Deep Shah's Blog

LetsBuy.com A Perfect Example of How Not to Run an Online Shopping Website

2012-01-22T13:56:00.001+05:30

Today I am going to narrate a story, A real story.

In this story, lead role is played by Deep Shah (name not changed) and His Friends. Every story has good guys and the bad guys right, in this story the bad guys are LetsBuy.com (name not changed) as a company (nothing personal against anyone). There are some side hero's in the story as well, I will introduce them as and when they appear.I am all exited to tell the story to you guys, so here it goes!

The Plot:

There was a guy called Deep Shah - A Software engineer living in Pune. He wanted to buy a new mobile, specs he wanted were:

Android OS
Dual core CPU
Minimum 4" Screen
1GB RAM
Budget max Rs. 20K (approx $400)

For a very long time none of the phones met all his criteria. Samsung Galaxy R (around Rs. 25K), Samsung Galaxy S-II (around Rs. 28K) and Samsung Galaxy Note (around Rs. 33K) would match the technical specs but were totally out of his budget.

Then one day, one of his friend pointed out that prices of Samsung Galaxy R have fallen to about Rs. 21K. He decided to go for it!

The Buy - The Good:

Just like any other middle class person, he decided to get the best deal (read that as cheapest deal) for his Samsung Galaxy R. He checked out the prices at a few local stores, but the best deal he could find was around Rs. 22K.

The online price for Samsung Galaxy R was around Rs. 21K. Deep Shah had ordered quite a few things (some of them expensive as well) online in foreign countries and they all were delivered just fine without any problems. But in India, his online shopping experience was limited to books and some not so expensive stuff. He really wanted to give Indian Online Shopping websites a go (actually he really wanted to save the 1000 bucks!).

Hence, he checked out few online shopping websites, that were selling The Samsung Galaxy R. I am not going to mention other online shopping websites names (since this post is not about them). He decided to order the phone from a website called LetsBuy.com who were selling the Samsung Galaxy R at Rs. 20990.

Why did he choose LetsBuy.com?

They had a decent web site
They had a call center where one could call for enquirers between 8 am to 9 pm, 7 days a week
They were giving away free branded accessories worth Rs. 1200 absolutely free
He had found a discount coupon for LetsBuy.com which would give him an additional Rs. 1000 off!

Obviously, the last two reasons were the most important reasons to choose LetsBuy.com, the phone was costing Rs. 19990 after all the discount!

OK so the choice of the website was made, now it was time to actually place an order. The moment of choosing the payment mode had arrived. There were mainly two options Cash on Delivery and Pay by Credit Card.

The choice was simple, it was Cash On Delivery, however, there was a catch, the discount coupon wont work for the Cash on Delivery payment mode. Damn, if he choose Cash On Delivery the phone would cost 1000 bucks more. You guys must have guessed his choice of Payment Mode by now, yes he choose Pay be Credit Card. Unfortunately to save 1000 bucks he risked loosing Rs. 19990!

The order was placed on LetsBuy.com on 30th December 2011 a Friday - Lets buy order number: LSB-757065 (actual order number). LetsBuy.com claimed that they will delivery the phone in 2-4 business days. Deep Shah was expecting the delivery latest by Thursday. Now it was time to wait anxiously for the delivery.

The Confusion - The Bad:

He checked the order status on LetsBuy.com and found that that, the order was going to be dispatched in two parts, one was the actual phone and other was the accessories. Over the weekend he saw that his order was partly dispatched from LetsBuy.com. The phone was on its way!

By Monday (02/01/2012), LetsBuy.com had updated the shipment number and the courier company's name on the order. The phone was being dispatched by a company called Aramex (name not changed), the shipment number was 6273338610 (actual shipment number).

Checking the current status on the shipment on Aramex.com, showed that shipment was on its way to Pune. However, one thing caught his attention, the shipment was to be delivered in Pune - 411013, but, when Deep Shah had placed the order the address mentioned as Pune - 411016.

Hmm, how could the pin code be different. He and His Friends smell something fishy.

He decided to call up the courier company to find out whats wrong with the order. The courier company Aramex was pretty prompt and accurate in their response. He could talk to an executive within 30 seconds. Deep Shah asked the executive to track the shipment 6273338610, to his surprise (read that as "shock"), executive informed that, the Order was in the name of Mr. Amer (name changed) and order was to be delivered in Magarpatta, Pune - 411013.

Clearly, something was messed up. Initial guess about the problem with Deep Shah's order was:

Probably incorrect shipment number was updated against the order on LetsBuy.com

He asked the courier company executive to find any shipment against his phone number (the number that he had given while booking the order). But unfortunately no shipment was found against his phone number.

Next, he decided to call LetsBuy.com call center to get some clarification. This was the first time that Deep Shah would talk to LetsBuy.com call center executive.

He had to wait a full 7 minutes before he could actually speak to someone at LetsBuy.com.

Deep Shah explained the situation with his order and asked for an explanation. The executive calmly said that may be an incorrect shipment number was updated for his order (which was his guess anyways). Executive requested Deep Shah to wait for 24-48 hrs and he will get this error resolved. Executive registered a complaint (no: cc120139286 - actual complaint number). This was on 02/01/2012

The Followup - The Ugly:

Next day (Tuesday 03/01/2012 - i.e. after 24 hrs), he logged on to LetsBuy.com to see if the shipment number was updated. But, there was no change in the shipment number. He decided to wait for one more day.

Next day (Wednesday - 04/01/2012), ditto, no change in the order information.

He decided to call LetsBuy.com again to find out whats going on for his complaint. Again, it was a long wait for 9 minutes before which he could speak to someone at LetsBuy.com.

Deep Shah had to repeat the entire story again to this new person. Executive informed that they were working on Deep Shah complaint and he has no update on the status of the complaint. Executive requested Deep Shah to wait for another 24 hrs. On this day, Deep Shah received the free accessories (not the phone).

Next day (Thursday - 05/01/2012), there was an unpleasant change on the order status, the Order status showing as Delivered on LetsBuy.com. That freaked out Deep Shah a bit.

He tracked the shipment no: 6273338610 on Aramex.com website. Turned out that it was actually delivered to Mr. Amer. But no, Deep Shah had not received the phone yet.

Shipment Delivery to Mr. Amer

He called LetsBuy.com, after a long wait for 8 minutes, he could speak to someone at LetsBuy.com.

He had to explain the entire story again (which was very frustrating). You must have guessed the result of this conversation by now, executive asked Deep Shah to wait for another 24 hrs as they had no update on his complaint/order. Executive even assured him that by the next day, in most likelihood Deep Shah should receive the product.

Next day (Friday - 06/01/2012) nothing happened. Following up with LetsBuy.com was frustrating and pointless. Every time, Deep Shah had to wait for long durations and then explain someone the same problem all over again and get the same response.

LetsBuy.com executives would not give their direct contact numbers or their email address. This means

Long wait before one could speak to a LetsBuy.com executive
Explaining the situation all over again (to different executives) and eventually get the same response - Please wait for 24 hrs.

Wait for 24hrs and do what? Expect some miracle?

On Sunday - 08/01/2012, i.e. Full 10 days after placing the order, Deep Shah still didn't have any trace of his phone.

He had received the free accessories, as one of my friend keeps saying,

Its like giving an hungry dog a rubber bone!

Deep Shah called up LetsBuy.com again and this time demanded to speak to some manager. He got to speak with Mr. Laxman Patil (name changed) a supervisor at LetsBuy.com. He seemed like a nice guy, heard Deep Shah out, understood the situation and promised to look into the matter himself. By this time all Deep Shah wanted was a regular feedback on his order.

Being a software engineer, he understands that all systems are pron to programming/human bugs/errors. But customers should be given regular feedback on what action was taken to rectify the mistake. Even after contacting LetsBuy.com, like 10 times, LetsBuy.com didn't have enough courtesy to call Deep Shah and update him with the complaint status.

BTW the order status on LetsBuy.com was Delivered, they didn't bother to change it, even after they agreed that it not yet delivered.

Order Status is Delivered Even though no phone was delivered

Next day (Monday - 09/01/2012), no news from LetsBuy.com, no news from Mr. Laxman Patil. During this period Deep Shah wrote numerous mails on customercare@letsbuy.com but LetsBuy.com didn't respond to even one.

During these days, Deep Shah and His Friends tried our different (random) tracking numbers, which were around the tracking number that was given to Deep Shah. They found quite a few tracking numbers that had funny shipment history. For e.g. Tracking Number 6273338621 was delivered twice once in Panji, India and once in Bombay India.

Shipment Delivered Twice

What did this signify? Does this mean Aramex was reusing shipment numbers?

We called up Aramex to ask whether they were reusing shipment number? Of course they said shipment numbers are unique. They cannot be duplicate.

Deep Shah and his friends could find multiple instances where there were two deliveries made on one shipment number. If you notice the tracking history of Deep Shah's shipment number you can see, even that one has two shipments moving.

One shipment arrived in Pune on 31/12/2011 at about 1845 hrs and went on hold.
Other shipment started from Banglore on 02/01/2012 and was delivered on 05/01/2012 to Mr. Amer

Notice two shipments moving on one shipment number

In this shipment, notice that, there are two products involved but only one delivery was made. While in all other shipments we noticed that two products were moving on the same shipment number but two deliveries were also made.

By this time, Deep Shah and His Friends were sure that his shipment was lying in Aramex Pune office but was never delivered.

What was so special about this shipment? Why others got delivered but this one was not delivered?

The interesting thing to notice about all other shipments with duplicate shipment numbers was that, they were to be delivered in different cities. for e.g. in the image shown above, one product was to be delivered in Panji Goa and other one in Bombay. But, in Deep Shah's case both the shipments were to be delivered in Pune.

This was the turning point. After some motivation from Deep Shah's Friends, they decided to visit Aramex Pune office and figure out if their theory was correct.

On Wednesday - 11/01/2012, There was still no updates from LetsBuy.com. Even after giving numerous calls to LetsBuy.com no one would every get back to give any update.

Deep Shah would call LetsBuy.com and insists to speak to Mr. Laxman Patil, he would have to wait for full 15-20 minutes, eventually Mr. Laxman Patil would inform that, he has escalated the case to higher authorities and action will be taken. It was too late and too little and very very frustrating.

Deep Shah and His Friends visited the Aramex office in the evening and explained them the situation. Initially they had tough time explaining them that two shipments were moving on one shipment number. If they looked up the shipment number in their system it would always bring up details of Mr. Amer's shipment. As there cannot be two shipments with same shipment number. Entire logistics business is based on this assumption.

Once a logistics expert (who had around 15 years of experience in Logistics) told me that,

We cannot have duplicate shipment number being generated in production, this would have catastrophic effects.

Now, I truly understand what he meant by that!

Anyways, In Aramex Pune office one guy (A very good guy) did agree that, two shipments were moving on one shipment number. He and his team manually checked all the shipments that were present in his warehouse.

Guess what, they did find a box that was shipped by LetsBuy.com and was with the same shipment number i.e. 6273338610. One very smart person at Aremex Pune office had kept this shipment aside since he saw two similar shipments with same shipment number had landed in his warehouse.

It seemed like the mystery was getting resolved, it was clear that the box had the phone for Deep Shah, but, Aramex would not deliver it to Deep Shah since, in their system, this shipment belonged to Mr. Amer. Funny part was, the shipment was already delivered to Mr. Amer.

Aramex guys requested Deep Shah to wait for one last day so that they could get a confirmation from LetsBuy.com about the shipment and delivery it to him.

Next day, Thursday - 12/01/2012 - Deep Shah called up LetsBuy.com to inform them about what he had found out, he had also written an email to inform about this new development.

Again as usual they asked Deep Shah to wait for 24-48 hrs. LetsBuy.com call center is like an IVR system which is programmed to listen to what customer has to say and always respond by saying we will get back to you in 24-48 hrs and never bother call the customer back. Worst part is they will ask whether they can help with anything else?, what a joke!

In the evening at about 1730 hrs, the smart guy from Aramex Pune office called up Deep Shah and asked him to come and collect his phone. Finally a sign of relief, Deep Shah did get his phone after 14 days of placing his order! Aramex Pune office folks actually managed to resolve the issue within 24 hrs as promised by them. They really were the people who saved the day in the end.

Out of curiosity, Deep Shah the smart guy at Aramex, Why duplicate shipment numbers are getting generated?

Turned out, a range of shipment numbers have been allotted to LetsBuy.com and they are not supposed to repeat them. But, because of a bug in LetsBuy.com's system, same shipment numbers were getting used for two different shipments.

Well, I can understand that mistakes do occur when you are running a business, but the quality of service is judged by how well you handle those erroneous cases. Letsbuy.com failed miserably in handling this case.

On a scale of 1 to 5, I would rate my shopping experience with LetsBuy.com at -5. I hope not every customer is treated like this at LetsBuy.com else its just a matter of time before which they perish!

What Should have been done:

Customers should not suffer because of some goof up in the system
Customers should always be keep informed about the status of their complaint
Customers should not be waiting for 7-15 min before which they can speak to a customer care executive
Customers should not be speaking to different executives every time they call.

I am sure LetsBuy.com knows much better how to run an online shopping website. However, things that I have pointed out are just common sense items which should not have been ignored.

End Result:

Deep Shah and His Friends promised themselves never to buy any thing from LetsBuy.com and hopefully people who read this post would do the same.

Moral of the story:

Never judge a website by the amount of discount they are offering on a product
Choose the one that has a better reputation and better and responsive customer care
Always always always opt for Cash On Delivery payment method

Standard Disclaimer:

Every incident mentioned in this story is real. All order numbers and shipment number are real order/shipment numbers.

NHibernate throws InvalidCastException in DEBUG mode for entities with composite keys

2011-12-31T01:36:00.001+05:30

Recently we faced a very weird issue with NHibernate. The issue was, on one developer's machine, when we ran an Integration test to fetch an entity (using its primary key), NHibernate threw an InvalidCastException exception.
The issue was *not* reproducible on any other developer's machine. This got me thinking, there must be some issue with the mappings which is causing NHibernate to throw the exception. The developer must have made some local changes to the mapping files, that is causing the exception.

Investigation:

I tried various things to figure out what was causing the problem. Here is the list of things that I tried.

Check mapping files for modifications:

I checked and rechecked the mapping files for any local modifications, but, No, there was no local modification. I even checked-out all the mapping files just to be sure that there was no modification. But still no luck, the issue was reproducible on her machine and not on any other machine.

Tried to reproduce the issue on my machine by connecting to other developers Database:

After the first failed attempt, next thought that crossed my mind was, may be, there was something wrong with entity data which is causing the exception. I decided to run the same test on my machine by connecting to other developers database.

Surprisingly, I was still not able to reproduce the issue. This was totally weird. I was 99% positive that, issue would get reproduced once I connect to the order developers database.

I was running out of options. How do proceed further? What should be the next step?

Stick to basics - Debug with Elimination:

I was back to the drawing board. Decided to use Elimination to narrow down the issue. The question that was bothering me was, Whats so special about the entity we are trying to load?

The entity we were trying to load was called EmpDept, it represents the Employee Department information (actual names of the entities have been changed). So what was so special about this table?

This table was a legacy table, it did not have a primitive primary key, it had a composite key made of UserID and DeptID.

So what? A lot of legacy databases have composite keys, NHibernate works well when it comes to mapping the composite keys, in fact, NHibernate has no problems loading the same entity on my machine!

I decided to, removed all other mapped properties and just kept the mappings for the composite key and try to fetch the entity. This time, it did fetch the entity without any problems on the other developers machine!

We were starting to get somewhere now!

I started adding back the mappings to find out which one was causing the issue. Finally, we did find the minimal possible mapping that would reproduce the issue.

The Cause:

We deduced that, the issue occurs when the following conditions are satisfied

Entity is mapped with composite primary key
Entity has a one-to-many relationship with other entity
The one-to-many relationship is not based on the composite primary key, instead its with property-ref
And the data type of property-ref is different from the id property

Wow! I know, those are too many conditions to be true at one time!

Let's look at the minimal mapping file that reproduced the issue
What do we have here? We have a very simple mapping between two entities.

Entity EmpDept is mapped to a table called EmpDepts
EmpDept uses a composite primary key to class EmpDeptIdentifier
The composite primary key consists of UserID and DeptID
EmpDept entity as another property called Email
There is a one-to-many relationship between EmpDept and Task entity i.e. One EmpDept can have Many Tasks
Note that the relationship is not based on the composite primary key
The relationship is based on a non identifier property called Email
The data type of the Email property is string while the data type of the composite primary key is EmpDeptIdentifier

By now, it must be clear that all of the required conditions for NHibernate to throw an exception have been satisfied, because of this NHibernate throws the InvalidCastException.

So far we had figured out, what conditions were required for NHibernate to throw the exception. However the real mystery was, Why was the exception occurring only on one developers machine? Why not on other developers machines?

