I have a collection of Boost unit tests I want to run as a console application.
When I'm working on the project and I run the tests I would like to be able to debug the tests, and I would like to have the console stay open after the tests run.
I see that if I run in release mode the console window stays up after the program exits, but in debug mode this is not the case.
I do not want to add 'system("pause");' or any other hacks like reading a character to my program. I just want to make Visual Studio pause after running the tests with debugging like it would if I were running in release mode. I would also like it if the output of tests were captured in one of Visual Studio's output windows, but that also seems to be harder than it should be.
How can I do this?
This question is related to
c++
visual-studio
unit-testing
boost
console
Or you could use boost_test "Test Log Output."
http://www.boost.org/doc/libs/1_47_0/libs/test/doc/html/utf/user-guide/test-output/test-log.html
Then it won't matter whether the console window shows up at all AND your build logging can preserve the unit testing output as an artifact for examination on failed builds...
Just use a logging library, like log4net, and have it log to a file appender.
You could also setup your executable as an external tool, and mark the tool for Use output window. That way the output of the tool will be visible within Visual Studio itself, not a separate window.
Do a readline at the end (it's the "forma cochina", like we say in Colombia, but it works):
static void Main(string[] args)
{
.
.
.
String temp = Console.ReadLine();
}
Just use a logging library, like log4net, and have it log to a file appender.
It would actually be more effort, but you could just build in VS.Net, run it from the regular command line (cmd.exe), and then attach to the process after it starts running. This is probably not the solution you are looking for however.
You could also setup your executable as an external tool, and mark the tool for Use output window. That way the output of the tool will be visible within Visual Studio itself, not a separate window.
Adding the following line will do a simple MS-DOS pause
displaying no message.
system("pause >nul | set /p \"=\"");
And there is no need to Ctrl+F5 (which will make your application run in Release Mode)
You say you don't want to use the system("pause")
hack. Why not?
If it's because you don't want the program to prompt when it's not being debugged, there's a way around that. This works for me:
void pause () {
system ("pause");
}
int main (int argc, char ** argv) {
// If "launched", then don't let the console close at the end until
// the user has seen the report.
// (See the MSDN ConGUI sample code)
//
do {
HANDLE hConsoleOutput = ::GetStdHandle (STD_OUTPUT_HANDLE);
if (INVALID_HANDLE_VALUE == hConsoleOutput)
break;
CONSOLE_SCREEN_BUFFER_INFO csbi;
if (0 == ::GetConsoleScreenBufferInfo (hConsoleOutput, &csbi))
break;
if (0 != csbi.dwCursorPosition.X)
break;
if (0 != csbi.dwCursorPosition.Y)
break;
if (csbi.dwSize.X <= 0)
break;
if (csbi.dwSize.Y <= 0)
break;
atexit (pause);
} while (0);
I just paste this code into each new console application I'm writing. If the program is being run from a command window, the cursor position won't be <0,0>, and it won't call atexit()
. If it has been launched from you debugger (any debugger) the console cursor position will be <0,0> and the atexit()
call will be executed.
I got the idea from a sample program that used to be in the MSDN library, but I think it's been deleted.
NOTE: The Microsoft Visual Studio implementation of the system() routine requires the COMSPEC environment variable to identify the command line interpreter. If this environment variable gets messed up -- for example, if you've got a problem in the Visual Studio project's debugging properties so that the environment variables aren't properly passed down when the program is launched -- then it will just fail silently.
Set a breakpoint on the last line of code.
Adding the following line will do a simple MS-DOS pause
displaying no message.
system("pause >nul | set /p \"=\"");
And there is no need to Ctrl+F5 (which will make your application run in Release Mode)
It would actually be more effort, but you could just build in VS.Net, run it from the regular command line (cmd.exe), and then attach to the process after it starts running. This is probably not the solution you are looking for however.
Try to run the application with the Ctrl + F5 combination.
In the older versions it would default to the console subsystem even if you selected "empty project", but not in 2010, so you have to set it manually. To do this select the project in the solution explorer on the right or left (probably is already selected so you don't have to worry about this). Then select "project" from the menu bar drop down menus, then select "project_name properties" > "configuration properties" > "linker" > "system" and set the first property, the drop down "subsystem" property to "console (/SUBSYSTEM:CONSOLE)". The console window should now stay open after execution as usual.
