Related
Why does the following code sometimes causes an Exception with the contents "CLIPBRD_E_CANT_OPEN":
Clipboard.SetText(str);
This usually occurs the first time the Clipboard is used in the application and not after that.
This is caused by a bug/feature in Terminal Services clipboard (and possible other things) and the .NET implementation of the clipboard. A delay in opening the clipboard causes the error, which usually passes within a few milliseconds.
The solution is to try multiple times within a loop and sleep in between.
for (int i = 0; i < 10; i++)
{
try
{
Clipboard.SetText(str);
return;
}
catch { }
System.Threading.Thread.Sleep(10);
}
Actually, I think this is the fault of the Win32 API.
To set data in the clipboard, you have to open it first. Only one process can have the clipboard open at a time. So, when you check, if another process has the clipboard open for any reason, your attempt to open it will fail.
It just so happens that Terminal Services keeps track of the clipboard, and on older versions of Windows (pre-Vista), you have to open the clipboard to see what's inside... which ends up blocking you. The only solution is to wait until Terminal Services closes the clipboard and try again.
It's important to realize that this is not specific to Terminal Services, though: it can happen with anything. Working with the clipboard in Win32 is a giant race condition. But, since by design you're only supposed to muck around with the clipboard in response to user input, this usually doesn't present a problem.
I know this question is old, but the problem still exists. As mentioned before, this exception occurs when the system clipboard is blocked by another process. Unfortunately, there are many snipping tools, programs for screenshots and file copy tools which can block the Windows clipboard. So you will get the exception every time you try to use Clipboard.SetText(str) when such a tool is installed on your PC.
Solution:
never use
Clipboard.SetText(str);
use instead
Clipboard.SetDataObject(str);
I solved this issue for my own app using native Win32 functions: OpenClipboard(), CloseClipboard() and SetClipboardData().
Below the wrapper class I made. Could anyone please review it and tell if it is correct or not. Especially when the managed code is running as x64 app (I use Any CPU in the project options). What happens when I link to x86 libraries from x64 app?
Thank you!
Here's the code:
public static class ClipboardNative
{
[DllImport("user32.dll")]
private static extern bool OpenClipboard(IntPtr hWndNewOwner);
[DllImport("user32.dll")]
private static extern bool CloseClipboard();
[DllImport("user32.dll")]
private static extern bool SetClipboardData(uint uFormat, IntPtr data);
private const uint CF_UNICODETEXT = 13;
public static bool CopyTextToClipboard(string text)
{
if (!OpenClipboard(IntPtr.Zero)){
return false;
}
var global = Marshal.StringToHGlobalUni(text);
SetClipboardData(CF_UNICODETEXT, global);
CloseClipboard();
//-------------------------------------------
// Not sure, but it looks like we do not need
// to free HGLOBAL because Clipboard is now
// responsible for the copied data. (?)
//
// Otherwise the second call will crash
// the app with a Win32 exception
// inside OpenClipboard() function
//-------------------------------------------
// Marshal.FreeHGlobal(global);
return true;
}
}
Actually there could be another issue at hand. The framework call (both the WPF and winform flavors) to something like this (code is from reflector):
private static void SetDataInternal(string format, object data)
{
bool flag;
if (IsDataFormatAutoConvert(format))
{
flag = true;
}
else
{
flag = false;
}
IDataObject obj2 = new DataObject();
obj2.SetData(format, data, flag);
SetDataObject(obj2, true);
}
Note that SetDataObject is always called with true in this case.
Internally that triggers two calls to the win32 api, one to set the data and one to flush it from your app so it's available after the app closes.
I've seen several apps (some chrome plugin, and a download manager) that listen to the clipboard event. As soon as the first call hits, the app will open the clipboard to look into the data, and the second call to flush will fail.
Haven't found a good solution except to write my own clipboard class that uses direct win32 API or to call setDataObject directly with false for keeping data after the app closes.
Use the WinForms version (yes, there is no harm using WinForms in WPF applications), it handles everything you need:
System.Windows.Forms.Clipboard.SetDataObject(yourText, true, 10, 100);
This will attempt to copy yourText to the clipboard, it remains after your app exists, will attempt up to 10 times, and will wait 100ms between each attempt.
Ref. https://learn.microsoft.com/en-us/dotnet/api/system.windows.forms.clipboard.setdataobject?view=netframework-4.7.2#System_Windows_Forms_Clipboard_SetDataObject_System_Object_System_Boolean_System_Int32_System_Int32_
This happen to me in my WPF application. I got OpenClipboard Failed (Exception from HRESULT: 0x800401D0 (CLIPBRD_E_CANT_OPEN)).
i use
ApplicationCommands.Copy.Execute(null, myDataGrid);
solution is to clear the clipboard first
Clipboard.Clear();
ApplicationCommands.Copy.Execute(null, myDataGrid);
The difference between Cliboard.SetText and Cliboard.SetDataObject in WPF is that the text is not copied to the clipboard, only the pointer. I checked the source code. If we call SetDataObject(data, true) Clipoard.Flush() will also be called. Thanks to this, text or data is available even after closing the application. I think Windows applications only call Flush() when they are shutting down. Thanks to this, it saves memory and at the same time gives access to data without an active application.
