAMP because it didn’t work for me. Aptana doesn’t seem to be able to see the files in the htdocs folder. It’s better to put your work outside the MAMP app, in case it got corrupted and you have to delete it.

Open MAMP, go to “Preferences”, under “Apache” tab, change the Document Root to somewhere else. I like to put it under

/Users/your_username/Sites

This folder is automatically created prior to Mountain Lion. If you purchased the MacBook with Mountain Lion then most probably this folder doesn’t exist. You can always create it or choose another folder. 🙂

When you launch http://localhost:8888 the next time, it will read directly from this folder.

Additional info:
If you have subdirectories inside this folder, simply use http://localhost:8888/subfolder

6. Managing MySQL database

The best tool to manage MySQL database is the tool made by Oracle themselves. They have clients for many platforms, including Windows, Mac and some Linux.

MySQL Workbench: a very nice tool to manage MySQL databases. http://dev.mysql.com/downloads/tools/workbench/

7. FTP Client

There’re many free FTP clients available in the market but these are the few that I have been using and I find them easy to use.

  1. Cyberduck (http://cyberduck.ch/)
  2. Filezilla (https://filezilla-project.org/)

Warning: Be aware that by default Filezilla stores and transmit your password in plain text. I’m not sure if this has been resolved but it has certainly compromised some of my sites before. There’s a way to “fix” it but it sounds too troublesome.

Cyberduck on the other hand stores your password using Mac OS’ keychain. One slight inconvenience in Cyberduck is that it doesn’t have the feature to only upload files that have changed. Filezilla is better in that sense.

More FTP clients here: http://mac.appstorm.net/roundups/internet-roundup/top-7-free-ftp-clients-for-mac/

8. LESS CSS compiler

LESS is a dynamic stylesheet language which extends CSS with dynamic behavior such as variables, mixins, operations and functions. If you want to learn more about LESS, visit www.lesscss.org.

LESS makes CSS logically easy to write and reusable. The only troublesome part, in my opinion, is that you need to compile LESS script into CSS format. The team that came up with LESS has provided a command-line tool to do that. Command-line?! Sounds complicated? Fortunately, there’re many GUI tools that can help us with that. The 2 that I have been using are simple and free of charge.

  1. SimpLESS (http://wearekiss.com/simpless) – SimpLESS was a very easy-to-use and straight-forward GUI compiler. However, I was having a lot of trouble trying to compile Twitter Bootstrap v3. So I switched to…
  2. Koala (http://koala-app.com/) – Koala is not just a LESS compiler, it also supports Sass, Compass and CoffeeScript. It can run on Windows, Mac and Linux. The GUI is also very straight-forward. The only thing that’s currently quite annoying is that it doesn’t allow user to add a single file to compile. You have a add the whole folder, then delete all the files that you don’t want to compile. Other than that, the functionality is comparable to SimpLESS.

For more GUI compilers, you may want to read up here: https://github.com/less/less.js/wiki/GUI-compilers-that-use-LESS.js

9. Configure AMPPS

  1. Download the latest version from http://www.ampps.com/downloads (as of writing it was version 2.0)
  2. Launch the app and Start both Apache and MySQL.
  3. Launch a browser and go to address http://localhost/
  4. You will be asked if you would like to make AMPPS secure. Go ahead and enter a password.
  5. The root folder is by default at /Applications/AMPPS/www/
  6. If you create a subfolder under the root folder, you can access your subfolder site using the URL http://localhost/subfolder

I’m trying out AMPPS now and considering if I should migrate all my projects from MAMP to AMPPS. The main advantage for me is that AMPPS supports both Windows and Mac, while MAMP is only for Mac. Other than that, I don’t feel much difference between the 2, as of now.