If it is a console application, use Ctrl + F5.
Try to run the application with the Ctrl + F5 combination.
Do a readline at the end (it's the "forma cochina", like we say in Colombia, but it works):
static void Main(string[] args)
{
.
.
.
String temp = Console.ReadLine();
}
You could also setup your executable as an external tool, and mark the tool for Use output window. That way the output of the tool will be visible within Visual Studio itself, not a separate window.
Set a breakpoint on the last line of code.
In the older versions it would default to the console subsystem even if you selected "empty project", but not in 2010, so you have to set it manually. To do this select the project in the solution explorer on the right or left (probably is already selected so you don't have to worry about this). Then select "project" from the menu bar drop down menus, then select "project_name properties" > "configuration properties" > "linker" > "system" and set the first property, the drop down "subsystem" property to "console (/SUBSYSTEM:CONSOLE)". The console window should now stay open after execution as usual.
Prompt for user input.
https://www.youtube.com/watch?v=NIGhjrWLWBo
shows how to do this for C++. For Node.js, this is taken right from the docs (and it works):
'use strict';
console.log('Hello world');
const readline = require('readline');
const rl = readline.createInterface({
input: process.stdin,
output: process.stdout
});
rl.question('Press enter to continue...', (answer) => {
rl.close(); /* discard the answer */
});
You could also setup your executable as an external tool, and mark the tool for Use output window. That way the output of the tool will be visible within Visual Studio itself, not a separate window.
I would use a "wait"-command for a specific time (milliseconds) of your own choice. The application executes until the line you want to inspect and then continues after the time expired.
Include the <time.h>
header:
clock_t wait;
wait = clock();
while (clock() <= (wait + 5000)) // Wait for 5 seconds and then continue
;
wait = 0;
I start the app with F11 and get a breakpoint somewhere in unit_test_main.ipp (can be assembly code). I use shift-f11 (Step out) to run the unit test and get the next assembly instruction in the CRT (normally in mainCRTStartup()). I use F9 to set a breakpoint at that instruction.
On the next invocation, I can start the app with F5 and the app will break after running the tests, therefore giving me a chance to peek at the console window
It would actually be more effort, but you could just build in VS.Net, run it from the regular command line (cmd.exe), and then attach to the process after it starts running. This is probably not the solution you are looking for however.
Set a breakpoint on the last line of code.
If it is a console application, use Ctrl + F5.
In Boost.Test there is the --auto_start_dbg
parameter for breaking into the debugger when a test fails (on an exception or on an assertion failure). For some reason it doesn't work for me.
For this reason I have created my custom test_observer that will break into the debugger when there is an assertion failure or an exception. This is enabled on debug builds when we are running under a debugger.
In one of the source files of my unit test EXE file I have added this code:
#ifdef _DEBUG
#include <boost/test/framework.hpp>
#include <boost/test/test_observer.hpp>
struct BoostUnitTestCrtBreakpointInDebug: boost::unit_test::test_observer
{
BoostUnitTestCrtBreakpointInDebug()
{
boost::unit_test::framework::register_observer(*this);
}
virtual ~BoostUnitTestCrtBreakpointInDebug()
{
boost::unit_test::framework::deregister_observer(*this);
}
virtual void assertion_result( bool passed /* passed */ )
{
if (!passed)
BreakIfInDebugger();
}
virtual void exception_caught( boost::execution_exception const& )
{
BreakIfInDebugger();
}
void BreakIfInDebugger()
{
if (IsDebuggerPresent())
{
/**
* Hello, I know you are here staring at the debugger :)
*
* If you got here then there is an exception in your unit
* test code. Walk the call stack to find the actual cause.
*/
_CrtDbgBreak();
}
}
};
BOOST_GLOBAL_FIXTURE(BoostUnitTestCrtBreakpointInDebug);
#endif
You say you don't want to use the system("pause")
hack. Why not?
If it's because you don't want the program to prompt when it's not being debugged, there's a way around that. This works for me:
void pause () {
system ("pause");
}
int main (int argc, char ** argv) {
// If "launched", then don't let the console close at the end until
// the user has seen the report.
