Debugging PHP

As an old fashioned programmer I grew up with debugging methods like post-mortem traces and trace statements.

Today however we have and are used to GUI’s for debugging and can single step code or even re-compile code and retry the operation. This is all nice in environments where applications can be frozen. If not, like web pages and applications depending on real-time communication with devices the single stepping alone ruins the applications workings and thus the debugging process.

Here old fashioned trace message and a viewer for them come in handy again. Normally on Microsoft Windows one uses the OutputDebugString() API. For PHP this API call was missing so I implemented a simple PHP extension that wrapped the API in two ways. One is just the call and the other way is as member function of an object. 

Big advantage of the OutputDebugString() API is that if there is no viewer active, the output is just ignored and vanishes into thin air, leaving no traces like massive log files. Also good to know is that it’s impossible to ruin http headers etc as the output is redirected to something else then the web browser.

The extension was written in Borland Delphi using the easy to use Php4Delphi library. As viewer one can use the free DbgView from Sysinternals.

The result is a very easy to use extension that can be left in the code for as long as one want/needs.

The following snippet test of the module is indeed loaded properly by the PHP interpreter:

   1: $module = "log";
   3: if(!extension_loaded($module)) {
   4:   echo "Log Module not Loaded";
   5:   exit;
   6: }


This snippet uses the php_log class:

   1: $log = new php_log();
   2: $log->cleardebugwindow();
   3: $log->outputdebugstring("PHP test log class", $log->info);


The cleardebugwindow() method send a special message to DbgView clearing the display. Outputdebugstring() takes two parameters, the message and a severity string. This last parameter is handy for grouping the messages or be able to search on certain types. It is not necessary to use the built-in types like info, warning or error, any tag is allowed.

The following code is not using classes:

   1: outputdebugstring("PHP test module", 'error');


The sources can be downloaded from this link. In order to compile it, you’ll also need to download php4delphi and configure it correctly for your PHP version and off-course a Borland Delphi version.

Posted in Uncategorized | Leave a comment

How to automate inclusion of versioning info in Java beans

This post is about how to solve a problem that bugged me for while. 

When developing Portlets for Liferay one always wonders what exact version is actually running on the various servers of the development chain (local/test/integration/production) and what sources it was compiled from.

  • The first solution (when still using CVS/SVN):

When using CVS or SVN this can easily be solved by using keyword expansion and have either one of these systems update the MANIFEST.MF file containing the keyword placeholders.

For SVN adding keywords to the MANIFEST.MF file and simply adding a couple of lines to the build.xml file that change the MANIFEST.MF file (so it gets marked for check-in every time), the requested information can be included pretty automatically.

   1: <?xml version="1.0"?>

   2: <!DOCTYPE project>


   4: <project name="my-portlet" basedir="." default="deploy">

   5:     <import file="../build-common-portlet.xml" />

   6:     <tstamp>

   7:         <format property="TimeDate.Now" pattern="yyyy-MM-dd HH:mm:ss" />

   8:     </tstamp>

   9:     <manifest file="docroot/META-INF/MANIFEST.MF" mode="update">

  10:         <attribute name="Ant-Build-Stamp" value="${TimeDate.Now}" />

  11:     </manifest>

  12: </project>

In the MANIFEST.MF file one needs to include the following placeholders:

   1: Svn-Revision: $Revision$

   2: Svn-Author: $Author$

   3: Svn-Date: $Date$

Finally in Eclipse one needs to add the keywords to the MANIFEST.MF file (this feature is hidden in the MANIFEST.MF context menu under TEAM|Set Property…. In the resulting dialog choose ‘svn:keywords’ and enter ‘Revision Date Author‘ on a single line.

CVS always expands the keywords if present so only the Ant script is needed.

Getting hold of this information and making use of it is similar to the Mercurial solution presented next.

  • The second solution (after switching to Mercurial):

This method adds Mercurial information to the MANIFEST.MF unfortunately this is immediately outdated the moment one does a commit/push (because then a new revision is created).

