I've used a free Beanstalk account for hobby development the past couple of years. I recommend it. :thumbsup: You can pick between Subversion, Git, and Mercurial. Subversion is more than fine for a single developer. The single repository is not much of a limitation since you can just put each project in a subfolder.

I'm trying to understand form.resx files. Particularly what is put in them and why. I have found plenty of information on resx files in general, but I am looking specifically to understand the relationship between the Visual Studio form designer and the form.resx files. Most of the little I have seen on the topic says that the file is to facilitate localization and something about storing images for the form. The link to localization is definitely real, since when I select a different location I get another resx file for that location. Based on that, I assume that the default resx file is tied to the default location. The specific issue that has me interested in this topic is as follows: I have created a customizable form framework that allows users to filter a huge list of properties down to the ones that they care about. The initial list of properties is generated by reflecting on the business object that the form edits. Another form allows them to shift properties from a list to show and a list to hide. A custom attribute called ControlDescription can decorate the business objects' properties to specify the type of control to show, and additional information such as a domain to populate values in a drop-down list. A FormConfig object keeps all of the form customization settings, which consist of a series of show and hide property lists for the various business objects. So what does all this have to do with resx files? Well, the forms are serializing the FormConfig object into the resx file. More accurately it is serializing each item in each list into the resx file. These lists are pretty long because the business objects have so many properties, which is why all this madness was developed in the first place. Anyway, I can't figure out why it serializes this particular object into the resx. It seems to be causing problems, because the resx files are getting corrupted on a fairly regular basis. A typical (slightly absurd) error is: Object of type 'Us.Fed.Fs.Nris.Nrvs.BusinessObjects.ControlDescription' cannot be converted to type 'Us.Fed.Fs.Nris.Nrvs.BusinessObjects.ControlDescription'. My hope is, that if I understood what the resx is doing and why, that I can either prevent it from serializing these objects by restructuring the application, or avoid whatever is causing the corruption of the resx files. Thanks, Preston

Thank you for the thorough reply. You have confirmed what I suspected. I will just duplicate the common code into both application projects. It seems like the simplest way to go. - Preston

I have some a solution with a class library project containing some common code and a couple of Windows application projects. I have referenced the Common class library using a project reference. The copy local option is set to true. Setup this way the app.exe file produced by the Windows application projects will only run if the common.dll file is in the same directory. Is there a way, or what is the best way, to compile the common.dll into the exe so that it will be self sufficient and can be distributed without the common.dll?

I have a simple form containing only a DataGridView control. I would like the form to grow or shrink to fit the contents of the DataGridView. Of course the DataGridView would also need to grow and shrink to fit it's contents. Any suggestions?

Well, I actually expected it was the case that I could not fully duplicate the external behavior of a C++ dll, but I wanted to make sure. The C++ wrapper dll occurred to me and would be possible since we have C++ developers, but it is not a good solution. Ultimately the problem was solve in another manner. We are communicating through he standard output stream and it works great. Thanks for everyones comments.

There is a program calling a C++ dll named "ffisamp.dll". The C++ code is as follows: ***** //Simple C function that just puts something in a string // and returns the length of that string both as the RC and as a param // Keeping the C interface as simple as possible. __declspec( dllexport ) int PopulateString ( char *FormsBuffer, int *BuffLen) { int LocalLength; strcpy(FormsBuffer,"A Fixed string from within the C program"); *BuffLen = strlen(FormsBuffer); LocalLength = strlen(FormsBuffer); return LocalLength; } ***** I need C# code that will compile to a dll that mimics the sample dll such that no modifications will need to be made to the calling program. I need to be able to just replace the old dll with the new one. I am most perplexed that the C++ method is not in a class. That might be that the source code is just a snippet of the whole dll's source code... since it comes from a sample, but I have the impression from the documentation that is all of it. Here is a link to the article I am working with. I am trying to communicate with an Oracle 10g form using methods to call C functions. http://www.oracle.com/technology/products/forms/htdocs/webutil/howto_ffi.html[^] I've tried a few things, but so far the calling application doesn't even find the function. :(( I am not familiar with C++, MFC etc... so I am sure I am missing some fundamentals.