Copy to clipboard:
IDataObject CopyStringToClipboard(string s)
{
var dataObject = new DataObject(s);
Clipboard.SetDataObject(dataObject, false);
return dataObject;
}
Code when app or window is closed:
try
{
if ((clipboardData != null) && Clipboard.IsCurrent(clipboardData))
Clipboard.Flush();
}
catch (COMException ex) {}
clipboardData is a window class field or static variable.
That's not a solution, just some additional information on how to reproduce it when all solutions work on your PC and fail somewhere else. As mentioned in the accepted answer - clipboard can be busy by some other app. You just need to handle this failure properly, to explain user somehow why it does not work.
So, just create a new console app with few lines below and run it. And while it is running - test your primary app on how it is handles busy clipboard:
using System;
using System.Runtime.InteropServices;
namespace Clipboard
{
class Program
{
[DllImport("user32.dll")]
private static extern bool OpenClipboard(IntPtr hWndNewOwner);
[DllImport("user32.dll")]
private static extern bool CloseClipboard();
static void Main(string[] args)
{
bool res = OpenClipboard(IntPtr.Zero);
Console.Write(res);
Console.Read();
CloseClipboard();
}
}
}
I want to make a C# program that can be run as a CLI or GUI application depending on what flags are passed into it. Can this be done?
I have found these related questions, but they don't exactly cover my situation:
How to write to the console in a GUI application
How do I get console output in C++ with a Windows program?
Jdigital's answer points to Raymond Chen's blog, which explains why you can't have an application that's both a console program and a non-console* program: The OS needs to know before the program starts running which subsystem to use. Once the program has started running, it's too late to go back and request the other mode.
Cade's answer points to an article about running a .Net WinForms application with a console. It uses the technique of calling AttachConsole after the program starts running. This has the effect of allowing the program to write back to the console window of the command prompt that started the program. But the comments in that article point out what I consider to be a fatal flaw: The child process doesn't really control the console. The console continues accepting input on behalf of the parent process, and the parent process is not aware that it should wait for the child to finish running before using the console for other things.
Chen's article points to an article by Junfeng Zhang that explains a couple of other techniques.
The first is what devenv uses. It works by actually having two programs. One is devenv.exe, which is the main GUI program, and the other is devenv.com, which handles console-mode tasks, but if it's used in a non-console-like manner, it forwards its tasks to devenv.exe and exits. The technique relies on the Win32 rule that com files get chosen ahead of exe files when you type a command without the file extension.
There's a simpler variation on this that the Windows Script Host does. It provides two completely separate binaries, wscript.exe and cscript.exe. Likewise, Java provides java.exe for console programs and javaw.exe for non-console programs.
Junfeng's second technique is what ildasm uses. He quotes the process that ildasm's author went through when making it run in both modes. Ultimately, here's what it does:
The program is marked as a console-mode binary, so it always starts out with a console. This allows input and output redirection to work as normal.
If the program has no console-mode command-line parameters, it re-launches itself.
It's not enough to simply call FreeConsole to make the first instance cease to be a console program. That's because the process that started the program, cmd.exe, "knows" that it started a console-mode program and is waiting for the program to stop running. Calling FreeConsole would make ildasm stop using the console, but it wouldn't make the parent process start using the console.
So the first instance restarts itself (with an extra command-line parameter, I suppose). When you call CreateProcess, there are two different flags to try, DETACHED_PROCESS and CREATE_NEW_CONSOLE, either of which will ensure that the second instance will not be attached to the parent console. After that, the first instance can terminate and allow the command prompt to resume processing commands.
The side effect of this technique is that when you start the program from a GUI interface, there will still be a console. It will flash on the screen momentarily and then disappear.
The part in Junfeng's article about using editbin to change the program's console-mode flag is a red herring, I think. Your compiler or development environment should provide a setting or option to control which kind of binary it creates. There should be no need to modify anything afterward.
The bottom line, then, is that you can either have two binaries, or you can have a momentary flicker of a console window. Once you decide which is the lesser evil, you have your choice of implementations.
* I say non-console instead of GUI because otherwise it's a false dichotomy. Just because a program doesn't have a console doesn't mean it has a GUI. A service application is a prime example. Also, a program can have a console and windows.
Check out Raymond's blog on this topic:
https://devblogs.microsoft.com/oldnewthing/20090101-00/?p=19643
His first sentence: "You can't, but you can try to fake it."
http://www.csharp411.com/console-output-from-winforms-application/
Just check the command line arguments before the WinForms Application. stuff.
I should add that in .NET it is RIDICULOUSLY easy to simply make a console and GUI projects in the same solution which share all their assemblies except main. And in this case, you could make the command line version simply launch the GUI version if it is launched with no parameters. You would get a flashing console.
There is an easy way to do what you want. I'm always using it when writing apps that should have both a CLI and a GUI. You have to set your "OutputType" to "ConsoleApplication" for this to work.
class Program {
[DllImport("kernel32.dll", EntryPoint = "GetConsoleWindow")]
private static extern IntPtr _GetConsoleWindow();
/// <summary>
/// The main entry point for the application.
/// </summary>
[STAThread]
static void Main(string[] args) {
Application.EnableVisualStyles();
Application.SetCompatibleTextRenderingDefault(false);
/*
* This works as following:
* First we look for command line parameters and if there are any of them present, we run the CLI version.