BUT when the portlet is build for deployment the information in the MANIFEST.MF is updated to the correct values prior to compiling. The deployment build itself does not commit any files so no new revision is created. So when deployed one can easily find out what exact source code is used and compiled.

The first part is to prep the build.xml file so it retrieves and writes/updates this information into the MANIFEST/MF file.

A plain Liferay generated build.xml file looks like:

   1: <?xml version="1.0"?>

   2: <project name="my-portlet" basedir="." default="deploy">

   3:     <import file="../build-common-portlet.xml" />


   5: </project>

By inserting/include the following xml just after the import task we retrieve the wanted information and write/update it into the MANIFEST.MF file. This file and the docroot/META-INF directory where it should reside are also created if missing.

   1: <import file="manifest.xml" />


The manifest.xml should contain the following Ant script:

   1: <?xml version="1.0"?>

   2: <project>

   3:     <if>


   5:         <!-- Execute ony once, saves time -->


   7:         <not>

   8:             <isset property="hgrevision" />

   9:         </not>


  11:         <then>


  13:             <!-- Test Operating Systems -->


  15:             <condition property="isUnix">

  16:                 <os family="unix" />

  17:             </condition>


  19:             <condition property="isWindows">

  20:                 <os family="windows" />

  21:             </condition>


  23:             <!-- Set Mercurial Executable -->


  25:             <if>

  26:                 <isset property="isUnix" />

  27:                 <then>

  28:                     <property name="mercurial" value="hg" />

  29:                 </then>

  30:             </if>


  32:             <if>

  33:                 <isset property="isWindows" />

  34:                 <then>

  35:                     <property name="mercurial" value="${env.ProgramFiles}/Mercurial/hg.exe" />

  36:                 </then>

  37:             </if>


  39:             <echo message="Using: ${mercurial}" />


  41:             <!-- Run Mercurial for Information -->


  43:             <exec executable="${mercurial}">

  44:                 <arg value="id" />

  45:                 <arg value="-n" />

  46:                 <redirector outputproperty="hgrevision" />

  47:             </exec>


  49:             <!-- Trim any trailing + sign -->


  51:             <script language="javascript">

   2:                 var hgrevision =

   3:                 project.getProperty("hgrevision");

   4:                 project.setProperty("hgrevision",

   5:                 hgrevision.replaceAll("[\+]", ""));



