How to customize right click with already existing instance of a program? - c++

I need to customize right click so that I can scan a directory with my anti-virus. I know how to do that using registry keys, but the problem is that I don't want to start a new instance of my program every time I want to scan a directory. My anti-virus needs to load some signature databases so it will take around 15 seconds for the program to load those. I need to use the instance of the program which I have already opened and is running for scanning the directory. How can I do that?
I am using C++Builder.
Thanks.

Considering you already know how to add the item to the right click contextual menu, I suggest implementing a client/server set of applications:
A server that loads up when you turn your computer on and does the scanning, and
The client that tells it what to do using IPC - inter-process communication.
You then add the client application to various contextual menus, passing it arguments that indicate what it should get the server to do depending on what you right-clicked on.
IPC is a bit of a pain in the butt, the easiest way is to use TCP/IP to and do local networking using a network library. There are many out there, however given you'll likely want to have other features such as UI elements and a tray icon, I suggest you look at Qt, namely the following components:
QtNetwork: For performing communication between the client and the server executable.
QSystemTrayIcon: For displaying a small icon on the tray.
There are quite a few other little bits of Qt you'll no doubt encounter (like all the fabulous UI stuff), and fortunately Qt is well documented and help is always available here, and from the Qt Developer Network. You can get started with Qt by downloading and installing the SDK:
http://qt.nokia.com/downloads/
Best of luck :).

Implement a DDE server in your anti-virus, and then add a ddeexec subkey to your Registry key. Alternatively, add an OLE Automation object to your app that implements the IDropTarget interface, and then add a DropTarget subkey to your Registry key that specifies the object's CLSID.
Either way, whenever your menu item is then invoked, Windows will call into your existing app instance if it is already running, otherwise it will launch a new instance and then call into it. Either way, Windows is handling all of that for you. All you are doing is providing an entry point for Windows to call into.
I would suggest the IDropTarget method, because DDE is deprecated, and because IDropTarget is more flexible. While your app is running, you could re-use the same IDropTarget object to handle OLE Drag&Drop operations on your app's UI window and Taskbar button, and support automated invokations of your scanner by other apps.

Related

Control c++ application remotely via web interface

I created a C++ QT application on Linux with a simple visual interface with three buttons. This application runs on a machine with a certain IP address in my wifi network. I want to be able to access such visual interface from the browser of a smartphone and click the buttons.
To accomplish this I used a remote desktop connection, but it's just a temporary solution as I want to be able to access my GUI from any smartphone without the need of installing anything, and also without offering other functionalities... the client should be able to press the three buttons and nothing more.
In other words, I would like to be able to do the following:
After I type the IP address of the linux machine in the browser of my smartphone, a html page opens up with my visual interface with three buttons;
When I press one button, my C++ code starts running in the background;
When I press again the buttons, the C++ app receives the commands and acts accordingly.
And so on until I click the button that closes my C++ App.
Now, is there a way to accomplish this? In a few words, is it possible to have a web interface to act as a GUI for my C++ app? I must admit I am really ignorant in web applications :) But maybe you know about a QT widget that solves my problems.
Thanks!
There are a lot of ways to accomplish this.
A simple way is to use QtHttpServer to talk to the objects inside your Qt Application to do the work you asked it to.
You can quickly get started by adapting the example here to your use case.
You may want to check out https://sourceforge.net/projects/conair/ - this lib allows you to communicate with your c++ qt code via javascript. You would, however, have to replicate your existing qt gui with a traditional web front-end technology - some straight forward HTML and javascript would probably suffice.

Managing Windows Though QT Application in Linux (C++)

I've been making a simple application which is able to launch a variety of other applications through QT5 (using QProcess class) but I've been running into a few key issues with the design. Specifically, it seems that Qprocess cannot set focus to windows that have been created via QProcess' start() function. This means that once a user opens more than one window, it can never return to the previous window that has been opened. After looking further into this dilemma, it has become clear that my program will need to be able to handle basic window management in order to fulfill my specifications.
I've decided that the best example to study for my program is Docky, which is capable of opening, closing and switching windows. Looking at the source code for that project was helpful, but there were many C# system calls that were used for fetching the list of client-windows which aren't available for my C++ program.
How can I get a list of all the X11 Windows that the client is running and provide basic window management (Switch To/ Open / Close Window) using C++? Is there a cross platform way of doing this through QT? Can I get this information directly using XServer?

Launching a windowed application from a console application?