* If there are no parameters, we try to find out if we are run inside a console and if so, we spawn a new copy of ourselves without a console.
* If there is no console at all, we show the GUI.
* We make an exception if we find out, that we're running inside visual studio to allow for easier debugging the GUI part.
* This way we're both a CLI and a GUI.
*/
if (args != null && args.Length > 0) {
// execute CLI - at least this is what I call, passing the given args.
// Change this call to match your program.
CLI.ParseCommandLineArguments(args);
} else {
var consoleHandle = _GetConsoleWindow();
// run GUI
if (consoleHandle == IntPtr.Zero || AppDomain.CurrentDomain.FriendlyName.Contains(".vshost"))
// we either have no console window or we're started from within visual studio
// This is the form I usually run. Change it to match your code.
Application.Run(new MainForm());
else {
// we found a console attached to us, so restart ourselves without one
Process.Start(new ProcessStartInfo(Assembly.GetEntryAssembly().Location) {
CreateNoWindow = true,
UseShellExecute = false
});
}
}
}
I think the preferred technique is what Rob called the devenv technique of using two executables: a launcher ".com" and the original ".exe". This is not that tricky to use if you have the boilerplate code to work with (see below link).
The technique uses tricks to have that ".com" be a proxy for the stdin/stdout/stderr and launch the same-named .exe file. This give the behavior of allowing the program to preform in a command line mode when called form a console (potentially only when certain command-line arguments are detected) while still being able to launch as a GUI application free of a console.
I hosted a project called dualsubsystem on Google Code that updates an old codeguru solution of this technique and provides the source code and working example binaries.
Here is what I believe to be the simple .NET C# solution to the problem. Just to restate the problem, when you run the console "version" of the app from a command line with a switch, the console keeps waiting (it doesn't return to the command prompt and the process keeps running) even if you have an Environment.Exit(0) at the end of your code. To fix this, just before calling Environment.Exit(0), call this:
SendKeys.SendWait("{ENTER}");
Then the console gets the final Enter key it needs to return to the command prompt and the process ends. Note: Don't call SendKeys.Send(), or the app will crash.
It's still necessary to call AttachConsole() as mentioned in many posts, but with this I get no command window flicker when launching the WinForm version of the app.
Here's the entire code in a sample app I created (without the WinForms code):
using System;
using System.Windows.Forms;
using System.Runtime.InteropServices;
namespace ConsoleWriter
{
static class Program
{
[DllImport("kernel32.dll")]
private static extern bool AttachConsole(int dwProcessId);
private const int ATTACH_PARENT_PROCESS = -1;
[STAThread]
static void Main(string[] args)
{
if(args.Length > 0 && args[0].ToUpperInvariant() == "/NOGUI")
{
AttachConsole(ATTACH_PARENT_PROCESS);
Console.WriteLine(Environment.NewLine + "This line prints on console.");
Console.WriteLine("Exiting...");
SendKeys.SendWait("{ENTER}");
Environment.Exit(0);
}
else
{
Application.EnableVisualStyles();
Application.SetCompatibleTextRenderingDefault(false);
Application.Run(new Form1());
}
}
}
}
Hope it helps someone from also spending days on this problem. Thanks for the hint go to #dantill.
/*
** dual.c Runs as both CONSOLE and GUI app in Windows.
**
** This solution is based on the "Momentary Flicker" solution that Robert Kennedy
** discusses in the highest-rated answer (as of Jan 2013), i.e. the one drawback
** is that the console window will briefly flash up when run as a GUI. If you
** want to avoid this, you can create a shortcut to the executable and tell the
** short cut to run minimized. That will minimize the console window (which then
** immediately quits), but not the GUI window. If you want the GUI window to
** also run minimized, you have to also put -minimized on the command line.
**
** Tested under MinGW: gcc -o dual.exe dual.c -lgdi32
**
*/
#include <windows.h>
#include <stdio.h>
static int my_win_main(HINSTANCE hInstance,int argc,char *argv[],int iCmdShow);
static LRESULT CALLBACK WndProc(HWND hwnd,UINT iMsg,WPARAM wParam,LPARAM lParam);
static int win_started_from_console(void);
static BOOL CALLBACK find_win_by_procid(HWND hwnd,LPARAM lp);
int main(int argc,char *argv[])
{
HINSTANCE hinst;
int i,gui,relaunch,minimized,started_from_console;
/*
** If not run from command-line, or if run with "-gui" option, then GUI mode
** Otherwise, CONSOLE app.
*/
started_from_console = win_started_from_console();
gui = !started_from_console;
relaunch=0;
minimized=0;
/*
** Check command options for forced GUI and/or re-launch
*/
for (i=1;i<argc;i++)
{
if (!strcmp(argv[i],"-minimized"))
minimized=1;
if (!strcmp(argv[i],"-gui"))
gui=1;
if (!strcmp(argv[i],"-gui-"))
gui=0;
if (!strcmp(argv[i],"-relaunch"))
relaunch=1;
}
if (!gui && !relaunch)
{
/* RUN AS CONSOLE APP */
printf("Console app only.\n");
printf("Usage: dual [-gui[-]] [-minimized].\n\n");
if (!started_from_console)
{
char buf[16];
printf("Press <Enter> to exit.\n");
fgets(buf,15,stdin);
}
return(0);
