darkstalker

this is so old, and finally on master :3

if you're gonna have a dedicated pc why not use linux? windows is terrible for servers

Is it possible to run builtin-ACE configure in the cmake configure step (when running 'cmake ..'). That's what currently causes problems while doing multi-jobs builds (make -jN).

That external project thing doesn't fit well with the whole cmake build system.

Values displayed in client are wrong, like mana/energy costs, casting times. It displays like you had no talents, but they're working server side. For most effects it's harmless, but some of them are used for client-side spell cast checking.

Mangos Version: 11751

Custom Patches: none (tested on R2 and clean mangos, was the same)

SD2 Version: none

Database Name and Version : YTDB 604

How it SHOULD work: spell mods affect both client and server effects

How it DOES work: spell mods are affecting only server side effects, client side effect is missing.

Can easily test with a shaman spec'd into Improved Ghost Wolf. If you have the talent, client displays cast time on Ghost Wolf and doesn't let you cast while moving, but if you stand still you can cast it and it goes instant cast.

MaNGOS doesn't have SpellFamilyFlags.test<> stuff.

And the way chosen is wrong.

that's patch for R2's extended ClassFamilyMask

we should stop using printf-like functions and use stream-type ones (like std::cout)

blizz servers run with a different version of the dbc. Shipped one with client is only meant for displaying spell data on interface.

Or maybe they create/modify spell entries in run-time.

new version of patch here, for anyone still interested in bit access and enumeration of spell family masks.

Visual studio doesn't follow standards and compiles wrong code in several cases. Better use gcc as reference.

Added sample implementation in base of mangosR2 branch

https://github.com/darkstalker/mangos/commit/eb8400dcdf5de4fa63ebd3d4edc937c86c6476b9

R2 contains all stuff on rsa branch? or it's totally different?

comments would be interesting if they wouldn't be on russian ><

You can reproduce it like this:

1. Mount up

2. Enter a battleground (like WSG)

3. Try passing through a door (you're too fat)

Seems to be caused by map switching while mounted.

Cooldown system should be rewritten with millisecond precision, so can properly implement spells like 70811

whats the "R2" repo?

Most dupe methods use the "item save fail" to distribute items without losing the original copy.

missing this check:

diff --git a/src/game/Spell.cpp b/src/game/Spell.cpp
index d1961ee..b19a05d 100644
--- a/src/game/Spell.cpp
+++ b/src/game/Spell.cpp
@@ -5862,6 +5862,10 @@ SpellCastResult Spell::CheckRange(bool strict)
                if (target == m_caster)
                    return SPELL_CAST_OK;

+                if (m_caster->GetTypeId() == TYPEID_PLAYER &&
+                    (m_spellInfo->FacingCasterFlags & SPELL_FACING_FLAG_INFRONT) && !m_caster->HasInArc(M_PI_F, target))
+                    return SPELL_FAILED_UNIT_NOT_INFRONT;
+
                float range_mod = strict ? 0.0f : 5.0f;
                float base = ATTACK_DISTANCE;
                if (Player* modOwner = m_caster->GetSpellModOwner())

It the spawn is on terrain, the Z coordinate is the ground level (map data)

no info on your report, needs at least to include a crash dump

IsNoStackSpellDueToSpell data should be stored on a db table..

Character should stay ingame for 20 seconds after a forced disconnect (matches the logout timer)

can't use boost instead of ACE?

Maybe the requires files are only in the GM client now

There is a few spells that shouldn't proc on absorb, like priest Inner Fire

i know that Gear Score addon influences target selection in some way.