The Enlightenment:

Our curiosity levels were at all time high! We decided to debug the NHibernate code to figure out why was exact same code working perfectly fine on one machine and failing on other.

After some intense NHibernate code debugging we found the NHibernate code that was throwing the exception.
Notice the line that throws the InvalidCastException in the above code? This is where the exception occurs.

The above method is called to build a string that will be used for logging. We traced back the call to this method, it was getting called from the MessageHelper.InfoString method, which in-turn was getting called from the LoadContexts.LocateLoadingCollection method.

We now knew the place from where the exception was thrown, but we still didn't know why it was thrown only on one machine.

Looking through the NHibernate code we realized that, the method MessageHelper.InfoString is only called if NHibernate is running with DEBUG log level.

That ladies and gentlemen, was the "Aha Moment!" for us. This was the missing piece of the puzzle. The developer on whose machine the test was failing was running NHibernate at log level DEBUG, while I was running the test at log level INFO. Because of this the method MessageHelper.InfoString was never invoked on my machine and thus no exception!

We Googled a little and found that there is an official bug reported for this scenario.

Eventually it turned out that, what we initially though as a big issue was really a non issue, since NHibernate will not be configured in DEBUG mode 99% of the time (because it logs hell lot of information). The issue we were facing is very unlikely in real systems. Although it was not such a big issue after all, this experience taught us quite a few things about NHibernate!

Integrating Spring.NET with ASMX WebServices

2011-11-13T12:55:00.001+05:30

In one of the previous post, we saw how easily we can Enable Dependency Injection for Web Forms Web Pages and User Controls. In this post we will cover the last topic in the Dependency Injection series i.e. Enabling Dependency injection in ASMX WebServices using Spring.NET.

Its not at all difficult to Enable Dependency Injection in ASMX WebServices. Lets look at the steps we will need to it

Configure the Spring Context in web.config
Add the Spring WebSupportModule Module to the list of httpModules
Configure the WebServiceHandlerFactory to intercept any path that matches the pattern *.asmx
Configure the WebService class in spring config file with its dependencies

Let's look at each step in a little more detail.

Step - 1 - Configure the spring context in web.config

First we need to add spring config section handler in the configSections. The updated web.config looks like
Next we need to add the spring config section in web.config. This section will inform Spring.NET as to where should it look for its configuration file. Add the following section to web.config
Nothing special about the above piece of configuration. We are informing Spring.NET that its configuration file is located in ~/Config/ directory and its called webservice.xml.

Step - 2 - We need to add the Spring WebSupportModule module to the list of httpModules

In this step, we will need to add the WebSupportModule to the list of httpModules in the web.config. This is how the updated web.config looks like
Step - 3 - Configure the WebServiceHandlerFactory to intercept any path that matches the pattern *.asmx

Again, all we need to do here is add an Handler in the httpHandlers section in web.config. Update web.config looks like
Step - 4 - Configure the WebService class in spring config file with its dependencies

We are almost there. All that is left now is to configure the WebService class in the Spring.NET configuration file webservice.xml.

Let's say that we want to develop a HelloWorld webservice, that has one method called GreetUser. This method takes in the user name as argument. We will inject the greeting to be used using Spring.NET.

The code of the HelloWorld webservice looks like

Nothing to explain here, all we are doing is prepending the Greeting Property to the userName passed and returning the value as the result.

Notice that this HelloWorldService class is a POCO and there is nothing web service specific to it.

The spring configuration that will inject the Greeting into the HelloWorldService looks like
Notice that, we are configuring the implementation of WebService with the ID HelloWorldServiceImpl. However, the actual WebService is exposed via Spring.NET. The Spring.Web.Services.WebServiceExporter acts as a proxy over the actual WebService implementation.

All set!

Lets run the server and browser the URL which would look like http://localhost:<your-port-number>/HelloWorldService.asmx?op=GreetUser. This should open up a page which looks like this

Page which will let us invoke the HelloWroldService.GreeUser WebService

This is the page from which we could invoke the HelloWorldService.GreetUser WebService.

Let's type in Git Shah in the user name field and click Invoke. Here is what we get
As you can see, the greeting message Good Morning was injected in the HelloWorldServiceImpl class via Spring.NET.

That's all folks, we have successfully Integrated Spring.NET with ASMX WebServices!

Making operator precidence more explicit with NHibernate Linq Provider

2011-11-06T13:03:00.002+05:30

Recently, while writing one of the NHibernate queries, using the NHibernate LINQ provider we faced a very interesting problem.

The issue was the NHibernate LINQ provider was not adding appropriate parenthesis to make the operator precedence more explicit.

The query was simple enough, we wanted to get a list of records based on certain condition. Let's say, for the sake of this post, consider that there is an employees table with ID, Name and Salary columns. The DDL to create the employees tables on SQLServer looks like
NHibernate class against which this table is mapped is called Employee. The Employee class and its mapping file looks like
As you can see there is nothing fancy about the table or the mapping.

Let's talk a bit about the query we want to fire against the employees table. We want to get list of employees whose name equals "Git" OR whose salary is between 1000 and 10000.

The SQL that would do the job for us would look like:

Again nothing remotely fancy about the SQL.

Let's try and write the same query using, NHibernate LINQ provider and lets look at the generated SQL. The NHibernate LINQ query to get similar results would look like:
The where condition is the exact replica of the SQL where condition that we had seen earlier. Only difference this time around is, we are using the LINQ syntax.

So far so good, the problem starts when we look at the SQL that is generated. The simplified version of the SQL generated from the above LINQ query looks like

I have removed all the named parameters that NHibernate adds, but essentially the query looks like the one shown above.

Whats wrong with the query?

Notice that there are no parenthesis to enforce the operator precedence! Because of this, it might get confusing to identify the order in which the operators will be process.

Consider a case when you have a complex query with many conditions and if proper parenthisis are not applied then it could get confusing.

I tried various permution and combination to force the NHibernate LINQ provider to spitout the parenthesis correctly, however I was not successful in doing so.

Finally, I decided to switch to the QueryOver API provided by NHibernate. QueryOver API is a Fluent wrapper over the NHibernate Criteria API. Using the QueryOver API the above code would look like

As you can see the nothing much has changed. The only changes are, .Query() call has been changed to .QueryOver() and instead of .ToList() we have to call .List().

Let's look at the generated SQL Query:
As you can see, it correctly adds the parenthesis to make the operator precedence more explicit!

Enabling Dependency Injection In Web Forms Pages and User Controls using Spring.NET

2011-10-28T16:08:00.000+05:30

In one of the previous post, I mentioned, How MVC3 and Web Forms can exist side by side in one project. We also saw, how we could Enable Dependency Injection using Spring.NET in an MVC3 Application.

But, What about the old Web Form pages and user controls? Can we Enable Dependency Injection in them, so that we can create loosely coupled code?

Surely there is a way of doing this using Spring.NET. In this post we will carry on from where we left and we will try to enable Dependency Injection even for Web Pages and User Controls. Before moving any further, if you have not already seen my previous post of integrating Spring.NET with MVC3 Application, now is a good time to see it.

Without wasting any more time, let's see how we can enable Dependency Injection in Web Forms and User Controls using Spring.NET

How Do they Do it?

For sake of this post, let's say we have a web page called Default.aspx under the Pages directory. The structure of the project looks somewhat like this

The project structure with Default.aspx

Our task is to inject the AccountService into the Default.aspx page. As mentioned in the previous post AccountService is a class that was already configured in Spring.NET, it has one method UserCount and we are going to show the user count on the Default.aspx page.

The steps to enabled dependency injection in web-forms web pages using Spring.NET are as follows:

Add HttpHandler, called Spring.Web.Support.PageHandlerFactory for path *.aspx
Make code changes to Default.aspx to fetch the UserCount from AccountService
Configure the dependencies of Default.aspx in web.xml (web.xml is a spring config file which we created in the previous post)

Let look at each step in more detail.

Step - 1 - Add HttpHandler Spring.Web.Support.PageHandlerFactory for path *.aspx

This step is pretty straight forward all we need to do is, Add an HttpHandler, called Spring.Web.Support.PageHandlerFactory. We need to configure it against the path *.aspx. This is required so that, Spring.NET gets an opportunity to perform Dependency Injection when a request comes for any web-forms web page.

We need to add the handler to the <httpHandlers> tag in the web.config. Updated web.config looks like

Step - 2 - Make code changes to Default.aspx to fetch the UserCount from AccountService

Lets look at the Default.aspx page
As you can see the html code of Default.aspx is very simple. It just has a label called userCount which will hold the value of number of users in the system.

We will get the value of userCount from AccoutService. Let's look at the code of Default.aspx.cs fileThe code of Default.aspx.cs has no surprises, it has a property AccoutService. We will plan to injet an implementation of AccoutService using Spring.NET. In the Page_Load method, we simply set the userCount label with the value returned by the UserCount method of AccoutService.

The only question left to answer is, how do we inject the AccoutService into the Default.aspx.cs.

Step - 3 - Configure the dependencies of Page Default.aspx in web.xml

For configuring the AccoutService dependency in Default.aspx we have to add the following object declaration in web.xml
Notice that we simply have to add the declaration of Default.aspx with proper path in web.xml. Since, Default.aspx is located in the Pages directory, we have to refer it as ~/Pages/Default.aspx. We have to add all the dependencies for the page, just like how we do it for other Spring.NET objects.

All set, lets test our setup and browser the page Default.aspx to see the result.

Enabled Dependency Injection in web.forms using Spring.NET

As you can see, we are able to see the user count value on the page. We have successfully Enabled Dependency Injection in Web-Forms Web Pages.

Lets quickly see how we can enable dependency injection in User Controls. Steps to do that are as follows:

Create a user control called UserCount.ascx that invokes AccountService to get the user count
Update the Default.aspx to use the UserCount.ascx user control
Configure the UserCount.ascx in web.xml

Step - 1 - Create a user control called UserCount.ascx that invokes AccountService to get the user count

We are going to create a very simple user control. Let's call it UserCont.ascx. This user control will simply fetch the user count using the AccountService and show it on the UI.
The code of UserCount.ascx is very simillar to what we have already seen in Default.aspx. We are setting the value of userCount label by invoking the UserCount() method of AccountService instance.

Notice that, we are not instantiating AccountService anywhere in the user control. We will inject it using Spring.NET. Before we look at, how exactly we do that lets update the Default.aspx to make use of the UserCount.ascx user control.

Step - 2 - Update the Default.aspx to use the UserCount.ascx user control

Let look at the updated Default.aspx
Nothing very alarming here. We simply are including the user control UserCount.ascx and showing it on the html.

Step - 3 - Configure the UserCount.ascx in web.xml

The final step is to configure the UserCount.ascx in web.xml. Here is how we do it.
Notice that we registered the UserCount.ascx in Spring.NET configuration file. We simply have to specify the complete path of the UserCount.ascx. Notice that we didn't have to change the configuration for Default.aspx at all. Spring.NET is smart enought to inject the correct instance of UserCount.ascx in Default.aspx.

Another important thing to notice is that we have to add abstract="true" to the declaration of UserCount.ascx.

All set, let's see the setup in action

Enabled Dependency Injection in user controls using Spring.NET

As we can see, we see two lines on the page showing the UserCount, first one is getting the user count directly from Default.aspx web page, while the second line is coming from the UserCount.ascx user control.

Mission Accomplished!

2300 Miles US Road Trip - Part - 2

2011-10-22T12:32:00.004+05:30

As mentioned in the previous post, we recently took a vacation in the US. The previous post is about, my experiences on the trip in the first week. In this post, I plan to share my experiences on the trip for the second week.

On a nice Sunday morning we decided to start our road trip.

Destination Niagara:

We started our road trip from Chicago towards Niagara. The distance of about 570 miles. We were 4 adults and one kid traveling in an Honda Odyssey. The drive took around 10 hours. On the way we saw some very beautiful landscapes. Here are a few images

Giant White cloud

Moon in all its glory

Metal Bridge on the way

Finally, at about 2030 hrs we reached our hotel at Niagara Falls. I was really exited to finally see the Niagara Falls, I had read about them seen them on Discovery Channel, but as they say

Seeing is believing!

We decided to freshen up and grab a quick dinner and go straight to the falls to enjoy the night view of the falls. I could hear the falls when we were at a distance of 0.5 miles (approx), simply thinking about the force of the falls was making me even more curious and exited.

When we reached the falls, I could not believe my eyes! The falls were Beautiful, Gigantic, Fast and I don't have enough adjectives to describe them. During the night, the falls are illuminated with lights of different color. The lights, along with the mist created by the falls, make falls an amazing site to see. Here are a few images

Fast moving falls along with very powerful floodlights

The lights keep changing their color which makes the scene even more beautiful.

On the other side where you see a tower like structure is Canada. The two friendly countries are seperated by one bridge of around 0.2 miles.

The Bridge That connects US and Canada

The side from where the lights are flashed is Canada

The big tower on the Canadian side is a revolving hotel

After spending quite some time at the falls, we decided to head back to the hotel so that we can catchup on some sleep, the next day was going to be a long day

The next day, we start on a mission to explore Niagara Falls. We decided to see a documentary on the history and mystery of Niagara Falls. It was quite an interesting movie, did you know that, One Ms. Annie Edson Taylor, locked herself and her pet cat in a barrel and fell over the Niagara Falls and survived the fall! The movie was well scripted and was a fun watch!

Next we took a ride on the Maid of the Mist. Its basically a boat ride that takes you very near the falls so that you can enjoy the view of the falls from its basin.

The Maid of the Mist boat that takes you near the fall

The American Falls And Bridal Veil Falls

The Horseshoe Falls

The boat ride was fun not only because we got to enjoy the beauty of the falls from very near but also because of hundreads of seagulls were flying around the falls. The entire landscape was really beautiful!

Seagulls flying around the falls

Next adventure was to go as near the fall as possible. This time, a walk in the Cave of the Winds.

The Americans have build a wooden platform, just about 3 meters away from the falls, once you are there, you are bound to get completely drenched. We experienced the force of falls pounding on top of our heads even when we were 3 meters away!

Cave of the Winds - The wooden platform 3 meters from the falls

Nearest point - You can't go any closer to the falls. The force of falls was enormous.

As you can see from the images there is no way one can get out dry even with a raincoat on. It was total fun!

We decided to spend the evening at the falls and enjoy its beauty, did a nature trail around the falls, overall had a fun filled day. Here are some evening snaps of the Niagara Falls.

You can even see the dark spots on the moon

Destination Virginia Beach:

Next day, we decided to continue our trip. Today's destination was Virginia Beach. The distance of Virginia Beach from Niagara Falls was about 640 miles. The journey took around 12 hours.

Our Hotel room had an awesome view. The hotel was right in front of the beach. There was nothing in between the hotel room and the beach! We could see the beach from our hotel room terrace. It was an amazing site!

Snap click from the hotel terrace

We could hear the waves from our room, the feeling was fabulous! Next morning, I woke up early so that I could click some snaps of the sunrise

Beautiful dawn at Sea

The sunrise and the reflection of sun in the ocean

Watched people feeding the seagulls, cycling at the beach, we even saw a few dolphins

The lady feeding the Seagulls

People cycling by the beach side

The small black spots that you see, are dolphins

I was pretty interested in doing some adventure activities at the beach, one activity that I wanted to do since a long time was Parasailing, we found a place and went for it. The best part of Parasailing activity was when they dip you into the ocean!

Getting our feet wet while parasailing

Once we were back on solid ground, I saw my second love

Nice and Shining Harley Davidson

Yes, Harley Davidson, I had an opportunity to ride this bike once and from there on, I have fallen in love with it!

Destination Washington D.C.:

On the same day at about 1700 hrs we started our journey towards Washington D.C. A journey of about 230 miles took us around 5 hours (caught in a jam on our way).

Next day we had a big list of places to visit. We started early we saw the Capitol Building, White House, Lincoln Memorial, World war II Memorial, Washington Monument and finally the Air and Space Museum.

All these places are pretty close to each other. One can easily walk and visit them. But at the end of the day, I felt like I had no legs, I was dead tired.

The Capitol Building

The White House

Lincoln Memorial

Big Statue of Abraham Lincoln in the Lincoln Memorial

World War - II Memorial

Washington Monument

Air and Space Museum

Out of all these places, I liked the Capitol building and Air and Space Museum.

End of Journey:

Could not believe that it was already time to head home. After a week of fantastic trip and wonderful memories it was time we start our journey towards Chicago. A journey of about 750 miles took us around 13 hours. On the way back, we also got a ticket for not wearing seat belts!

Had lots of fun with some very different/beautiful experiences (one of the "different" experience on the way back was, I lost my checked-in bag!)

Overall it was a fantastic and memorable trip!