}
/* GUI mode */
/*
** If started from CONSOLE, but want to run in GUI mode, need to re-launch
** application to completely separate it from the console that started it.
**
** Technically, we don't have to re-launch if we are not started from
** a console to begin with, but by re-launching we can avoid the flicker of
** the console window when we start if we start from a shortcut which tells
** us to run minimized.
**
** If the user puts "-minimized" on the command-line, then there's
** no point to re-launching when double-clicked.
*/
if (!relaunch && (started_from_console || !minimized))
{
char exename[256];
char buf[512];
STARTUPINFO si;
PROCESS_INFORMATION pi;
GetStartupInfo(&si);
GetModuleFileNameA(NULL,exename,255);
sprintf(buf,"\"%s\" -relaunch",exename);
for (i=1;i<argc;i++)
{
if (strlen(argv[i])+3+strlen(buf) > 511)
break;
sprintf(&buf[strlen(buf)]," \"%s\"",argv[i]);
}
memset(&pi,0,sizeof(PROCESS_INFORMATION));
memset(&si,0,sizeof(STARTUPINFO));
si.cb = sizeof(STARTUPINFO);
si.dwX = 0; /* Ignored unless si.dwFlags |= STARTF_USEPOSITION */
si.dwY = 0;
si.dwXSize = 0; /* Ignored unless si.dwFlags |= STARTF_USESIZE */
si.dwYSize = 0;
si.dwFlags = STARTF_USESHOWWINDOW;
si.wShowWindow = SW_SHOWNORMAL;
/*
** Note that launching ourselves from a console will NOT create new console.
*/
CreateProcess(exename,buf,0,0,1,DETACHED_PROCESS,0,NULL,&si,&pi);
return(10); /* Re-launched return code */
}
/*
** GUI code starts here
*/
hinst=GetModuleHandle(NULL);
/* Free the console that we started with */
FreeConsole();
/* GUI call with functionality of WinMain */
return(my_win_main(hinst,argc,argv,minimized ? SW_MINIMIZE : SW_SHOWNORMAL));
}
static int my_win_main(HINSTANCE hInstance,int argc,char *argv[],int iCmdShow)
{
HWND hwnd;
MSG msg;
WNDCLASSEX wndclass;
static char *wintitle="GUI Window";
wndclass.cbSize = sizeof (wndclass) ;
wndclass.style = CS_HREDRAW | CS_VREDRAW;
wndclass.lpfnWndProc = WndProc;
wndclass.cbClsExtra = 0 ;
wndclass.cbWndExtra = 0 ;
wndclass.hInstance = hInstance;
wndclass.hIcon = NULL;
wndclass.hCursor = NULL;
wndclass.hbrBackground = NULL;
wndclass.lpszMenuName = NULL;
wndclass.lpszClassName = wintitle;
wndclass.hIconSm = NULL;
RegisterClassEx (&wndclass) ;
hwnd = CreateWindowEx(WS_EX_OVERLAPPEDWINDOW,wintitle,0,
WS_VISIBLE|WS_OVERLAPPEDWINDOW,
100,100,400,200,NULL,NULL,hInstance,NULL);
SetWindowText(hwnd,wintitle);
ShowWindow(hwnd,iCmdShow);
while (GetMessage(&msg,NULL,0,0))
{
TranslateMessage(&msg);
DispatchMessage(&msg);
}
return(msg.wParam);
}
static LRESULT CALLBACK WndProc (HWND hwnd,UINT iMsg,WPARAM wParam,LPARAM lParam)
{
if (iMsg==WM_DESTROY)
{
PostQuitMessage(0);
return(0);
}
return(DefWindowProc(hwnd,iMsg,wParam,lParam));
}
static int fwbp_pid;
static int fwbp_count;
static int win_started_from_console(void)
{
fwbp_pid=GetCurrentProcessId();
if (fwbp_pid==0)
return(0);
fwbp_count=0;
EnumWindows((WNDENUMPROC)find_win_by_procid,0L);
return(fwbp_count==0);
}
static BOOL CALLBACK find_win_by_procid(HWND hwnd,LPARAM lp)
{
int pid;
GetWindowThreadProcessId(hwnd,(LPDWORD)&pid);
if (pid==fwbp_pid)
fwbp_count++;
return(TRUE);
}
I have written up an alternative approach which avoids the console flash. See How to create a Windows program that works both as a GUI and console application.
Run AllocConsole() in a static constructor works for me
In my application I have a lot of Console#WriteLine or Console#Write calls. This are so many, that the application is really slower than necessary.
So I'm looking for an easy way to remove or at least disable all Console#WriteLine and Console#Write calls.
What's the easiest way to achieve that?
Here's a even quicker quick fix I implemented.
...
static int Main(String[] args)
{
...
#if !DEBUG
Console.SetOut(TextWriter.Null);
Console.SetError(TextWriter.Null);
#endif
...
}
HTH
Maybe Find&Replace function in any code editor?