  52:             <echo message="Local Revision: ${hgrevision}" />


  54:             <exec executable="${mercurial}">

  55:                 <arg value="id" />

  56:                 <arg value="-t" />

  57:                 <redirector outputproperty="hgtags" />

  58:             </exec>

  59:             <echo message="Tag: ${hgtags}" />


  61:             <exec executable="${mercurial}">

  62:                 <arg value="id" />

  63:                 <arg value="-b" />

  64:                 <redirector outputproperty="hgbranch" />

  65:             </exec>

  66:             <echo message="Branch: ${hgbranch}" />


  68:             <exec executable="${mercurial}">

  69:                 <arg value="log" />

  70:                 <arg value="-r${hgrevision}" />

  71:                 <arg value="--template" />

  72:                 <arg value='&quot;{date|isodate}&quot;' />

  73:                 <redirector outputproperty="hgdate" />

  74:             </exec>

  75:             <echo message="Date: ${hgdate}" />


  77:             <exec executable="${mercurial}">

  78:                 <arg value="log" />

  79:                 <arg value="-r${hgrevision}" />

  80:                 <arg value="--template" />

  81:                 <arg value='&quot;{node}&quot;' />

  82:                 <redirector outputproperty="hgnode" />

  83:             </exec>

  84:             <echo message="Node: ${hgnode}" />


  86:             <exec executable="${mercurial}">

  87:                 <arg value="log" />

  88:                 <arg value="-r${hgrevision}" />

  89:                 <arg value="--template" />

  90:                 <arg value='&quot;{node|short}&quot;' />

  91:                 <redirector outputproperty="hgsnode" />

  92:             </exec>

  93:             <echo message="Short Node: ${hgsnode}" />


  95:             <exec executable="${mercurial}">

  96:                 <arg value="log" />

  97:                 <arg value="-r${hgrevision}" />

  98:                 <arg value="--template" />

  99:                 <arg value='&quot;{author}&quot;' />

 100:                 <redirector outputproperty="hgauthor" />

 101:             </exec>

 102:             <echo message="Author: ${hgauthor}" />


 104:             <exec executable="${mercurial}">

 105:                 <arg value="log" />

 106:                 <arg value="-r${hgrevision}" />

 107:                 <arg value="--template" />

 108:                 <arg value='&quot;{repo}&quot;' />

 109:                 <redirector outputproperty="hgrepo" />

 110:             </exec>

 111:             <echo message="Repository: ${hgrepo}" />


 113:             <!-- Ant Built-in properties -->


 115:             <echo message="Java Version: ${java.runtime.version}" />

 116:             <echo message="Java Vendor: ${java.vendor}" />


 118:             <!-- Liferay properties (see of SDK) -->


 120:             <echo message="Liferay-Version: ${lp.version}" />


 122:             <!-- Ant Build timestamp -->


 124:             <tstamp>

 125:                 <format property="TimeDate.Now" pattern="yyyy-MM-dd HH:mm:ss" />

 126:             </tstamp>


 128:             <!-- Create META-INF if missing -->


 130:             <if>

 131:                 <not>

 132:                     <available file="docroot/META-INF" type="dir" />

 133:                 </not>

 134:                 <then>

 135:                     <mkdir dir="docroot/META-INF" />

 136:                 </then>

 137:             </if>


 139:             <manifest file="docroot/META-INF/MANIFEST.MF" mode="update">


 141:                 <!-- Mercurial Information -->


 143:                 <attribute name="Hg-Revision" value="${hgrevision}" />

 144:                 <attribute name="Hg-Tags" value="${hgtags}" />

 145:                 <attribute name="Hg-Branch" value="${hgbranch}" />

 146:                 <attribute name="Hg-Date" value="${hgdate}" />

 147:                 <attribute name="Hg-Node" value="${hgnode}" />

 148:                 <attribute name="Hg-Short-Node" value="${hgsnode}" />

 149:                 <attribute name="Hg-Author" value="${hgauthor}" />

 150:                 <attribute name="Hg-Repository" value="${hgrepo}" />


 152:                 <!-- Ant Built-in properties -->


 154:                 <attribute name="Java-Version" value="${java.runtime.version}" />

 155:                 <attribute name="Java-Vendor" value="${java.vendor}" />


 157:                 <!-- Liferay properties (see of SDK) -->


 159:                 <attribute name="Liferay-Version" value="${lp.version}" />


 161:                 <attribute name="Ant-Build-Stamp" value="${TimeDate.Now}" />

 162:             </manifest>

 163:         </then>

 164:     </if>

 165: </project>


  • The above script expects hg Mercurial executable to be on the path under Linux and to be installed in “c:\program files\Mercurial” when running under Windows.
  • The Short-Node consists of the first 12 characters from the full Mercurial ”’Node”’ hash. This corresponds to the change set value that Source forge displays when browsing the repository.
  • The Revision is the local revision within a project repository (in contrast to the system wide values of node and short-node). It corresponds to the change set number that Source forge displays when browsing the repository.
  • Because of the small piece of JavaScript to cleanup the revision number, Java 1.6 or later is required.

The resulting MANIFEST.MF looks like:

   1: Manifest-Version: 1.0

   2: Ant-Version: Apache Ant 1.7.1

   3: Created-By: 20.4-b02 (Sun Microsystems Inc.)

