In RAD Studio XE6 there appeared an opportunity to develop applications for iOS and Android on C++. Let’s consider the peculiarities of database application development on C++ using UniDAC.

UniDAC is Universal Data Access Components that allow to connect to SQLite, Oracle, MySQL, PostgreSQL, and InterBase (since version XE3) from iOS and Android (you can use UniDAC for connection from Windows to almost all the known databases, but this is beyond the scope of this blog).

Connection to database

Work with databases on iOS and Android in general is no different from the one on Windows, but there are some nuances when setting connection and deploying files to a mobile device if you work with a local DB. We will consider here how to establish connection to each supported database.

Design-time

Let’s create a new Android application on C++ working with a MySQL server. For this, in the File|New menu select FireMonkey Mobile Application – C++ Builder.

In the appeared dialog select Blank Application.

Then place the TUniConnection and TMySQLUniProvider components onto the form, which will be named as UniConnection1 and MySQLUniProvider1 respectively. Set up the UniConnection1 component by setting the ProviderName property to MySQL and filling in the required properties: UserName, Password, Server, Port.

If the listed properties are set correctly, then you can select a value for the Database property from the combobox. Now you can test connection by setting the Connected property to True. Take into account, that this test connection is performed on Win32 for XE6, not for the mobile application you are developing.

Compilation

Add the libmyproveder200.a file to the project, as it is shown below:

See the correspondence of object module names to database servers in the table below:

If you don’t add this file, you will get an error message during compilation:

“[ldandroid Error] C:\Users\Public\Documents\Embarcadero\Studio\14.0\PlatformSDKs\android-ndk-r9c\toolchains\arm-linux-androideabi-4.6\prebuilt\windows\bin\arm-linux-androideabi-ld.exe: .\Android\Debug\Unit1.o: in function _ZTX6TForm1:.\Android\Debug\Unit1.o.ll(.data.rel.ro._ZTX6TForm1+0×6): error: undefined reference to ‘vtable for Mysqluniprovider::TMySQLUniProvider’”

Compile the project.

Run-Time

You should place those providers onto the form, which you will use, and add object modules of these providers (as it was shown in the design-time sample). But take into account that for Android and iOS these modules have the same names, but are located in different folders:

“C:\Program Files\Devart\UniDAC for RAD Studio XE6\Lib\Android\”
“C:\Program Files\Devart\UniDAC for RAD Studio XE6\Lib\iOSDevice\”

You should either place the TUniConnection component onto the form or add the following lines to the header file:

#include "DBAccess.hpp"
#include "Uni.hpp"

and the following – to the cpp file:

#pragma link "DBAccess"
#pragma link "Uni"

If you are going to open access to a local database on a mobile device, then add the line to the header file:

#include <System.IOUtils.hpp>

in order to get access to IOUtils namespace.

SQLite

If you don’t deploy the database file to a mobile device, you should set:

SpecificOptions->Values["ForceCreateDatabase"] = "True"

In this case, on the first application launch, a database file will be created automatically.

Sample

  TUniConnection * Connection;

  Connection = new TUniConnection(Form1);

  try {
	Connection->ProviderName = "SQLite";
	Connection->SpecificOptions->Values["ForceCreateDatabase"] = "True";
	Connection->Database = System::Sysutils::IncludeTrailingPathDelimiter(
	  System::Ioutils::TPath::GetDocumentsPath()) + "db.sqlite3";
	Connection->Connect();
	ShowMessage("Connected successfully");
  }
  __finally {
	Connection->Free;
  }

Oracle

It is impossible to install Oracle client to a mobile device, because Oracle client for Android and iOS simply doesn’t exist. Therefore the only way to establish connection to Oracle from a mobile device is to connect directly via TCP/IP. For this, the Direct option should be set to True:

SpecificOptions->Values["Direct"] = "True";

In addition, the server name must be set correctly, since, if we have no client – we have no tnsnames.ora file with the server list as well. Therefore, to establish connection from Android and iOS, we need to know the server Host and Port, as well as its SID or Service Name.

To connect via the SID, the server should be set in the following way:

Server = "Host:Port:sid=SID";

or a simplified way:

Server = "Host:Port:SID";

To connect via the Service Name – as follows:

Server = "Host:Port:sn=ServiceName";

In other words, the ‘sid=’ prefix of the third parameter indicates that connection is established via the SID, and the ‘sn=’ prefix indicates that connection is established via the Service Name. If no prefix is specified, then, by default, it is considered, that we want to establish connection via the SID.

The majority of Oracle servers have the same SID and Service Name, so you, most likely, won’t have to go into such nuances, since you can learn more about this in the Oracle documentation.

