Been using VS2010 express. Writing in C# which i figured was a good middle step towards object oriented languages. One thing that has made my code somewhat repetitive is the inability to use multiple class inheritance in C#. I.e. I can't say class A inherits from class B and class C.
c# is great because you can quickly and easily get windows with buttons textboxes and dials up and running. This is not available in C++ in the express version since the MFC libraries are not included.
Now, I have thought of just desinging a C# front end which saves parameters to a file then execute a c++ which reads the file, runs and then saves a file which i open with another(or the same) c# backend exec to read and play ard with the results. But this would make it cumbersome always executing the whole sequence again if you want to change something. Not to mention debugging, will probably need to have to instances running.
Reading on of the Visual studio 2012 Express for desktop announcment, it stated that "You can also combine C++, C#, and Visual Basic projects into a single solution, making it easy to write a single application using any of the available languages." http://blogs.msdn.com/b/visualstudio/archive/2012/09/12/visual-studio-express-2012-for-windows-desktop-is-here.aspx?PageIndex=3
Now I would be happy with that, after all I dont expect and dont need at this stage to do any wizbank stuff i.e special button functionality/design which is easier and provided with the MFC in C++.
My question is: Has anyone tried this in Visual Studio 2012 in the Express for windows desktop version in Windows 7''? I.e can you combine a c++ and a c# projects which interact, trace the code form one project to another when debugging for example? Are there any special restrictions? I mean if its combining executables only its not much use, but i expect its more than that, but how much more? For example, can an object designed in C# instanciate an object designed in c++ pass it a reference to other objects like classes which hold inputs and outputs or data which are proccessed in the c++ class and still accesible in the c# code to display results etc?
I am asking this before downloading the new express version because I expect it will set me back a couple of months since going from C# to C++ i would think is like going from Visual Basic to C. I wouldnt want to get into all the trouble (i dont mind really but it would be huge step back) to find out i cant "seemlesly" integrate a c# front end with a C++ proccesing solution.
You have three options for interfacing C++ with C#:
pinvoke: You specify function signatures in C# and give them an attribute specifying what DLL they reside in. This is fairly painful to do if you need to pass around any complex types at all.
COM: A C++ DLL would implement a COM object called by the C# code.
C++/CLI: Allows mixing of managed and unmanaged C++ code in a single C++ project. It is very nice for interfacing with other libraries but, in VS 2010 at least, lacks helpful features such as intellisense. If you really wanted to go this route, I would write three projects: Your core C++ code as a static library, your C++/CLI DLL to wrap it, and your C# application.
Related
I am attempting to develop a GUI application for Tails. I'm doing the initial development on Debian 8 since development directly in Tails can be a pain.
I started out using Anjuta, but the documentation is essentially non-existent. The Anjuta website has nothing at all about how Glade is integrated or how to use it. I can't even track down documentation on how to change the main window title. The only tutorial I found has you start a project and build it using the default files that are generated for a GTKmm project.
Is there a good book or online tutorial out there for doing GUI development in Anjuta?
This is maybe not a complete answer, but it's too large to put in as a comment. I use Anjuta fairly regularly, but I share your feeling about the missing documentation (which is, by the way, not unique for Anjuta). I appreciate Anjuta (and Glade) very much, so don't take the following as criticisms on either program.
I would recommend you consider using PyGTK for GUI creation. It is a lot more productive. You can design the GUI in Glade - exactly the same way you would do for C/C++ - and then implement the code in Python, which you can also edit and manage from Anjuta. There are plenty of code examples, for example on the nullege code search engine.
About the work flow in Anjuta (for C/C++). It is based mainly on the Autotools system, so you should really read up a little on make, Makefile, and related tools. Though in principle Anjuta manages this, you will, sooner or later hit a problem, and some knowledge about Autotools will help you a long way (also this tutorial or this one. This slide series is interesting - probably because it is more graphical. There are even some video tutorials, like this one.).
There is no real necessity to use Glade from inside Anjuta. In fact, Glade has passed a long process distancing itself from 'code generation'. It now only contains an XML generator, which can be called separately. I find the screen space left for Glade inside Anjuta insufficient for comfortable work anyway.
So, in conclusion: If you mainly need a GUI, consider Python + Gtk. If you do need C or C++, Anjuta is a great IDE, but look at Gtk Development examples (like this one). Following those, the use of Anjuta should be a lot clearer.
