NetSky

What about Poisons? they will get removed at vanish if i interpret your change correctly... i think it would be better to empty all threat lists where the related player is listed...so if creature should aggro again and player is not visible he should be ignored by ai ...

work is in progress but possessed will be finished first before summon possessed will be completed(eye of kilrogg, eye of acherus,etc.) btw this quest also need eventai or sd2 support i got it working so far but removing m_mover and setting back to charmer does not work at the moment :-/

del... double post caused by laggs :-/

it is not blizzlike yet and is still missing some features ... so i wont think it will get implemented that soon... btw... diff --git a/src/game/Player.cpp b/src/game/Player.cpp index 7a3d09e..020313e 100644 --- a/src/game/Player.cpp +++ b/src/game/Player.cpp @@ -6380,7 +6380,7 @@ void Player::CheckDuelDistance(time_t currTime) bool Player::IsOutdoorPvPActive() { - return (isAlive() && !HasInvisibilityAura() && !HasStealthAura() && (HasFlag(PLAYER_FLAGS, PLAYER_FLAGS_IN_PVP) || sWorld.IsPvPRealm()) && !HasUnitMovementFlag(MOVEMENTFLAG_FLYING2) && !isInFlight()); + return (isAlive() && !HasInvisibilityAura() && !HasStealthAura() && (HasFlag(PLAYER_FLAGS, PLAYER_FLAGS_IN_PVP) || sWorld.IsPvPRealm()) && !HasMovementFlag(MOVEMENTFLAG_FLYING2) && !isInFlight()); } void Player::DuelComplete(DuelCompleteType type) if you want this to get implemented provide some useful information for further development

i will see if i find some time to make trade/follow/ etc work two sided at the moment i am working @ fixes for main core and new anti cheat system note: very important is the issue to prevent mixed group sign up at bg arena!

i also do not think it is a good idea to change all values in db by formula... in the end this will just mess up the database

wips spirit(visual) is bugged again ... but i did not find the reason yet

as i see you use package compression... try it without package compression or lower the compression rate

could you post a full backtrace?

@ derex sry i was wrong this time http://github.com/mangos/mangos/commits/72768af990ded10591958d0fe5a601552a2deeb7 - buf changes were causing this server/client crashes ... i reverted the wrong apend calls and changed buf size - everything is running well again

analyzing my crash dumps further i realized that my las crashes were related to buffer overflows caused by worldsocket somehow... i think this change could be responsible because buffer is set to big : http://github.com/mangos/mangos/commit/128e3d09a55113577139de4613a20dfa194b9697 i will revert and try again.... i hope someone will read this... @ waza seems your crashes got less... at least if you applied the posted patch this should fix your instance crashes... the other one is related to your freeze handler somehow could also be buffer overflow.... it would help alot if you post more than this but in the end there seems vmaps or bufferoverflow in ace netcode be responsible...

Revision : 7911 Patches : SD2 1106 UDB : 380 ACID : 29 http://pastebin.com/m36ee12f1 this could be caused by faked packages because i get empty vectors in socket?? - thats strange.... @ Waza try this : diff --git a/src/game/ObjectAccessor.cpp b/src/game/ObjectAccessor.cpp index b235e5a..7b87383 100644 --- a/src/game/ObjectAccessor.cpp +++ b/src/game/ObjectAccessor.cpp @@ -383,7 +383,7 @@ ObjectAccessor::Update(uint32 diff) { Object* obj = *i_objects.begin(); i_objects.erase(i_objects.begin()); - if (!obj) + if (!obj || !obj->IsInWorld()) continue; _buildUpdateObject(obj, update_players); obj->ClearUpdateMask(false);

all of you crashes appear at gridunload have you ever tried disabling grid unload in mangosd.conf? at least you should have enough RAM for doing this.....

hmm its not blizzlike but awesome idea anyways for all those who like soccer or football (I do not but I still think it is a nice Idea...) what do you want to use as ball? a gnome?

bool Map::UnloadGrid(const uint32&, const uint32&, bool)" #1 0x08217a9e in Map::UnloadAll (this=0x1354dd00, pForce=true) at ../../../src/game/Map.cpp:1041 grid = (NGridType &) @0x0: <error reading variable> disable unloadgrid.... look where the variable was initialized... and compare create/write method variable for this method (output) with bool unloadall used variables

maybe using more network threats helps.... are you using vmaps? collision detection?