I may be approaching this in completely the wrong way as I am pretty new to the C++ language and the overall way this kind of application should be structured, but I hope to confirm the correct method here.
Essentially, I have one cpp file which operates as a console application & a separate cpp file which runs as a windowed application. I want to be able to launch the windowed application when a certain point is reached within the console application. Is this possible? If so, how would I go about doing this?
Some more detail - The console application acts as a 'server' using winsock to communicate with another console application (the client). When the console server application reaches a certain point (a client connects with it) I wish to then launch the other windowed application I have created which will render certain graphics onscreen using Directx. Currently I have both these cpp files as separate projects in a single C++ 2010 Express solution. Currently, there are no links between the two cpp files and they both operate correctly when run separately.
If any more specifics are required, I can provide them but I really want to find out if this approach in general will work.
Thanks.
If you are not running a managed C++ application, then CreateProcess is the canonical WIN32 system call to use.
Do you just want to run the exe from another exe?
System::Diagnostics::Process::Start("C:\\Folder\\file.exe");

Make Non-GUI server application in Visual Studio

I did a project of Server Socket in Visual studio in C++ MFC based. Now, after debugging the project, server GUI opens, then after clicking on the CONNECT button on server GUI, you can connect the clients to that server and so on.
Now I want to use that server exe file in some other computer. So that whenever that computer starts, that server exe automatically starts. so for this i need to disable the connect button, so that after debugging, server GUi opens and connected automatically. But i don't want that server GUI opens in another computer in autostart as well. i want to disable that server GUI.
I got an idea of modalless dialog to work on it. Is it good or what approach should I use ?
You probably want to separate GUI part and server part of your application. Ideally, if your server is actually a server, you should start it as service. Then you will have separate GUI tool to control it.
Another approach is to have command line argument that determines whether server should be started with GUI enabled or disabled.
The only professional and stable solution of such an application is, to splitt it in a console part, which you put under the control of service control and a gui part wich the user can start when he wants.
I tried solutions like yours and so I can tell you from my own experience, that you will face a lot of problems.
However, a possible solution would be to hide the window and lay the app down to systray and this is a very interesting discussion about hinding windows.
Additionaly I have two good advices in case of MFC:
Never ever just "copy-past" code without to know what MFC is doing in the background (Win32api).
Do not use MFC. Have a closer look at QT or wxWidges when you need windows, to shortcut encapsulation of win32api also have a look at boost library. It is realy worth the time you spend on!
Good luck!

C++ Event Hooks

I'm trying to understand event hooks in C++. I know what an event is, I've used them a lot in Java, C# and Javascript.
What I'm having trouble with is finding the documentation, and tutorials on stuff like global hooks, dll injection, global hooks without a DLL.
Lets say that I wanted to iterate through the browser tabis in FireFox .. would I need to hope that FireFox has an API for C++? Or lets say I wanted to do something when a user opens a new tab would I need to use a hook that FireFox would provide in their API?
The above is just an example so people know what I'm trying to learn/understand. Am I thinking on the right ines?
I seen a post on a forum and for the past 2 hours I've took an interest. I always say that a tricky challange, or a new challange, makes a stronger programmer.
Any resources, or any help, would be very much appreciated.
C++ itself does not have events or hooks, but a lot of C++ libraries and frameworks implement them. For an example of generic events library, see Boost.Signals.
Some of the implementations allow their events to be seen by other applications, but the API is application-specific (e.g. for Firefox, see XPCOM).
Windows has a mechanizm of hooks that allows to monitor various events in its windowing system. However, it is an OS feature, not related to C++. As it's a system mechanizm, all Windows applications are affected even if they don't do anything for it. The documentation for Windows hooks can be found here. Also, since you mentioned "global hooks without a DLL", see SetWinEventHook, which is a higher-level API than Windows hooks linked above and can be used with hook functions both implemented in DLLs or EXEs.
Look up MSDN for SetWindowsHookEx. It should be your entrance in Windows hooks. If you ar etargetting a parituclar window for mthe system then a less intrusive option is SetWindowLongPtr. For the first API you are going to need some Dll injection - which gets automatically for you by the system. Just follow these steps:
Create a Dll that exports a HOOKPROC function (actual type dependent upon the hook tpe - read in the docs)
Load that Dll in your application and retrieve a pointer to the HOOKPROC function. LoadLibrary / GetProcAddress APIs may be used for this.
From your application, make a call to SetWindowsHookEx feeding in the appropriate parameters - this will inject the dll in the target process. So, the dll is now loaded into both your app's process and in the target process. So you will need a mechanism to IPC between the two processes probably. Lots of ways here - sockets, pipes, shared segment in DLL, filesystem, windows messages, COM servers + events, etc etc.
The former API, while less powerful, does not require DLL injection.
Choose wisely & good luck!
I dont think firefox would be having a C++ aPI to find the open tabs....
If you want to find out open tabs or whenever a new tab is open , you can basically hook the firefox window and get all events happening on that window to your hook procedure.
If you open spy++ in VC++ and track firefox window , you can see a new MozillaContentWindowClass gets created every time whenever a new tab is opened. So you can basically iterate through window handles and get information about open tabs.
You can use SetWindowLongPtr to set the subclass procedure for that window.