[BUG] memory leak detected when calling AsyncRedisCluster::redis(sw::redis::StringView const&, bool)

[mroland91]

Describe the bug
I have a WIP project where I am integrating redis++ (thanks for making it BTW!).
After upgrading to the latest 1.3.11 Valgrind reported a memory error. I've also used 1.3.8 and 1.3.10 for a longer period and this problem was not present on those releases.

To Reproduce

auto event_loop = std::shared_ptr<sw::redis::EventLoop>();
sw::redis::ConnectionOptions opts;
opts.keep_alive = true;
opts.host = "localhost";
opts.port = 6379;
opts.resp = 3;
auto redis_cluster = std::make_shared<sw::redis::AsyncRedisCluster>(opts, sw::redis::ConnectionPoolOptions{}, sw::redis::Role::MASTER, event_loop);
sw::redis::Future<std::string> ping_future = redis_cluster->redis("key1", false).ping();
...

Environment:

• OS: Rocky Linux 8
• Compiler: gcc 8.5.0
• hiredis version: 1.2.0
• redis-plus-plus version:1.3.11

Additional context

==6011== Memcheck, a memory error detector
==6011== Copyright (C) 2002-2022, and GNU GPL'd, by Julian Seward et al.
==6011== Using Valgrind-3.19.0 and LibVEX; rerun with -h for copyright info
==6011== Command: /redis-client-test
==6011== Parent PID: 6010
==6011== 
==6011== 
==6011== HEAP SUMMARY:
==6011==     in use at exit: 21,523 bytes in 30 blocks
==6011==   total heap usage: 3,648 allocs, 3,618 frees, 546,766 bytes allocated
==6011== 
==6011== 2,510 (552 direct, 1,958 indirect) bytes in 1 blocks are definitely lost in loss record 21 of 22
==6011==    at 0x4C388C3: operator new(unsigned long) (vg_replace_malloc.c:422)
==6011==    by 0x6B06E31: allocate (new_allocator.h:111)
==6011==    by 0x6B06E31: allocate (alloc_traits.h:436)
==6011==    by 0x6B06E31: __allocate_guarded<std::allocator<std::_Sp_counted_ptr_inplace<sw::redis::AsyncConnection, std::allocator<sw::redis::AsyncConnection>, (__gnu_cxx::_Lock_policy)2> > > (allocated_ptr.h:97)
==6011==    by 0x6B06E31: __shared_count<sw::redis::AsyncConnection, std::allocator<sw::redis::AsyncConnection>, sw::redis::ConnectionOptions&, sw::redis::EventLoop*> (shared_ptr_base.h:675)
==6011==    by 0x6B06E31: __shared_ptr<std::allocator<sw::redis::AsyncConnection>, sw::redis::ConnectionOptions&, sw::redis::EventLoop*> (shared_ptr_base.h:1342)
==6011==    by 0x6B06E31: shared_ptr<std::allocator<sw::redis::AsyncConnection>, sw::redis::ConnectionOptions&, sw::redis::EventLoop*> (shared_ptr.h:359)
==6011==    by 0x6B06E31: allocate_shared<sw::redis::AsyncConnection, std::allocator<sw::redis::AsyncConnection>, sw::redis::ConnectionOptions&, sw::redis::EventLoop*> (shared_ptr.h:708)
==6011==    by 0x6B06E31: make_shared<sw::redis::AsyncConnection, sw::redis::ConnectionOptions&, sw::redis::EventLoop*> (shared_ptr.h:724)
==6011==    by 0x6B06E31: sw::redis::AsyncConnectionPool::_create() (async_connection_pool.cpp:273)
==6011==    by 0x6B078B0: sw::redis::AsyncConnectionPool::fetch() (async_connection_pool.cpp:120)
==6011==    by 0x6B0071D: sw::redis::GuardedAsyncConnection::GuardedAsyncConnection(std::shared_ptr<sw::redis::AsyncConnectionPool> const&) (async_connection.cpp:586)
