[oi-dev] build issue in percona-server-57

[Jim Klimov]

On November 6, 2018 7:28:43 AM UTC, Michal Nowak <mnowak at startmail.com> wrote:
>On 11/06/18 01:07 AM, Rich Reynolds wrote:
>> On 11/5/18 2:38 PM, Bob Friesenhahn wrote:
>>> On Mon, 5 Nov 2018, Rich Reynolds wrote:
>>>
>>>>
>>>> [ 32%] Building CXX object 
>>>> sql/CMakeFiles/binlog.dir/binlog_crypt_data.cc.o
>>>>
>/codebase/oi-userland/components/database/percona-server-57/percona-server-5.7.22-22/sql/binlog_crypt_data.cc:
>
>>>> In member function 'void Binlog_crypt_data::free_key(uchar*&, 
>>>> std::size_t&)':
>>>>
>/codebase/oi-userland/components/database/percona-server-57/percona-server-5.7.22-22/sql/binlog_crypt_data.cc:62:37:
>
>>>> error: 'memset_s' was not declared in this scope
>>>>     memset_s(key, 512, 0, key_length);
>>>>                                     ^
>>>> make[3]: *** [sql/CMakeFiles/binlog.dir/build.make:336: 
>>>> sql/CMakeFiles/binlog.dir/binlog_crypt_data.cc.o] Error 1
>>>>
>>>>
>>>>
>>>> this a patching, upstream or makefile issue?  ill investigate if no
>
>>>> one knows off the top of head
>>>
>>> This appears to be a C'11 function.  See 
>>> https://en.cppreference.com/w/c/string/byte/memset.  Depending on
>C'11 
>>> at this time is very optimistic.
>>>
>>> What version of GCC are you using?
>>>
>>> Bob
>> cmake was fine with 6.4.0
>> 
>> -- Running cmake version 3.11.3
>> -- Found Git: /usr/bin/git (found version "2.19.1")
>> -- Configuring with MAX_INDEXES = 64U
>> -- The C compiler identification is GNU 6.4.0
>> -- The CXX compiler identification is GNU 6.4.0
>> -- Check for working C compiler: /usr/gcc/6/bin/gcc
>> -- Check for working C compiler: /usr/gcc/6/bin/gcc -- works
>> -- Detecting C compiler ABI info
>> -- Detecting C compiler ABI info - done
>> -- Detecting C compile features
>> -- Detecting C compile features - done
>> -- Check for working CXX compiler: /usr/gcc/6/bin/g++
>> -- Check for working CXX compiler: /usr/gcc/6/bin/g++ -- works
>> -- Detecting CXX compiler ABI info
>> -- Detecting CXX compiler ABI info - done
>> -- Detecting CXX compile features
>> -- Detecting CXX compile features - done
>> 
>> and cuz more information is better than less:
>> 
>> -- CMAKE_BUILD_TYPE: RELEASE
>> -- COMPILE_DEFINITIONS: 
>>
>_FILE_OFFSET_BITS=64;__EXTENSIONS__;_POSIX_PTHREAD_SEMANTICS;_REENTRANT;_PTHREADS;BOOST_GEOMETRY_SQRT_CHECK_FINITENESS;HAVE_CONFIG_H;HAVE_LIBEVENT1
>
>> 
>> -- CMAKE_C_FLAGS: -m32 -O3 -Wall -Wextra -Wformat-security -Wvla 
>> -Wwrite-strings -Wdeclaration-after-statement
>> -- CMAKE_CXX_FLAGS:  -m32 -O3 -Wall -Wextra -Wformat-security -Wvla 
>> -Woverloaded-virtual -Wno-unused-parameter
>> -- CMAKE_C_LINK_FLAGS:  -m32
>> -- CMAKE_CXX_LINK_FLAGS:  -m32
>> -- CMAKE_C_FLAGS_RELEASE: -O3 -DNDEBUG -DDBUG_OFF -DNDEBUG
>> -- CMAKE_CXX_FLAGS_RELEASE: -O3 -DNDEBUG -DDBUG_OFF -DNDEBUG
>> -- Configuring done
>> 
>> rich
>> 
>
>Three months ago illumos added support for memset_s(3C) 
>(https://www.illumos.org/issues/8548) before that it likely wasn't
>detected.
>
>memset_s(3C) seems to require
>
>   #define __STDC_WANT_LIB_EXT1__ 1
>   #include <string.h>
>
>I don't know how to do that properly in cmake/C++.
>
>If anyone does know, I'll fix it in userland when updating Percona 
>Server to 5.7.23-23.
>
>This works for cmake to use GCC 7.3:
>
>   GCC_VERSION = 7
>
>Michal
>
>_______________________________________________
>oi-dev mailing list
>oi-dev at openindiana.org
>https://openindiana.org/mailman/listinfo/oi-dev

I would expect adding these lines to individual source files using the routine, maybe after other headers just in case so the macro does not add surprise when processing other headers that do work today :)

On userland side, extract this change into an automatically applied patch-file and the common web of makefiles should do the magic.

Jim
--
Typos courtesy of K-9 Mail on my Android