gbd/lldb使用2

      继续前文，加入list 对象。

      

  1 #include<list>

  2 #include<iostream>

  3 

  4 using namespace std;

  5 

  6 struct particle

  7 {

  8         int id;

  9 };

 10 

 11 

 12 int main()

 13 {

 14         int count = 4;

 15         particle* pp = new particle[count];

 16         list<particle> lsp;

 17 

 18         for(int i=0; i<count; i++)

 19         {

 20                 pp[i].id = i+1;

 21                 lsp.push_back(pp[i]);

 22                 pp++;

 23         }

 24 

 25         for(list<particle>::iterator it=lsp.begin(); it !=lsp.end(); it++)

 26         {

 27                 cout << "current particle id is "  << it->id << endl;

 28         }

 29 

 30         return 0;

 31 }       

~      

   分别使用 apple llvm 5.1(clang-503.0.40) + lldb  平台 和 red hat 4.47-4(gcc) + gdb 平台   

   lldb运行 ： 

(lldb) run

Process 8437 launched: '/Users/Desktop/cpp 2014/basic_100/LLDB/PA' (x86_64)

current particle id is 1

current particle id is 1504

Process 8437 stopped

* thread #1: tid = 0xb865b, 0x0000000100001658 PA`main + 984 at ParticleAssignment.cpp:41, queue = 'com.apple.main-thread', stop reason = EXC_BAD_ACCESS (code=1, address=0x1100000012)

    frame #0: 0x0000000100001658 PA`main + 984 at ParticleAssignment.cpp:41

  gdb运行：

(gdb) run

Starting program: /ifs/user/GNU/PA

current particle id is 1

current particle id is 2

current particle id is 3

current particle id is 4

Program received signal SIGSEGV, Segmentation fault.

0x0000003bfd07b55c in free () from /lib64/libc.so.6

首先，lldb内存溢出检查更严格？输出循环第二步就报错了。redhat gdb 程序结束才返回错误。 

    两个gdb最终都指向程序的最后一行（cpp:41)，但是这行 "return 0;"  不会错啊！ 

    

另外，lldb给出 stop reason = exc_bad_access, 程序里是有内存越界访问的；同样check一下 free() from /lib64/libc.so.6，是说“指针越界，析构无法完成”，即，同样指出内存越界访问。

lldb走不下去了，继续在 gdb中调试， 在 cpp:41 往前看，实际修改内存的部分设置断点 ( b 27)

(gdb) b 27

Breakpoint 1 at 0x400b28: file ParticleAssignment.cpp, line 27.

(gdb) run

Starting program: /ifs/user/MPITest/GNU/PA 

particle->id 1

particle->id 0

particle->id 2

particle->id 0

这个输出，显然不是想要的。第二个就错了，check list的单元组成， --line21 lsp.push_back(pp[i]);

(gdb) b 22

Breakpoint 1 at 0x400b09: file ParticleAssignment.cpp, line 22.

(gdb) run

Starting program: /ifs/user/zhengjia/MPITest/GNU/PA 

particle[i].id 1

Breakpoint 1, main () at ParticleAssignment.cpp:22

22 pp++;

Missing separate debuginfos, use: debuginfo-install glibc-2.12-1.132.el6_5.4.x86_64 libgcc-4.4.7-4.el6.x86_64 libstdc++-4.4.7-4.el6.x86_64

(gdb) print/x *pp

$1 = {id = 0x1}

(gdb) continue

Continuing.

particle[i].id 2

Breakpoint 1, main () at ParticleAssignment.cpp:22

22 pp++;

(gdb) print/x *pp

$2 = {id = 0x0}

(gdb) print/x pp

$3 = 0x602014

(gdb) continue

Continuing.

particle[i].id 3

Breakpoint 1, main () at ParticleAssignment.cpp:22

22 pp++;

(gdb) print/x *pp

$4 = {id = 0x2}

(gdb) print/x pp

$5 = 0x602018

这里，指针pp是连续，但所指内容不连续 1-0-2-0。 

引入pp，是希望采用指针访问particle对象集合（list），但这里pp又充当数组名了。

修改方法一，删去 pp++; 

修改方法二，全部（赋值，压栈）都使用指针

  pp->id = i+1;

  lsp.push_back(*pp);

  pp++;

运行： 

current particle id is 1

current particle id is 2

current particle id is 3

current particle id is 4

   this is a stupid sample, just have fun with gdb, while what about lldb, i even can't find continual breakpoint there, shit !