For example find all
Console.WriteLine
and replace with
//Console.WriteLine
To switch off Console.WriteLine() output in runtime temporarily use
// backup the previous output handler connected to Console
TextWriter backupOut = Console.Out;
// activate a null handle
Console.SetOut(TextWriter.Null);
// this console output will be invisible
Console.WriteLine("Hidden output on Console.");
// restore the previous handle
Console.SetOut(backupOut);
// this console output will be visible
Console.WriteLine("Showing output on Console.");
Similarly, you can disable program Debug.WriteLine() output in runtime, also while compiling in Debug mode, by using a dummy TextWriterTraceListener, like so
using System.Diagnostics;
using System.IO;
//...
public class MyDummyListener: TextWriterTraceListener
{ }
// ......
[STAThread]
static void Main()
{
// Back up the old one
DefaultTraceListener[] debuglisteners = {(DefaultTraceListener) Debug.Listeners[0]};
// Plug in dummy listener
TextWriterTraceListener[] dummylisteners = {new MyDummyListener()};
Debug.Listeners.Clear();
Debug.Listeners.AddRange(dummylisteners);
// this is an invisible debug message
Debug.WriteLine("This one is invisible");
// ..
// to activate output again, plug in the previous listener
Debug.Listeners.Clear();
Debug.Listeners.AddRange(debuglisteners);
Debug.WriteLine("This one is visible again");
// ....
}
Note: I tested above code in .NET Framework 4.7.2 in Debug mode, using the Winforms platform.
If they are non-essential (logging) then you should have used System.Diagnostics.Debug.Print() to start with.
Luckily WriteLine() is compatible with Debug.Print() so that's an easy S&R. And fix some usings maybe.
Replacing Console.Write() might be a little trickier.
To be complete: The Debug.Print() statements can be turned on/off with a checkbox in project|Properties.
Use a simple, non regex, Find/Replace dialog, and replace every Console.WriteLine( by Debug.WriteLine(.
You can then keep the possibility to track what was previously outputted directly to the console window, still optimizing performance in release mode.
In Visual Studio, use the Tools|Macros|Record temporary macro option to record a macro that does a Find of "Console.Write", and deletes the line. That is:
Ctrl+F to find "Console.Write"
then Ctrl+L to delete the line.
Save the macro and then run it against every file in the project that has the offending lines.
Should take about two minutes.
I would suggest, however, that you back up your source first. Just in case.
I have a simple application that I would like to sort of automate via switches. But when I do run it via switches I don't really want a user interface showing. I just want it to run, do it's job, print stuff out in the console, and exit. On the other hand if I don't run it with any switches I want the user interface to pop up. And in this case I don't really want a console window hanging around in the background.
Is there any way I can do this, or do I have to create two separate projects, one Console Application and one Windows Application?
Whilst not exactly what you have asked, I've achieved the appearance of this behaviour in the past by using the FreeConsole pInvoke to remove the console window.
You set the output type of the project to be a console application. You then define the extern call to FreeConsole:
[DllImport("kernel32.dll", SetLastError=true)]
private static extern int FreeConsole();
Then, in you Main method you switch based on your conditions. If you want a UI, call FreeConsole before opening the form to clear the console window.
if (asWinForms)
{
FreeConsole();
Application.Run(new MainForm());
}
else
{
// console logic here
}
A console window does briefly appear at startup, but in my case it was acceptable.
This is a bit of a hack though and has a bad smell, so I'd seriously consider whether you do want to go down this route.
From 'The Old New Thing'
How do I write a program that can be run either as a console or a GUI application?
You can't.
(I'll let you click on the article for the details of how to fake it)
Sure, just put a switch statement (or if else construction) in the static main(string[] args), based on arguments passed in command line. I also do this to switch between executing as a service or as a console...
NOTE: Set project type as Console App
[DllImport("kernel32.dll", SetLastError=true)]
private static extern int FreeConsole();
[STAThread]
static void Main(string[] args)
{
if (args.Length == 0 && args[0] == "C") // Console
{
// Run as console code
Console.WriteLine("Running as Console App");
Console.WriteLine("Hit any Key to exit");
Console.ReadLine();
}
else
{
//Console.SetWindowSize(1,1);
//string procName = Assembly.GetExecutingAssembly().FullName;
//ProcessStartInfo info = new ProcessStartInfo(procName );
//info.WindowStyle = ProcessWindowStyle.Minimized;
// EDIT: Thanks to Adrian Bank's answer -
// a better approach is to use FreeConsole()
FreeConsole();
Application.Run(new MyForm());
}
}
EDIT: Thanks to Adrian Bank's answer, FreeConsole() is much better approach to "dispense" with Console window than just minimizing it...
They are two different paradigms, I don't think that using a command line switch like this is a good idea. Why not build the core logic into a console application and then call that from the GUI when needed? This would nicely separate the UI from the implementation but would still provide a way to use the Console app stand alone when needed.
I believe the answer is no, or it was last time I looked into this problem.
The executable is marked as either a windowed application or a console application. You can see this in the properties for you project in Visual Studio, under application, Output type
You could simulate the behavior by having two application, a console application that if executed with no arguments launches the GUI application. You may see a console window flash, unless you ran in from an already open console.
Without implementing your own version of a console window the answer is no. When Windows loads you executable it decides whether or not to give you a console window based on data in the PE header. So you can make a windowed app not have a window but you can't make a windoed app have a console.