   4: Hg-Revision: 27

   5: Hg-Tags: tip

   6: Hg-Branch: my-portlet

   7: Hg-Date: 2012-08-14 16:41 +0200

   8: Hg-Node: cdd2c1242f3e3f528ff1e71022570d57d1b9a342

   9: Hg-Short-Node: cdd2c1242f3e

  10: Hg-Author: me

  11: Hg-Repository: 9876543219101112131415161718192021222324252621

  12: Java-Version: 1.6.0_33-b03

  13: Liferay-Version: 6.0.12

  14: Ant-Build-Stamp: 2012-08-15 12:17:12

  15: Java-Vendor: Sun Microsystems Inc.

  16: Class-Path: 

The lines between Created-By: and Class-Path: are originating from added the Ant build.xml script.

  • Finally how to retrieve this information in out Java code:

We need to add some utility methods to our code:

   1: /**

   2:  * Retrieves a property from /META-INF/MANIFEST.MF without a prefix.

   3:  * 

   4:  * @param property

   5:  *            the property to retrieve

   6:  * @return the property value, 'na' or 'error'.

   7:  */

   8: public static String getManifestProperty(final String property) {

   9:     if (_log == null) {

  10:         _log = LogFactoryUtil.getLog(MyPortlet.class);

  11:     }


  13:     Properties prop = new Properties();


  15:     try {

  16:         prop.load(FacesContext.getCurrentInstance().getExternalContext()

  17:                 .getResourceAsStream("/" + JarFile.MANIFEST_NAME));


  19:         String rev = prop.getProperty(property);


  21:         return rev == null ? "na" : rev.trim();

  22:     } catch (IOException e) {

  23:         _log.error("Error retrieving " + property + " Property from /"

  24:                 + JarFile.MANIFEST_NAME + " (" + e.getMessage() + ").");

  25:     }


  27:     return "error";

  28: }


  30: /**

  31:  * Retrieves a property from /META-INF/MANIFEST.MF with an optional prefix.

  32:  * 

  33:  * @param prefix

  34:  *            a prefix like 'Svn-' or 'Hg-' in our case.

  35:  * @param property

  36:  *            the property to retrieve

  37:  * @return the property value, 'na' or 'error'.

  38:  */

  39: public static String getManifestProperty(final String prefix,

  40:         final String property) {

  41:     return getManifestProperty(prefix + property);

  42: }


  44: /**

  45:  * Retrieves a 'Hg-' prefixed property from /META-INF/MANIFEST.MF and

  46:  * cleans the $Keyword$ definition from the result.

  47:  * 

  48:  * @param property

  49:  *            the property to retrieve

  50:  * @return the property value, 'na' or 'error'.

  51:  */

  52: public static String getManifestHgProperty(final String property) {

  53:     return getManifestProperty("Hg-", property)

  54:             .replace("$" + property + ": ", "").replace(" $", "").trim();

  55: }


  57: public static String LPad(final String str, final int length) {

  58:     return LPad(str, length, ' ');

  59: }


  61: public static String LPad(final String str, final int length, final char car) {

  62:     return str

  63:             + String.format("%" + (length - str.length()) + "s", "")

  64:                     .replace(" ", String.valueOf(car));

  65: }


  67: public static String RPad(final String str, final int length) {

  68:     return RPad(str, length, ' ');

  69: }


  71: public static String RPad(final String str, final int length, final char car) {

  72:     return String.format("%" + (length - str.length()) + "s", "").replace(

  73:             " ", String.valueOf(car))

  74:             + str;

  75: }

and call use code for example in the constructor of our Java backing bean like:



   3:"Class:", 32) + getClass().getName());


   5:"Revision:", 32)

   6:         + GroupwallHelpers.getManifestHgProperty("Revision"));

   7:"Node:", 32)

   8:         + GroupwallHelpers.getManifestHgProperty("Node"));

   9:"Short-Node:", 32)

  10:         + GroupwallHelpers.getManifestHgProperty("Short-Node"));

  11:"Tags:", 32)

  12:         + GroupwallHelpers.getManifestHgProperty("Tags"));

  13:"Branch:", 32)

  14:         + GroupwallHelpers.getManifestHgProperty("Branch"));

