REM Change to local working copy directory
cd c:\myproject

REM Run SVN update to verify working copy code is current
svn update 

REM Run merge target
ant -f c:\myproject\_build\build.xml email-merge-status

My Ant script simply runs the merge and sends me an email. See my previous post (from 2006!) about sending email via Ant if you need help configuring that. All my properties ${something} are defined in an external property file but you should get the idea.

<target name="email-merge-status" depends="-init" description="Send email with merge status">
<exec executable="svn" outputproperty="svnmergeoutput">
<arg line="merge ${svn.trunk} ${path.branch} --dry-run" />
</exec>

<mail mailhost="${mail.server}" mailport="${mail.port}" subject="${ant.project.name} Merge Status">
<from address="${mail.default.from}" />
<to address="${mail.default.to}" />
<message>
Merge status for: ${ant.project.name}
Running:  merge ${svn.trunk} ${path.branch} --dry-run

=================================================================

${svnmergeoutput}
</message>
</mail>
</target>

I’m using the –dry-run SVN option with the merge just so I can get an idea of how many conflicts I can expect. If there aren’t many I will remove that and automate the merge and possibly the commit as well.

Now a few times I day I get a nicely formatted email that lets me know if I need to take action to update my branch:

Merge status for: IRBIS
Running: merge https://svn.myproject.some.url/svn/myproject/trunk/ c:/myproject –dry-run
=================================================================
— Merging r703 through r708 into ‘c:\myproject’:
U c:\myproject\admin_buckets_search.cfm
U c:\myproject\admin_gen_attachment_types.cfm
U c:\myproject\event_attachments.cfm
U c:\myproject\admin_gen_detail.cfm
U c:\myproject\wireframe\balsamiq_files\popup_existing_responses.bmml

I quickly ran into an issue with SQL Server integrated authentication.

My first stab was to simply provide a username and password:

	<target name="test">
		<sql driver="com.microsoft.sqlserver.jdbc.SQLServerDriver"
			url="jdbc:sqlserver://localhost:1433;databaseName=test"
			userid="myUsername"
			password="myPassword"
			print="TRUE">
			SELECT CURRENT_TIMESTAMP
		</sql>
	</target>

That failed with a username, password error.

A quick Google later I discovered the ‘integratedSecurity’ attribute. Note you now leave the username/password blank:

<target name="test">
	<sql driver="com.microsoft.sqlserver.jdbc.SQLServerDriver"
		url="jdbc:sqlserver://localhost:1433;databaseName=test;integratedSecurity=true"
		userid=""
		password=""
		print="TRUE">
		SELECT CURRENT_TIMESTAMP
	</sql>
</target>

But that still bombed out. I was getting this error:

SQL Exception: com.microsoft.sqlserver.jdbc.SQLServerException: This driver is not configured for integrated authentication.
WARNING: Failed to load the sqljdbc_auth.dll

Digging around I discovered this file is included in the JDBC driver download:

C:\Downloads\Microsoft SQL Server JDBC Driver 3.0\sqljdbc_3.0\enu\auth\x64\sqljdbc_auth.dll

There are several versions provided (IA64, x64, x86). Pick the one for your environment and copy it to your C:\Windows\system32 directory.

Your script should now work!

Buildfile: C:\build\db.build.xml
test:
      1 Executing commands
      1 2011-06-17 15:49:50.853
      1 0 rows affected
      1 1 of 1 SQL statements executed successfully
BUILD SUCCESSFUL
Total time: 1 second

AntForm is a java tool designed to add interaction to Ant scripts through graphical forms.

Setup is like any other Ant tas. Download the .jar file, add a task definition and be on your way. Installation is covered on the AntForm site.

Within the download file AntForm includes a test.xml file which I’d suggest you start with. It has several good examples that can help you get started. It gives a good example of the many controls available which allow you to build some very robust interfaces which I’ve captured in the screenshot below. You can see some examples of password dialogs, tabs, radio buttons and more.

AntForm Examples

I’ve certainly got some ideas on where I can use this! If you haven’t checked out Ant yet check out my Ant wiki for some great resources to help you get started!

Today we’ll look at how you can implement this using Ant.

First we’ll update our index.cfm file and add some code to check a value and show our friendly message:

<cfset maintenanceMode = false>
<cfif maintenanceMode IS true>
	<h2>Site Maintenance</h2>
	<p>The application is currently unavailable. Please check back in a few minutes (or hours if we really screw something up).</p>
	<p>We apologize for the inconvenience.</p>
	<cfabort>
</cfif>

So now we need to flip the bit on maintenanceMode.