// (See the MSDN ConGUI sample code)
//
do {
HANDLE hConsoleOutput = ::GetStdHandle (STD_OUTPUT_HANDLE);
if (INVALID_HANDLE_VALUE == hConsoleOutput)
break;
CONSOLE_SCREEN_BUFFER_INFO csbi;
if (0 == ::GetConsoleScreenBufferInfo (hConsoleOutput, &csbi))
break;
if (0 != csbi.dwCursorPosition.X)
break;
if (0 != csbi.dwCursorPosition.Y)
break;
if (csbi.dwSize.X <= 0)
break;
if (csbi.dwSize.Y <= 0)
break;
atexit (pause);
} while (0);
I just paste this code into each new console application I'm writing. If the program is being run from a command window, the cursor position won't be <0,0>, and it won't call atexit()
. If it has been launched from you debugger (any debugger) the console cursor position will be <0,0> and the atexit()
call will be executed.
I got the idea from a sample program that used to be in the MSDN library, but I think it's been deleted.
NOTE: The Microsoft Visual Studio implementation of the system() routine requires the COMSPEC environment variable to identify the command line interpreter. If this environment variable gets messed up -- for example, if you've got a problem in the Visual Studio project's debugging properties so that the environment variables aren't properly passed down when the program is launched -- then it will just fail silently.
Try to run the application with the Ctrl + F5 combination.
Prompt for user input.
https://www.youtube.com/watch?v=NIGhjrWLWBo
shows how to do this for C++. For Node.js, this is taken right from the docs (and it works):
'use strict';
console.log('Hello world');
const readline = require('readline');
const rl = readline.createInterface({
input: process.stdin,
output: process.stdout
});
rl.question('Press enter to continue...', (answer) => {
rl.close(); /* discard the answer */
});
I would use a "wait"-command for a specific time (milliseconds) of your own choice. The application executes until the line you want to inspect and then continues after the time expired.
Include the <time.h>
header:
clock_t wait;
wait = clock();
while (clock() <= (wait + 5000)) // Wait for 5 seconds and then continue
;
wait = 0;
You say you don't want to use the system("pause")
hack. Why not?
If it's because you don't want the program to prompt when it's not being debugged, there's a way around that. This works for me:
void pause () {
system ("pause");
}
int main (int argc, char ** argv) {
// If "launched", then don't let the console close at the end until
// the user has seen the report.
// (See the MSDN ConGUI sample code)
//
do {
HANDLE hConsoleOutput = ::GetStdHandle (STD_OUTPUT_HANDLE);
if (INVALID_HANDLE_VALUE == hConsoleOutput)
break;
CONSOLE_SCREEN_BUFFER_INFO csbi;
if (0 == ::GetConsoleScreenBufferInfo (hConsoleOutput, &csbi))
break;
if (0 != csbi.dwCursorPosition.X)
break;
if (0 != csbi.dwCursorPosition.Y)
break;
if (csbi.dwSize.X <= 0)
break;
if (csbi.dwSize.Y <= 0)
break;
atexit (pause);
} while (0);
I just paste this code into each new console application I'm writing. If the program is being run from a command window, the cursor position won't be <0,0>, and it won't call atexit()
. If it has been launched from you debugger (any debugger) the console cursor position will be <0,0> and the atexit()
call will be executed.
I got the idea from a sample program that used to be in the MSDN library, but I think it's been deleted.
NOTE: The Microsoft Visual Studio implementation of the system() routine requires the COMSPEC environment variable to identify the command line interpreter. If this environment variable gets messed up -- for example, if you've got a problem in the Visual Studio project's debugging properties so that the environment variables aren't properly passed down when the program is launched -- then it will just fail silently.
I start the app with F11 and get a breakpoint somewhere in unit_test_main.ipp (can be assembly code). I use shift-f11 (Step out) to run the unit test and get the next assembly instruction in the CRT (normally in mainCRTStartup()). I use F9 to set a breakpoint at that instruction.
On the next invocation, I can start the app with F5 and the app will break after running the tests, therefore giving me a chance to peek at the console window
Set a breakpoint on the last line of code.
