How to reference a .NET assembly

What version of .NET is your assembly written in and what version of TC are you running?  There are limitations as to what assemblies can be viewed/utilized based upon version of TC.

Hello Robert,



thanks for the quick reply. I am using current Test Complete version 8.50.618.7.

Actually I am trying to deploy the .dlls described in this thread

http://smartbear.com/forums/forum/post/?mode=singleThread&thread=56c6c0f2-9f49-49ac-9c33-5a8f4d12a1d7

but they do not appear in the intellisense after referencing... Some other assemblies do appear correcly - I just checked in Test Complete.

Maybe I am missing a step somewhere...



Thanks,



Marin

The reason I asked for the version of .NET that the assembly is written in is because TC 8.50 has discontinued support for .NET 1.0 and 1.1.



http://smartbear.com/support/viewarticle/22809/#DiscontinuedSupport



I'm not sure if that affects whether or not you can access assemblies written in those versions.  I do know that, using TC 7, I had difficulties at one point in time accessing .NET 4 assemblies because they weren't supported in TC 7 so perhaps you're running into a similar problem.

Hi Marin,

Although I have added a reference to a .NET assembly (Project Properties -> CLR Bridge -> Browse Files...) and saved the project, I do not see assembly as namespace displayed in intellisense when typing "dotNET. "...

This can happen in the following cases:

* The assembly implements only interfaces and no classes.

* The assembly references another assembly that can be found neither in the GAC, nor in the main assembly's folder, nor in other assembly probing locations.

Actually I am trying to deploy the .dlls described in this thread http://smartbear.com/forums/forum/post/?mode=singleThread&thread=56c6c0f2-9f49-49ac-9c33-5a8f4d12a1d7 but they do not appear in the intellisense after referencing...

Please check out my last answer in this thread and try configuring the CLR Bridge as described. It should work fine after that.

Hello Helen/Robert,



many thanks for your replies.

Unfortunately I could not make it work like Helen suggested - no matter where the files are and how they are referenced, I can not see the classes.



After some futile search on the net, I decided to build my own PDFBox library .NET. I did it with latest available PDFBox 1.6.0 and IKVM 0.46.0.1. Now when referencing the .dll in CLR bridge, I can see all classes and, can instantiate objects and call mehods.This could point in the direction that Robert was mentioning - that version 0.7.3 is not compatible with current version of Test Complete

 

Now, everything seems to work, I can call the methods described in the other thread, except that my stripped text is not a plain text but a consists only of special chars (see attached image). The string length is correct. My first thought was that maybe wrong encoding is causing this, but no matter what encoding is used (UTF-8, UTF-16, ISO-xxxx-x) nothing seems to work... Maybe it is a problem within PDFBox?

Do you have any suggestions?



Many thanks for your help, much appreciated.



Marin

Hi Marin,

Unfortunately I could not make it work like Helen suggested - no matter where the files are and how they are referenced, I can not see the classes.

Perhaps, in your case, TestComplete couldn't find some additional assemblies, to which the PDFBox-0.7.3.dll assembly refers. I just placed the following assemblies (taken from http://sourceforge.net/projects/pdfbox/files/PDFBox/PDFBox-0.7.3/PDFBox-0.7.3.zip/download) to the same folder, then specified a reference to PDFBox-0.7.3.dll in the CLR Bridge project properties and this worked fine with my sample PDF document:



   * PDFBox-0.7.3.dll

   * FontBox-0.1.0-dev.dll

   * IKVM.GNU.Classpath.dll

   * IKVM.Runtime.dll



It is possible that some additional assemblies are required in your case. You can try to place all the assemblies from the PDFBox package to the folder with your PDFBox-0.7.3.dll library.



Furthermore, PDFBox ver. 0.7.3 is quite outdated. I would recommend that you download a later version of the PDFBox package (for instance, ver. 1.6.0 from http://pdfbox.lehmi.de) or build it by yourself as it is described at http://pdfbox.apache.org/userguide/building_pdfbox.html (what you have actually done).



Moreover, since you are using TestComplete ver. 8.50, you can try to use the original Java PDFBox library with your TestComplete project instead of using its .NET version. For this purpose, you should specify the Java Bridge project settings. For more information, please see the Calling Functions From Java Classes help topic.

except that my stripped text is not a plain text but a consists only of special chars (see attached image). The string length is correct. My first thought was that maybe wrong encoding is causing this, but no matter what encoding is used (UTF-8, UTF-16, ISO-xxxx-x) nothing seems to work...

What language is your stripped text written in? The point is that the Watch List panel in TestComplete is actually non-Unicode. Please try specifying the needed language in the Language for non-Unicode programs system option (see Setting the System Language Manually and Setting the System Language From Scripts for details). Does this help?

Maybe it is a problem within PDFBox?

You can try to check this by creating a simple application in C# or VB.NET that loads your PDFBox assembly and performs the same actions as your script code does. Then, you can try to debug your application, take a look at the results of the text processing and compare them with the results you see in the Watch List panel of TestComplete.

Hello Vladimir,

many thanks for your reply and for taking the time to check the deployment of PDFBox.

Unfortunately I can not make it work using any known .NET assemblies out there, also my own built assembly will not work.

I am now trying to use the Java pre-built .jar package from PDFBox site that you suggested.

It is referenced it in JavaBridge as shown in attached screenshot javaBridge.jpg, but still text will not be stripped from PDF document - text.length() will always be 0...

I wonder if you could verify that code in the first two functions is correct (third one will throw an error as specified in the comment) - see attched functions.jpg



Many thanks again for your attention,



Marin

Hi Marin,



Please try the following code snippet:

function foo()

{

  var doc = JavaClasses.org_apache_pdfbox_pdmodel.PDDocument.load_3("C:\\temp\\TEST.pdf");

  var pdfStripper = JavaClasses.org_apache_pdfbox_util.PDFTextStripper.newInstance();

  var text = pdfStripper.getText_2(doc);

  Log.Message(text.length());

}

It works fine with my sample PDF document. Does it help in your case?

Hello Vladimir,



your version does it: now I get the length correctly - thank you very much for the hint.

But: variable "text" will still be empty in Log window when trying to otput it like in (see attached screenshot how an odd tiny rectangle is showing there when hovering over with the mouse):

function foo()

{   

    var doc = JavaClasses.org_apache_pdfbox_pdmodel.PDDocument.load_3("C:\\temp\\TEST.pdf");

    var pdfStripper = JavaClasses.org_apache_pdfbox_util.PDFTextStripper.newInstance();

    var text = pdfStripper.getText_2(doc);

    Log.Message(text.length()); 

    Log.Message(text);

} 

How can I actually see what is the value in the "text" varable?



Many thanks again,



Marin

Hi Marin,

variable "text" will still be empty in Log window when trying to otput it

Actually, the value returned by the getText_2 method is a String object with a number of properties, fields and methods. TestComplete also adds the OleValue special property to this object. I recommend that you call this property to obtain the string stored in your text object. For example:

function foo()

{   

    var doc = JavaClasses.org_apache_pdfbox_pdmodel.PDDocument.load_3("C:\\temp\\TEST.pdf");

    var pdfStripper = JavaClasses.org_apache_pdfbox_util.PDFTextStripper.newInstance();

    var text = pdfStripper.getText_2(doc);

    Log.Message(text.length());

    Log.Message(text.OleValue);

} 

Does this help?



BTW, you can set a breakpoint on one of the last two lines in your foo script routine and try to debug it. When the script execution is stopped at the breakpoint, you can easily inspect your text object via the Evaluate and Inspect dialogs and explore the set of available object properties, fields and methods.