You could access the server, open the file with an editor and hope you don’t screw something up while editing the file… or you could do it from the comfort of your IDE with Ant. We will create a macrodef so we can use the same code to turn on/off maintenance mode in all our environments.

The macrodef code:

<macrodef name="maintenance-mode" description="Enable or disable maintenance mode">
	<attribute name="mode" default="false"/>
	<attribute name="dest"/>
	<sequential>
		<echo>
			Setting mainenance mode to: @{mode} on ${path.@{dest}}
		</echo>
		<replaceregexp file="${path.@{dest}}/www/index.cfm" match="maintenanceMode = (\w+)" replace="maintenanceMode = @{mode}"/>
	</sequential>
</macrodef>

First we have defined some attributes we’ll pass in:

• mode = true or false
• dest = destination – where the file lives we need to modify

We will add an echo in there so we can see what’s going on when we run things, and finally we’ll use the ReplaceRegexp task to search for our string. We simply need to look for “maintenanceMode = whatever” and that’s easy to do with a regular expression. Then we simply update that string with the value we’ve passed in as ‘mode’.

Tip: If you have problems with Regular Expressions check out the RegEx Tester plugin for Eclipse.

Finally we will create two new targets to turn things on and off:

<target name="DEVELOPMENT-MAINTENANCE-ON">
	<maintenance-mode mode="true" dest="development"/>
</target>

<target name="DEVELOPMENT-MAINTENANCE-OFF">
	<maintenance-mode mode="false" dest="development"/>
</target>

Here we define our ‘mode’ and ‘destination’ attributes. The ‘development’ destination is simply a property I have mapped in Ant which points to my web files. Depending on how you access your files you may have to just through more hoops here.

path.development = \\\\serverName\\Inetpub\\wwwroot\\www\\${ant.project.name}

Using a combination of properties and attributes: ${path.@{dest} this will give us the proper path to the file we need to change.

You can easily create additional targets for production, etc.

And finally running this in Eclipse we can see:

Buildfile: D:\workspace\_build\build.xml
DEVELOPMENT-MAINTENANCE-OFF:
     [echo]
     [echo] Setting mainenance mode to: false on \\serverName\Inetpub\wwwroot\www\myproject
     [echo]
BUILD SUCCESSFUL
Total time: 1 second

For more on Ant check out my Ant wiki.

I then looked to the Ant wiki and much to my surprise there were no MySQL specific entries!

After a bit of Googling and tinkering however I’ve come up with a script that does simple dump/restore functions and should provide the basics for someone to expand on if they need more.

The script is available on my wiki: http://thecrumb.com/wiki/ant/database

One interesting thing I ran into that had me stumped for an afternoon was defining my user/password as properties and passing them into my script.

I usually define all my properties in external files.  In this example I had the following defined:

• db.local.username = myusername
• db.local.password = password

And was using it like:

<exec executable="${path.mysql}/mysqldump" input="dbscripts/@{filename}">
<arg value="--user=${db.@{server}.username}"/>
<arg value="--password=${db.@{server}.password}"/>
<arg value="--host=${db.@{server}}"/>
<arg value="--port=3306"/>
<arg value="@{database}"/>
</exec>

When I ran this it bombed and I got a MySQL error.  I initially thought it was a MySQL permission error so I tested it via the command line and everything worked.

OK.

Then I swapped out the property values in my Ant script with hard coded values:

<exec executable="${path.mysql}/mysqldump" input="dbscripts/@{filename}">
<arg value="--user=root"/>
<arg value="--password=password"/>
<arg value="--host=${db.@{server}}"/>
<arg value="--port=3306"/>
<arg value="@{database}"/>
</exec>

I expected that to fail as well but it worked!  Weird.  I then tried all sorts of various combinations.  Eventually I tried using property names that did NOT include ‘username‘ and ‘password‘.

• db.local.un = myusername
• db.local.pw = mypassword

<exec executable="${path.mysql}/mysqldump" input="dbscripts/@{filename}">
<arg value="--user=${<strong>db.@{server}.un</strong>}"/>
<arg value="--password=${<strong>db.@{server}.pw</strong>}"/>
<arg value="--host=${db.@{server}}"/>
<arg value="--port=3306"/>
<arg value="@{database}"/>
</exec>

And that worked! Even when defined in an external property file.

I have no idea if ‘username‘ and ‘password‘ are some sort of reserved word.  I tried it in both Ant 1.7 and  1.8 with the same results.

The scripts on the wiki work in my environment  (Ant 1.7, Windows XP and MySQL 5.1.x) .

If you modify the scripts and do something interesting with them – please let me know!

And if you blow something up – please don’t blame me!

This DZone Refcard will provide you with the perfect resource to help you automate software build processes with Apache Ant. Included is a description of the Anatomy of an Ant Script along with a number of tables detailing, among others: Core Java Related Tasks, Infrastructure Tasks, and SCM Related Tasks.