2300 Miles US Road Trip - Part - 1

2011-10-09T17:48:00.000+05:30

Recently me and my family had an opportunity to visit the US for two weeks. I had some official work for about 4 days. We had planned a vacation for the rest of the time. Naturally, I wanted to visit lots of places, do lots of stuff when I was in the US. Some of the top places I wanted to visit were Niagara Falls, The White House, Virginia Beach, Pennsylvania, Florida, Yellowstone National Park, Hoover Dam, Las Vegas, The Grand Canyon, Sears Tower (Willis Tower) etc.

The list was endless and I knew we cannot cover all these places in just seven days. We had to decide on places that were practically possible and onces that I wanted to visit the most.We decided to do a road trip to some of the most appealing places (read that as places that were practically possible :)). We ended up doing a 2300 miles fantastic road trip, here is a post that summarizes my experiences on the entire trip and things that I loved the most.

Our journey started from the Pune airport. We were supposed to fly from Pune to Delhi to Chicago. We were visiting the Delhi Airport for the first, we had approximately 5 hours between flights, I thought that will be a good time to see the Delhi Airport. I had heard lots of good things about it and boy, I was in for a surprise.

The Delhi Airport is one of the best International Airport that I have ever been to. By the way, I have been to a lot of international airports (Dubai, Bharain, Addis Ababa, Israel, Frankfurt, Paris, Berlin, Zurich, Chicago, Minneapolis, Amsterdam etc.) and I can be sure that, this one was one of the best.

There were things to do, places to visit, shops to be explored, right at the Airport itself. There are huge corridors, walkways, benches with palm trees around them, tint of indian culture etc.

Benches with palm trees around them

Huge corridors, walkways

Hand postures representing a dance form called Kathak

Out of all the shops at the Airport one that appealed the most to me had a healing pool in it. Its like a small pool which has water till your ankles, people are supposed to remove their footwear and walk inside the pool. The tiles in the pool would pressurize certain areas on your sole which gives you a very refreshing feeling. This science is called Acupressure. I did try it and felt totally incredible, felt really energized.

The healing pool with lots of bells hanging


As you can see people could walk in the pool and get rejuvenated

In the same shop there was live entertainment going on.

Person on left is playing flute and the other guy is playing tabla

The person with a long bamboo stick is playing a musical instrument called Flute and the other guy is playing a traditional Indian instrument called Tabla. The sound of these instruments is very soothing and relaxing.

All around the Airport, there are huge beautiful metal sculptures which depict Indian culture

The Sculpture of lord Shiva

Delhi Airport had a western look as well. There were recliner chairs with charging pods in the chair itself. Nice, Shiny Mercedes was also lying around at the airport. Over all it was a surprising experience.

Mercedes at the Delhi Airport

Next stop Chicago:

At Chicago on Monday - Labor Day - morning we visited a Swami Narayan temple near by. The temple had a rich white facade with lush green lawn and blue fountains. Under the clear blue sky with sparse white cloud cover it looked really beautiful!

The Swami Narayan Temple with white facade and lush green lawns

Blue colored fountains

Next stop Minneapolis:

As I mentioned earlier I had 4 days of official work. I had to leave for Minneapolis on Monday. But my trip does not stop here. Even when I was at Minneapolis, my friends at work took me to some very nice restaurants in the evenings.

Pat sitting on the left hand side and Paul on the right

But by far the single biggest attraction of my stay at Minneapolis was the Base ball game! My friends at Minneapolis had arranged tickets for a Base Ball game!

The game was between the home team i.e. Twins and Chicago White Sox. Although the home team was not having a great phase they did win this game! We had a very good view, our seats were right behind the batter, I could click some very good snaps

The road to the office was scenic, our office was around an artificial lake because of this the entire area had very good green cover with beautiful flowers and lovely trees!

Path way to the office

One evening one of my friend took me to the Mall of America! Its supposed to be the second largest mall in the world and boy it was big! There were hundreads of shops in fact it was so big that they had a mini amuzement part right inside the mall! I did sit in one of the roller coster which looked the most scary roller coster it had a drop of around 93 degrees!

Back to Chicago:

After fishing my work in Minneapolis, I came back to Chicago on Saturday morning, we decided to visit Chicago down town in the evening. Chicago downtown is just like any other city with lots of big buildings and beautiful lake near by.

We visited the Cloud Gate, Buckingham Fountain, Sears Tower, Navy Pier and oh yes, saw the fireworks at Navy Pier

The Cloud Gate is like a giant silver bean, it was quite a popular spot, we saw at least 6-7 newly wed couples near it.

The Cloud Gate

Snaps from the Buckingham Fountain, sears tower and navy pears

Buckingham Fountain

View from the Sears Tower

Fireworks at Navy Pier

At Sears Tower, there is glass compartment which is suspended, which means there is nothing below it - The Sky Deck! You are standing on glass 110 stories high with nothing below you, Scary!

While heading home we even saw a opera it was quite interesting actually!

By this time one week had already passed. I had lot of fun and some very first experiences in this memorable one week.

Next day we were supposed to leave for our road trip. In my next post I will share my experiences about the places that we visited in the coming week!

Stay tuned folks!

Strongly Typed Session Using MVC3 and Spring.NET - Part - 3

2011-10-02T15:26:00.000+05:30

This is the last post in the series of posts to show How we can achieve Strongly Typed Session using MVC3 and Spring.NET. In previous two posts (first and second) we saw what were the benefits of strongly typed session and how we could enable the session scope using Spring.NET 1.3.1.

The question that we had left unanswered in the previous post was:

What is "session" scope supposed to do with Strongly typed sessions?

Well the session scope is the key, in this post we will see how it enables us to create a strongly typed session.

The session scope enables us, to register objects with scope=session. Spring.NET will create one instance per user Session for objects that are configured with scope=session. This means effectively we can use such an object to hold information that would otherwise be saved in the HTTP Session. Since, these objects are configured in Spring.NET, they can be injected as dependencies in any other class.

Where would we want to inject such a class?

MVC Controllers are a good place to do so.

Why?

Because, controllers are the class that would do some business processing and then would want to set some information in the HTTP Session.

For e.g. Lets consider this scenario, we have a AuthenticationController which has a Login action. This method takes in the username and password as argument. This method calls the AuthenticationService to find whether the username and password combination is valid. If the username and password are valid, AuthenticationController wants to set the username in the session so that it could be used by other pages to show user information on the page.

In our case we do not want to use the HTTP Session. Hence, our AuthenticationController will put the username in some object, let's call it UserContext, which has been configured in Spring.NET with scope="session".

But wait,

What if two users are trying to login at the same time? Will one thread overwrite the username saved in the UserContext by the other thread?

As the UserContext object is configured with scope="session", Spring.NET is smart enough to understand that, two requests are coming to AuthenticationController are coming from two different user sessions, it injects two different instances of UserContext in the AutheticationController for two different user sessions. Hence, one thread will *not* overwrite the username set in the UserContext by the other thread.

So far so good.

What happens when, user who had logged in earlier (i.e. UserContext has the UserName property set), sends another request to the server? How do we ensure that correct UserContext instance is injected in the Controller so that previously set UserName could be retrieved?

Again, Spring.NET is smart enough to understand that, for the given user session an instance of UserContext was already created, Spring.NET will find the correct UserContext instance associated with the user session and injects it in other Controllers/objects that need it. This effectively eleminates the direct use of HTTP Session for us. We can be sure that, whatever information we set in UserContext will be specific for a given User Session and correct UserContext instance will be injected in the controller whenever we need it.

I know, at first it all looks very confusing, but lets look at the code to make things clear.

These are the steps required to configure the UserContext and AuthenticationController

Register the UserContext in Spring.NET configuration file with scope="session"
Write the implementation for the UserContext
Register the AuthenticationController in Spring.NET configuration flie with singleton="false". Add to this the dependency of UserContext.
Write implementation of Login Action to set the UserName Property in the UserContext. Write another action method let's call it UserInfo which will retrieve the UserName Property from the UserContext object and display it on the UI
Write the RazorIndex view to show the UserName from the ViewBag
Test that the setup is working with two different browser windows trying to Login and access the UserInfo at the same time. We should see correct user names being displayed based on where the request is coming from

Lets look at individual steps in detail

Register the UserContext in Spring.NET configuration file with scope="session"

Update the web.xml and add the configuration for UserContext as follows
Notice that we have registered the UserContext object with scope="session". Spring.NET will understand that the UserContext object is registered with singleton but in session scope i.e. One UserContext object per user session.

Write the implementation for the UserContext

The UserContext class would be simple enough.
Register the AuthenticationController in Spring.NET configuration flie with singleton="false". Add to this the dependency of UserContext

Update the web.xml to add the configuration for AuthenticationController as follows.
Write implementation of Login Action to set the UserName Property in the UserContext

AuthenticationController will have a Login Action and UserInfo Action. Login Action will set the UserName in the UserContext. For demonstration purpose we will append the SessionId to the UserName. This will show us that, when we request the login action method from two different browser windows we will see different user name on the UI.
Write the RazorIndex view to show the UserName from the ViewBag

The RazorIndex.cshtml will simply show the UserName from the ViewBag if its not null. Else it will show no user logged in yet message.
Test that the setup is working with two different browser windows trying to login and access the UserInfo at the same time

Enough theory, lets see the setup in action. As they say,

Real test is in the real world!

Let's try and browse the URL Authentication/UserInfo from FireFox and Internet Explorer and see what happens

User not logged in - FireFox

User not logged in -IE

As the UserName is not set in UserContext we see no user name set message.

Now in IE, lets browse the URL Authentication/Login and see what happens.

User Successfully Logged in. UserName and SessionID shown - IE

As we can see, the page redirect to the Authentication/UserInfo User name with session id is shown.

Lets reload the Authentication/UserInfo the page on FireFox. Remember that, we have not yet visited the Login page on FireFox, but we have visited the Login page on IE.


Even after user is logged in on IE, he/she is not logged in on FireFox

Yep even though we have set the UserName in the user session started from IE, it does not effect the UserContext reference in the User Session started from FireFox.

Now lets visit the Authentication/Login URL in FireFox as well to see what happens.


Now user is logged in even in FireFox

As you can see it redirects to Authentication/UserInfo with UserName set. Notice that the session id in FireFox window is different from the session id in IE.


Notice that the Http SessionId are different in two images

This effectively proves that UserContext reference plugged in by Spring.NET when request comes from IE is different from the one that gets plugged in when request comes in from FireFox i.e. We have successfully injected different UserContext reference per User Session. UserContext object can now be used to hold any information that would have be placed in HTTP Session. This effectively eliminates the use of HTTP Session and gets us a strongly typed object which can be used as a replacement to the HTTP Session.

That's it! we have successfully achieved Strongly Typed Session in an MVC3 application using Spring.NET.

Strongly Typed Session Using MVC3 And Spring.NET - Part - 2

2011-09-27T05:00:00.000+05:30

In my previous post, we saw why its a good idea to have strongly typed HTTP Session. In this post we will do the real thing (well at least first part of the real thing).

Using Sping.NET we can control the objects lifecycle. By default the Objects configured in Spring.NET are Singleton for the entire application (singleton for scope="application"). Singleton for the entire application means, Spring.NET will create only one instance of the given object for the life-cycle of the application. Service classes, DAO classes etc are a good fit for Singleton Pattern.

However, there are other Object Scopes that can be configured in Spring.NET. These Scope's tell Spring.NET that, A given object is singleton but in a given scope.

One of the interesting scope is session. Session scope is only available if we use a web-aware Spring IApplicationContext implementation. For such objects, Spring.NET would create one instance per HTTP Session. In this post we will see how we could take advante of the session scope, to create Strongly typed session in an MVC3 application.

Enough blabber. Let see how exactly its done.

How do they do it!

As we had seen in one of the previous post to Integrate Spring.NET with MVC3 application, we had used the Spring.Context.Support.MvcContextHandler. This class instantiates Spring.Context.Support.MvcApplicationContext, which in turn initializes the spring application context.

Support for session scope is porivded by spring.web.dll. Unfortunately, as of Spring.NET 1.3.1, spring.web.mvc.dll does not depend on spring.web.dll, which effectively means, MvcApplicationContext is unaware of the session scope.

Hence, using just the MvcApplicationContext there is no way we can eable the session scope for a certain objects. Its a bug you can find more details here. But where there is will, there is a way! There is obviously a way around this problem.

Here are the steps to enable full session scope support for any object in Spring.NET 1.3.1

Update the web.config to use a custom WebContextHandler
Register the spring WebSupportModule in web.config under the <httpModule> section for IIS6 and under the <modules> section for IIS7+
Implement the custom WebcontextHandler. Override the property DefaultApplicationContextType. This property informs Spring.NET about the class type of ApplicationContext that Spring.NET should instantiate. This will be again a custom class that extends from Spring.Context.Support.MvcApplicationContext.
Implement the custom class that inherits from Spring.Context.Support.MvcApplicationContext. This class will basically override two methods

CreateXmlObjectDefinitionReader to return the instance of WebObjectDefinitionReader
CreateObjectFactory to return the instance of WebObjectFactory

That's it! We have now successfully enabled the session scope for Spring.NET 1.3.1. Let look at the actual code.

Step - 1 - Update the web.config to use a custom WebContextHandler

Update the web.config configSection to look like

The class aspnetwebFormsWithMvc3App.Context.MvcWebContextHandler is our custom class where we will override the property DefaultApplicationContextType.

Step - 2 - Register the spring WebSupportModule in web.config under the <httpModule> section for IIS6 and under the <modules> section for IIS7+

Adding the spring module for IIS6 and IIS7

Step - 3 - Implement the custom WebcontextHandler

Lets look at the custom WebcontextHadler class aspnetwebFormsWithMvc3App.Context.MvcWebContextHandler

As you can see this class is pretty simple, it just overrides the property DefaultApplicationContextType to return the class type of custom WebApplicationContext.

Step - 4 - Implement the custom class that inherits from Spring.Context.Support.MvcApplicationContext

This is the class where the real work is done.
That's it! We have successfully enabled the session scope in MVC3 application using Spring.NET.

Question is: What is all this supposed to do with Strongly typed sessions?

In the next and the final post in this series, we will see how we can take advantage of session scoped objects to create the Strongly Typed Session in an MVC3 application using Spring.NET.

Strongly typed Session with MVC3 and Spring.NET - Part - 1

2011-09-21T23:38:00.000+05:30

In every Web project I have worked so far, always there is a need to store some or the other information in HTTP Session.

I am very finicky about putting things in the HTTP Session. IMHO HTTP Session should be kept as light as possible. There should be a very very strong and compelling reason to put a certain value in the HTTP Session. Having said that, I do agree that there are times when we want to put certain information in the HTTP Session like user name, user id, etc.

With me so far? Till now we understand that, we do not like to put things in HTTP Session, however, we understand that there is a definite need to put certain things in HTTP Session for the app to work correctly.

Question is, Is there an alternate to putting things in HTTP Session?

In this series of three posts we will try to answer this question. We will try to use MVC3 and Spring.NET to eliminate the need of putting stuff in HTTP Session.

But wait a minute, let's first answer this question,

Why is putting things in HTTP Session is such a bad idea?

Lets start by looking at the code that will put UserName into the HTTP Session.
Does it remind you of code that you have seen in one of your applicatoin? It certainly reminds me of code that is used all over the place in one of the projects that I had worked.

I totally hate to see code like this. Why whats wrong with such code?

Session is used like a Hash Table:

In our example, key of the hash table would be UserName and value will be Git Shah. Imagine while putting/retrieving values to/from the HTTP Session, if incorrect keys are used, results could be catastrophic. Even a small typo like one shown in the example below would yield incorrect results
Ofcourse there are ways to get around this problem. Simplest of them is to use a Static Class that holds all the keys as constants.

However, you can still not gaurentee that every programmer on your team will always always use constant keys from the static class. Believe me, because of human error or some other unexplainable reason, someone would use hardcoded inline key to put value in the Session and rest of the team will curse him/her for eternity!

Session values are not strongly typed:

We have inserted a string value Git Shah in the HTTP Session against the key UserName but when we want to retrieve the value from the session we will have to typecast to its correct type. In our example to retrieve the value of UserName key in HTTP Session code would look like
I have a name for code like that, I call it A Time Bomb!

Its just a matter of time before which you start getting a class cast exception.

Dude, whats in my Session? Where are these values used?

Can we say for sure what all values are stored in our session?

Another important question to answer is, Can we say for sure, from what all places in the code are these values accessed?

These are a very important questions. Answers to these questions are absolutely vital for the health of your application.

In one of the projects that I had worked, we were extremely confused for a very long period of time. The question that was bothering us was, who was setting a specific value in the session and in what condition?

Consider an example

UserName key is populated in the LoginController.Login method. Lets look at another place where UserName is used.
AccoutController.UpdateUser updates the value of the UserName key in the HTTP Session to the new value passed as argument and AccountController.GetUserDetails method fetches the value of the UserName key from the HTTP Session and sets the ViewBag.UserName property.

