Tur*_*rbo 2 c++ exception-handling stack-trace
我有这个二进制文件通过抛出 std:string 类型的异常而崩溃。
来自剥离二进制文件的堆栈跟踪:
terminate called after throwing an instance of 'std::string'
*** Aborted at 1309483487 (unix time) try "date -d @1309483487" if you are using GNU date ***
PC: @ 0x3fb0c30155 (unknown)
*** SIGABRT (@0xd54) received by PID 3412 (TID 0x40d03940) from PID 3412; stack trace: ***
@ 0x3fb180de70 (unknown)
@ 0x3fb0c30155 (unknown)
@ 0x3fb0c31bf0 (unknown)
@ 0x2aaaaab80cc4 (unknown)
@ 0x2aaaaab7ee36 (unknown)
@ 0x2aaaaab7ee63 (unknown)
@ 0x2aaaaab7ef4a (unknown)
@ 0x4c2622 XYZ::connect()
@ 0x4c3e0f XYZ::refresh()
@ 0x3fb18062f7 (unknown)
@ 0x3fb0cd1e3d (unknown)
Run Code Online (Sandbox Code Playgroud)
现在的问题是,refresh() 函数确实尝试捕获 std::string。看起来像:-
bool XYZ::refresh() {
try {
connect();
} catch (string& s) {
return false;
}
return true;
}
Run Code Online (Sandbox Code Playgroud)
知道为什么它没有被抓住吗?或者我读错了堆栈跟踪?
也许部分或所有模块是用-fno-exceptions? 有关如何更改异常行为的详细信息,请参阅http://gcc.gnu.org/onlinedocs/libstdc++/manual/using_exceptions.html。
例如,以下短程序在以下情况下显示"terminate called after throwing an instance of 'std::string'":
foo()是用-fno-exceptions, 和foo()调用抛出类型异常的东西std::string(所有其他模块都是用 编译的-fexceptions)
#include <string>
#include <iostream>
using namespace std;
int foo();
int main()
{
try {
foo();
}
catch (string& s) {
std::cout << "caught it: \"" << s << "\"" << endl;
}
return 0;
}
Run Code Online (Sandbox Code Playgroud)请注意,如果我简单地foo.cpp使用 -fexceptions(g++ 的默认值)重新编译并重新链接,程序将显示:
caught it: "the string exception"
Run Code Online (Sandbox Code Playgroud)
正如预期的那样。
或者也许某些中间函数有一个没有列出的 throw 规范std::string?
例如,这个程序:
#include <string>
#include <iostream>
using namespace std;
int Hunc() throw(int); // can only throw int (?)
int main()
{
try {
Hunc();
}
catch (string& s) {
std::cout << "caught it: \"" << s << "\"" << endl;
}
return 0;
}
int Hunc() throw(int)
{
throw string("the string exception");
}
Run Code Online (Sandbox Code Playgroud)
还显示“在抛出 'std::string' 实例后调用终止”。这两个示例都在装有 MinGW 4.5.1 的 Windows 机器上进行了测试。