EDIT:
Very useful answer. I have some underlying legacy code that has to be
C++. Is there a way to mix Python and C++ in Anjuta, or do you know of
any guideposts or tutorials for such?
You can open a C++ project in Anjuta - maybe even import you legacy code directly as a Makefile project. You can also add new files to your C/C++ project and create them as Python files. I've never tried to do that though, and I'm not sure how Anjuta would treat them, for example, in the Makefile(s). I don't have large projects mixing languages at the moment, but for small projects, I like 'Geany', because it doesn't get in the way. You do have to maintain the Makefiles manually.
I'm working on a wrapper to compile c++ code having MFC and windows API calls into their linux versions.
The c++ code has the following characteristics:
No GUI component present.
Has a maximum of about 10 MFC classes used mostly for string parsing.
It has lots of windows specific constants used such as HINSTANCE, LPCTSTR and so on.
I'm not allowed to compile using wine in linux. Till now i've come across wxwidgets, it seems quite vast, i doubt if i'll be needing all it's components.
Please share your ideas in creating the wrapper, is there any specific code that is already available which does this task or part of this task ?
There is no automatic or even semi-automatic way to convert an MFC application to wxWidgets. It is, of course, perfectly possible to do it and many, many projects have gone through this transformation but you just need to do it.
See MFC page of wxWiki for some starting points.
I am working on a VC++ DLL that uses SWIG (Simplified Wrapper and Interface Generator) for C#. The DLL does not use ATL or MFC, it is set to use only the Standard Windows Libraries. Now because of some memory issues, I want to output formatted messages from functions to a log file. What are the different way to do that? Is there a pre-existing trace class that I can use?
I know just 3 different ways to do it:
Use one of the existing log libraries for C++. This is the most flexible approach since you can profit from many nice features like formatting, different appenders etc
Use OutputDebugString function. This function sends message to the debugger, but does not put it to any file by default.
Implement your own simple logger
I hope this helps.
I opened hello.cpp. Why isn't there File -> Compile -> hello.cpp? What is an other easy way, if any?
The file must be part of a project before you can compile/build it. Starting with VS 2005 (or maybe 2008), you can use File -> New -> Project From Existing Code to wrap a project around source files.
Or, you could use something like notepad++ and write a script that compiles with the windows 7/Visual studio C++ compile (CL.exe). That could work as well and do what you describe. C++ was meant for medium and large size projects that require performance and space efficiency. If you want to practice you could use the method I described but otherwise I'd recommend using something like Python/Java/PHP/etc to crank out applications that can take the performance hit of an interpreted/JIT compiled language. This will allow you to develop at a rapid pace. I prefer C++ myself but I use the approach very frequently.
The best way its to create an empty project add your hello.cpp then Visual Studio will build your program.
I think that java executables (jar files) are trivial to decompile and get the source code.
What about other languages? .net and all?
Which all languages can compile only to a decompile-able code?
In general, languages like Java, C#, and VB.NET are relatively easy to decompile because they are compiled to an intermediary language, not pure machine language. In their IL form, they retain more metadata than C code does when compiled to machine language.
Technically you aren't getting the original source code out, but a variation on the source code that, when compiled, will give you the compiled code back. It isn't identical to the source code, as things like comments, annotations, and compiler directives usually aren't carried forward into the compiled code.
Managed languages can be easily decompiled because executable must contain a lot of metadata to support reflection.
Languages like C++ can be compiled to native code. Program structure can be totally changed during compilation\translation processes.
Compiler can easily replace\merge\delete parts of your code. There is no 1 to 1 relationship between original and compiled (native) code.
.NET is very easy to decompile. The best tool to do that would be the .NET reflector recently acquired by RedGate.
Most languages can be decompiled but some are easier to decompile than others. .Net and Java put more information about the original program in the executables (method names, variable names etc.) so you get more of your original information back.
C++ for example will translate variables and functions etc. to memory adresses (yeah I know this is a gross simplification) so the decompiler won't know what stuff was called. But you can still get some of the structure of the program back though.
VB6 if compiled to pcode is also possible to decompile to almost full source using P32Dasm, Flash (or actionscript) is also possible to decompile to full source using something like Flare