As you can see in our simplistic example UserName key of the HTTP Session is used at three places. In real life, it would be used in many more places than this. The code that accesses the UserName key from the HTTP Session would be littered all over the place!

How will we find the exact places where UserName key is populated/fetched into/from the HTTP Session?

There is no easy way to find the exact places where UserName key of the HTTP Session is used. We can extract a constant for the UserName and then search for its usages, but still this is not a full proof way, as I mentioned earlier, there are many fools out there who will not bother to use the constant and use an Hardcoded string key instead.

These are few of the problems that are a big concern to me. Question to be answered is,

How can we avoid such code?
How can we achieve an alternate option that will discourage people from writing such code?

Before we dig deeper into the solution, lets first focus on some of the design goals that this new solution should have.

Things that we put into and fetch from the session should be strongly typed objects. There should not be any cast of any sort. This will avoid Time Bomb like code.
Session should not be used as a Hash Table. Whatever we put into the session should become a property of the session which can be accessed from anywhere in the code. Moreover, because it becomes a property of the HTTP Session, we would get all sorts of intelli-sense on it. There is no chance of any typo in this case. For e.g. the existing code

Should no longer be required instead we should have code that looks like

At any point in time, it should be easy enough to find the exact places where a certain property of HTTP Session is used. In some ways this point is related to the previous point. If we are able to convert Keys of session into strongly typed Properties then we can just search for the usages of those properties and we know exactly where those properties were used!

Enough goals to start with. Question is, Can these goals be implemented in real life?

Well, the good news is, these goals are definitely implementable. In fact its not too difficult to implement these goals when we use Spring.NET with MVC3.

What is the bad news? You will have to wait for my next posts to see how exactly we could implement these goals :)

Stay tuned folks in the next couple of posts we are going to have some fun with HTTP Session Management!

Calling a Stored Procedure with OUT parameters using NHibernate

2011-09-20T23:08:00.000+05:30

Hibernate/NHibernate is one product that I appreciate a lot. I absolutely love the flexibility that NHibernate offers.

Lots of ORM products out there work well when we are developing application from scratch, however they fail miserable when we have to work with existing legacy database. Hibernate/NHibernate on other end, works very well when application are build from scratch as well as when we have to work with legacy database. Hibernate/NHibernate provides lots of hooks because of which, using it with any legacy database is completely painfree.

In this post, we are going to see how we could use NHibernate to get around one such legacy issue. One of the common technique employed in legacy databases to exchange information between the DB and Application layer is, having stored procedures with OUT parameters. OUT parameters are special type of parameters in which the stored procedure returns values which can then be accessed by the Application layers.

We will not go into details whether OUT parameters are good or bad but, we will see how exactly we could use NHibernate to invoke stored procedures with OUT parameters.

Calling stored procedures without any OUT parameters using NHibernate is fairly straight forward, but calling stored procedures having OUT parameters using NHibernate is a bit tricky. Without wasting any more time lets straight away dive into the solution.
How do they do it?

To call stored procedures with OUT parameters we have to fallback to the System.Data.SqlClient.SqlCommand (assuming that the database is MSSQL) to get the job done.

Let's say that,

We have a stored procedure called encrypt that is supposed to encrypt any string
It takes INPUT parameter called stringToEncrypt
It returns the encrypted string as an OUTPUT parameter called encryptedValue

We want to invoke this stored procedure using NHibernate. The steps to do this would be as follows

Get an instance of Session using the NHibernateSessionFactory (or get it some how injected into the DAO class using Dependency Injection)
Start the transaction on the NHibernate session using BeginTransaction method
Create a new instance of System.Data.SqlClient.SqlCommand
Set the Connection property of the SqlCommand instance using the NHibernate Session's Connection property.
Set the SqlCommand instances CommandType property as CommandType.StoredProcedure
Set the SqlCommand instances CommandText property as the stored procedure name, in our case encrypt
Add the INPUT parameter to the SqlCommand instance. In our case, adding the parameter with name stringToEncrypt and setting the value of the parameter to the string we want to encrypt
Add the OUTPUT parameter to the SqlCommand instance with Direction property set to ParameterDirection.Output. In our case, add the parameter encryptedValue. Since, this is an output parameter we will have to specify its SqlDbType as SqlDbType.NVarChar and size as 255. But the most important thing here is, to set the Direction property of SqlParameter to ParameterDirection.Output. This tells the SqlCommand that this parameter is an output parameter. After executing the stored procedure we will be able to get the value out of this parameter.
Enlist the SqlCommand instance with the NHibernate transaction. This is another crucial step. If this is not done SqlCommand will throw an error saying "ExecuteNonQuery requires the command to have a transaction when the connection assigned to the command is in a pending local transaction. The Transaction property of the command has not been initialized"
Execute the command using ExecuteNonQuery method on the command instance.
Get the output parameter value from the SqlCommand instance
Commit the NHibernate transaction and Close the NHibernate Session.

As you can see there are quite some steps involved while invoking a stored procedure with OUT parameters. But looking at the code would definately make things easier.
And that is how its done!

Fixing the The virtual path '/test/currentcontext.dummy' maps to another application, which is not allowed

2011-09-05T19:39:00.000+05:30

We have recently integrated Spring.NET in an MVC3 application. In one of my previous post, I have already documented the detailed process of integrating Spring.NET with MVC3 application. Spring.NET is really helpful when it comes to writing Modular, Testable code. I totally recommend Spring.NET as Dependency Injection framework of choice!

We did a fresh deployment of our application on a new server. On this computer, when we tried to access any of the pages of our application we got an error which looked like

Virtual Path Maps to Another Application, is not allowed

That was not very encouraging. How can an application that was working correctly on all other servers stared giving exceptions for this particular server?

We tried to find out what were the differences between the other server and this new server. We narrowed it down to one very interesting difference, we used to deploy our project as a website in IIS on all other server. While on the server that was giving us exceptions, we deployed our application as a virtual directory under the default website. This very curious difference was giving us exceptions while initializing the spring context.

I Googled a little and found out that this is a known issue which has been reported here and here.

The workaround to this issue was simple enough, we just have to give a name the spring context in web.config.

Would any random name work? Do we have to follow some convention?

Our Spring.NET section in web.config file looked like
By now it should be clear that, we have to add a name attribute to node. But the question is, what value should it hold?

The answer is pretty simple, the Spring.NET context should be named after the virtual directory under which the application is deployed. For example, if the virtual directory under which the application is deployed is called springmvc. The updated web.config should look like

This did the trick! After making this change we were able to browse all pages in our application and Spring.NET integration with MVC3 application was working well!

Small issues like these have very simple solutions but, one might end up spending quite some time before finding a satisfactory solution. I decided to document my finding for the benefit of the community!

Integrating NHibernate with Spring.NET and MVC3

2011-08-28T22:14:00.005+05:30

We have seen in one of the previous post that its not very difficult to integrate Spring.NET with MVC3 application. Almost every real project has a need to talk to a database for doing stuff. When it comes to interacting with the Database, NHibernate does an Awesome job.

I have used NHibernate with legacy database as well as databases that were designed from scratch. IMHO, other competing products have a long way to catchup before they can be compared with the flexibility that NHibernate offers!

Recently, in one of MVC3 project we integrated NHibernate using the Fluent API with Spring.NET. This process is not very difficult but it involves a few steps. I decided to document the steps for the benefit of the community.

We will continue from where we left in the previous post, if you did not read the previous post already then, I strongly recommend reading it to understand how you can integrate Spring.NET with MVC3 application.

Very well than, Lets start!

To integrate NHibernate using the Fluent API in a Spring.NET MVC3 application one has to do the following

Make the Web.config changes to include the Parsers config section in the spring configuration.
Register the OpenSessionInView HTTP Module in web.config. We will be using single session per request, NHibernate session handling strategy.
Add the reference to the required DLL's
Register the DB provider in the spring configuration file.
Register the NHibernate Session Factory in the spring configuration file
Write a custom implementation of the NHibernate Session Factory that loads and registers the Assembly that has Fluent NHibernate mappings in it.
Register the Transaction Manager in the spring configuration file
Enable the Spring AOP Transactions using attributes in the spring configuration file
Test out the NHibernate integration

Lets execute the steps one by one. Please remember I will be continuing from where I left in the previous post. Hence, I will be assuming that you already have Spring.NET correctly integrated in your MVC3 application.

Step - 1 - Make the Web.config changes to include the Parsers config section in the spring configuration

To integrate only Spring.NET with MVC3 application we had to add a configurationSection called context. To integrate NHibernate with Spring.NET in a MVC3 application we have to add another configurationSection called parser. This section will help Spring.NET understand the Database and Transaction configuration. Lets look at how its done.

Step - 2 - Register the OpenSessionInViewModule HTTP Module in web.config. We will be using single session per request, NHibernate session handling strategy

Why do we need the OpenSessionInViewModule HTTP Module?

This will ensure that before the request is processed by any Controller, NHibernate session is opened and attached to the current thread. And just before sending the response to the client, the NHibernate Session will be flushed and closed. This enables us to do lazy loading even when the View is being rendering. Lets look at the web.config change required to enable this:

Step - 3 - Add the reference to the required DLL's

For integrating NHibernate with Spring.NET and MVC3 application we need to add the references of following spring DLL's

Spring.AOP.dll -> Used to enable AOP using Spring.NET. This will be heavily used when we enable transactions in our application.
Spring.Data.dll -> Used to enable the Spring.NET Database module
Spring.Data.NHibernate30.dll -> Used to integrate Spring.NET with NHibernate
NHibernate.dll -> The core NHibernate DLL
Iesi.Collections.dll -> Provides the ISet interface and its implementation
FluentNHibernate.dll -> Used to enable FluentNhibernate configuration for NHibernate
log4net.dll -> Used by NHibernate for logging. Lot of effort has been put into making the NHibernate log as detail as possible and as readable as possible. This is a great tool when it comes to debugging something that you cannot understand easily
NHibernate.ByteCode.Castle.dll -> Used by NHibernate to enable Dynamic Proxies
Castle.Core.dll -> Used by NHibernate to enable Dynamic Proxies using Castle

Step - 4 - Register the DB provider in the spring configuration file

This step is pretty simple. We need to inform NHibernate about the Database. This is done by registering a DbProvider in the spring configuration file. We will be registering the DbProvider in the web.xml (web.xml was the file we created in the previous post. This hold all the Spring.NET configuration). Add following configuration to the web.xml

Notice that we have to define the db name space in the objects root tag. Please update the above configuration with the DB Credentials that you are using.

Step - 5 - Register the NHibernate session Factory in the spring configuration file

Session Factory is a heavy weight NHibernate object. Only one instance of SessionFactory is create per database. Lets look at the Spring.NET configuration file for registering the NHibernate Session Factory.

The above configuration registers a NHibernate session factory which is of type aspnetwebFormsWithMvc3App.NHibernate.FluentNhibernateLocalSessionFactoryObject and is found in the assembly aspnetwebFormsWithMvc3App. This is a custom class that we will create later in the post. It has various properties like

Which DbProvider to use
What are the FluentNhibernateMappingAssemblies. This is a list of assembly names which our NHibernate Session factory class will search for to locate Fluent Nhibernate mappings
What are the MappingAssemblies. This is a list of assembly names which has normal NHibernate mappings
What are the HibernateProperties. These are the key value pairs of NHibernate properties that we would want to use. We are setting the connection provider, dialect, driver_class and sql_sql values

Step - 6 - Write a custom implementation of the NHibernate session factory that loads and registers the Assembly that has Fluent NHibernate mappings in it

We have to provide a custom implementation of NHibernate Session Factory class so that we could load our Fluent NHibernate assemblies. These assemblies will have the Fluent NHibernate mapping classes. The code of our Custom NHibernate Session Factory looks like this

Step - 7 - Register the Transaction manager in the spring configuration file

NHibernate always interacts with the database using Transactions. We would definitely want to control how and when the transactions are started and ended. For that to happen we need to register the transaction manager in the spring configuration file. Add the following to the web.xml spring configuration file.

Step - 8 - Enable the Spring AOP transactions using attributes in the spring configuration file

We would want to govern how and when the transactions are started using Attributes. For this to happen we need to inform Spring.NET that we are going to use Attributes to define the transactional behavior we need. Spring.NET reads these attributes and starts/commits/rollbacks transactions using AOP. To inform spring.net about our intentions add the following line in web.xml spring configuration file.

Notice that we added the line tx:attribute-driven but we also added the namespace tx in the objects root node declaration.

If you have followed all the steps so far then you have successfully Integrated NHibernate with Spring.NET in an MVC3 application.

Step - 9 - Test out the NHibernate integration

To test out our brand new NHibernate integration with Spring.NET in an MVC3 application, lets change the implementation of the AccountService.UserCount method. Currently the AccountService.UserCount method returns a hard coded value of 5454. We will now replace it with the correct implementation, we will get the record count from the User table.

We will have to create an NHibernate entity that maps to the User table. Lets create the User class that will map to the User table. We will also create a class called UserMap class that will hold the NHibernate Fluent Mappings for the User class.

The details are embedded in comments. We have the User entity class and the UserMap mapping class to hold Fluent NHibernate mappings.

Now, its time to actually use the User class to get the count of number of users in the database. To do that, we will have to inject the SessionFactory instance in the AccountService. Updated Spring.NET configuration file web.xml would look like

After injecting the SessionFactory its time now to change the implementation of the UserCount method to return the count of users from the users table. Updated AccountService looks like

Since we have registered the OpenSessionInViewModule HTTP Module the GetCurrentSession method of SessionFactory instance will return the same session instance that was created by the HTTP Module. This is absolutely vital for the Lazy loading to work correctly.

Also notice that we have added the [Transaction] attribute to the method UserCount with TransactionPropagation as Required. This will make sure that whenever the method UserCout is invoked either a new transaction is started or the method will join any existing transaction started earlier in the same thread.

All set! Its time to run the server and browse the /Home/SpringIndex URL to see if our integration with NHibernate works!

Showing the actual user count using NHibernate

As show in the image we are successfully getting the UserCount from the database. We have successfully integrated NHibernate with Spring.NET in an MVC3 application!

India Against Corruption

2011-08-21T16:39:00.000+05:30

No matter who you are or what you do, everyone in India has heard about Anna Hazare. Team Anna Hazare, has been protesting at ramlila maidan (in New Delhi) against the Government and is trying to push the Jan Lokpal Bill.

Over the years, Corruption in India has grown by leaps and bounds. Everywhere there is some or the other form of corruption. From getting a motor driving license to getting admission in a school/college, we i.e. the common man has to face corruption at all levels.

At bigger levels there is bigger corruption. Indian politicians have found a way to exploit the system in the way they want it. May it be, organizing Common Wealth Games or building roads, corruption is everywhere.

Corruption has made our nation handicapped. Team Anna hopes that by implementing the Jan Lokpal Bill we can curb the corruption at high level by 60-65%. I certainly hope that it gets implemented.

But what about us?

At times, in many situations, we feel helpless against the system. To get around it, we ourselves fuel the corruption. Its high time that we change ourselves. How many times we have given 100 Rs to the traffic cop and not taken a receipt? I am sure most of us have done this, I certainly have done it.

Corruption is a two way process, person who is taking money is corrupt and person who is giving money is also corrupt. We common man have to make sure that from now on, we will never every fuel corruption ourselves. If we can do that, corruption can be reduced even further. This is the first step!

Team Anna is taking great pains to push the Jan Lokapl Bill, the least we could do is change ourselves and stop fueling corruption.

In my entire life span, I have never seen India as united as it's today! People of India want to get rid of corruption.

A few days back, if we asked who he/she was, they would say they are Hindu, Muslim, Christian etc. but today we are all Indians! Indians who want a change! We have forgotten our petty differences and have reunited. It feels like a nation with 1 Billion+ people, aiming for one thing i.e. get rid of corruption!

My only message to government is, you better pay attention to our demands, else wait and perish!

There are very few times when we get to be part of history, in India today is that day! Lets support Team Anna in this historic movement to remove corruption from our country!

Charity begins at home, lets take a oath today that we will not fuel corruption in any way! No more bribes!

Lets do whatever we could to support this movement in whatever way we could! Jai Hind!

Integrating Spring.net with MVC3

2011-08-13T14:35:00.001+05:30

If you are doing any kind of serious project, you sure want to use some sort of dependency injection.

Why do I need dependency injection?

Lot's of people have written lot's about why dependency injection is required. I am not going to repeat all that, just a few points that I think, cannot be ignored

Configuration flexibility - alternative implementations of a given service can be used without recompiling code
Testability - Since we can provide alternative implementation of any given service at any time, any class can be unit tested very easily and effectively. Its easy to inject a fack implementation of a service in the class we want to unit test
Reduction of Boiler plate code - Since we no longer need to create instances of classes and inject them with proper dependencies
Promotes loose coupling between classes and subsystems - Since classes are not dependent on a specific implementation, they are dependent on the contract that the other classes expose.

