Gacko

Stupid question - have you patched your client? And did you re-run the connection_patcher after updating to 6.2.3?

Hey, this looks like linker errors, so cpp files are already build at this time. I don't think it's related to custom code (if there is some...) Regards, Gacko

clang: warning: argument unused during compilation: '-pthread' 0 0x1001e1724 __assert_rtn + 144 1 0x10021625e archive::File<x86_64>::makeObjectFileForMember(archive::File<x86_64>::Entry const*) const + 1118 2 0x100215c3b archive::File<x86_64>::justInTimeforEachAtom(char const*, ld::File::AtomHandler&) const + 139 3 0x10022a3fe ld::tool::InputFiles::searchLibraries(char const*, bool, bool, bool, ld::File::AtomHandler&) const + 210 4 0x100231181 ld::tool::Resolver::resolveUndefines() + 189 5 0x1002331a5 ld::tool::Resolver::resolve() + 79 6 0x1001e1b17 main + 669 A linker snapshot was created at: /tmp/worldserver-2013-09-27-195633.ld-snapshot ld: Assertion failed: (memberIndex != 0), function makeObjectFileForMember, file /SourceCache/ld64/ld64-224.1/src/ld/parsers/archive_file.cpp, line 355. clang: error: linker command failed with exit code 1 (use -v to see invocation) So you think using SDK 10.8 works? Hm, I'll try. Edit: FUCK YOU APPLE! Thanks Jeremy, it works with SDK 10.8

... you're using gcc. So, I guess you don't have a solution for this, too. Right? (I'm gonna post the error in a few minutes.)

Clang is still on 5.0.0, so there is not 5.0.x Never the less: Which OS are you working with? This would be important to me as I'm still getting a error during linking.

jtth, you probably mean this: diff --git a/src/server/shared/Cryptography/OpenSSLCrypto.cpp b/src/server/shared/Cryptography/OpenSSLCrypto.cpp index 417be81..e3fc28d 100644 --- a/src/server/shared/Cryptography/OpenSSLCrypto.cpp +++ b/src/server/shared/Cryptography/OpenSSLCrypto.cpp @@ -25,35 +25,35 @@ std::vector<ACE_Thread_Mutex*> cryptoLocks; static void lockingCallback(int mode, int type, const char* /*file*/, int /*line*/) { - if (mode & CRYPTO_LOCK) - cryptoLocks[type]->acquire(); - else - cryptoLocks[type]->release(); +// if (mode & CRYPTO_LOCK) +// cryptoLocks[type]->acquire(); +// else +// cryptoLocks[type]->release(); } static void threadIdCallback(CRYPTO_THREADID * id) { - CRYPTO_THREADID_set_numeric(id, ACE_Thread::self()); +// CRYPTO_THREADID_set_numeric(id, ACE_Thread::self()); } void OpenSSLCrypto::threadsSetup() { - cryptoLocks.resize(CRYPTO_num_locks()); - for(int i = 0 ; i < CRYPTO_num_locks(); ++i) - { - cryptoLocks[i] = new ACE_Thread_Mutex(); - } - CRYPTO_THREADID_set_callback(threadIdCallback); - CRYPTO_set_locking_callback(lockingCallback); +// cryptoLocks.resize(CRYPTO_num_locks()); +// for(int i = 0 ; i < CRYPTO_num_locks(); ++i) +// { +// cryptoLocks[i] = new ACE_Thread_Mutex(); +// } +// CRYPTO_THREADID_set_callback(threadIdCallback); +// CRYPTO_set_locking_callback(lockingCallback); } void OpenSSLCrypto::threadsCleanup() { - CRYPTO_set_locking_callback(NULL); - CRYPTO_THREADID_set_callback(NULL); - for(int i = 0 ; i < CRYPTO_num_locks(); ++i) - { - delete cryptoLocks[i]; - } - cryptoLocks.resize(0); -} No newline at end of file +// CRYPTO_set_locking_callback(NULL); +// CRYPTO_THREADID_set_callback(NULL); +// for(int i = 0 ; i < CRYPTO_num_locks(); ++i) +// { +// delete cryptoLocks[i]; +// } +// cryptoLocks.resize(0); +} This makes it compiling if tools are deactivated. Otherwise you have to "remove" a log output in MapBuilder.cpp: diff --git a/src/tools/mmaps_generator/MapBuilder.cpp b/src/tools/mmaps_generator/MapBuilder.cpp index 9c2dda5..9872cf3 100644 --- a/src/tools/mmaps_generator/MapBuilder.cpp +++ b/src/tools/mmaps_generator/MapBuilder.cpp @@ -340,7 +340,7 @@ namespace MMAP /**************************************************************************/ void MapBuilder::buildMap(uint32 mapID) { - printf("[Thread %u] Building map %03u:n", uint32(ACE_Thread::self()), mapID); + //printf("[Thread %u] Building map %03u:n", uint32(ACE_Thread::self()), mapID); std::set<uint32>* tiles = getTileList(mapID); I got it working under Mountain Lion. But it broke again under Mavericks... as I already told

This is not a problem of OpenSSL. Is it that hard to interpret what your compiler tells you? It is an Apple / ACE problem. The given variable is a simple long on Windoof and Linux iirc. But in Mac it's a struct. You'll get the same problem when building TC with tools as the mmap tools uses ACE threads. And before I forget: Why are you using GCC? I entered "Mac World" in Januar 2013. All I had to do was building ACE and readline on my own. There is a package for cmake and Apples ancient OpenSSL was supported by TC at this time. The only problem I had was a typedef redefinition in DetourNavMesh.h which I solved with renaming the new type name. And this worked for several months. But all worked fine with clang. After some months without TC I tried to build it again still on Mountain Lion and I got the problem you're describing here. I searched for the commit which introduced it and as it (the commit) is just fixing crashes I never had I commented those lines. Oh, I forgot, I built OpenSSL 1.0.1e to get it running and defining OPENSSL_LIBRARIES does not work as this variable is overwritten by a composition of OPENSSL_SSL_LIBRARIES and OPENSSL_CRYPTO_LIBRARIES so I had to define those instead. At the moment I'm having a linker error during build, which occurs since I'm running Mavericks.

Please check your database, I guess you did not apply all sql updates from the updates directory in the sql directory.