「try{} catch(…){}」的评论 https://doyj.com/2006/09/11/try-catch/ Laputa - 远方地平线，闪耀着光芒 Thu, 11 Sep 2014 06:16:14 +0000 hourly 1 https://wordpress.org/?v=6.7.2 评论者：my3439955 https://doyj.com/2006/09/11/try-catch/comment-page-1/#comment-20352 Thu, 11 Sep 2014 06:16:14 +0000 http://www.doyj.com/2006/09/11/try-catch/#comment-20352 之前用vc9测试的。我用vc6又测试了一下，果然如楼主所说。

]]> 评论者：my3439955 https://doyj.com/2006/09/11/try-catch/comment-page-1/#comment-20351 Thu, 11 Sep 2014 06:11:01 +0000 http://www.doyj.com/2006/09/11/try-catch/#comment-20351 非法的内存访问属于Win32异常范畴，C++语言的try catch根本就不会去处理这种异常，也没有这种能力。

Release也不会那样去优化，可以做如下代码验证：
try
{
*(int *)NULL = 11;
throw 111;
}
catch (…)
{
cerr<<"excetion"<<endl;
}
这里是有throw的，Release不会优化掉了，但是仍旧崩溃。

这段代码在Debug下也崩溃，博主能捕获到异常可能是因为那个地址真的就可写。最好直接用地址0来做测试。

]]> 评论者：tenlywu https://doyj.com/2006/09/11/try-catch/comment-page-1/#comment-8881 Sun, 12 Apr 2009 16:12:17 +0000 http://www.doyj.com/2006/09/11/try-catch/#comment-8881 刚好写GA代码文章的时候，提到了try catch

]]> 评论者：hanbger https://doyj.com/2006/09/11/try-catch/comment-page-1/#comment-1659 Sun, 08 Apr 2007 18:26:21 +0000 http://www.doyj.com/2006/09/11/try-catch/#comment-1659 :em21: good example~~~

]]> 评论者：空心人 https://doyj.com/2006/09/11/try-catch/comment-page-1/#comment-42 Wed, 13 Sep 2006 05:24:45 +0000 http://www.doyj.com/2006/09/11/try-catch/#comment-42 以前曾有过这个的疑问，赶紧在Linux上试了一下，Segment fault之后也不能被catch住。到网上查了一下，要想catch住这种系统信号，可以在信号处理函数中加上throw语句，并且在用g++编译时，加上non-call-exceptions

Since processor exceptions may occur anywhere in the code, the stack un-
winding libraries must be prepared to handle exceptions within any context.
This functionality is enabled by a special GNU g++ compiler °ag, \-fnon-call-
exceptions". Without the use of this °ag, only explicitly thrown exceptions are
handled. This option is typically used from user-space code that attempts to
throw exceptions from signal handlers. In the kernel, this allows an exception to
be correctly dispatched even in the absence of an explicit throw call.

]]>