Enough said about dependency injection. At this point if you are still not convinced that you need dependency injection, then I strongly recommend reading the full series of posts on Dependency Injection from Fabien Potencier here

Alright, so we need dependency injection, what next?

How do we implement dependency injection? Should we hand code the logic and create our own way of doing dependency injection?

Well, take my advice, never ever, never every do that! Always look for options, you will surely find one depdendency injection framework/container that suites your requirements. Even if it does not fit your requirement completely, I am sure the framework/container provides enough hooks, based on which you can customize it to you taste!
In the year 2011 and beyond, you cannot be serious if you are wasting your time and someone Else's money writing a dependency injection framework (I know a few people who have done that, can you believe it!)

What is Spring.NET?

Spring.NET is an awesome dependency injection framework (actually its lot more than that). Its open-source and its design is based on the proven Java based Spring Framework. The breath of functionality in Spring .NET spans application tiers which allows you to treat it as a "one stop shop" but that is not required. The best thing about Spring .NET is, its not an all-or-nothing solution. You can use the functionality in its modules independently. Spring.NET integrates with a lot of other projects like NHibernate, NUnit, Active MQ, Quartz .NET, Velocity etc.

Enough about Spring.NET, if you want to know more details about Spring.NET, I strongly recommend visiting this site, it has all the info you need. It has links to the latest documentation, blogs, articles, faq's etc.

By this time we agree that, we should use a dependency injection framework and for the sake of this post, I am going to show how easily we can integrate the Spring.NET dependency injection framework in our MVC3 application. Lets not waste any more time and jump into the code straightaway to get this part started, shall we!

How do they do it?

To Integrate Spring.NET into the MVC3 application you will need to do the following steps

Add the Spring.NET dependencies to the project.
Make the web.cofig changes to register the spring context
Declare the dependencies in an xml file
Write the code to interact with the dependencies in the HomeController
Inherit the Global class in the Global.asax.cs file from Spring.Web.Mvc.SpringMvcApplication

Lets have a look at each step one after the other.

Add the spring.net dependencies to the project

For integrating Spring.NET in your MVC3 application you will need to add references of following Spring.NET assemblies.

Spring.Core.dll
Spring.Web.dll
Spring.Web.Mvc.dll

These dll's can be found in the downloaded Spring.NET zip package from this link.

Make the web.config changes to register the spring context

In this step we add a <configSection> for the Spring.NET configuration. In the spring config section we would like to register the spring context. Update web.config should look like

What we are doing here is, telling Spring.NET to look for a file called web.xml inside the Config directory.

What will the web.xml file contain?

It will contain the dependency injection configurations for the classes in our application.

Declare the dependencies in an xml file

Spring will search for a file named web.xml inside the Config directory, directly under the root of you application. Lets make sure, Spring finds the file. Create a directory called Config at the root level of your MVC3 application and create a file called web.xml inside it. The project structure should look somewhat like this

Project Structure with web.xml included in the project

Now, its time to register our dependencies in the web.xml. We are going to add a very simple dependency of IAccountService that will give us total number of users in the system. The web.xml would look like this

The web.xml is just a plain XML file that uses Spring.NET xml schema to inject the AccountService into the HomeController. We have given an id to the HomeController its Home in our case. We are also declaring that HomeController has a property Called AccountService. This property will be auto injected by the Spring.NET container with the reference of type AccountService.

Notice that the HomeController is not a singleton class. MVC3 framework creates a new instance of the Controller every time a request comes for that controller. Hence, Spring needs to know that, the HomeController is not supposed to be a singleton class. This is done by adding the attribute singleton="false" in the declaration of the HomeController.

The next declaration in the web.xml is that of AccountService. For the sake of this example this class has no dependencies. For our example this class is going to be very simple it will return some hard coded values.

Notice that the singleton attribute has not be specified for the declaration of AccountService. By default the singleton value is true, this means that the AccountService is a singleton class. Spring.NET will instantiate only one instance of the AccountService for the entire application!

Write the code to interact with the dependencies in the HomeController

As seen in Step - 2, HomeController is going to have a property of type AccountService. We are going to query the AccountService to get the user count which will be eventually be shown on the UI.

As you can see there is no magic in our classs. HomeController simply calls the AccountService to get the UserCount and sets the value in the ViewBag. The AccountService implementation is extremely simple, it returns a hardcoded value of 5454.

Lets look at the view SpringIndex.cshtml that will show the UserCount from the ViewBag.

As you must have guessed, the view could not have been any simpler. It simply shows the message with the UserCount in it.

Inherit the Global class in the Global.asax.cs file from Spring.Web.Mvc.SpringMvcApplication

We are almost done with all the plumbing required to integrate Spring.NET with MVC3 but, one last thing is left. We need to inherit our Global class in the Global.asax.cs file from Spring.Web.Mvc.SpringMvcApplication instead of System.Web.HttpApplication. This is required so that Spring registers the controller factory to instantiate the Controllers when the request comes. This is done so that Spring gets an opportunity to dependency inject the controllers.

And we are done! Lets run the application and browse the URL /Home/SpringIndex to see what happens. If you have followed the post till this point, browsing the url will result in a page like this

SpringIndex renders with UserCount value

The User count 5454 is coming from the AccountService which we successfully injected into the HomeController.

When the request for SpringIndex action of the HomeController comes to the server following things occur

Spring intercepts the call and instantiates HomeController
Injects all the dependencies of HomeController i.e. AccountService in our case
Calls the action method SpringIndex to complete the request
SpringIndex method calls the AccountService to get the UserCount and sets it in the ViewBag
SpringIndex.cshtml is rendered which shows UserCout!

We have successfully integrated Spring.NET with a MVC3 application! That's all folks!

Asp.Net MVC3 and Asp.Net Web Forms Side-By-Side in One Project - With Razor Views

2011-08-06T12:50:00.001+05:30

In The previous post, I explained how we could use MVC3 and Asp.Net Web Forms Side-By-Side in One Project. In that post, we were rendering MVC3 pages using the ASPX View Engine. ASPX View Engine is the legacy view engine, Razor View Engine has significant advantages over the ASPX View Engine. In this post we will see how we could use the Razor View Engine instead of the ASPX View Engine.

The Asp.Net MVC 3 has always supported the concept of view engines – which are the pluggable modules that implement different template syntax options. Razor view engine is optimized around HTML generation using a code-focused templating approach.

Why Razor View engine?

Razor View engine is build with the following design goals (read the detail design goals here)

Compact, Expressive, and Fluid: Razor minimizes the number of characters and keystrokes required in a file, and enables a fast, fluid coding workflow. Unlike most template syntax's, you do not need to interrupt your coding to explicitly denote server blocks within your HTML. The parser is smart enough to infer this from your code. This enables a really compact and expressive syntax which is clean, fast and fun to type.
Easy to Learn: Razor is easy to learn and enables you to quickly be productive. You use all your existing language and HTML skills.
Is not a new language: Razor enables developers to use their existing C#/VB (or other) language skills and deliver a template markup syntax that enables an awesome HTML construction workflow with language of our choice.
Works with any Text Editor: Razor doesn’t require a specific tool and enables you to be productive in any plain old text editor (notepad works great). That said, I have to mention this, Resharper 6.0 has one of the best support for Razor Syntax. Its a paid tool but its definitely worth the price.
Has great Intellisense: While Razor has been designed to not require a specific tool or code editor, it will have awesome statement completion support within Visual Studio.
Unit Testable: This is one of the most important Design Goal. Razor implementation supports the ability to unit test views (without requiring a controller or web-server and can be hosted in any unit test project – no special app-domain required).

If you want to see a detailed reasoning of why Razor is better then ASPX View Engine, I strongly recommend vising this link

Now that we are convinced that, Razor View Engine is better than ASPX View Engine, lets integrate the Razor views in our Asp.Net Web Forms application with MVC3 integrated in it.

How Do they do it?

Initially, I though there must be setting somewhere in the project properties which will enable us to use the Razor View Engine, however, after several wasteful minutes of searching and researching in the project properties I found that, there is no such setting, we can directly go ahead and create the Razor view in the project.

Razor views have an extension of cshtml. Lets create our first Razor view called RazorIndex.cshtml right next to the Index.aspx that we had created in the previous post. This view will have nothing special, it will display one message that we set in the ViewBag.

Notice that we are displaying the UserName from the ViewBag. Notice that, in the new Razor syntax, we have to use the @ sign to access any objects and properties.

Let's add one more method in our HomeController which will render the RazorIndex view and set the UserName in the ViewBag. The controller code will look no different from how it looked for rendering ASPX pages.

This is a very good example of showing, how well the Seperation of Concerns Design Pattern has been implemented in MVC3. Even though we want to use a different View Engine, controller code remains unchanged. In fact, Controller has no idea about which View Engine will be used to render the view. This is an excellent example of why things should be loosely coupled to achieve modular and clean code!

Before we can see if our Razor view is rendered correctly or not let's not forget to add one more route to our routes collection in Global.asax.cs file

We have just added one route called RazorHome. All set! Lets build the project and access the URL /home/razorindex and see whats the outcome.

First attempt to render Razor View results in an compilation error

OK does not look too good. The error states that The type or namespace name 'Helpers' does not exist in the namespace 'System.Web' To fix this we will have to do a change in the web.config. We will need to add a few assemblies in the <system.web><compilation> section. The updated web.config would look like

After this lets build and access the URL /home/razorindex and les see what happens

Second attempt at rendering the Razor view also results in compilation error, this time it cribs for ViewBag

Still not quite there, but getting closer. This time the error says that The name 'ViewBag' does not exist in the current context. The problem is that the compiled class RazorIndex_cshtml is extending System.Web.WebPages.WebPage class. As we had seen in the previous post, MVC views needs to extend from System.Web.Mvc.WebViewPage and not System.Web.WebPages.WebPage.

How do we fix this issue?

We need to some how tell the framework that the Razor pages should extend System.Web.Mvc.WebViewPage class. This is done by adding another web.config in the Views directory. The project structure would now look like this

Project Structure with two web.config's

The content of this new web.config should look like

The element <pages pageBaseType="System.Web.Mvc.WebViewPage"> tells that, the Razor views will have base class of System.Web.Mvc.WebViewPage.

Lets Rebuild the app and browse the URL /home/razorindex and see what happens.

Success! Razor view renders correctly

Bingo! We have an Asp.Net Web Forms application with MVC3 Razor views integrated in it!

PS: Now you have no reason to not move to ASP.NET MVC3 from Web Forms, we have proved that MVC3 can co-exist with a Web Forms in the same project!

Asp.Net MVC3 and Asp.Net Web Forms Side-By-Side in One Project.

2011-07-31T12:16:00.007+05:30

You must have known from my previous post that, we have moved to MVC3 instead of the Web Forms approach.

While I was suggesting the clients to adopt the new MVC3 framework, the obvious question from their side was, Can the Web Forms and MVC3 co-exist? Can the new MVC code reside in the Web Forms project?

I Googled a little, found that it was very much possible, however we needed to spike it out to be completely sure. After some trial and errors, we were able to make the Web Forms and MVC3 code co-exist in one project. The entire process was not very straight forward hence, I decided to write a post on how to do it.

Let's first create a bare bone ASP.NET Web Forms application. This can be done by using the Create New Project Wizard of the Visual Studio 2010

Create new Asp.Net Web Forms Project

The Wizard generates the web.config file, master page and two pages About.aspx and Default.aspx. Running the app at this stage should open up a page like this

Asp.Net Web Forms Default Page

Now we have an asp.net web forms application up and running. Lets try and add the MVC3 code in the existing project and lets get it running. This would effectively prove that Web Forms and MVC3 can co-exist.

Integrating MVC3 into the Web Forms application - How Do They Do It!

MVC3 is build on Convention Over Configuration. Controllers in MVC3 are put in a special directory called Controllers, the views go into the Views directory. Lets create these two directories at the root level.

Adding Directories Controllers and View

Let's add a very simple controller class called HomeController to the project. All MVC3 Controllers need to extend the System.Web.Mvc.Controller class. To do this, you will need to add the reference of the System.Web.Mvc.dll into the project.

Lets look at the code that goes into the HomeController

The HomeController has a very simple Index Action method. This Index action is simply going to render the Index view. The Action method also adds a property to the ViewBag called UserName. We will use the UserName Property in the view.

The Index view needs to be created. Let's create a file called Index.aspx in the Views/Home/ directory. The Index.aspx file looks like

Notice that, the aspx page inherits the class System.Web.Mvc.ViewPage and not System.Web.Page. All MVC3 views must extend the ViewPage class.

Also Notice, how we are showing the UserName from the ViewBag using the <%= %> Syntax. Unlike traditional ASPX pages there is no code behind for this ASPX file.

Curious thing about Index.aspx file is that, it does not inherit from the HomeController class. Then how can the view access the ViewBag property UserName which was set in the Controller?

Well, the MVC3 framework implements Separation of Concerns between the View and the Controller. But it surely needs to pass information between the Controller and View, ViewBag is just one way of doing this. Whatever we set in the ViewBag from the Controller is available in the Views! This is just one of the way in which you can pass information between Controller and View.

The project structure at this stage looks like


Project Structure after adding the HomeController and Index.aspx page

At this stage, if we could open up the URL http://<host name>:<port>/<your application name>/home/index then we can be sure that we have successfully integrated MVC3 in a Web Forms Project. Lets try and browse the URL and see what we get.

404 Error When we try to browse /home/index URL

We get a The resource cannot be found (404) error.

Why?

This is because, we have not yet informed the framework that, when someone hits the URL /home/index then how should it be served. This can be done by registering a new route in the Global.ascx.cs file.

This file has an Application_Start method. This method is called when the application is started. In this method we have to inform the framework that, when someone sends a request for the URL /home/index then, route that URL to the HomeController. This is done using the following code

The complete code of Global.ascx.cs file looks like

Lets run the application and browse the URL /home/index lets see what happens now?

Parser Error - Could not load type 'System.Web.ViewPage'

Nope, not there yet.

Turns out that for the page to get compiled and render correctly we need to add the System.Web.Mvc in the system.web/compilation/assemblies section in the web.config. The updated web.config looks like

The only change here was to add the assembly System.Web.Mvc to the list of assemblies under the compilation section. This will make sure that this assembly is available to the aspx view when its compiled.

After making this change lets run the app and browse the URL /home/index and lets see what happens.

Index.aspx is rendered Successfully using MVC3

Success! We are able to render the page using MVC3 in a Web Forms application!

Notice that, currently we are using the aspx rendering engine for rending MVC3 views, in the next post I will explain how we could take advantage of the new Razor Rendering Engine in a Web Forms application.

Showing an Image Gallery for HTML5 Mobile Web Site

2011-07-24T16:15:00.004+05:30

Last Friday, I got task that was to be done as quickly as possible.

What Was the Task?

There was a HTML5 based mobile web site that few people from my company were working on. On a certain page, we were showing thumbnail images


Page with Thumbnail images on it

If user wants to see a bigger image, they can click on one of the thumbnail images, a new page opens up which shows only the image that user had clicked.

New page is opened,
it shows only the image that was clicked

If user wants to see the next image, they have click back button on the browser and then click the other image in the thumbnail list. Certainly this was not very user friend behavior. There was no easy way to see all the images one after the other. User had to navigate between the pages to see all the images.

Hence, I was given the task to implement some sort of a image gallery which will enable users to navigate between the images easily.

The Approach

Initially, I thought that the task was simple. The possible approaches that immediately came to my mind where

Full Home Grown Solution - This meant, writing some JavaScript on my own to show Next and Previous links (on the page where we are showing the image) and then cycle through the thumbnail images.
Reuse an existing image gallery JavaScript library - This meant, Googling a little, finding what image gallery JavaScript libraries are available and using the one that fits the our requirements.

I am firm believer that, one should not reinvent the wheel! Hence, for obvious reasons, I ruled out the first approach. I narrowed down my task, to find the Image Gallery JavaScript that would suite our requirements.

Lets look at the HTML used to show the thumbnails and then the bigger images.

As you can see, the HTML used to show the thumbnails and full images is pretty standard. Thumbnail images are wrapped with an Anchor (<a>) tag with href attribute pointing to the full image. Because of this, when user's click on the thumbnail, a new page with full image URL opens up.

I was in no mood to change the HTML or use completely different approach to show the Image Gallery. I wanted to do as little as possible to get the desired results, I am a lazy developer you see!

We were using jquery on our HTML5 Mobile Website. I was thinking, there should be a ready made jquery plugin, which would show the images one after the other and I would have to do little or no work. Googling a little, I found many JQuery based Image Gallery JavaScript that looked promising. I will list down a few of my not so successful attempts at showing the Image Gallery and then finally the one worked for me.

