Re: [Cppunit-devel] REQUEST: throw user-defined errors in setUp() and tearDown()

[Baptiste L. <gai...@fr...>]

RE: [Cppunit-devel] REQUEST: throw user-defined errors in setUp() and =
tearDown()You can use a TestListener and track startTestRun() and =
endTestRun() to open/close the DB connection.

You still need to find a way to 'pass' it to the test case =
(singleton...).

While I'm at it, does anybody see a need for something like:

void setUp( TestResult *controller )
{
    DbConnection *connection =3D reinterpret_cast<DbConnection *>(=20
                    controller->getNamedResource( "DbConnection1" ) );
}

Baptiste.

  ----- Original Message -----=20
  From: Moran Ben-David=20
  To: 'Michel Armin' ; cpp...@li...=20
  Sent: Thursday, July 11, 2002 4:19 PM
  Subject: RE: [Cppunit-devel] REQUEST: throw user-defined errors in =
setUp() and tearDown()


  While we are on the subject.. Is there a way to set up a TestFixture =
once for all tests?  For example, if my SetUp creates a database =
connection required by the tests in my fixture, it would be ideal if it =
only did so once.

  Moran=20

  -----Original Message-----=20
  From: Michel Armin [mailto:Arm...@si...]=20
  Sent: Thursday, July 11, 2002 3:23 AM=20
  To: cpp...@li...=20
  Subject: RE: [Cppunit-devel] REQUEST: throw user-defined errors in=20
  setUp() and tearDown()=20



  Is there a reason why we do not support assertion in setUp() and =
tearDown()?=20
  The code could be easily modified to support this.=20

  IMHO the general question is, what should be done in setUp() and =
tearDown()=20
  respectively. Personally, I use them to provide everything that is =
common to=20
  all test methods in a class. This also includes setting up a =
particular=20
  state of my (product code) environment that is a precondition for all =
the=20
  tests in a testcase class. And of course, this will lead to code that =
might=20
  fail (for whatever reason). Therefore I like the idea of being able to =
have=20
  asserts in the setUp() and tearDown() methods, because these would =
lead to a=20
  more precise error/failure message, why the test-methods couldn't be=20
  executed successfully. And locating bugs in our programs is the =
primary goal=20
  of this project, isn't it?=20

  Any other opinions?=20




  Regards=20

  Armin.=20

  > -----Original Message-----=20
  > From: Baptiste Lepilleur [mailto:gai...@fr...]=20
  > Sent: Mittwoch, 10. Juli 2002 19:44=20
  > To: John Lam; cpp...@li...=20
  > Subject: Re: [Cppunit-devel] REQUEST: throw user-defined errors in=20
  > setUp() and tearDown()=20
  >=20
  >=20
  > It will be done. It goes along with the following TODO features:=20
  >=20
  >   - Provide a mean for the user to catch 'custom' exception=20
  > in TestCase::run=20
  >     (exception that do not subclass std::exception, such as=20
  > MFC CException,=20
  > or=20
  >     RogueWave RWXMsg).=20
  >=20
  > That feature will likely result in a piece of code that will=20
  > run a specific=20
  > method, catching exceptions and assertion failures. This=20
  > should avoid that=20
  > so very long method...=20
  >=20
  > Though, notes that you usually try to avoid putting assertion=20
  > in setUp() and=20
  > tearDown() (though the memory/resource leaks detection is a=20
  > good candidate=20
  > for a tearDown assertions).=20
  >=20
  > Baptiste.=20
  >=20
  > ----- Original Message -----=20
  > From: "John Lam" <jl...@iu...>=20
  > To: <cpp...@li...>=20
  > Sent: Friday, June 21, 2002 10:57 PM=20
  > Subject: [Cppunit-devel] REQUEST: throw user-defined errors=20
  > in setUp() and=20
  > tearDown()=20
  >=20
  >=20
  > If my setUp() or tearDown() methods fail, I would like to=20
  > have something=20
  > more specific than "setUp() failed ..." showing up in my=20
  > logs. I'd like=20
  > to propose the following change to TestCase.cpp:=20
  >=20
  > It currently reads:=20
  >=20
  >   try {=20
  >   setUp();=20
  >=20
  >   try {=20
  >     runTest();=20
  >   }=20
  >   catch ( Exception &e ) {=20
  >     Exception *copy =3D e.clone();=20
  >     result->addFailure( this, copy );=20
  >   }=20
  >   catch ( std::exception &e ) {=20
  >     result->addError( this, new Exception( e.what() ) );=20
  >   }=20
  >   catch (...) {=20
  >     Exception *e =3D new Exception( "caught unknown exception" );=20
  >     result->addError( this, e );=20
  >   }=20
  >=20
  >   try {=20
  >     tearDown();=20
  >   }=20
  >   catch (...) {=20
  >     result->addError( this, new Exception( "tearDown() failed" )=20
  > );=20
  >   }=20
  >   }=20
  >   catch (...) {=20
  >   result->addError( this, new Exception( "setUp() failed" ) );=20
  >   }=20
  >=20
  > I would like to add:=20
  >=20
  >   catch ( const Exception &e ) {=20
  >     Exception *copy =3D e.clone();=20
  >     result->addError( this, copy );=20
  >   }=20
  >=20
  > to the outer most try block so that I can get more meaningful output =

  > when errors occur.=20
  >=20
  > Similarly, I would like to add the same code to the try block that=20
  > guards the call to tearDown().=20
  >=20
  > Does this sound reasonable to the community?=20
  >=20
  > Cheers,=20
  >=20
  > -John=20
  > http://www.iunknown.com=20
  >=20
  >=20
  >=20
  >=20
  > -------------------------------------------------------=20
  > Sponsored by:=20
  > ThinkGeek at http://www.ThinkGeek.com/=20
  > _______________________________________________=20
  > Cppunit-devel mailing list=20
  > Cpp...@li...=20
  > https://lists.sourceforge.net/lists/listinfo/cppunit-devel=20
  >=20
  >=20
  >=20
  >=20
  > -------------------------------------------------------=20
  > This sf.net email is sponsored by:ThinkGeek=20
  > Two, two, TWO treats in one.=20
  > http://thinkgeek.com/sf=20
  > _______________________________________________=20
  > Cppunit-devel mailing list=20
  > Cpp...@li...=20
  > https://lists.sourceforge.net/lists/listinfo/cppunit-devel=20
  >=20



  -------------------------------------------------------=20
  This sf.net email is sponsored by:ThinkGeek=20
  PC Mods, Computing goodies, cases & more=20
  http://thinkgeek.com/sf=20
  _______________________________________________=20
  Cppunit-devel mailing list=20
  Cpp...@li...=20
  https://lists.sourceforge.net/lists/listinfo/cppunit-devel=20