  15:"Repository:", 32)

  16:         + GroupwallHelpers.getManifestHgProperty("Repository"));

  17:"Date:", 32)

  18:         + GroupwallHelpers.getManifestHgProperty("Date"));

  19:"Author:", 32)

  20:         + GroupwallHelpers.getManifestHgProperty("Author"));


  22:"Ant-Build: ", 32)

  23:         + GroupwallHelpers.getManifestProperty("Ant-Build-Stamp"));


  25:"Java-Version: ", 32)

  26:         + GroupwallHelpers.getManifestProperty("Java-Version"));

  27:"Java-Vendor: ", 32)

  28:         + GroupwallHelpers.getManifestProperty("Java-Vendor"));

  29:"Ant-Build: ", 32)

  30:         + GroupwallHelpers.getManifestProperty("Liferay-Version"));




Final Note: A better type of output would be to use _log.debug() instead of so the Tomcat log files are not polluted to much.

Posted in Uncategorized | Leave a comment

Command-line scanning a directory or file with Microsoft Security Essentials

After doing the usual web search with Google and Bing I found only sites claiming doing a scan with Microsoft Security Essentials of a file or directory from the command-line was not possible. Most sites just say it’s possible to initiate a quick or full scan or even update signatures from the command-line. But as Microsoft Security Essentials has real-time protection there is no need to scan manually (as a lot of us are used to) to make sure the file is scanned.

But on Windows 7 (x64 version confirmed) the tools in %ProgramFiles%\Microsoft Security Client contain some files that look promising. My first guess was to look into %ProgramFiles%\Microsoft Security Client\msseces.exe, but that program only pops-up the user-interface and worse case starts a default scan.

As Microsoft Security Essentials is able to scan manually (it has a explorer context menu, located in %ProgramFiles%\Microsoft Security Client\shell.ext.dll)

A far shot was to search for rundll32 to be used to fire the context menu but all I found was references to viruses and trojans doing the same (so not the best way to walk)

Finally I accidently fired up one of the other executables in %ProgramFiles%\Microsoft Security Client with a promising name (MpCmdRun.exe) with a switch –h and voila a long description with the answer tucked inside.

By issuing the command

“%ProgramFiles%\Microsoft Security Client\MpCmdRun.exe” -Scan -ScanType 3 –File “<file or folder to scan>”

one is able to start the command-line version of Microsoft Security Essentials, make it perform a file or folder scan and thus integrate it with popular tools like Winrar like:


Note: although %ProgramFiles% points to c:\program files at both Windows x86 and x64, not all applications will expand it properly to something like c:\program files.

Posted in Uncategorized | Leave a comment

Getting the Name of a C# Component.

Today I had the need to address a C# Component I wrote and that is part of a larger C# UserControl by name (so by a String). I know there are other ways to address components but for some reason I will not explain in this blog post I needed this feature.

The main problem is that Components do not have a Name property although in the Visual Studio Property Inspector components do show a ‘(Name)’ property containing the value I was after.

I started searching the internet but without luck. Plenty of solutions for Controls but none for Components except for the obvious ‘why do need this anyway’ type of non-solutions and a fair number of unanswered questions.

Some months ago I already solved a minor piece of the puzzle by finding a way to get the Component’s Name at Design Time.

A simple ToString() and some string cutting was enough to get the value of the myserious ‘(Name)’ property. This with one limitation, the Component should not have it’s own ToString() method overriding the default one.  The following code does the trick:

   1: /// <summary>

   2: /// GetName() only works at Design Time.

   3: /// </summary>

   4: /// <returns>The Name of the Component as shown in the Designer</returns>

   5: public string GetName()

   6: {

   7:     int split = ToString().IndexOf(' ');

   8:     return ToString().Substring(0, split);

   9: }

However nice this works on Design-Time, on Run-Time this code fails completely as the ToString() is empty or at best contains the class name of the Component.

So how do we get this Design-Time value available at Run-Time? The answer was a specially defined property that behaves differently at Design and Run-Time.