The JQuery LightBox Plugin

I am a fan of this plugin. I had used this plugin in the past for few of my other projects. I like everything about this plugin, the ease of use, the end result etc, its completely awesome! Hence, I decided to give this plugin a go.

I opened up the JQuery LightBox examples URL on an iPhone to see how it behaves on the phone. This is what I get

LightBox examples URL opened in iPhone 3Gs

There is no doubt that it works on the iPhone. But somethings were not right.

The preview image is still very small
Next, Previous and Close links are difficult to find and click
One needs to Zoom In on the page to actually see the bigger preview image or click next, previous links.
Does not support swipe to change the image.
Clicking the Browser Back button takes the users to the Previous page in the browser history, it should have closed the full image preview dialog and stayed on the same page.

I realized that I was trying to fit one solution to all problems. Its evident that although the JQuery LightBox does an awesome job at showing the Image Gallery for web based application, I needed a JavaScript library that was build/optimized for Mobile devices.

Please note that, the issues that I just mentioned are not JQuery LightBox specific issues. There is no doubt that the authors of JQuery LightBox have done a fantastic job at developing an Image Gallery JavaScript library, that does an awesome job at showing Image Gallery on web pages build for Computer and not for mobile devices.

I decided to search for Image Gallery JavaScript libraries that were build for the Mobile devices.

Touch Gallery

Finally, I found a JQuery based Image Gallery JavaScript library that was build for mobile devices. The Usage was simple enough, all I had to do was

Add the <meta> tag for the viewport
Include the TouchGallery javascript library
Select the appropriate Anchor (<a>) elements and call touchGallery JavaScript function on them.

Updated HTML looks like

JavaScript to integrate the Touch Gallery looks like

So far so good, lets look at the rendered result

Preview image opens up in full screen
using Touch Gallery

Better? Yes, the result is definitely better then the JQuery LightBox plugin. Lets look at things that I like and dislike about this plugin.

The Good

The preview image is bigger and occupies full screen width
User can change images using a swipe

The Bad

There are no buttons to go Back and Forward. User might not know in the first go that, they could swipe to change the image.
Big black sections at the top and bottom of the preview image
There is no close link, which users can click to go back to the page which shows all the thumbnails
Clicking the Browser Back button, takes the users to the Previous page in the browser history, it should have closed the full image preview dialog and stayed on the same page.

Seems like we were able to solve a few problems but still there are a few negatives.

Now what? Write a custom solution? NO! Never do that! Obviously people must have faced this problem before and there has to be a JQuery based JavaScript library that fits my requirements, after all, its not Rocket Science!

I decided to find a better solution!

PhotoSwipe

PhotoSwipe is a FREE HTML/CSS/JavaScript based image gallery specifically targeting mobile devices. It provides visitors with a familiar and intuitive interface allowing them to interact with images on your mobile website. So far it seems to fit the bill. Lets integrate it and see the results.

Integrating it was not all that difficult, all I had to do was

Add the <meta> tag for the viewport
Add the PhotoSwipe and dependent JavaScript files.
Add the PhotoSwipe css
Select the appropriate Anchor (<a>) elements and call photoSwipe on them.

Update HTML looks like

The Script to integrate the PhotoSwipe JavaScript library looks like

Looks simple right, lets look at the results

PhotoSwipe in action
Portrait mode with navigation links

PhotoSwipe in action
Ladscape mode with navigation links

As you can see, the results are wonderful.

The Good

The preview image is bigger and occupies full screen width
User can change images using a swipe
There are navigation links using which users can go back, forward, close the dialog and even play the images as slide show
Clicking the back button of the browser show the original page with thumbnails on it!

PhotoSwipe fits the bill completely!

IMHO there is no one solution that fits in all situations, but certainly there is a solution out there in the wild, that will fit your requirements! If you don't find it in the fist go, try harder!

Asp.Net MVC vs Asp.Net Web Forms

2011-07-17T19:09:00.003+05:30

We were working on this project which was build using C#, .Net 4.0 and Asp.NET Web Forms.

I come from a Java background where MVC pattern has been around since ever. There are tons of Java based web frameworks out there that implement the MVC pattern. It looked quite surprising to me that, till very recently there was no equivalent of MVC pattern in the .Net world. But, recently the things have changed, Microsoft has come up with its own implementation of the MVC pattern called Asp.NET MVC3.

I certainly like the MVC approach over the Web Forms approach for multiple reasons, I had to present these reasons to the client and convince them to move to Asp.Net MVC3 instead of the current Asp.Net Web Forms approach. Finally, after quite some talking and presenting and discussions we managed to convenience the client to adaopt the Asp.Net MVC 3 approach over the Web Forms approach.

This blog post highlights the points why I think, MVC makes a better choice as the presentation framework over Web Forms.

Separation of Concerns (SoC) And Single Responsibility Principle

If you ask me, what are the two most important principles which helps in keeping the code clean, modular, readable and maintainable?

My Answer would be, Separation of Concerns (SoC) and Single Responsibility Principle.

Its extremely important that every class in the system has a well defined job for itself. No class should try and do too many things, in fact it should do only *one* thing and do it well (Single Responsibility Principle). If one class tries to do too many things, then the code of the class would be messy, difficult to understand and maintain.

The code behind classes of the Web Forms approach are victim of this problem. Lot of times, I have seen the code behind classes are bloated and are longer than 2000 lines. Such classes are difficult to maintain and understand. The code looks convoluted.

Why does this happen?

Because, code behind classes have lots of responsibility. They deal with

Showing the view
Dynamically rendering JavaScript
Updating the model
Validating the values entered by the user
Implementing some business logic
Making a call to the service
Differentiate the Post and Get requests etc.

End result is a truly complex and gumped up code behind.
MVC on the other hand, implements these principle very well.

Controller classes are only responsible to update the model, decided which view to render and that's about it.
Views only responsibility is to use the model to render itself.
Model just holds the business realted information.

End result of all this is clean, modular, readable and maintainable code.

MVC gives full control over the rendered HTML

In the Web Forms approach, we were using third part controls called Telerik controls. At times its really difficult to control the HTML that gets rendered from these controls. Many times the HTML that is rendered is more complex than what would be required. Because of this, designers find it difficult to style it the way they want it.

With MVC, you get full control of the exact HTML that gets rendered. This is an absolute boon for the designers and developers

Unit Testing

I don't want to talk about how difficult it is to unit-test web forms code behind classes. The web form code behind classes are tightly coupled with things like Session, ViewState Request, Response etc. This makes them extremely bad candidates for unit testing. To unit test web forms code behind classes one needs a very strong will and extremely good mocking and Unit testing skills.

On the other hand, MVC has been build with Unit Testing in mind. One can easily do Test Driven Development with MVC controllers. Controllers are not coupled with Session or Request or Response object which makes them easily unit testable.

Search Engine Optimizations and RESTful URL's

Web Forms application URL's end with .aspx extension. They represent a physical file in the file system. The MVC URL's do not have any extensions. They do not logically map to a file on the file system. Instead they map to a controller and an action method inside the controller. The URL of MVC application look like

http://localhost:8080/appname/customer/create
http://localhost:8080/appname/customer/1
http://localhost:8080/appname/customer/list
http://localhost:8080/appname/customer/edit

Integration with Third Party JavaScript libraries

Its difficult to integrate the web forms application seamlessly with a JavaScript library. For e.g. If you wanted to do Ajax with jQuery in a web forms application, imagine how much code you will have to write. How will you render the dynamic HTML in this case. How will you preserve the view state? Its all very messy.

With MVC, we do not have a problem like this ever. By default, MVC application comes integrated with jQuery. However, you could go ahead and integrate any other JavaScript library of your choice. Since view is not tightly coupled with controller integrating any JavaScript library is extremely easy.

No IsPostBack

I think there is no Asp.Net application in the world that does not have code which looks like

One class/method to handle both the get and post requests. At every step we have to decide whether its a get request or a post request and change the business processing.

This is infrastructure code. We should not be bothering about checking whether the request is a get request or a post request. Framework should do this for me.

With MVC, we do not have to keep checking whether its a get request or a post request. We can simply put an attribute to a method marking it only accessible via the GET request or only accessible via the POST requests. With MVC the above code would look like

Notice that we simply have to mark which methods are invokable over which protocol using either the HttpGet or HttpPost attributes. End result, simplified and easy to understand code.

No View State

Web Forms provides a wrapper over the stateless HTTP protocol to make it state-full. It achieves this by adding a hidden field called __VIEWSTATE__ to the rendered page. If not handled correctly and especially when you are using some third party control library like Telerik controls ViewState can easily get out of hands. I have see pages that have ViewState that goes in MB's. This ViewState will be sent/received to/from the server for every Post requests. Its sent/received to/from the server even when you do a Ajax request. The performance of such pages goes for a toss!

MVC has no concept of ViewState. MVC does not add a wrapper over the stateless HTTP Protocol. In fact it harnesses the HTTP protocol. Hence, MVC pages are much faster and they can easily out perform the Web Forms counter parts on any day.

These are some of the points why I think, MVC is definitely a better choice over Web Forms for building Customer Facing, fast Web 2.0 applications.

This list is by not means an extensive comparision of Web Forms vs the MVC appraoches. Just some little peaces which I though were really important.

We have made our choice to move to MVC, have you?

Lohagad - Nature at its best - Rejuvenate yourself

2011-07-09T21:19:00.001+05:30

It all started with me and my friend Mahendra, talking about how the weather in Pune is perfect for a trek. We have been talking about doing a trek since one year now, but because of some or the other reason, never managed to actually do it. But this time, I really really wanted to do it! After some initially talks, we decided to do a trek to Lohagad.

I asked around to see if others were interested. Total 12 people agreed for the trek. We decided to start at 0630 hrs on Saturday and come back by afternoon.

Staring from 2000 hrs on Friday evening, I started getting SMS'es from people saying they cannot make it. By morning, I was thinking it will be just me and my wife who will be going for the trek :). But in the end, we were 5 adults and 4 kids (all of them 8 years old) made it!

As planned we started off at about 0630 hrs. I had been to Lohagad in the past, at that time, there were no shops at the top, i.e. no food and no water at the top. Hence, we wanted to have heavy breakfast before starting. We tried a few good restaurants but none of them were open that early. Finally, we settled for a dhaba type place to have breakfast. Although the ambiance of the place was not all that great, the food was good! We had Misal, Poha and Tea. After filling our tanks, it was time to hit the road again.

We reach the base village Malawali at about 0830 hrs, it was a good surprice to see a very well organized pay and park facility at Malawli. We parked our vehciles there and started the climb.

The Way Up:

Right at the beginning of our ascent we saw two waterfalls.

As we climbed up, breathtaking views would unfold in front of our eyes. The natural beauty on the way was simply amazing.

A few images to illustrate what I mean

At one point right over our head we saw a big rock which looked like a gigantic ship. It would immediately remind you of the famous scene from "Titanic" where Leonardo and Kate are standing at the tip of the titanic to feeling the breeze.

The snaps does not do justice to the view that we saw!

For the first time in my life I saw blue beetles

After walking some more, we saw another natural miracle. We saw a big rock hanging by the side of the cliff. The rock and cliff would be attached only at the ends. The space between the center of the rock and the cliff formed a shape which looked similar to Shiva's third eye! Its difficult to explain, a picture would clarify things.

Finally, after about 1.5 hours of walking and enjoying the natural marvels, we reached the base of the Fort. There we were greeted by our ancestors

There are lots of monkeys in and around the Lohagad fort. They were waiting to get one opportunity to snatch anything that looked like food.

There are proper steps created to go from the base of the fort to the fort itself.

During the monsoons, small stream of water keeps flowing down the steps. On our way into the fort, we saw a small lake formed by the back waters of the Pawna dam.

On the way up, we saw ruins of few small cannons. We tried to lift them just to see how heavy they were, but they wont move even a inch!

A few more steps and we saw a gigantic Maratha gate! It had massive chains and sharp spikes all over it. Needless to say that one person cannot open or close the gate. The spikes made sure that, it would be difficult even for a fully grown elephant, to break open it.

Last few steps, took us into the fort. No one wanted to walk any further without having a bite. We decided to stop there and have something to eat. We had carried lots to eat, Sandwiches, Chips, Sabzi Roti etc. When we were half way through our lunch, one monkey located us and came extremely close to us, looking for an opportunity to snatch something. We reacted quickly, we put everything back in the bags and started walking. 5 adults and 4 kids against 1 monkey, believe me, monkey was not at all afraid of any of us. In fact he was actually trying to scare us. We had to finally give him a peace of sandwich.

We decided to move ahead since we could not continue our lunch at the same place. After walking for 5 min we saw a Dargah and a small pond, we decided to finish our lunch there.

After lunch, it was time to sit and relax and do nothing absolutely nothing! Cool breeze was blowing continuously.

I was at "peace" with myself!

I was not thinking about anything, nothing absolutely nothing at all! Not thinking about the city roads, traffic jams, home, projects to be delivered, meeting the deadlines, nothing!

My 10 gallon head was completely empty. Simply, sitting and enjoying the natural beauty. I have to admit, I had not felt like this for quite some time now. I am not a person who would keep thinking and worry about stuff, but still the feeling that I got at that point was beyond anything that I had experienced before.

In the movie The Pursuit of Happyness, right before he gets the job as a Broker, Chris Gardner says,

The next day, after work, we just went to the beach. Far away from anything, everything, just Christopher and me. Far away from buses, and noise, and my constant disappointment in my ten gallon head, in myself

At that moment, I was able to understand what he was talking about. What is that feeling and I can tell you one thing, it surely feels great! That moment for me was the best moment in the entire trek. It was totally worth it!

After relaxing for about an hour or so, we decided to explore the, West side of Lohagad fort to see The Vinchukata(Scorpion Tail).

On the way to Vinchukata one has to climb down an almost straight wall made of rocks.

It was pretty steep, but in the end I did make it. From the tip of the fort you could see the entire landscape. There was no obstruction. On one side you could see magnificent Sahyadri range and on the other side you could see the Pune-Mumbai express way. Clouds were passing right through us. Its was simply beautiful!

Later, at about 1400 hrs, we decided to start our descent. While coming down one of our friend, twisted her ankle. Luckily, it was not that bad. After some rest she was able to walk normally. We reached the base village at about 1545 hrs.

It was trek a wonderful trek with breathtaking views on the way and at the top. But the most important USP for me were, those moments at the top when I was at peace with myself!

I would strongly recommend everyone to do treks on a regular basis. Oh! almost forgot, all 4 kids did manage to do the entire trek by themselves. No one had to carry them, in fact the kids carried their bags all by themselves!

If 8 year old kids can do the trek without any problems, so can you! Go for it, I can assure you, you will return completely rejuvenated!

PS: To view all pictures visit this link

How to show turn by turn directions on Android

2011-06-25T13:12:00.001+05:30

As you can probably make out, these days I am working on an Android project. And I will tell you one thing, I am loving it!
I have worked on iOS, BREW and Simbian OS in the past but the productivity, while developing on Android is far better than the other platforms!

Excellent job Android guys!

Coming back to the post topic, recently I in one of the Android projects, we had a requirement that we wanted to show turn by turn directions between two geo points.

We are all familier with the excellent support for Turn by turn directions on the native Google Maps application.

Although, Android API permits us embed a maps view in the application, but it does not provide a way to show turn by turn directions between two geo points.

In early days of Android (SDK 1.0), Google had provided the DrivingDirections class. Using this class it was possible to get the driving directions between two geo points. However, since SDK 1.1, Google removed the DrivingDirections class from the API.

Why would they remove something as useful as this?

Because it violated some of the legal agreements made by Google. Extract from the Google Maps terms states that

8.3 Content License. Subject to these Terms (including but not limited to Section 9 (License Requirements) and Section 10 (License Restrictions)), Google gives you a personal, worldwide, royalty-free, non-transferable, non-assignable, and non-exclusive license to access, use, publicly perform and publicly display the Content in your Maps API Implementation, as the Content is provided in the Service, and in the manner permitted by the Terms. Specifically, you understand the following:

(a) Content (including but not limited to map data, traffic, directions, and places) is provided for planning purposes only. You may find that weather conditions, construction projects, closures, or other events may cause road conditions or directions to differ from the results depicted in the Content. You should exercise judgment in your use of the Content.
(b) Certain Content is provided under license from third parties, including Tele Atlas B.V. ("Tele Atlas"), and is subject to copyright and other intellectual property rights owned by or licensed to Tele Atlas and/or such third parties. You may be held liable for any unauthorized copying or disclosure of this content. Your use of Tele Atlas map data and certain other Content (including certain business listings Content) is subject to additional restrictions located in the Legal Notices page.