I just copied from http://social.msdn.microsoft.com/forums/en-US/Vsexpressvc/thread/1555ce45-8313-4669-a31e-b95b5d28c787/?prof=required:
The following works for me :-)
/////////////////////////////////////////////////////////////////////////////////////
Here is another reason the console may disappear. And the solution:
With the new Visual Studio 2010 you might see this behavior even when you use Ctrl + F5 aka "start without debugging". This is most likely because you created an "empty project" instead of a "Win32 console application". If you create the project as a "Win32 console application" you can disregard this as it does not apply.
In the older versions it would default to the console subsystem even if you selected "empty project", but not in Visual Studio 2010, so you have to set it manually. To do this select the project in the solution explorer on the right or left (probably is already selected so you don't have to worry about this).
Then select "project" from the menu bar drop down menus, then select "project_name properties" ? "configuration properties" ? "linker" ? "system" and set the first property, the drop down "subsystem" property to "console (/SUBSYSTEM:CONSOLE)". The console window should now stay open after execution as usual.
/////////////////////////////////////////////////////////////////////////////////////
Or you could use boost_test "Test Log Output."
http://www.boost.org/doc/libs/1_47_0/libs/test/doc/html/utf/user-guide/test-output/test-log.html
Then it won't matter whether the console window shows up at all AND your build logging can preserve the unit testing output as an artifact for examination on failed builds...
It would actually be more effort, but you could just build in VS.Net, run it from the regular command line (cmd.exe), and then attach to the process after it starts running. This is probably not the solution you are looking for however.
Just use a logging library, like log4net, and have it log to a file appender.
In Boost.Test there is the --auto_start_dbg
parameter for breaking into the debugger when a test fails (on an exception or on an assertion failure). For some reason it doesn't work for me.
For this reason I have created my custom test_observer that will break into the debugger when there is an assertion failure or an exception. This is enabled on debug builds when we are running under a debugger.
In one of the source files of my unit test EXE file I have added this code:
#ifdef _DEBUG
#include <boost/test/framework.hpp>
#include <boost/test/test_observer.hpp>
struct BoostUnitTestCrtBreakpointInDebug: boost::unit_test::test_observer
{
BoostUnitTestCrtBreakpointInDebug()
{
boost::unit_test::framework::register_observer(*this);
}
virtual ~BoostUnitTestCrtBreakpointInDebug()
{
boost::unit_test::framework::deregister_observer(*this);
}
virtual void assertion_result( bool passed /* passed */ )
{
if (!passed)
BreakIfInDebugger();
}
virtual void exception_caught( boost::execution_exception const& )
{
BreakIfInDebugger();
}
void BreakIfInDebugger()
{
if (IsDebuggerPresent())
{
/**
* Hello, I know you are here staring at the debugger :)
*
* If you got here then there is an exception in your unit
* test code. Walk the call stack to find the actual cause.
*/
_CrtDbgBreak();
}
}
};
BOOST_GLOBAL_FIXTURE(BoostUnitTestCrtBreakpointInDebug);
#endif
I just copied from http://social.msdn.microsoft.com/forums/en-US/Vsexpressvc/thread/1555ce45-8313-4669-a31e-b95b5d28c787/?prof=required:
The following works for me :-)
/////////////////////////////////////////////////////////////////////////////////////
Here is another reason the console may disappear. And the solution:
With the new Visual Studio 2010 you might see this behavior even when you use Ctrl + F5 aka "start without debugging". This is most likely because you created an "empty project" instead of a "Win32 console application". If you create the project as a "Win32 console application" you can disregard this as it does not apply.
In the older versions it would default to the console subsystem even if you selected "empty project", but not in Visual Studio 2010, so you have to set it manually. To do this select the project in the solution explorer on the right or left (probably is already selected so you don't have to worry about this).
Then select "project" from the menu bar drop down menus, then select "project_name properties" ? "configuration properties" ? "linker" ? "system" and set the first property, the drop down "subsystem" property to "console (/SUBSYSTEM:CONSOLE)". The console window should now stay open after execution as usual.
/////////////////////////////////////////////////////////////////////////////////////
Source: Stackoverflow.com