MessageQ inside member function
-
Hi All, I have created a window using the Win32 api functions. Now I created a message loop inside a member function of the class like so:
{ UINT uMsg = 0; // Start another thread??? while (true) { if (::PeekMessage ((LPMSG) (&uMsg), m_hWnd, 0, 0, PM_REMOVE)) { ::TranslateMessage((LPMSG) (&uMsg)); ::DispatchMessage((LPMSG) (&uMsg)); } else { if (MessageQueueEmpty () == FAILURE) { return FAILURE; } } } return SUCCESS; }However after the first iteration,thisbecomesNULL. If I makeuMsgstatic, it works fine. Why?? :confused: regards, Rich "Programming today is a race between software engineers striving to build bigger and better idiot-proof programs, and the Universe trying to produce bigger and better idiots. So far the Universe is winning." -- Rich Cook -- modified at 10:24 Thursday 2nd March, 2006 -
Hi All, I have created a window using the Win32 api functions. Now I created a message loop inside a member function of the class like so:
{ UINT uMsg = 0; // Start another thread??? while (true) { if (::PeekMessage ((LPMSG) (&uMsg), m_hWnd, 0, 0, PM_REMOVE)) { ::TranslateMessage((LPMSG) (&uMsg)); ::DispatchMessage((LPMSG) (&uMsg)); } else { if (MessageQueueEmpty () == FAILURE) { return FAILURE; } } } return SUCCESS; }However after the first iteration,thisbecomesNULL. If I makeuMsgstatic, it works fine. Why?? :confused: regards, Rich "Programming today is a race between software engineers striving to build bigger and better idiot-proof programs, and the Universe trying to produce bigger and better idiots. So far the Universe is winning." -- Rich Cook -- modified at 10:24 Thursday 2nd March, 2006This code has a few problems. Here’s is an edited down version of your code which shows the biggest:
UINT uMsg = 0; ::PeekMessage((LPMSG)(&uMsg), m_hWnd, 0, 0, PM_REMOVE)The first parameter to
PeekMesageis aLPMSGbut your variable is of typeUINT. You would get a compiler error if you tried the following code:::PeekMessage(&uMsg, m_hWnd, 0, 0, PM_REMOVE)You "solve" this by adding the following C-style cast:
(LPMSG)This will never work, all you've done with this cast is say to the compiler, "Yeah I know a
LPMSGisn't the same as aUINT*and you normally wouldn't compile this code but I'm the man, just blindly push ahead and compile it anyway". The problem is aUINTisn't aMSG, cast or no cast. Here's what it should look like (edited down version):MSG m; ::PeekMessage(&m, m_hWnd, 0, 0, PM_REMOVE)A word of advice: Never use C-style casts (a type in brackets). Always use function style casts instead. When you find yourself tempted to write
(UINT)iwritestatic_cast<UINT>(i)instead. If the cast doesn't make sense the compiler will complain (you can still get into trouble but the chances are a lot less). If you want to make the compiler do what you say as a C-style cast does use areinterpret_cast. Casts are ugly (but occasionally necessary) and generally indicate a design flaw. The function style casts syntax allows you to spot casts more easily. Also C-style casts are not fine grained enough where as function style casts come in four types:dynamic_cast,static_cast,reinterpret_cast,const_castso it is clearer and harder to abuse. Steve -
This code has a few problems. Here’s is an edited down version of your code which shows the biggest:
UINT uMsg = 0; ::PeekMessage((LPMSG)(&uMsg), m_hWnd, 0, 0, PM_REMOVE)The first parameter to
PeekMesageis aLPMSGbut your variable is of typeUINT. You would get a compiler error if you tried the following code:::PeekMessage(&uMsg, m_hWnd, 0, 0, PM_REMOVE)You "solve" this by adding the following C-style cast:
(LPMSG)This will never work, all you've done with this cast is say to the compiler, "Yeah I know a
LPMSGisn't the same as aUINT*and you normally wouldn't compile this code but I'm the man, just blindly push ahead and compile it anyway". The problem is aUINTisn't aMSG, cast or no cast. Here's what it should look like (edited down version):MSG m; ::PeekMessage(&m, m_hWnd, 0, 0, PM_REMOVE)A word of advice: Never use C-style casts (a type in brackets). Always use function style casts instead. When you find yourself tempted to write
(UINT)iwritestatic_cast<UINT>(i)instead. If the cast doesn't make sense the compiler will complain (you can still get into trouble but the chances are a lot less). If you want to make the compiler do what you say as a C-style cast does use areinterpret_cast. Casts are ugly (but occasionally necessary) and generally indicate a design flaw. The function style casts syntax allows you to spot casts more easily. Also C-style casts are not fine grained enough where as function style casts come in four types:dynamic_cast,static_cast,reinterpret_cast,const_castso it is clearer and harder to abuse. SteveHi Steve, Thanks for the answer. Very helpful even though I can't believe I didn't see it. thanks again, Rich "Programming today is a race between software engineers striving to build bigger and better idiot-proof programs, and the Universe trying to produce bigger and better idiots. So far the Universe is winning." -- Rich Cook