   1: [DesignerSerializationVisibility(DesignerSerializationVisibility.Visible)]

   2: [Browsable(false)]

   3: public String ColumnName

   4: {

   5:     get

   6:     {

   7:         //If at designTime we store GetName into fColumnName.

   8:         if (ToString().IndexOf(' ') != -1)

   9:         {

  10:             fColumnName = GetName();

  11:         }

  12:         return fColumnName;

  13:     }

  14:     set

  15:     {

  16:         //If at designTime we store fColumnName and ignore value.

  17:         if (ToString().IndexOf(' ') != -1)

  18:         {

  19:             fColumnName = GetName();

  20:         }

  21:         else

  22:         {

  23:             fColumnName = value;

  24:         }

  25:     }

  26: }

  27: private String fColumnName;


Above code will use part of the GetName() method’s code to distinguish between Design-Time (ToString()’s value is a combination of Design-Time name and ClassName separated by a space) and Run-Time where ToString() is just the ClassName.

The getter and setters work differently at Design-Time and Run-Time.

Both getter and setter set the private storage field ‘fColumnName’ at Design-Time with the value returned by GetName() and the getter off-course returns this value. So on Design-Time the setter totally ignores the ‘value’ parameter passed.

On Run-Time the getter does return the private storage field and the setter does set it with value as a normal property would do.

Finally, the attributes of the property make sure it is not visible in the property inspector but will be saved into the *.designer.cs file properly. I experimented with the ReadOnly attribute too but that only leads to the form designer not saving the property into the *.designer.cs file. If you remove the ‘Browsable’ attribute or set it to true you can see the ColumnName property at Design-Time.

Basically what happens is that the Design-Time value end up in the private storage field of a property that is saved into the *.designer.cs file. At Run-Time GetName() is totally ignored and the value retrieved from the *.designer.cs file is used as if it where a normal property.

The result is a property that changes nicely when a Component is renamed and cannot be edited at Design-Time but is serialized into the *.designer.cs file as it should.

Note: the code presented in this blog post only works on Components you write yourself and frees you from maintaining a separate Name property that has no actual link to the Components Design-Time name nor will be adjusted when renaming the Component.

Maybe someone can bolt it on with an extension method.

Posted in Uncategorized | Leave a comment

Static properties on design-time

A couple of weeks ago I was working on a component to use & generate html help files. The idea was to drop a component on each form where it would do it’s work:

  1. Generate files for Microsoft’s HtmlHelp Workshop when executed from Visual Studio.
  2. Show all kind of context sensitive help when executed outside Visual Studio.

During the writing I needed an Enable property visible in the property inspector of Visual Studio that was global to all instances of my component.

I used the following (simple) construct:


   2: public Boolean Enabled {

   3:   get { 

   4:     return fEnabled; 

   5:   }

   6:   set { 

   7:     fEnabled = value; 

   8:   }

   9: }


  11: private static fEnabled = true;


This code shows a Enabled property in the property inspector, but when the value is of a single component instance is changed, the changed value in all other instances too.

Posted in C#, Programming | Leave a comment

RadioCheck for MenuToolStrip

Today I wanted to code a submenu with menu items that are checked like a radio button group. So once a single menu item has been checked, there is exactly one item checked all the time.

The task proved a bit harder than I expected. The RadioCheck property is no longer present in MenuToolStrip and ToolStripMenuItem as it was in the old MenuItem class.

So time to code one!

Basically the code below assumes you have a ToolStripMenuItem which contains all ToolStripMenuItems that are part of the radio button group in its DropDownItems collection.

Optionally you may check one ToolStripMenuItem as startup. Once a ToolStripMenuItem is checked there is no way to uncheck them all.

Each of these ToolStripMenuItems must have the CheckOnClick property set to true and have a event handler attached like:

   1: tmi.CheckedChanged += new EventHandler(tmi_CheckedChanged);


The event handler code below unchecks all other ToolStripMenuItems in the DropDownItems collection in such way that one does not end up with a stack overflow exception.

