I was checking links on one of my HTML pages by quickly clicking on each link and then using Alt-Lft to get back to the previous page. I did this for about half a dozen links and then got an Assert: Line 176 of histbld.cpp in winfe // Adds a history item to the passed in menu. Also attaches a numerica accelerator to the menu item // if nPosition is less than 10. BOOL CHistoryMenuBuilder::AddHistoryToMenu(HMENU hMenu, int nPosition, BOOL bCurrent, History_entry* pEntry, int nMaxMenuItemLength) { char* lpszText; /*>*/ ASSERT(pEntry->title); if (!pEntry->title) return FALSE; <Snip> pEntry->title is 0x00ee5810, and points to "" The title pointer is non-zero, so I don't understand why it is asserting. ASSERT causes a break only when the argument is equal to zero, right? pEntry->address = "http://www.genealogy.org/~bryce/Harrington/Benjamin/HarringtonRoyalDescent.txt" pEntry->referer = "http://www.genealogy.org/~bryce/Harrington/Benjamin/index.html" Stack Trace: ------------ 1. CHistoryMenuBuilder::AddHistoryToMenu(HMENU__ * 0x0000096c, int 0, int 1, _History_entry * 0x011719d0, int 45) line 176 + 41 bytes 2. CHistoryMenuBuilder::FillItems(HMENU__ * 0x0000096c, _XP_List * 0x01171bd0, int 0, int 45, _History_entry * 0x011719d0) line 238 3. CHistoryMenuBuilder::Fill(HMENU__ * 0x0000096c, int 45, eFillEnum eFILLALL) line 106 4. CMainFrame::BuildHistoryMenu(CMenu * 0x00ee39a0 {CMenu}) line 600 5. CMainFrame::OnInitMenuPopup(CMenu * 0x00ee39a0 {CMenu}, unsigned int 3, int 0) line 540 6. CWnd::OnWndMsg(unsigned int 279, unsigned int 2412, long 3, long * 0x00cbf95c) line 1786 7. CWnd::WindowProc(unsigned int 279, unsigned int 2412, long 3) line 1522 + 30 bytes 8. AfxCallWndProc(CWnd * 0x00d6d1b4 {CMainFrame hWnd=0x00000ce8}, HWND__ * 0x00000ce8, unsigned int 279, unsigned int 2412, long 3) line 210 + 26 bytes 9. AfxWndProc(HWND__ * 0x00000ce8, unsigned int 279, unsigned int 2412, long 3) line 363 + 25 bytes 10. AfxWndProcBase(HWND__ * 0x00000ce8, unsigned int 279, unsigned int 2412, long 3) line 203 + 21 bytes 11. KERNEL32! bff93d4b() 12. KERNEL32! bff73663() 13. KERNEL32! bff928e0() 14. 00cb8ae4() 15. 058f64c9()

It's possible that the title was filled by a different thread while the ASSERT dialog was up. It is a mystery to my why the author of the code believed it necessary to throw an assert if checking for NULL anyhow.

Old bug... Marking Verified/Won't Fix