In simple terms, Google has tied up with many Third party companies to get the mapping and other information. The terms of the agreement do not allow them to expose those functionality in an open source project. This is the reason why they removed the DrivingDirections API since SDK 1.1

All right fine! Google could not expose that as API's, but we sill have this requirement to fulfill. We want to show turn by turn directions between two geo points on Android platform.

How do we do it?

As we all know that the native Google Maps application (which comes pre-installed on all Android devices) does a great job of showing Turn by Turn driving directions. The updated application also provides a link to Google Navigate (beta) which shows 3D maps and updates the route as the user location changes.

We should definitely try and make use of this feature for our requirement. After all, why reinvent the wheel!

Hence now, our tasks are:

To invoke the native google maps application
Show turn by turn directions between two geo points on the native google maps application

One very big plus of the Android platform is that, it has multitasking support. Because of this, when user clicks the back button on the phone, they will come back to our application, user does not feel they are navigating away from our application. This enhances the user experience quite a lot!

So the questions is how do we invoke the native Google Maps application to show turn by turn direction between two geo points!

Inter process/activity communication on Android platform is done via an Intent. Hence the answer to our question is: Via an Intent!

Yes and that's the correct answer! But what Intent, what will be the URL, what should be the parameters?

Slow down, slow down, lets tackle one problem at a time!

This is how they do it!

Looking at this URL it seems that we can invoke the native maps application using the following URLs

One line description for the above url's is

Opens the Maps application to the given location or query. The Geo URI scheme (not fully supported) is currently under development.
The z field specifies the zoom level. A zoom level of 1 shows the whole Earth, centered at the given lat,lng. A zoom level of 2 shows a quarter of the Earth, and so on. The highest zoom level is 23. A larger zoom level will be clamped to 23.

Basically, the description says that, The above URLs will show the maps application which is centered around the given geo point

Hence,

This code will actuall open up the native map application which is centered around the given geo point. We have solved one task from our task list. We have managed to invoke the native google maps application!

But wait! We want to show two points not just one point on the map. We also want to show turn by turn directions between those points!

Using the URL i.e. the geo:x:y in the above format, we cannot specify two geo points. Damn!

I tried many permutations & combinations to specify the second geo point, but nothing worked.

Finally, after a lot of trials and errors I found a workaround to this problem.

This is how its actually done!

The URL geo:x,y does not support the second geo point, hence we can't use that URL. What is the other URL we could use, which will show turn by turn directions between two geo points?

After some more googling, I found that, we might be able to use the URL format used by maps.google.com website. The maps.google.com follows a simple url pattern

The maps.google.com uses two parameters "saddr" (source address) and "daddr" (destination address). If we pass those two parameters correctly it shows, turn by turn directions between those points.

Web Based Turn-By-Turn Directions on Google Maps

Why are we even bothered about viewing turn by turn directions on the web?

Because guess what, we can use the same web based URL pattern on the Android platform!

What? Yes, we can use the URL http://maps.google.com/maps?saddr=43.0054446,-87.9678884&daddr=42.9257104,-88.0508355 to open up the Google Maps application to see turn by turn directions between two geo points!

The above code surely works! It shows the user, turn by turn directions between the geo points [43.0054446,-87.9678884] and [42.9257104,-88.0508355].

There is only one problem though, Android platform identifies that the above URL can be processed by two Activities!

The Browser Activity - Since the URL starts with http://
The Google Maps Activity - Since the address we are opening up is maps.google.com

Whenever Android platform has more than one options to process an intent, it leaves the choice to users, gives them an option of both the applications and lets them chose the application to fulfill the intent.

User gets an option to chose the application for the processing the Maps URL

User can chose either application, both show driving directions between those two geo points.

Driving directions using the browser application

Driving directions using the Native Google Maps application

How to make this solution better:

Although we have achieved what we intended for. But there is one extra step that user has to perform, i.e. select the application of his choice to see the driving directions. At this point, user might notice that he is going out of the current application and into a different application. There is nothing wrong with this approach but, it does not integrate the Google Maps application seamlessly with our application.

How do we eliminate this extra step of choosing the application to view turn by turn directions?

The answer is simple and we can do this by providing a hint in the Intent we are firing.

The line intent.setClassName("com.google.android.apps.maps", "com.google.android.maps.MapsActivity"); does the trick. We are passing additional information in the Intent. We are informing that, the class name of the activity that should fulfill the intent is "com.google.android.apps.maps/com.google.android.maps.MapsActivity"

That's it! Now Android platform knows that, user wants to see the turn by turn directions on the native google maps application and not on the browser. The extra step of chosing the application is eliminated.

User directly goes to this view

There are many advantage of this approach

User can choose to start navigation using Google Navigate (beta) (without us doing anything).
User can click back button and return to our application as well!

As we can see, Android platform is pretty awesome, when it comes to ease of development! Something as complex as showing turn by turn driving directions between two geo points can be achieved by writing 3 lines of code!

How to change the text color of Android options menu

2011-06-22T09:00:00.005+05:30

As promised in the previous post, here comes the second installment of styling the options menu. In the previous post we saw how we could change the background color of the options menu. This time we are going to look at how we could change the text color of Android options menu.

We already know that styling the options menu is not really easy. As shown in the previous post, its a eight step process. To change the text color of the options menu we need to put in another hack. We need to add a ninth step. Lets look at all nine steps once again.

Setting the custom LayoutInflater.Factory instance.
Implementing the onCreateView method that will actually set the background selector for the options menu.
In the implementation of onCreateView, check if the view that is to be created is of type com.android.internal.view.menu.IconMenuItemView. Yes, I know, its an *internal* class!
If the view is of type IconMenuItemView then create the view using the LayoutInflater.
At this point we have an instance of IconMenuItemView but we cannot change the background color of the view directly. Even if we change the background color at this point, framework updates it again and our changes are overwritten
We have to change the background color after the view has rendered, using the Handler API.
Change the background color in run method of the Runnable instance that we pass to the Handler's post method.
Change the text color using reflection. This is the new step. We need to change the text color using reflection.
Catch all exceptions that could occur in the entire process. Notice that, we are using an internal class there are high changes that its behavior could change without any notice. In fact this approach does not work on Android 2.3. In this case we have to fallback to the default menu styling.

Lets look at code to make things clear

The menu_selector.xml and colors.xml have been kept unchanged from the previous post. The comments are embedded in the code. One new methods has been added setTextColor. This method does the job for us!

The above screenshot shows the menu item color changes when its pressed.

That's it! We have successfully changed the background color and text color of Android options menu.

How to change the background color of Android options menu

2011-06-18T17:21:00.001+05:30

I have been trying my hand at developing apps with Android. I have had my fair share of leanings. Today I am going to share my experience around styling the options menu in Android.

What is the options menu?

Options menu is the menu that is shown to the user when the Menu key is pressed on the Android phone.

Lets look at an example, that will make things easier to understand.

The layout file.

The above code shows the following GUI

As you can see from the above screenshot the background color of the Activity is greenish while the default Color of the selected element of the options menu is Orange.

It definitely looks inconsistent and ugly. This is the reason we need to style the options menu to match the rest of the application theme.

Android platform is highly configurable. One can change all aspects of look and feel on the android platform. Hence, when this task was given to me, I thought it would be a 2 min fix. But unfortunately it took much longer than I expected.

All I wanted to do is change the background selector of the options menu. Android does not provide a straight forward way to do this. In fact, it meant writing quiet some code. I decided to write a post on this, so that other can benefit from my learning's.

So back to the original question. How do we change the background color of Android options menu?

How do they do it?

Now there is nothing fancy about the above code. But to change the background color we have to write some fancy code :)

Changing the options menu background color is a 8 step process

Setting the custom LayoutInflater.Factory instance.
Implementing the onCreateView method that will actually set the background selector for the options menu.
In the implementation of onCreateView, check if the view that is to be created is of type com.android.internal.view.menu.IconMenuItemView. Yes, I know, its an *internal* class!
If the view is of type IconMenuItemView then create the view using the LayoutInflater.
At this point we have an instance of IconMenuItemView but we cannot change the background color of the view directly. Even if we change the background color at this point, framework updates it again and our changes are overwritten
We have to change the background color after the view has rendered, using the Handler API.
Change the background color in run method of the Runnable instance that we pass to the Handler's post method.
Catch all exceptions that could occur in the entire process. Notice that, we are using an internal class there are high changes that its behavior could change without any notice. In fact this approach does not work on Android 2.3. In this case we have to fallback to the default menu styling.

Yep, I know, it looks like a lot of work and pretty complex too. Lets look at the code to make things more clear.

The menu_selector is an XML file that choses the background color based on the state of the menu item. My menu_selector.xml looks like this

The colors.xml file looks like

The comments are embedded in the code. One new methods has been added called setMenuBackground. This method does the job for us!

The screenshots of the styled options menu

As we can see the Options Menu background color has been changed from Orange to greenish.

That's all folks!

We have successfully changed the background color of Android options men. But wait a minute, how to change the default text color of the Android options menu? If this question is bothering you stay tuned, in the next post I am going to explain exactly that!

Visual Studio 2010 Fixing "The referenced project XX does not exist" Error

2011-06-11T14:46:00.001+05:30

Recently one of my friend was facing a weird issue while building our application. The application was developed in C#. We were using MSBuild as the build tool.

What was the issue?

The issue was simple, she was not able to compile the application.

The issue was special in many ways because, all of us (i.e. her team members) were able to compile the project just fine. Even the continuous integration service - Jenkins (formally Hudson), was able to compile the application. We were not able to compile the application only on her machine. The build failed with the error

The project "AB.dosomeprojects.Services.FillImporter.Service.csproj" referred the project "abcdefgh.dosomeprojects.Common.csproj" which was in the path "..\..\..\abcdefgh.dosomeprojects\abcdefgh.dosomeprojects.Common". The path was relative to the Service project directory.

The directory structure was correct, Common project was correctly reffered from the Service project, but still we were not able to build the app on her machine.

Initially, we thought that there must me some local changes on her machine because of which, she was not able to run the build on her machine. Build was not passing neither with Visual Studio 2010 nor with MSBuild and that too only on her machine.

We looked for any local changes on her machine, but unfortunately her machine was clean. There were no local changes. We were simply clueless! Why does the project build fine on other developers machine while it does not build on her machine?

After some intense Googling, we found an interesting hidden treasure shipped silently by Microsoft Visual Studio 2010. I call such issues as hidden treasure because, I can hardly believe my ears when I hear/read about such an issue.

I am sure it will come as an surprise to you as well. The issue is:

The Real Problem:

Visual Studio 2010 fails to build the project when the following conditions are met:

Solution with multiple projects exists, where there exists dependency relationships among them
The sum of the following two path length is exactly added up to 259 characters (= MAX_PATH – 1)

The path of a referencing project’s directory.
The relative path to a referenced project from the current directory (= a referencing project’s directory).

Hmm, cryptic information lets decrypt it.

Condition 1: We have solution with multiple projects where there exists dependency relationships among them

Yes, we do have the solution with multiple projects in it and one project depends on the other. In our case the AB.dosomeprojects.Services.FillImporter.Service.csproj depends on the abcdefgh.dosomeprojects.Common.csproj project. Hence, condition 1 is met.

Condition 2: The sum of the following two path length is exactly added up to 259 characters (= MAX_PATH – 1)

This one is tricky. Which path sum are we talking about?

Condition 2.1: The path of a referencing project’s directory.

The path of the referencing project's directory, in our case the path of the AB.dosomeprojects.Services.FillImporter.Service.csproj on the file system, which was, "E:\projects\ABC\SVN\dosomeprojectsweb\trunk\AB.dosomeprojects.Services\AB.dosomeprojects.Services.FillImporter\AB.dosomeprojects.Services.FillImporter.Service" has total length of 158 chars.

Condition 2.2: The relative path to a referenced project from the current directory (= a referencing project’s directory).

The relative path of the referenced project from the current directory, in our case its the relative path of abcdefgh.dosomeprojects.Common.csproj, which is, "..\..\..\abcdefgh.dosomeprojects\abcdefgh.dosomeprojects.Common\abcdefgh.dosomeprojects.Common.csproj" has total length of 101 chars.

The sum total of these two path lengths: 158 + 101 = 259!

Wow! condition 2 is also met.

As I said, we have uncovered an hidden treasure of Visual Studio 2010. Its unable to build the project when the path length of referencing project + path length of relative path to a referenced project = 259!

If the sum total is one less or one greater than 259 it all works fine! On other developers machine the total path length was less than or greater than 259 but was never equal to 259 and because of this reason we were all able to build the project but she was not able to build the project on her machine! Weird isn't?

This issue occurs due to a bug in the Path.GetFullPath in .NET Framework library. This is a known issue in Visual Studio 2010.

The Fix:

The workaround is pretty straight forward, rename one of the folders a bit longer or a bit shorter and make sure that, the sum of the two path length is NOT equal to 259 chars.

For more information on the issue visit the offical Microsoft Knowledge base article here http://support.microsoft.com/kb/2516078

Now you know, we are having so much fun with .Net (sarcasm intended)!

How to build the web project on the CI server without Visual Studio 2010

2011-05-28T11:50:00.003+05:30

On one of the ASP.Net project, we were using Visual Studio 2010 IDE and MSBuild, to compile and build the application. Pretty standard setup for a web project right?. On developer machine all worked very smoothly, since each developer had Visual Studio 2010 installed locally on their machines.

We setup a continuous integration (CI) server Jenkins (formally Hudson) for this project. Naturally, we didn't wanted to install Visual Studio 2010 on the CI server. But we still wanted to build our app on the CI server. Hence, we installed Microsoft Windows SDK 7.1 and all its components on the CI server. After installing the SDK, we were able to build all projects except the web project on the CI server.

The error that we were getting was

Obviously, since the Visual Studio 2010 was not installed on the CI server, the path mentioned in the error didn't exist and that's why the web project was not able to build.

I opened the csproj file in Notepad++, to find the reference of "Microsoft.WebApplication.targets". I found a reference of "Microsoft.WebApplication.targets" in the csproj which looked liked

Indeed the web csproj file had an import statement, which added the reference to "Microsoft.WebApplication.targets". But I was not sure why this reference was needed in the first place.

My first attempt at fixing the CI server:

I removed the reference of "Microsoft.WebApplication.targets" from the web csproj file and tried building the application using MSBuild. To my surprise, the app build fine, it ran all the tests without any issue. My confidence grew a little, I thought I had the solution. I thought the reference was added to the csproj by Visual Studio 2010, but was never used.

I committed the fix, build went green on the CI Server as well!

After a few hours, some developers complained that, when they now open up the web project, Visual Studio 2010 prompted for converting the project into a Visual Studio 2010. Certainly, this behaviour was triggered after I removed the import statement from the web csproj file.

Its pretty clear now, why the import was required. Visual Studio 2010 identifies that, the project is a Visual Studio 2010 Web project using the import of "Microsoft.WebApplication.targets". Hence, although the build run's fine using MSbuild on developer machine as well as on CI server this is not the right fix!

My second attempt at fixing the CI server:

Googling a little I found that, there are only two options to fix this particular issue.

Install Visual Studio 2010 on the CI server
Copy the "Microsoft.WebApplication.targets" file manually from any developer machine to the CI server

It was sad to see that there was no standalone installer that installed the "Microsoft.WebApplication.targets" at the desired location. We had to either install the full Visual Studio 2010 or copy the file manually.

I decided to chose the lesser evil of the two. We copied the file "Microsoft.WebApplication.targets" from my machine (located at path C:\Program Files (x86)\MSBuild\Microsoft\VisualStudio\v10.0\WebApplications on a Windows 7 64 Bit installation) onto the CI server (at the path C:\Program Files\MSBuild\Microsoft\VisualStudio\v10.0\WebApplications. CI server has Windows 2003 OS installed on it.) and then ran the build. Yep! It worked! The build went green on the CI server.

I know, its just an hack, but for now I had no other option!

iOS: Fixing the WiFi connection closed after 30 minutes

2011-05-21T11:35:00.000+05:30

One nice Saturday afternoon when I was doing almost nothing, I got a phone call. The number was big and didn't look like it was from India. I picked up the call and said "Hello...", a familiar voice replied back. The voice was of one of my good friend. This friend of mine had gone to Germany for a month, his iOS project was going live. He had called me that afternoon to discuss about a problem that he was facing in his app.

The Problem

The application was to be deployed on iPad. Application used to poll the server for updates every 10 seconds. For the application to work correctly, it was necessary that the polling should work flawlessly.

This polling worked well, however, he was facing one ugly little problem. The issue was, after about 30 minutes, the polling would stop and because of this the application would not function correctly after 30 minutes. This would happen only if the iPad is not connected to the charger or a computer. If the iPad was connected to the charger or a computer polling would function correctly for any duration.

