計劃在bash運行良好，但不具有的valgrind

Question

我用C語言編寫的程序，這在bash運行良好運行，但下的valgrind給人奇怪的結果，而Valgrind的報告內存泄漏。

運行在bash：

：〜/沙/ binofino $ ./a.out

24 = 3 + 21 
24 = 3 + 21 
24 = 3 + 8 + 13 
24 = 1 + 2 + 8 + 13 
24 = 1 + 2 + 3 + 5 + 13 
24 = 1 + 2 + 21 

下的valgrind：

>:~/sandbox/binofino$ valgrind ./a.out 
==20116== Memcheck, a memory error detector 
==20116== Copyright (C) 2002-2009, and GNU GPL'd, by Julian Seward et al. 
==20116== Using Valgrind-3.6.0.SVN-Debian and LibVEX; rerun with -h for copyright info 
==20116== Command: ./a.out 
==20116== 
==20116== Invalid read of size 4 
==20116== at 0x804857A: main (in /home/..../sandbox/binofino/a.out) 
==20116== Address 0x41a402c is 0 bytes after a block of size 4 free'd 
==20116== at 0x4025016: realloc (vg_replace_malloc.c:525) 
==20116== by 0x804879F: get_fibo_index (in /home/..../sandbox/binofino/a.out) 
==20116== by 0x804854F: main (in /home/..../sandbox/binofino/a.out) 
==20116== 
24 = 1 + 1 
==20116== Invalid read of size 4 
==20116== at 0x80485CE: main (in /home/..../sandbox/binofino/a.out) 
==20116== Address 0x41a402c is 0 bytes after a block of size 4 free'd 
==20116== at 0x4025016: realloc (vg_replace_malloc.c:525) 
==20116== by 0x804879F: get_fibo_index (in /home/..../sandbox/binofino/a.out) 
==20116== by 0x804854F: main (in /home/..../sandbox/binofino/a.out) 
==20116== 
24 = 1 
==20116== Invalid free()/delete/delete[] 
==20116== at 0x4024B3A: free (vg_replace_malloc.c:366) 
==20116== by 0x804867E: main (in /home/..../sandbox/binofino/a.out) 
==20116== Address 0x41a4028 is 0 bytes inside a block of size 4 free'd 
==20116== at 0x4025016: realloc (vg_replace_malloc.c:525) 
==20116== by 0x804879F: get_fibo_index (in /home/..../sandbox/binofino/a.out) 
==20116== by 0x804854F: main (in /home/..../sandbox/binofino/a.out) 
==20116== 
==20116== 
==20116== HEAP SUMMARY: 
==20116==  in use at exit: 8 bytes in 1 blocks 
==20116== total heap usage: 4 allocs, 4 frees, 20 bytes allocated 
==20116== 
==20116== LEAK SUMMARY: 
==20116== definitely lost: 8 bytes in 1 blocks 
==20116== indirectly lost: 0 bytes in 0 blocks 
==20116==  possibly lost: 0 bytes in 0 blocks 
==20116== still reachable: 0 bytes in 0 blocks 
==20116==   suppressed: 0 bytes in 0 blocks 
==20116== Rerun with --leak-check=full to see details of leaked memory 
==20116== 
==20116== For counts of detected and suppressed errors, rerun with: -v 
==20116== ERROR SUMMARY: 5 errors from 3 contexts (suppressed: 11 from 6) 

它不僅報告內存泄漏，它也報告錯誤，程序的輸出是完全錯誤的。

爲什麼？

Answer 1

因爲你有一個bug（實際上是幾個bug）。

特別是，在get_fibo_index你realloc某些內存，然後在main讀取正確的分配緩衝區結束（產生完全未定義的結果）。

與-g重建程序，Valgrind的下重新運行它，並修復它找到的所有「無效」的錯誤。

（與-g重建會給你的文件和行信息，這將使修復錯誤更容易。）