   1: /// <summary>

   2: /// Simulate RadioCheck MenuItems.

   3: /// </summary>

   4: /// <param name="sender"></param>

   5: /// <param name="e"></param>

   6: void tmi_CheckedChanged(object sender, EventArgs e)

   7: {

   8:     ToolStripMenuItem tmi = (ToolStripMenuItem)sender;

   9:     ToolStripItemCollection items = ((ToolStripMenuItem)(tmi.OwnerItem)).DropDownItems;


  11:     //Make sure that one menu item is always checked.

  12:     if (tmi.Checked)

  13:     {

  14:         //Uncheck all other menu items.

  15:         foreach (ToolStripMenuItem tmi2 in items)

  16:         {

  17:             if (!tmi.Equals(tmi2) && tmi2.Checked)

  18:             {

  19:                 tmi2.Checked = false;

  20:             }

  21:         }

  22:     }

  23:     else

  24:     {

  25:         //If anything else is already checked, 

  26:         //bail out to prevent a stack overflow.

  27:         foreach (ToolStripMenuItem tmi2 in items)

  28:         {

  29:             if (tmi2.Checked)

  30:             {

  31:                 return;

  32:             }

  33:         }


  35:         //If nothing is checked, check ourselves again.

  36:         tmi.Checked = true;

  37:     }

  38: }

/// <summary>
/// Simulate RadioCheck MenuItems.
/// </summary>
/// <param name="sender"></param>
/// <param name="e"></param>
void tmi_CheckedChanged(object sender, EventArgs e)
    ToolStripMenuItem tmi = (ToolStripMenuItem)sender;
    ToolStripItemCollection items = 
((ToolStripMenuItem)(tmi.OwnerItem)).DropDownItems; //Make sure that one menu item is always checked. if (tmi.Checked) { //Uncheck all other menu items. foreach (ToolStripMenuItem tmi2 in items) { if (!tmi.Equals(tmi2) && tmi2.Checked) { tmi2.Checked = false; } } } else { //If anything else is already checked, //bail out to prevent a stack overflow. foreach (ToolStripMenuItem tmi2 in items) { if (tmi2.Checked) { return; } } //If nothing is checked, check ourselves again. tmi.Checked = true; } }
Posted in C#, Programming, Windows | Leave a comment

Access a local Tomcat through Apache

Heard a nice ‘trick’ some time ago I need to write about (so I do not forget it ;-).

I wanted to access a tomcat web application on a server that was mainly running a Drupal CMS.

As I do not like shooting holes in firewalls I dislike the use of port 8080 too. The solution I heard was both simple and elegant.

Basically one installs and configures a Tomcat server and deploy a web application on this Tomcat server listening top a port that can only be reached on the server itself (8081 in the example below).

In the Apache ‘httpd.conf’ file you enable mod_proxy by uncommenting the line:

   1: LoadModule proxy_module modules/

Then you define a mapping of the tomcat URL to a URL inside the apache servers web space.


   2: # http://<ip>:8081//?">http://<ip>:8081/<Webapp>/<Webappurl>?<parameters>

   3: # is mapped mapped to: 

   4: # http://<ip>/<Webappurl>?<parameters> 


   6: ProxyPass /OAIHandler

   7: ProxyPassReverse /OAIHandler

In above example:

  • 8081 is the local port to the Tomcat server, not reachable from the outside.
  • is the local address of the server, not reachable from the outside.
  • OAIHandler is a URL inside the web application that does the work (and I re-used it as URL inside the Drupal webspace). URL parameters are nicely appended.
  • LiLiTarget is the name of the web application inside Tomcat’s ‘webapps’ directory.
  • The name of the web application is not used as part of the URL.

Nice side effect of this approach is that you easily swap Tomcat servers or have multiple Tomcat servers running, each with their own port number and each running only a single webapp deployed. This way you can nicely restart/maintain a single Tomcat without affecting the reset of the instances or have multiple Tomcat versions running on a single server.

