Visual leak detector (VLD) displays empty call stack
Asked Answered
F

2

5

I'm using visual leak detector. It displays the call stack properly for some leaks, but I get some outputs that the call stack is empty like this:

---------- Block 131148 at 0x69B3AF68: 144 bytes ----------
  Leak Hash: 0xCA7D251C, Count: 1, Total 144 bytes
  Call Stack (TID 1600):
  Data:
    B8 0A 80 01    01 00 00 00    7F 7D 00 00    00 00 00 00     ........ .}......
    00 00 00 00    A4 81 1A 00    17 00 00 00    A8 AE EA 4E     ........ .......N
    40 EE 22 53    00 00 70 42    FB 19 CC BD    D4 E7 2E 00     @."S..pB ........
    CD CD CD CD    80 2D FB 6E    A0 4A B2 69    20 8D 00 54     .....-.n .J.i...T
    D8 6B 6A 54    00 CD CD CD    38 DB 81 01    40 EE 22 53     .kjT.... 8...@."S
    CD CD CD CD    CD CD CD CD    B0 AF B3 69    CD CD CD CD     ........ ...i....
    60 DB 81 01    40 EE 22 53    E8 03 00 00    CD CD CD CD     `...@."S ........
    C8 AF B3 69    CD CD CD CD    CD CD CD CD    CD CD CD CD     ...i.... ........
    CD CD CD CD    CD CD CD CD    00 00 00 00    CD CD CD CD     ........ ........

I don't know how to track issues like this.

Any help would be appreciated.

Front answered 3/4, 2016 at 7:20 Comment(0)
M
8

You should try to change the output format from ascii to unicode in vld.ini.

Your stack information must have unicode caracters that cannot be converted to ascii.

If VLD is compiled in debug, this results in assertions.

See this related topic on VLD web site: https://vld.codeplex.com/workitem/10606

I posted an alternative to changing vld.ini in a comment (need to modify vld sources and recompile).

Melmela answered 3/4, 2016 at 10:33 Comment(0)
G
0

Changing ascii to unicode in vld.ini does not work for me.

But moving my project from directory that contains unicode chars (such as Chinese Characters) to not-containing-unicode-chars-directory, vld does show call stack.

Gravity answered 3/12, 2020 at 2:57 Comment(0)

© 2022 - 2024 — McMap. All rights reserved.