Sample

  TUniConnection * Connection;

  Connection = new TUniConnection(Form1);

  try {
	Connection->ProviderName = "Oracle";
	Connection->SpecificOptions->Values["Direct"] = "True";
	Connection->Server = "server:1521:orcl";
	Connection->Username = "user_name";
	Connection->Password = "password";
	Connection->Connect();
	ShowMessage("Connected successfully");
  }
  __finally {
	Connection->Free;
  }

MySQL

MySQL client software for Android and iOS also doesn’t exist, therefore connection to MySQL server will also be established directly via TCP/IP. For this, let’s set the corresponding option:

SpecificOptions->Values["Direct"] = "True";

Sample

  TUniConnection * Connection;

  Connection = new TUniConnection(Form1);

  try {
	Connection->ProviderName = "MySQL";
	Connection->SpecificOptions->Values["Direct"] = "True";
	Connection->Server = "server";
	Connection->Port = 3306;
	Connection->Username = "user_name";
	Connection->Password = "password";
	Connection->Connect();
	ShowMessage("Connected successfully");
  }
  __finally {
	Connection->Free;
  }

PostgreSQL

With PostgreSQL, everything is more simple. Since UniDAC only allow establish direct connection via TCP/IP, it is enough for us to specify the Server and Port and perform Connect.

Sample

  TUniConnection * Connection;

  Connection = new TUniConnection(Form1);

  try {
	Connection->ProviderName = "PostgreSQL";
	Connection->Server = "server";
	Connection->Port = 5432;
	Connection->Database = "database_name";
	Connection->SpecificOptions->Values["Schema"] = "schema_name";
	Connection->Username = "user_name";
	Connection->Password = "password";
	Connection->Connect();
	ShowMessage("Connected successfully");
  }
  __finally {
	Connection->Free;
  }

InterBase

Using InterBase ToGo, you can connect to both local or remote DB.

To connect to a local db, just the path to the local DB on the device should be set:

Database = System::Sysutils::IncludeTrailingPathDelimiter(
	     System::Ioutils::TPath::GetDocumentsPath()) + "db.gdb";

If you need to establish connection to a remote server, you should specify not only the database, but the server as well:

UniConnection.Server = "server";
UniConnection.Database = "C:\db.gdb";

Please note that the System::Sysutils::IncludeTrailingPathDelimiter (System::Ioutils::TPath:: GetDocumentsPath) prefix should be specified when connecting to a local DB, and it is not needed for connection to a remote database.

Local database sample

  TUniConnection * Connection;

  Connection = new TUniConnection(Form1);

  try {
	Connection->ProviderName = "InterBase";
	Connection->Database = System::Sysutils::IncludeTrailingPathDelimiter(
	  System::Ioutils::TPath::GetDocumentsPath()) + "db.gdb";
	Connection->Username = "user_name";
	Connection->Password = "password";
	Connection->Connect();
	ShowMessage("Connected successfully");
  }
  __finally {
	Connection->Free;
  }

Remote database sample

  TUniConnection * Connection;

  Connection = new TUniConnection(Form1);

  try {
	Connection->ProviderName = "InterBase";
	Connection->Server = "server";
	Connection->Database = "C:\db.gdb";
	Connection->Username = "user_name";
	Connection->Password = "password";
	Connection->Connect();
	ShowMessage("Connected successfully");
  }
  __finally {
	Connection->Free;
  }

Deployment to mobile device

The deployment to mobile device process is described in the How to Develop Android Database Applications in RAD Studio XE5″ article.

Difference between Android and iOS application deployment

Please note that the deployment path is different for Android and iOS. If you want to deploy your application to both platforms, then make sure the deployment paths are specified correctly for both of them.

NOTE: Don’t forget to change the Remote Path default value “.” with one of the described above.

In addition, thus the providers for Android and iOS have similar names, they differ from each other. The providers are supplied with UniDAC, but each of them is located in a folder corresponding to a platform.

Debug

We won’t focus on the debug process in this article, since it was described in details in the previous article: Remote Debug of Android Applications in RAD Studio XE5 via Wi-Fi

Conclusion

RAD Studio XE5 release allowed creating applications with Devart Data Access Components in Delphi for work with SQLite, Oracle, MySQL, PostgreSQL and InterBase from both mobile platforms: Android and iOS. RAD Studio XE6 release made it possible not for Delphi only, but for C++Builder as well. But we are not going to stop. Many users ask us to support connection to SQL Server from Android and iOS – and we will make every effort to give them such a capability.

During the development of Review Assistant, we encountered a significant performance problem with displaying 100+ comments in one list, as creation of each comment takes a considerable amount of time. At first sight, it seemed that enabling virtualization would fix the problem, but we faced the following obstacles:

1. When scrolling, the elements jump form one to another without any smoothness. In addition, the thumb height begins to change in size, what looks just weird.
2. When selecting IsPixelBased in true, the elements are scrolled smoothly, but a lot of them begin to lag. And the worst thing is that sometimes the application crashes with StackOverflowException. The crash is caused by the code in the VirtualizingStackPanel.MeasureOverrideImpl method, where the tail call is used, and the call depth is not limited by any means.

