为什么TRACE()会引发浮点下溢异常? [英] Why TRACE() throws Floating-point underflow exception?
本文介绍了为什么TRACE()会引发浮点下溢异常?的处理方法,对大家解决问题具有一定的参考价值,需要的朋友们下面随着小编来一起学习吧!
问题描述
为什么在TRACE中出现浮点下溢(参数:0x00000000)? 我在这里需要什么格式说明符?
Why do I get a Floating-point underflow (parameters: 0x00000000) in TRACE? What format specifier do I need here?
// show load progress by callback-Funktion (on Statusbar)
ULONGLONG len = 1000; // ar.GetFile()->GetLength();
ULONGLONG pos = 800; // ar.GetFile()->GetPosition();
double perc = (double)pos/(double)len*100;
TRACE("load from %X, Position: %ld, Length: %ld, Perc: %lf \n",
this, pos, len, perc );
更新
len
和pos
无关紧要,TRACE
在 Debug-Mode (调试模式)中始终抛出错误.
与使用多字节字符集一起编译.将代码从VS9升级到VS14.在发布模式中,一切正常.
Update
The len
and pos
does not care, TRACE
throws always an Error in Debug-Mode.
Compiled with Use Multi-Byte Character Set. Upgraded the code from VS9 to VS14. In Release-Mode it seems working all fine.
调试输出为:
Natvis: Parsing natvis xml file: C:\Program Files (x86)\Microsoft Visual Studio 14.0\Common7\Packages\Debugger\Visualizers\windows.media.natvis.
Natvis: Parsing natvis xml file: C:\Program Files (x86)\Microsoft Visual Studio 14.0\Common7\Packages\Debugger\Visualizers\windows.natvis.
Natvis: Parsing natvis xml file: C:\Program Files (x86)\Microsoft Visual Studio 14.0\Common7\Packages\Debugger\Visualizers\winrt.natvis.
Natvis: C:\Program Files (x86)\Microsoft Visual Studio 14.0\Common7\Packages\Debugger\Visualizers\atlmfc.natvis(9,28): Successfully parsed expression 'm_hWnd' in type context 'CWnd'.
Exception thrown at 0x5B4EE469 (ucrtbased.dll) in TestProgD.exe: 0xC0000093: Floating-point underflow (parameters: 0x00000000).
Unhandled exception at 0x5B4EE469 (ucrtbased.dll) TestProgD.exe: 0xC0000093: Floating-point underflow (parameters: 0x00000000).
推荐答案
格式说明符与参数类型不匹配,因此您具有不确定的行为.
The format specifiers do not match the argument types, so you have undefined behaviour.
-
%X
(unsigned int
)应该是%p
(指针). -
%ld
(long int
)应该是%llu
(unsigned long long
).在Windows平台上,long int
与int
相同,因此,无论代码是针对32位还是64位构建的,它始终为32位.只有long long
始终为64位.当然,您还希望匹配签名,这就是为什么必须使用u
说明符的原因.
%X
(unsigned int
) should be%p
(pointer).%ld
(long int
) should be%llu
(unsigned long long
). On the Windows platformlong int
is the same asint
, so it is always 32-bit, regardless whether the code is built for 32-bit or 64-bit. Onlylong long
is always 64-bit. Of course you also want to match the signedness, that's why theu
specifier must be used.
参考.
这篇关于为什么TRACE()会引发浮点下溢异常?的文章就介绍到这了,希望我们推荐的答案对大家有所帮助,也希望大家多多支持IT屋!
查看全文