Package: libstdc++5-3.3-dev
Version: 1:3.3.1-0pre0
Severity: normal
Tags: upstream
Hi, I think I've found a bug in the allocation of strings done by the
standard c++ library. Maybe you are already aware of it, but I thought
it might be useful to report it.
This small testcase:
-------------------
#include <string>
int main(int argc, char* argv[])
{
std::string stars = "*************";
return 0;
}
Is reported to have leaks with three different leak analyzers:
leaktracer (2.4-1.0)
Gathered 1 (1 unique) points of data.
(gdb)
#-- Leak: counted 1x / total Size: 1280
No source file for address 0x400a7efb.
ccmalloc (0.4.0-2.0)
(...)
.---------------.
|ccmalloc report|
=======================================================
| total # of| allocated | deallocated | garbage |
+-----------+-------------+-------------+-------------+
| bytes| 1280 | 0 | 1280 |
+-----------+-------------+-------------+-------------+
|allocations| 1 | 0 | 1 |
+-----------------------------------------------------+
(...)
*100.0% = 1280 Bytes of garbage allocated in 1 allocation
| |
| | 0x40121da6 in <???>
| |
| | 0x080490b5 in <main>
(...)
valgrind (20031012)
(...)
==19195== ERROR SUMMARY: 0 errors from 0 contexts (suppressed: 0 from 0)
==19195== malloc/free: in use at exit: 1280 bytes in 1 blocks.
==19195== malloc/free: 1 allocs, 0 frees, 1280 bytes allocated.
==19195==
==19195== searching for pointers to 1 not-freed blocks.
==19195== checked 4562000 bytes.
==19195==
==19195== 1280 bytes in 1 blocks are still reachable in loss record 1 of 1
==19195== at 0x4002CAFD: __builtin_new (vg_replace_malloc.c:172)
==19195== by 0x4002CB68: operator new(unsigned) (vg_replace_malloc.c:185)
==19195== by 0x402CEEFA:
std::__default_alloc_template<true,0>::_S_chunk_alloc(unsigned, int&) (in
/usr/lib/libstdc++.so.5.0.5)
==19195== by 0x402CEE0C:
std::__default_alloc_template<true,0>::_S_refill(unsigned) (in
/usr/lib/libstdc++.so.5.0.5)
==19195==
==19195== LEAK SUMMARY:
==19195== definitely lost: 0 bytes in 0 blocks.
==19195== possibly lost: 0 bytes in 0 blocks.
==19195== still reachable: 1280 bytes in 1 blocks.
==19195== suppressed: 0 bytes in 0 blocks.
(...)
Testing with other string constructors, such as
string stars(10, '*');
or passing a string as a parameter to a procedure, gives the same
results, while using the empty constructor (string stars;) shows no
leak at all.
At first I thought this might be a problem of the leak analyzers, but 3
analyzers having the same problem seems a bit too much.
I consulted some other programmers, and all agreed that this really
looked as a bug in the standard library.
I hope you can have a look at it.
Love,
Margarita Manterola
-- System Information:
Debian Release: testing/unstable
Architecture: i386
Kernel: Linux hormiga.amadeus 2.4.19-686 #1 Mon Nov 18 23:59:03 EST 2002 i686
Locale: LANG=es_AR, LC_CTYPE=es_AR
Versions of packages libstdc++5-3.3-dev depends on:
ii g++-3.3 1:3.3.1-0pre0 The GNU C++ compiler
ii gcc-3.3-base 1:3.3.1-0pre0 The GNU Compiler Collection (base
ii libc6-dev 2.3.2.ds1-10 GNU C Library: Development Librari
ii libstdc++5 1:3.3.2-4 The GNU Standard C++ Library v3
-- no debconf information