You can, but with some drawbacks:
You can prevent having this black window on start up if you compile for subsystem Windows.
But then you have to attach the process to the calling console (cmd.exe) manually via AttachConsole(-1)
http://msdn.microsoft.com/en-us/library/windows/desktop/ms681952%28v=vs.85%29.aspx
This alone does not do the job. You also have to redirect the three std streams to the console via these calls:
// redirect unbuffered STDOUT to the console
lStdHandle = (long)GetStdHandle(STD_OUTPUT_HANDLE);
hConHandle = _open_osfhandle(lStdHandle, _O_TEXT);
fp = _fdopen( hConHandle, "w" );
*stdout = *fp;
setvbuf( stdout, NULL, _IONBF, 0 );
fp = _fdopen( hConHandle, "r" );
*stdin = *fp;
setvbuf( stdin, NULL, _IONBF, 0 );
// redirect unbuffered STDERR to the console
lStdHandle = (long)GetStdHandle(STD_ERROR_HANDLE);
hConHandle = _open_osfhandle(lStdHandle, _O_TEXT);
fp = _fdopen( hConHandle, "w" );
*stderr = *fp;
setvbuf( stderr, NULL, _IONBF, 0 );
// make cout, wcout, cin, wcin, wcerr, cerr, wclog and clog
// point to console as well
ios::sync_with_stdio();
Sample from: http://cygwin.com/ml/cygwin/2004-05/msg00215.html
The problem with your WinMain call is that windows already has forked out your process so the calling cmd.exe console will have returned from your .exe already and proceed with the next command.
To prevent that you can call your exe with start /wait myexe.exe
This way you also get the return value of your app and you can check it with %errorlevel% as usual.
If there is a way to prevent that process forking with subsystem windows please let me know.
Hope this helps.
I want to make a C# program that can be run as a CLI or GUI application depending on what flags are passed into it. Can this be done?
I have found these related questions, but they don't exactly cover my situation:
How to write to the console in a GUI application
How do I get console output in C++ with a Windows program?
Jdigital's answer points to Raymond Chen's blog, which explains why you can't have an application that's both a console program and a non-console* program: The OS needs to know before the program starts running which subsystem to use. Once the program has started running, it's too late to go back and request the other mode.
Cade's answer points to an article about running a .Net WinForms application with a console. It uses the technique of calling AttachConsole after the program starts running. This has the effect of allowing the program to write back to the console window of the command prompt that started the program. But the comments in that article point out what I consider to be a fatal flaw: The child process doesn't really control the console. The console continues accepting input on behalf of the parent process, and the parent process is not aware that it should wait for the child to finish running before using the console for other things.
Chen's article points to an article by Junfeng Zhang that explains a couple of other techniques.
The first is what devenv uses. It works by actually having two programs. One is devenv.exe, which is the main GUI program, and the other is devenv.com, which handles console-mode tasks, but if it's used in a non-console-like manner, it forwards its tasks to devenv.exe and exits. The technique relies on the Win32 rule that com files get chosen ahead of exe files when you type a command without the file extension.
There's a simpler variation on this that the Windows Script Host does. It provides two completely separate binaries, wscript.exe and cscript.exe. Likewise, Java provides java.exe for console programs and javaw.exe for non-console programs.
Junfeng's second technique is what ildasm uses. He quotes the process that ildasm's author went through when making it run in both modes. Ultimately, here's what it does:
The program is marked as a console-mode binary, so it always starts out with a console. This allows input and output redirection to work as normal.
If the program has no console-mode command-line parameters, it re-launches itself.
It's not enough to simply call FreeConsole to make the first instance cease to be a console program. That's because the process that started the program, cmd.exe, "knows" that it started a console-mode program and is waiting for the program to stop running. Calling FreeConsole would make ildasm stop using the console, but it wouldn't make the parent process start using the console.
So the first instance restarts itself (with an extra command-line parameter, I suppose). When you call CreateProcess, there are two different flags to try, DETACHED_PROCESS and CREATE_NEW_CONSOLE, either of which will ensure that the second instance will not be attached to the parent console. After that, the first instance can terminate and allow the command prompt to resume processing commands.
The side effect of this technique is that when you start the program from a GUI interface, there will still be a console. It will flash on the screen momentarily and then disappear.
The part in Junfeng's article about using editbin to change the program's console-mode flag is a red herring, I think. Your compiler or development environment should provide a setting or option to control which kind of binary it creates. There should be no need to modify anything afterward.
The bottom line, then, is that you can either have two binaries, or you can have a momentary flicker of a console window. Once you decide which is the lesser evil, you have your choice of implementations.
* I say non-console instead of GUI because otherwise it's a false dichotomy. Just because a program doesn't have a console doesn't mean it has a GUI. A service application is a prime example. Also, a program can have a console and windows.
Check out Raymond's blog on this topic:
https://devblogs.microsoft.com/oldnewthing/20090101-00/?p=19643
His first sentence: "You can't, but you can try to fake it."
http://www.csharp411.com/console-output-from-winforms-application/
Just check the command line arguments before the WinForms Application. stuff.