It is well laid out (like all the Refcradz) and covers the latest Ant release – 1.8.

Turns out it was easy.

Crank up the Ant view and right click on your build file:

Run As

Click Run As and select External Tools Configuration

External Tool Dialog

Click the Main tab and enter your Ant arguments in the arguments dialog. What is neat is you can create multiple configurations so you can run your build easily with different arguments.  Enter a friendly name and notice it show up on the file list on the left.

You can then right click on that name and duplicate it and then modify the arguments in the new config.

Duplicate

Now if you return to the Ant view, right click and select Run As… you will get a dialog with the configurations you defined…

Options

My Ant target looks like:

And we’ll replace ${myValue} with an argument I defined in my configuration above (-DmyValue=”Jim Likes Ant”). Running my two options I defined will output

Buildfile: D:\workspace\ant\helloworld-1.xml
helloworld:
[echo] Hello World Jim Likes Ant
BUILD SUCCESSFUL
Total time: 122 milliseconds

Buildfile: D:\workspace\ant\helloworld-1.xml
helloworld:
[echo] Hello World Jim Hates Ant
BUILD SUCCESSFUL
Total time: 122 milliseconds

Growl is originally for the Mac but there is a port for Windows aptly named Growl For Windows.  Download and install.  I also added the install path to my Windows PATH so I can run the growlnotify script from the command line.

Once you’ve done this you can verify it’s working by popping open a DOS box and typing:

C:> growlnotify "THIS IS A TEST"

And you should get a popup:

Growl Popup

Next you’ll need to grab a few things so you can interact with Growl from Ant…

Download growlbuildlistener

An Ant BuildListener that displays notifications via Growl/JGrowl

Follow the installation instructions (in the wiki section).

This simply involves downloading two files that contain jar files – you’ll need:

• the growllistener jar  (growllistener.jar)
• the JGrowl jar (binaryblizzard.jar) (from JGrowl)

Drop those in your Ant lib folder, and in Eclipse make sure those jars are added to your classpath (Ant Home Entries).

Next you will need to add a listener so Ant can use the GrowlBuildListener.  Run the following line from a command prompt:

set ANT_ARGS=-listener net.slimeslurp.growl.GrowlListener

Next I had to add a new taskdef line to my build script:

And now you should be able to drop in a new GrowlEcho task in your build file wherever you want a popup to occur.

Now when I kick off my scripts I see something like this:

Growl Example

I did this mainly for fun but I can see some applicable uses for this. Growl also looks neat because it appears you can configure it to push notices over the network. So in theory I could setup everyone on my team with Growl and when someone kicked off a build we could all be notified.

If you come up with a creative use for this please let me know in the comments!

Since I’m a Linux geek I’m familiar with cron on Linux and went looking for a version for Windows.

Cron enables users to schedule jobs (commands or shell scripts) to run automatically at a certain time or date.

Turns out there are quite a few cron solutions available on Windows.  Most of them however were commercial solutions.  After a bit of digging I found nnCron ( http://www.nncron.ru/ ).  nnCron has a commercial version but they offer a free “lite” version which covered the basic features I needed:

nnCron LITE is a small, but full-featured scheduler that can start applications and open documents at a specified time or with specified frequency. nnCron LITE is a perfect freeware Windows clone of a well-known UNIX scheduler Cron (including all the useful Anacron features).

Here are a few of the features of nnCron LITE:

• it can be started as a system service or as a regular standalone application
• it understands cron table format (Unix) and is managed with easy-to-edit text crontab files that are fully compatible with Unix crontabs
• it can set and use environment variables
• it can run applications authorized as currently logged user

So far I’ve been using it to schedule a backup task I have for our JIRA and Confluence installs using SyncBack.  And I have a few system utilities setup (defrag, etc) that run occasionally.  Unlike Scheduled Task I’ve never had an issue with the nnCron service failing.

It’s been years since I’ve worked in a shop where we used FTP to deploy files.  I always felt that manual FTP was the worst possible way to move files around (this was before I discovered Ant).  I ended up moving files around with SVN (which I still do today but Ant is the middle man).

I also never used Dreamweaver. I went straight from Homesite to CFEclipse.

Glyn’s post got me thinking however.  Does Dreamweaver’s ‘site’ feature force you into FTP?  Do people open Dreamweaver, assume the ‘site’ feature is the only option and follow that path?

I’m seeing a lot of the same conversations resurrected with the release of ColdFusion Builder as I did when CFEclipse first appeared.  People struggling with Eclipse’s “project” concept. And the fact that FTP may not be the only option.