CXTPCommandBarsFrameHook crash

Message Topic Search Topic Options Post Reply Create New Topic Printable Version Translate Topic

   During shutdown we have experienced crashes in CXTPCommandBarsFrameHook code. The call to GetFramePaintManager has returned a null pointer. In particular, as we shut down the hook's RedrawFrame has been called and the GetFramePaintManager()->DrawFrame line crashed due to the null pointer.

I see that CXTPCommandBarsFrameHook::GetFramePaintManager() can return NULL as can the call to pPaintManager->GetFramePaintManager if it is called.

I also see that many of the calls to GetFramePaintManager in XTPFrameHook.cpp test the returned pointer. Only a few do not and I'm modifying those to do so.

Would be great if a CJ update did the same so I don't have to maintain this change.

Not sure how we get into this as it isn't happening a lot. We do use skinning depending on a user setting for a color theme choice but I run with the skin a lot and haven't crashed. So this change I'm making is a hope so. As in I hope it avoids the crash.

Author	Message Topic Search Topic Options Post Reply Create New Topic Printable Version Translate Topic
rdhd Members Profile Send Private Message Find Members Posts Add to Buddy List Senior Member Joined: 13 August 2007 Location: United States Status: Offline Points: 960	Post Options Post Reply Quote rdhd Report Post Thanks(0) Quote Reply Topic: CXTPCommandBarsFrameHook crash Posted: 18 February 2026 at 2:18pm
	During shutdown we have experienced crashes in CXTPCommandBarsFrameHook code. The call to GetFramePaintManager has returned a null pointer. In particular, as we shut down the hook's RedrawFrame has been called and the GetFramePaintManager()->DrawFrame line crashed due to the null pointer. I see that CXTPCommandBarsFrameHook::GetFramePaintManager() can return NULL as can the call to pPaintManager->GetFramePaintManager if it is called. I also see that many of the calls to GetFramePaintManager in XTPFrameHook.cpp test the returned pointer. Only a few do not and I'm modifying those to do so. Would be great if a CJ update did the same so I don't have to maintain this change. Not sure how we get into this as it isn't happening a lot. We do use skinning depending on a user setting for a color theme choice but I run with the skin a lot and haven't crashed. So this change I'm making is a hope so. As in I hope it avoids the crash.

agontarenko Members Profile Send Private Message Find Members Posts Add to Buddy List Moderator Group Joined: 25 March 2016 Status: Offline Points: 343	Post Options Post Reply Quote agontarenko Report Post Thanks(0) Quote Reply Posted: 7 hours 58 minutes ago at 4:09am
	Hello, I've added nullptr check. How I can to reproduce your problem? Regards, Artem Gontarenko

rdhd Members Profile Send Private Message Find Members Posts Add to Buddy List Senior Member Joined: 13 August 2007 Location: United States Status: Offline Points: 960	Post Options Post Reply Quote rdhd Report Post Thanks(0) Quote Reply Posted: 1 hour 28 minutes ago at 10:39am
	Hi Artem, I wish I knew a way to reproduce the issue. It was random from what I can tell (I just got dump files). We do use skinning and I have been closing the app when we got a deadlock situation as all threads were waiting on objects. I have waited to see if the app would "wake up" and finish but when it is locked that hasn't happened and I just kill it (stop the debugger). I have no ideal if the two are related. I added the pointer checks too so time will tell if this pops up elsewhere.