The workaround

To get the polling to work again user has to do one of the following

Connect the iPad to a power source i.e. a charger or a computer
Pressing the home button on the iPad and reopen the application

The Fix

I had not faced this issue earlier. In fact, I have not done much iPhone development (at least non yet), but I decided to give it a shot anyways. I decided to google about this issue and find a fix.

After some intense googling, I found that, iPad switches off the WiFi connection automatically after 30 minutes to preserve battery.

That's a good strategy alright but, what about the apps like my friends app? The app need to poll the server every 10 seconds to function correctly. How can we stop the iPad from switching off the WiFi connection automatically after 30 minutes?

The fix was pretty simple, One needs to set a property in the property list file. The property is called UIRequiresPersistentWiFi.

Quoting from the Apple developer documentation

UIRequiresPersistentWiFi (Boolean - iOS) specifies whether the application requires a Wi-Fi connection. iOS maintains the active Wi-Fi connection open while the application is running.

That was it! I knew that, I have found a fix to my friends problem. Asked him to add the property to the plist file and do a test again. As expected the problem was solved. Now, the app was able to poll the server for any duration without a problem even when the iPad was not connected to the power source!

Simple things like this waste a lot of developer time, but once you learn something the hard way, you will never forget it!

Fixing the "skia decoder->decode returned false" error while downloading images on Android

2011-05-14T15:39:00.004+05:30

I had to implement a common requirement in one of my Android project. I had to show a image that was hosted on the web, in my Android application. Now that is a very common situation right? I am sure you must have faced this situation before.

I would have imagined, showing the images from the web would be as simple as setting the correct image source URL to the ImageView. But as it turns out life is not always that simple.

So how do we show the images from the web using the ImageView?

We have to first download the image using the HttpClient
Decode the image data using BitmapFactory and convert it into Bitmap
Associate the Decoded Bitmap with the ImageView

Lets look at the code to show the web image in the ImageView

The above code works and works well, it does show the image hosted on the web in the ImageView. However, sometimes, I get an error in the log saying

Now, this is something unexpected. Whenever I get this error the image download stops and image is not shown in the ImageView.
The Cause:

Googled a little and as usual, I was not the only one to face this issue. There are lots of threads online, that mention this particular issue (here and here), "decoder returned false".

It seems that when skip method of the InputStream class does not skip the given number of bytes then, the BitmapFactory fails to decode the InputStream. Usually, the issue occurs on a slow internet connection. However, I cant confirm that, I have a 2Mbps dedicated internet connection, even on this connection issue is reproduicable!

The Fix:

There are two ways of fixing this issue.

Method 1 - Method that involves writing some custom code

This method makes sure that, the skip method of InputStream actually skip's the mentioned number of bytes.

To do this, we will need to wrap the InputStream into a custom class, override the skip method and ensure that we skip the given number of bytes no matter what. This can be achieved by subclassing the FilterInputStream class and overriding the skip method. Lets look at how the new implementation of the skip method.

As seen in the above code, the skip method simply makes sure that we skip the give number of bytes. The class FlushedInputStream now needs to be used while we decode the image. The updated code with the FilteredInputStream looks like

Method 2 - Method that involves using the BufferedInputStream

This method is very simple and straight forward we simply need to wrap the imageContentInputStream with the BufferedInputStream. This also seems to fix the problem. The updated code looks like

Yep, these two methods fix the "decode returned false" issue.

Is your Bank Account safe with ICICI bank?

2011-05-07T13:44:00.004+05:30

ICICI Bank is one of the biggest bank in India. Its renowned for having most number of ATM (after SBI) in the country. I always though that, ICICI Bank has a good security model to prevent online thefts.

No, this post is not an advertisement of ICICI bank, I wanted to share a weird experience that I had with ICICI Bank. This experience will expose the security model of ICICI Bank. Over and above it, I wanted to help them correct their mistake, but they didn't do anything as yet.

So what was the incident, how did I try to help and what was their response?

The Incident

One fine Monday morning, I got 4 SMS'es from ICICI Bank, informing me that for a certain account XXXXXXX12345

Address has been updated
ATM pin number has changed to 1234
Land line number has been changed to 020-30212345
Mobile number has been changed to 9912345678

When I got those SMS'es, I was a little surprised. I was holding an account with ICICI Bank but I never requested them to change all this information for my account. I had not called them for any such requests, then, who call them for changing all this information? Was my Bank Account hacked?

After some initial anxiety, I calmed down and rechecked those SMS'es. When I reread the messages, this time a little carefully, I realized that, my bank account number does not end with "12345". Analyzing further I realized, I do not have a land line whose number is 020-30212345. Surprising thing was, the mobile number was indeed correct. 9912345678 was indeed my mobile number. Rest of the details (except the mobile number) were no way related to me.

The Response

By now it was clear to me that there has been a mistake. I decided to call up the ICICI Bank call center to find out what was wrong. I cleared the IVR authentication, got my call transferred to the Customer Service Executive. First, I wanted to confirm that my Bank Account was not hacked. I asked them if there were any requests made for changing the address, phone number, atm pin of my account. Thankfully, there were no such requests made for my account.

Now the question that was bothering me was, who holds the account that ends with "12345" and why was it associated with my mobile number? I asked the Customer Service Executive about which accounts are associated with my mobile number?

She said, there was indeed one account associated with my mobile number that ends with "12345". Bingo!

Being a decent man, I told her that, the account that ends with "12345" is not my account. There seems to be a mistake, because of which my mobile number was associated with incorrect account. She told me, she can see that the account is held by some "Mr. Pathak". She also told me that the account was newly opened in the "Bhandarkar Road Branch - Pune". I requested her to kindly delink my mobile number from his account. She agreed and said this will be done in 2 days. I was relieved.

By now I was thinking, this is such a big problem. If someone gave an incorrect mobile number, or the bank's customer service executive made a mistake noting down the mobile number, All vital information will be SMS'ed to the incorrect mobile number. I knew the other guys ATM pin number, I knew his account number, I knew his address, his land line number. If this information falls in wrong hands, anything is possible.

By now you must be thinking that the issue must have been resolved. But unfortunately, the issue does not end here. After about 5 days I got an SMS informing that 25000 Rs. were credited into the account ending with "12345"

My first reaction to this was, WTF!

I called up the ICICI Bank call center again, to find out why my mobile number was not delinked. This time they informed me that they cannot delink my mobile number from "Mr. Pathak's" account on my request. I need to prove that, the mobile number under question is actually my mobile number. I asked them, what should I do to prove it. They said visit any nearest ICICI Bank branch with and ID proof and the post paid Bill of my mobile number.

My reply to that was, I have no motivation to go to the bank to prove this. Its a clear case of human error, they should have delinked my mobile number from Mr. Pathak's bank account.

I wanted to help them rectify their mistake, but they will not let me do so. Why the hell should I waste my time? I definitely have better things to do!

The conversation ended there. I was frustrated, why do people stop using their brains? Why are processes so rigorously implemented that people stop using common sense?

A few days passed, I got two more SMS informing that 25000 Rs were withdrawn from the account ending with "12345" and the account balance was 0.0 Rs. The other message informed that, the balance of bank account ending with "12345" had fallen below the allowed limit of 10000 Rs.

This was the limit, I decide to actually visit the ICICI Bank branch and once and for all get my mobile number delinked from Mr Pathak's bank account. I carried a photo identification proof (my PAN Card) and the post paid bill of my mobile number. I went to the "Shivaji Nagar - Pune Branch"

The response that I got from the branch was unbelievable. The lady at the bank said she cannot delink Mr. Pathak's account from my mobile number. I showed the bank lady, my pan card, even the post paid mobile bill for my mobile number with my name written all over it!

Unfortunately nothing helped, she said she cannot do it. She said it was against the processes of ICICI Bank!

I urged her that, there has to be a way to do it. Inform Mr. Pathak about the mistake, get him to rectify this problem. She assured me that something will be done. But till date nothing has been done. Till now I get SMS'es about Mr. Pathak's bank details on my mobile number!

How did I try to help

Called up the call center twice to inform them about the mistake
Went to the bank and proved that the mobile number indeed belonged to me but still they didn't delink it from Mr. Pathak's bank account.
Wasted 5 hrs of my life fixing a problem that didn't impact me much.

What should we learn from this incident

Software is meant to reduce human effort. Processes are meant to reduce human mistakes. This incident is a perfect example of what happens when people follow processes and use software without using their brains!

PS: All events and conversations mentioned in the post are real. However, all Names and Numbers have been changed to preserve the privacy of a certain person.

How to fix the “circular file references are not allowed” Error in ASP.Net

2011-04-30T21:27:00.004+05:30

On one of my .Net projects, I came across an interesting issue. I wasted couple of hours fixing it. Hence, I decided to share my findings, so that others do not have to waste their time fixing the same issue.

The issue

The issue was pretty simple, the app would not build. The error that I was getting while building a ASP.Net web project using MSBuild was:

Of course the error said there is some sort of circular reference in my code. I looked around to check and recheck, if I have created a circular reference by mistake. However, if there was any circular reference, the code would not compile. Code was compiling fine but it was failing when we ran aspnet_compiler.exe.

The ASP.Net Compilation tool (aspnet_compiler.exe) enables you to compile an ASP.Net Web application, this helps application performance because end users do not encounter a delay on the first request to the application.

I checked again, but certainly there was no code related circular dependency then why the aspnet_compiler.exe was complaining about the circular file references?

The Explanation

Googling a little, I found that, by default, in a Website Project, ASP.Net creates one DLL per folder. Hence, if you have the following setup:

User control ucA.ascx is present in the directory "A". ucA.ascx refers another user control ucB.ascx
User control ucB.ascx is present in the directory "B". ucB.ascx refers another user control ucC.ascx
User control ucC.ascx is present in the directory "A".

The folder A's DLL will reference the folder B's DLL, which will again reference the folder A's DLL, causing a "circular file reference".

This is the reason why the aspnet_compiler.exe fails with the "circular file reference" error.

The Fix

There are two ways this issue could be fixed

Rearrange the user controls (or MasterPages) to remove the circular references. Usually this means moving the user controls in separate directories. In our example, moving ucC.ascx to a new directory "C" (Preferred Solution).
Use batch=”false” in the compilation tag of the web.config file. This will cause a new DLL to be created for each control/page in the site. This should fix the error but is really lousy for performance, so it should be avoided.

I moved the ucC.ascx in a different directory and yes the error went away!

Fixing the java.lang.ClassNotFoundException com.test.model.SomeModelClass in loader dalvik.system.PathClassLoader on Android

2011-04-24T12:15:00.001+05:30

Recently, I had faced a very interesting issue on one of my Android project. The issue was the dreadful java.lang.ClassNotFoundException.

The Android App, that was facing the ClassNotFoundException was tested on multiple devices. Some of the devices that we tested the app on were

The app worked fine on all devices without any problems at all. We decided to go live with the App on Android Market.

Everything looked great till one day when one of the QA's downloaded the app on HTC MyTouch 4G. The app refused to work on this device. The app threw ClassNotFoundExceptions. The stack trace looked like

After looking at the stack trace, I was thinking, Hmm this is weird, the app works on all these devices without an issue. Moreover, the app cannot compile if the class com.test.model.SomeModelClass was not found. Then why it doesn't work on HTC MyTouch 4G?

My initial reaction was, may be the app was not downloaded correctly from the Market. May be the downloaded APK was corrupt?

I requested the QA's to uninstall the app and re-install it again from the market. I was hoping that this could fix the issue.

But software development has taught me one thing

Nothing changes, If Nothing changes

Yep, the reinstall didnt work. The app was still throwing the ClassNotFoundException.

We were using ProGuard. Proguard is a free Java class file shrinker, optimizer, obfuscator, and preverifier. Another thought that crossed my mind was, Is ProGuard removing the class com.test.model.SomeModelClass?

I created the APK without running ProGuard on it. Requested the QA to test it.

That didn't work either. ProGuard was not at fault.

By now, I was really confused. What could be the issue with the app? I thought, may be, if I try to load the class explicitly as the first thing in the app, then would it fix the issue?

I updated the APK, loaded the class explicitly when the app starts, requested the QA to retest the APK.

Nope, that didn't fix the issue. Interesting thing that I noticed about about this change was, the class loads successfully when I load it explicitly, but the class is not found when the app really needs it at a later stage!

History has shown us, class loader issues are not easy to fix.

I was quickly running out of options. People were downloading the app on their device from the Market and may be some of them were facing the ClassNotFoundException issue with the app. I needed a solution and that too quickly! We didn't want to make a wrong first impression. What should I try next?

At the back of my mind, I was sure that its some sort of a class loader issue. When I explicitly load the class its found, but when the class is needed later in the app, its not found. Surely something to do with a different class loader. I decided to go back to basics, let's to do it step by step.

First step,

When does the app need the class?

The app makes a HTTP call to get some information. The information is returned in form of JSON responses. As mentioned in my previous post, I was using Gson to parse the JSON response. Gson is an excellent library with it comes to JSON manipulation.

Gson converts JSON response into Java Model Objects. This is the point when the Java Model Class com.test.model.SomeModelClass is loaded. The problem lies here, Gson tries to load the class com.text.model.SomeModelClass dynamically and its unable to find it!

Why Gson does not find the class?

The real confusion in my mind was, why Gson is able to load the class on all other devices but HTC MyTouch 4G? My reasoning to that was, may be, Gson is loaded in a different class loader in HTC MyTouch 4G? Is that even possible?

After some intense googling and trying our numerous things, I finally figured out what was issue and what was solution.

This issue occurs only when all of the following conditions are met

The App uses Gson to manipulate JSON
The device is HTC Desire HD or MyTouch 4G on T-Mobile

The issue is HTC - The device manufacturer, also uses Gson library and they have made it public.

Somehow, when the JVM is loaded on such devices, the Gson is also loaded along with it. Because of this, the Gson library that is bundled with the app never gets loaded.

When the global Gson tries to load classes from our domain model, it fails to find them. This is because, the class loader that loaded the global Gson library does not have our app on its classpath!

That explained what the problem was, let's see how we can fix it!

The Fix

The fix was pretty simple. We need to change the package name of the Gson library so that the global Gson library does not interfere with the local Gson library. To change the package name of local Gson library, I used JarJar project.

I followed the following steps to change the package name of the local Gson library

Download jarjar project.
Put jarjar-1.1.jar and gson-1.6.jar in the same directory
Create a new text file in this directory, lets call it rules.txt.
Write the following line in the rules.txt file: rule com.google.gson.** com.google.myproject.@1
Open a command prompt and execute the command, java -jar jarjar.jar process rules.txt gson-1.6.jar myproject-gson-1.6.jar.
Replaced the gson library reference in the project with myproject-gson-1.6.jar
Update the imports to use the new package name
Compile and build the new APK
Requested the QA to test the updated APK

To look at more details visit this link.

And that was it, It worked!

It was such a relief! I am glad that, I faced this issue, makes me love the job that I am doing!

How to work with JSON on Android - Part - 2

2011-04-16T13:58:00.002+05:30

As mentioned in my previous post, Gson is an excellent lightweight opensource library to do JSON manipulations. I have used this library on the Android platform, it works great!

We have seen that, converting JSON to Java model objects and Java model Objects to JSON is pretty straight forward with Gson.

In this post lets look at the various hooks that Gson provides to deal with not so straight forward scenarios.

Lets take an example. Lets say you want to convert the JSON string value "08:00" into Java model Time object. A very valid scenario right?

Basically, we need to serialize and deserialize the Time object in a special way. Gson provides a way in which we can hookup our own serializer and deserializer for a given type. This serializer and deserializer will be invoked to convert the Java model object into JSON String and JSON string into Java model object.

The approach taken by Gson to do this task is pretty natural and easy to follow. Great job guys!

Lets look at the code to put things in perspective.

The Time class has two fields one to represent the hour and other represents the minute of the day. Let's say that, we want the JSON representation of Time objects to be of in the format "HH:MM". Let's first look at the code to convert Java model Time object into JSON

Gson provides an interface called JsonSerializer. We can provide an implementation of JsonSerializer to serialize the Time object in the way we want. The code to do this would look like

Ok so the serializer does not look very complex. How do we tell Gson to use TimeSerializer when it serializes the Time object?

Turns out, Gson providers a class called GsonBuilder. This class can be used to register custom serializers and deserializers with Gson. GsonBuilder class has the responsibility to construct an instance of Gson with registered serializers and deserializers. Let's look at the code

Does not look too difficult right?

Serialization done, let's look at deserialization. Any guesses?

I am sure you got it right. Gson provider an interface called JsonDeserializer. We have implement this interface to provide a custom deserializer for the Time Java model class. Let's look at the deserializer code.

The code of deserializer looks pretty simple. The code to integrate the deserializer is exactly the same as the code to integrate the serializer.

Thats all folks! Gson is a really good framework to manipulate JSON.

We have effectively used it in Android projects. If you have to do JSON manipulation in your project, I would suggest to have a look at Gson for sure!