i can remember this was already working without the patch

Rev : [7878] gdb crash logs : http://pastebin.com/m836c33e its obvious this is related to vmap calculations... anyways is there anyone working on optimizing vmaps ?if there is any project about optimizing vmaps i would love to join! just contact me... best regards netsky

Rev : [7878] server.log at ../../../../src/shared/Database/SqlDelayThread.cpp:38 s = <value optimized out> #3 0x084c5e01 in ACE_Based::Thread::ThreadTask (param=0x863bb30) at ../../../src/shared/Threading.cpp:159 No locals. #4 0xb7c46047 in ACE_OS_Thread_Adapter::invoke (this=0x863ba48) at ../../../../dep/ACE_wrappers/ace/OS_Thread_Adapter.cpp:90 hook = (class ACE_Thread_Hook *) 0x9ae0e58c func = ( ACE_THR_FUNC_INTERNAL) 0x84c5df0 <ACE_Based::Thread::ThreadTask(void*)> arg = (void *) 0x863bb30 status = (ACE_THR_FUNC_RETURN) 0xfffffdfc #5 0xb7bf2461 in ace_thread_adapter (args=0x863ba48) at ../../../../dep/ACE_wrappers/ace/Base_Thread_Adapter.cpp:124 status = (ACE_THR_FUNC_RETURN) 0xfffffdfc #6 0xb78da0bd in start_thread () from /lib/tls/libpthread.so.0 No symbol table info available. #7 0xb775901e in clone () from /lib/tls/libc.so.6 No symbol table info available. Thread 2 (Thread -1219540048 (LWP 18063)): #0 0xb78dfb3c in __nanosleep_nocancel () from /lib/tls/libpthread.so.0 No symbol table info available. #1 0x084c5ea3 in ACE_Based::Thread::Sleep (msecs=10) at ../../../dep/ACE_wrappers/ace/OS_NS_unistd.inl:918 No locals. #2 0x084be68b in SqlDelayThread::run (this=0x8639688) at ../../../../src/shared/Database/SqlDelayThread.cpp:38 s = <value optimized out> #3 0x084c5e01 in ACE_Based::Thread::ThreadTask (param=0x8639688) at ../../../src/shared/Threading.cpp:159 No locals. #4 0xb7c46047 in ACE_OS_Thread_Adapter::invoke (this=0x8630ac0) at ../../../../dep/ACE_wrappers/ace/OS_Thread_Adapter.cpp:90 hook = (class ACE_Thread_Hook *) 0x17381fe8 func = ( ACE_THR_FUNC_INTERNAL) 0x84c5df0 <ACE_Based::Thread::ThreadTask(void*)> arg = (void *) 0x8639688 status = (ACE_THR_FUNC_RETURN) 0xfffffdfc #5 0xb7bf2461 in ace_thread_adapter (args=0x8630ac0) at ../../../../dep/ACE_wrappers/ace/Base_Thread_Adapter.cpp:124 status = (ACE_THR_FUNC_RETURN) 0xfffffdfc #6 0xb78da0bd in start_thread () from /lib/tls/libpthread.so.0 No symbol table info available. #7 0xb775901e in clone () from /lib/tls/libc.so.6 No symbol table info available. Thread 1 (Thread -1219492160 (LWP 18059)): #0 0xb78db183 in pthread_join () from /lib/tls/libpthread.so.0 No symbol table info available. #1 0xb7c96a7a in ACE_Thread_Manager::wait_task (this=0xdcd1138, task=0xde9a1fc) at ../../../../dep/ACE_wrappers/ace/OS_NS_Thread.inl:2794 i = 0 copy_count = 1 result = <value optimized out> #2 0xb7c8ffea in ACE_Task_Base::wait (this=0x0) at ../../../../dep/ACE_wrappers/ace/Task.cpp:90 No locals. #3 0x08406174 in WorldSocketMgr::Wait (this=0xde9a1bc) at ../../../src/game/WorldSocketMgr.cpp:102 i = 1 #4 0x0812d899 in Master::Run (this=0x8631f50) at ../../../src/mangosd/Master.cpp:318 pidfile = {static npos = 4294967295, _M_dataplus = {<std::allocator<char>> = {<__gnu_cxx::new_allocator<char>> = {<No data fields>}, <No data fields>}, _M_p = 0x860d5d4 ""}} t = {m_iThreadId = 2918460336, m_hThreadHandle = 2918460336, m_task = 0xdb50eb0, static m_ThreadStorage = {_vptr.ACE_TSS = 0x8583348, keylock_ = {lock_ = {__m_reserved = 0, __m_count = 0, __m_owner = 0x0, __m_kind = 0, __m_lock = {__status = 0, __spinlock = 0}}, removed_ = false}, once_ = true, key_ = 3}, static m_TpEnum = { m_priority = {0, 0, 0, 0, 0, 0, 0}}} td2 = {m_iThreadId = 2910071728, m_hThreadHandle = 2910071728, m_task = 0xdc5ac38, static m_ThreadStorage = {_vptr.ACE_TSS = 0x8583348, keylock_ = {lock_ = {__m_reserved = 0, __m_count = 0, __m_owner = 0x0, __m_kind = 0, __m_lock = {__status = 0, __spinlock = 0}}, removed_ = false}, once_ = true, key_ = 3}, static m_TpEnum = { m_priority = {0, 0, 0, 0, 0, 0, 0}}} wsport = 8085 bind_ip = {static npos = 4294967295, _M_dataplus = {<std::allocator<char>> = {<__gnu_cxx::new_allocator<char>> = {<No data fields>}, <No data fields>}, _M_p = 0xdc5e51c "0.0.0.0"}} #5 0x0812d0b4 in main (argc=1, argv=0xbfb852f4) at ../../../src/mangosd/Main.cpp:163 cfg_file = 0x85120c8 "/home/dev/309/etc/mangosd.conf" c = <value optimized out> 143 uint32 GetEntry() const { return GetUInt32Value(OBJECT_FIELD_ENTRY); }