In addition, we wanted to display elements of different types (reviews and comments) in the list. It can be solved by the usage of TreeView instead of ItemsControl and by specifying several HierarchicalDataTemplates. But the above-mentioned problems still persist.

Read the rest of this entry »

This is the second article of the two-part series that explains implementation of the TFS custom check-in policy for pre-commit code review. We developed this policy for Review Assitant - our code review tool.

In the first article we explained what is a check-in policy and what is the implementation procedure. This article shows nuts and bolts behind our implementation of a check-in policy.

In this article, we shall discuss:

• The algorithm of our Pre-Commit Code Review policy;
• Main problems of implementing the algorithm;
• Restrictions of the policy implementation.

Read the rest of this entry »

This is the first article of the two-part series that explains implementation of the TFS custom check-in policy for pre-commit code review. We developed this policy for Review Assitant - our code review tool.

Idea Behind the Implementation

A while ago a user posted an idea on the Review Assistant forum on UserVoice:

“I setup a build management in TFS server and add a build check-in policy to TFS, and now every check-in has to have a successful build, otherwise you cannot check-in.

It’d be great if we can setup a custom policy to TFS that every check-in needs to pass a code review. So, instead of developers having to shelve their changes manually and assign the shelveset to a code review, if this tool can do that automatically for you in a check-in process, that’d be super.”

Other users supported this idea so we decided to implement it.
Read the rest of this entry »

The current version of dbForge Studio for MySQL does not allow to import the whole Access database at once. Instead, there is an option to migrate separate Access tables in MySQL format.

The article below describes the entire process of converting Microsoft Access tables to MySQL.

Read the rest of this entry »

Summary: This article contains information on how to integrate the Review Assistant code review tool into Jenkins by using MSBuild target.

This is the third article in a series of Integration Review Assistant with Continuous Integration Systems. Read the Integrating Review Assistant with MSBuild and Integrating Review Assistant with CCNet articles to get an additional information about the integration with continuous integration systems. Read the rest of this entry »

Summary: This article contains information on how to integrate the Review Assistant code review tool into CruiseControl.NET (CCNet) by using MSBuild target.

This is the second article in a series of Integration Review Assistant with Continuous Integration Systems. Read the Integrating Review Assistant with MSBuild and Integrating Review Assistant with Jenkins articles to get an additional information about the integration with continuous integration systems. Read the rest of this entry »

Summary: This article contains information on how to integrate the Review Assistant code review tool into an MSBuild-based build environment.

This is the first article in a series of Integration Review Assistant with Continuous Integration Systems.

Why Use Continuous Integration?

If several developers are working on the project, from time to time they need to integrate their changes to the project’s code base. The developers also need to run builds to check if the new or updated code can be build successfully. These builds are called integrated builds. Normally, they are performed with a specified schedule.
It is quite difficult to find errors that occur in integrated builds that are performed occasionally, since the number of code changes introduced between the previous and current builds may be huge. To reduce the number of bugs in integrated builds and to make it is easier to find errors, integrated builds should be performed as often as possible. At the best case, builds should be performed right after the source code has been modified and reviewed.
Read the rest of this entry »

It’s incorrect to consider a database as a stable unit, since various kinds of undesirable situations can appear over time — database performance degradation, malfunctions, etc.

To minimize occurrence of such situations, maintenance plans are created, that guarantee stability and optimal performance of any databases.

Among maintenance plan tasks, the following can be pointed out:

1. Index defragmentation
   
2. Statistics update
3. Backup and checking the logical and physical integrity of all the objects in database

Let’s consider automation of each task.

So, the first point…

In addition to file system and log file fragmentation, sizable impact on database performance is provided by fragmentation within data-files:

1. Fragmentation inside separate index pages

After record insert, update and delete operations, empty spaces inevitably appear on pages. There is nothing wrong with that, since the situation is regular enough, but there is a slight hitch…

A very important role is played by the string length. For example, if a string is longer than a half of a page, the free half of the page won’t be used. As a result, when increasing the number of strings, the unused space in the database will be increasing. Read the rest of this entry »

Convertion of columns into rows is widely used in real tasks. In this article we will talk about different implementations of T-SQL UNPIVOT transformations.

Let’s assume we have an aggregated table which contains data about results of the games played by each player. We have a task to convert columns into rows.

IF OBJECT_ID ('dbo.Players') IS NOT NULL
	DROP TABLE dbo.Players;

CREATE TABLE dbo.Players
(
	  PlayerID INT
	, Win INT
	, Defeat INT
	, StandOff INT
	, CONSTRAINT PK_Players PRIMARY KEY CLUSTERED (PlayerID) ON [PRIMARY]
);
INSERT INTO dbo.Players (PlayerID, Win, Defeat, StandOff)
VALUES
	(1, 7,  6,  9),
	(2, 12, 5,  0),
	(3, 3,  11, 1);

Read the rest of this entry »