==6011==    by 0x6B0FC28: construct<sw::redis::GuardedAsyncConnection, std::shared_ptr<sw::redis::AsyncConnectionPool>&> (new_allocator.h:136)
==6011==    by 0x6B0FC28: construct<sw::redis::GuardedAsyncConnection, std::shared_ptr<sw::redis::AsyncConnectionPool>&> (alloc_traits.h:475)
==6011==    by 0x6B0FC28: _Sp_counted_ptr_inplace<std::shared_ptr<sw::redis::AsyncConnectionPool>&> (shared_ptr_base.h:545)
==6011==    by 0x6B0FC28: __shared_count<sw::redis::GuardedAsyncConnection, std::allocator<sw::redis::GuardedAsyncConnection>, std::shared_ptr<sw::redis::AsyncConnectionPool>&> (shared_ptr_base.h:677)
==6011==    by 0x6B0FC28: __shared_ptr<std::allocator<sw::redis::GuardedAsyncConnection>, std::shared_ptr<sw::redis::AsyncConnectionPool>&> (shared_ptr_base.h:1342)
==6011==    by 0x6B0FC28: shared_ptr<std::allocator<sw::redis::GuardedAsyncConnection>, std::shared_ptr<sw::redis::AsyncConnectionPool>&> (shared_ptr.h:359)
==6011==    by 0x6B0FC28: allocate_shared<sw::redis::GuardedAsyncConnection, std::allocator<sw::redis::GuardedAsyncConnection>, std::shared_ptr<sw::redis::AsyncConnectionPool>&> (shared_ptr.h:708)
==6011==    by 0x6B0FC28: make_shared<sw::redis::GuardedAsyncConnection, std::shared_ptr<sw::redis::AsyncConnectionPool>&> (shared_ptr.h:724)
==6011==    by 0x6B0FC28: sw::redis::AsyncRedisCluster::redis(sw::redis::StringView const&, bool) (async_redis_cluster.cpp:49)
==6011==    by 0x49A9B8:  test() (test.cpp:891)
==6011==    by 0x453EC4:  Redis_Test::TestBody() (tc1.cpp:188)
==6011==    by 0x555AC2E: void testing::internal::HandleExceptionsInMethodIfSupported<testing::Test, void>(testing::Test*, void (testing::Test::*)(), char const*) (in /usr/lib64/libgtest.so.1.8.0)
==6011==    by 0x554FE31: testing::Test::Run() (in /usr/lib64/libgtest.so.1.8.0)
==6011==    by 0x554FF77: testing::TestInfo::Run() (in /usr/lib64/libgtest.so.1.8.0)
==6011==    by 0x5550054: testing::TestCase::Run() (in /usr/lib64/libgtest.so.1.8.0)
==6011==    by 0x5550524: testing::internal::UnitTestImpl::RunAllTests() (in /usr/lib64/libgtest.so.1.8.0)
==6011== 
==6011== LEAK SUMMARY:
==6011==    definitely lost: 552 bytes in 1 blocks
==6011==    indirectly lost: 1,958 bytes in 20 blocks
==6011==      possibly lost: 384 bytes in 1 blocks
==6011==    still reachable: 18,629 bytes in 8 blocks
==6011==         suppressed: 0 bytes in 0 blocks
==6011== Reachable blocks (those to which a pointer was found) are not shown.
==6011== To see them, rerun with: --leak-check=full --show-leak-kinds=all
==6011== 
==6011== For lists of detected and suppressed errors, rerun with: -s
==6011== ERROR SUMMARY: 1 errors from 1 contexts (suppressed: 0 from 0)

[sewenew]

Thanks for reporting the bug! I'll take a look into it.

Regards

[sewenew]

The problem has been fixed. Please check the latest code on master branch. I'll publish a new release after some more tests.

Sorry for the late reply. Too busy these days...

Regards

[mroland91]

Thanks, the issue is resolved on the latest master.