I should add that in .NET it is RIDICULOUSLY easy to simply make a console and GUI projects in the same solution which share all their assemblies except main. And in this case, you could make the command line version simply launch the GUI version if it is launched with no parameters. You would get a flashing console.
There is an easy way to do what you want. I'm always using it when writing apps that should have both a CLI and a GUI. You have to set your "OutputType" to "ConsoleApplication" for this to work.
class Program {
[DllImport("kernel32.dll", EntryPoint = "GetConsoleWindow")]
private static extern IntPtr _GetConsoleWindow();
/// <summary>
/// The main entry point for the application.
/// </summary>
[STAThread]
static void Main(string[] args) {
Application.EnableVisualStyles();
Application.SetCompatibleTextRenderingDefault(false);
/*
* This works as following:
* First we look for command line parameters and if there are any of them present, we run the CLI version.
* If there are no parameters, we try to find out if we are run inside a console and if so, we spawn a new copy of ourselves without a console.
* If there is no console at all, we show the GUI.
* We make an exception if we find out, that we're running inside visual studio to allow for easier debugging the GUI part.
* This way we're both a CLI and a GUI.
*/
if (args != null && args.Length > 0) {
// execute CLI - at least this is what I call, passing the given args.
// Change this call to match your program.
CLI.ParseCommandLineArguments(args);
} else {
var consoleHandle = _GetConsoleWindow();
// run GUI
if (consoleHandle == IntPtr.Zero || AppDomain.CurrentDomain.FriendlyName.Contains(".vshost"))
// we either have no console window or we're started from within visual studio
// This is the form I usually run. Change it to match your code.
Application.Run(new MainForm());
else {
// we found a console attached to us, so restart ourselves without one
Process.Start(new ProcessStartInfo(Assembly.GetEntryAssembly().Location) {
CreateNoWindow = true,
UseShellExecute = false
});
}
}
}
I think the preferred technique is what Rob called the devenv technique of using two executables: a launcher ".com" and the original ".exe". This is not that tricky to use if you have the boilerplate code to work with (see below link).
The technique uses tricks to have that ".com" be a proxy for the stdin/stdout/stderr and launch the same-named .exe file. This give the behavior of allowing the program to preform in a command line mode when called form a console (potentially only when certain command-line arguments are detected) while still being able to launch as a GUI application free of a console.
I hosted a project called dualsubsystem on Google Code that updates an old codeguru solution of this technique and provides the source code and working example binaries.
Here is what I believe to be the simple .NET C# solution to the problem. Just to restate the problem, when you run the console "version" of the app from a command line with a switch, the console keeps waiting (it doesn't return to the command prompt and the process keeps running) even if you have an Environment.Exit(0) at the end of your code. To fix this, just before calling Environment.Exit(0), call this:
SendKeys.SendWait("{ENTER}");
Then the console gets the final Enter key it needs to return to the command prompt and the process ends. Note: Don't call SendKeys.Send(), or the app will crash.
It's still necessary to call AttachConsole() as mentioned in many posts, but with this I get no command window flicker when launching the WinForm version of the app.
Here's the entire code in a sample app I created (without the WinForms code):
using System;
using System.Windows.Forms;
using System.Runtime.InteropServices;
namespace ConsoleWriter
{
static class Program
{
[DllImport("kernel32.dll")]
private static extern bool AttachConsole(int dwProcessId);
private const int ATTACH_PARENT_PROCESS = -1;
[STAThread]
static void Main(string[] args)
{
if(args.Length > 0 && args[0].ToUpperInvariant() == "/NOGUI")
{
AttachConsole(ATTACH_PARENT_PROCESS);
Console.WriteLine(Environment.NewLine + "This line prints on console.");
Console.WriteLine("Exiting...");
SendKeys.SendWait("{ENTER}");
Environment.Exit(0);
}
else
{
Application.EnableVisualStyles();
Application.SetCompatibleTextRenderingDefault(false);
Application.Run(new Form1());
}
}
}
}
Hope it helps someone from also spending days on this problem. Thanks for the hint go to #dantill.
/*
** dual.c Runs as both CONSOLE and GUI app in Windows.
**
** This solution is based on the "Momentary Flicker" solution that Robert Kennedy
** discusses in the highest-rated answer (as of Jan 2013), i.e. the one drawback
** is that the console window will briefly flash up when run as a GUI. If you
** want to avoid this, you can create a shortcut to the executable and tell the
** short cut to run minimized. That will minimize the console window (which then
** immediately quits), but not the GUI window. If you want the GUI window to
** also run minimized, you have to also put -minimized on the command line.
**
** Tested under MinGW: gcc -o dual.exe dual.c -lgdi32
**
*/
#include <windows.h>
#include <stdio.h>
static int my_win_main(HINSTANCE hInstance,int argc,char *argv[],int iCmdShow);
static LRESULT CALLBACK WndProc(HWND hwnd,UINT iMsg,WPARAM wParam,LPARAM lParam);
static int win_started_from_console(void);
static BOOL CALLBACK find_win_by_procid(HWND hwnd,LPARAM lp);
int main(int argc,char *argv[])
{
HINSTANCE hinst;
int i,gui,relaunch,minimized,started_from_console;