Rev : [7721] GetInstance: object Defias Companion(706), typeId 3, in world 1, should be in map 36,47 but that's not loaded yet. mangos-worldd: ../../../src/game/MapInstanced.cpp:126: Map* MapInstanced::GetInstance(const WorldObject*): Assertion `false' failed. Program received signal SIGABRT, Aborted. [switching to Thread -1375237200 (LWP 4594)] 0xb7741947 in raise () from /lib/tls/libc.so.6 maybe maps should be loaded earlier or we could implement something for loading maps at asserting fail while setting players to some kind of waiting list

ok this is kind of obvious this should be a constant if i look at variables and mapping variables. thx i will test it now! i think this should be added to git =) one question is left : i thought const is for methods that return explicit values but maybe i am wrong - at least i hope so =)

REV: [7740] Initializing Action Buttons for '15795' Program received signal SIGSEGV, Segmentation fault. [switching to Thread -1376261200 (LWP 2917)] std::_Rb_tree<unsigned char, std::Pair<unsigned char const, ActionButton>, std::_Select1st<std::Pair<unsigned char const, ActionButton> >, std::less<unsigned char>, std::allocator<std::Pair<unsigned char const, ActionButton> > >::find ( this=0xe576298, __k=@0xadf7e13b) at /usr/lib/gcc/i486-linux-gnu/4.1.2/../../../../include/c++/4.1.2/bits/stl_tree.h:1317 1317 if (!_M_impl._M_key_compare(_S_key(__x), __k)) so if you ask me this is caused by changing iterator to const_iterator or const_iterator to iterator in previous commits i will revert these suppose the crashes are gone.....

i tried to point out that if there are several spells/talents beeing applied at shapeshift there should be general local spellid/value vars and different placed cast method accesses using this variables/values. thats all. Edit : Did you realize Druids can access some talents even if they do not have the right shapeshifting form.... i realized this is possible with sprint...this is fixed with some checks... but i wonder if there could be more... did you realize anything concerning this issue?

yeah i just was talking bout this : there should be general additional cast methods so we do not have to add custom ones if there are several spells/talents like this they could all use the same cast access and you do not have to add different ones

thats right but additional local variables and checks within this shapshifting solution for just one ability is hacky if you ask me