Posted in Tomcat | Leave a comment

JScript parameters

Ever wondered how to get rid of those pesky *.js.php files where you need PHP to write a single variable into a JScript?

After a lot of searching I stumbled upon a nice solution (see that allows you to pass query parameters to a script in it the SCRIPT’s tag’s SRC attribute just like normal html pages. The problem inside the script is that there are potentially two sets of query data, one for the script (the one we’re after) and one for the page the script is a part of. This last set is access through the usual window.location object.

   1: var scripts = document.getElementsByTagName('script');

   2: var myScript = scripts[ scripts.length - 1 ];


   4: var queryString = myScript.src.replace(/^[^\?]+\??/,'');


   6: var params = parseQuery( queryString );


   8: function parseQuery ( query ) {

   9:    var Params = new Object ();

  10:    if ( ! query ) return Params; // return empty object

  11:    var Pairs = query.split(/[;&]/);

  12:    for ( var i = 0; i < Pairs.length; i++ ) {

  13:       var KeyVal = Pairs[i].split('=');

  14:       if ( ! KeyVal || KeyVal.length != 2 ) continue;

  15:       var key = unescape( KeyVal[0] );

  16:       var val = unescape( KeyVal[1] );

  17:       val = val.replace(/\+/g, ' ');

  18:       Params[key] = val;

  19:    }

  20:    return Params;

  21: }

After which you can simple use it with params[‘course’] or whatever you’re after.

Posted in Uncategorized | Leave a comment

Tomcat running on x64

Today I solved a problem that hunted me for the last couple of days. When I use Eclipse x64 and startup Tomcat, the Tomcat server runs in 32 bits mode.

So I started Googling for answers and found not much more than that it should be supported since version x. No luck thus.

Today I took another approach: I uninstalled all 32 bits Java versions and also the x64 JRE so I only had a single x64 JDK left on my machine. This off-course crashes the 32 bits Eclipse version I still have hanging around but that’s another story.

Starting Tomcat with this configuration still showed a 32 bits Java version. So I started examining the batch files used for Tomcat startup. There I found a curious setenv.bat containing code to reset the JRE_HOME after probing the bundled Tomcat JRE.

The resolution was simple (if the JAVE_HOME environment variable is set correctly to the x64 JDK location), just delete the bundles 32 bits JRE directory and Tomcat will happily start in 64-bits mode.

Nice side-effect is that I know only have to update a single Java version and am freed of yet another old bundled version that isn’t automatically updated.

Posted in Tomcat | Leave a comment

Command line completion for Windows

When using a command shell on Linux another Unix dialect you have the ability to complete a part of a path by pressing tab. I was always under the impression that windows did not have such thing (off-course it does have auto completion in the gui but not at the command line). Big was my surprise when I encountered an obscure option of cmd.exe (the Windows command interpreter/shell). The option, /f (or /f:on or /f:off) lets you toggle path and filename completion on and off under windows when inside a cmd console window. The default keys for completion are CTRL-F for filenames and CTRL-D for directories. The Windows command interpreter even automatically inserts or removes quotes when necessary it encounters spaces in a path or filename. In the registry one can even change the keys for both completions to both the same key (tab for instance to mimic Unix). It’s also possible to set these keys on either a user-by-user basis or system-wide. I only did not find a way to turn it on permanently, but the easiest way around this is to simply add /f:on to a cmd.exe shortcut. The registry keys are:

HKEY_LOCAL_MACHINE\Software\Microsoft\Command Processor\CompletionChar
HKEY_LOCAL_MACHINE\Software\Microsoft\Command Processor\PathCompletionChar

for the logged-on user and:

HKEY_LOCAL_MACHINE\Software\Microsoft\Command Processor\CompletionChar
HKEY_LOCAL_MACHINE\Software\Microsoft\Command Processor\PathCompletionChar

for all users. The values are the character codes in hex. For more information just Google.

Posted in Windows | Leave a comment