/*
** If not run from command-line, or if run with "-gui" option, then GUI mode
** Otherwise, CONSOLE app.
*/
started_from_console = win_started_from_console();
gui = !started_from_console;
relaunch=0;
minimized=0;
/*
** Check command options for forced GUI and/or re-launch
*/
for (i=1;i<argc;i++)
{
if (!strcmp(argv[i],"-minimized"))
minimized=1;
if (!strcmp(argv[i],"-gui"))
gui=1;
if (!strcmp(argv[i],"-gui-"))
gui=0;
if (!strcmp(argv[i],"-relaunch"))
relaunch=1;
}
if (!gui && !relaunch)
{
/* RUN AS CONSOLE APP */
printf("Console app only.\n");
printf("Usage: dual [-gui[-]] [-minimized].\n\n");
if (!started_from_console)
{
char buf[16];
printf("Press <Enter> to exit.\n");
fgets(buf,15,stdin);
}
return(0);
}
/* GUI mode */
/*
** If started from CONSOLE, but want to run in GUI mode, need to re-launch
** application to completely separate it from the console that started it.
**
** Technically, we don't have to re-launch if we are not started from
** a console to begin with, but by re-launching we can avoid the flicker of
** the console window when we start if we start from a shortcut which tells
** us to run minimized.
**
** If the user puts "-minimized" on the command-line, then there's
** no point to re-launching when double-clicked.
*/
if (!relaunch && (started_from_console || !minimized))
{
char exename[256];
char buf[512];
STARTUPINFO si;
PROCESS_INFORMATION pi;
GetStartupInfo(&si);
GetModuleFileNameA(NULL,exename,255);
sprintf(buf,"\"%s\" -relaunch",exename);
for (i=1;i<argc;i++)
{
if (strlen(argv[i])+3+strlen(buf) > 511)
break;
sprintf(&buf[strlen(buf)]," \"%s\"",argv[i]);
}
memset(&pi,0,sizeof(PROCESS_INFORMATION));
memset(&si,0,sizeof(STARTUPINFO));
si.cb = sizeof(STARTUPINFO);
si.dwX = 0; /* Ignored unless si.dwFlags |= STARTF_USEPOSITION */
si.dwY = 0;
si.dwXSize = 0; /* Ignored unless si.dwFlags |= STARTF_USESIZE */
si.dwYSize = 0;
si.dwFlags = STARTF_USESHOWWINDOW;
si.wShowWindow = SW_SHOWNORMAL;
/*
** Note that launching ourselves from a console will NOT create new console.
*/
CreateProcess(exename,buf,0,0,1,DETACHED_PROCESS,0,NULL,&si,&pi);
return(10); /* Re-launched return code */
}
/*
** GUI code starts here
*/
hinst=GetModuleHandle(NULL);
/* Free the console that we started with */
FreeConsole();
/* GUI call with functionality of WinMain */
return(my_win_main(hinst,argc,argv,minimized ? SW_MINIMIZE : SW_SHOWNORMAL));
}
static int my_win_main(HINSTANCE hInstance,int argc,char *argv[],int iCmdShow)
{
HWND hwnd;
MSG msg;
WNDCLASSEX wndclass;
static char *wintitle="GUI Window";
wndclass.cbSize = sizeof (wndclass) ;
wndclass.style = CS_HREDRAW | CS_VREDRAW;
wndclass.lpfnWndProc = WndProc;
wndclass.cbClsExtra = 0 ;
wndclass.cbWndExtra = 0 ;
wndclass.hInstance = hInstance;
wndclass.hIcon = NULL;
wndclass.hCursor = NULL;
wndclass.hbrBackground = NULL;
wndclass.lpszMenuName = NULL;
wndclass.lpszClassName = wintitle;
wndclass.hIconSm = NULL;
RegisterClassEx (&wndclass) ;
hwnd = CreateWindowEx(WS_EX_OVERLAPPEDWINDOW,wintitle,0,
WS_VISIBLE|WS_OVERLAPPEDWINDOW,
100,100,400,200,NULL,NULL,hInstance,NULL);
SetWindowText(hwnd,wintitle);
ShowWindow(hwnd,iCmdShow);
while (GetMessage(&msg,NULL,0,0))
{
TranslateMessage(&msg);
DispatchMessage(&msg);
}
return(msg.wParam);
}
static LRESULT CALLBACK WndProc (HWND hwnd,UINT iMsg,WPARAM wParam,LPARAM lParam)
{
if (iMsg==WM_DESTROY)
{
PostQuitMessage(0);
return(0);
}
return(DefWindowProc(hwnd,iMsg,wParam,lParam));
}
static int fwbp_pid;
static int fwbp_count;
static int win_started_from_console(void)
{
fwbp_pid=GetCurrentProcessId();
if (fwbp_pid==0)
return(0);
fwbp_count=0;
EnumWindows((WNDENUMPROC)find_win_by_procid,0L);
return(fwbp_count==0);
}
static BOOL CALLBACK find_win_by_procid(HWND hwnd,LPARAM lp)
{
int pid;
GetWindowThreadProcessId(hwnd,(LPDWORD)&pid);
if (pid==fwbp_pid)
fwbp_count++;
return(TRUE);
}
I have written up an alternative approach which avoids the console flash. See How to create a Windows program that works both as a GUI and console application.
Run AllocConsole() in a static constructor works for me