Tassader2

I don't think Tom_Rus intented to be rude by posting his 2010 version. The versions were fairly different, with dinstict approaches. Unluckily one approach is far better than the previous only because of structure, since it's more viable. Your version, Neo, was very good in ideas, coding and progress, so you don't need to feel bad for how the situation in this thread evolved.

I personally like the Neo2003's version more, both the code and the DB layout, but the neccessity to be able to run win32 code is fatal weakness. That could be avoided by using (few) pre-generated seeds and keys for each module, stored in db.

Hi TOM_RUS, firstly BIG thanks for amazing work.

I am testing on local, and i have questions, my gm account continue been banned by warden without any third program...

Here examples of this bans

2011-06-29 06:09:31 RESULT PAGE_CHECK fail, CheckId 800 account Id 5

2011-06-29 06:36:40 RESULT PAGE_CHECK fail, CheckId 261 account Id 7

2011-06-29 06:36:40 RESULT PAGE_CHECK fail, CheckId 88 account Id 7

2011-06-29 06:36:42 RESULT PAGE_CHECK fail, CheckId 261 account Id 5

2011-06-29 15:45:56 RESULT PAGE_CHECK fail, CheckId 799 account Id 5

2011-06-29 15:56:00 RESULT PAGE_CHECK fail, CheckId 799 account Id 5

2011-06-29 16:17:59 RESULT PAGE_CHECK fail, CheckId 261 account Id 5

2011-06-29 18:53:57 RESULT PAGE_CHECK fail, CheckId 88 account Id 5

2011-06-30 00:48:43 RESULT PAGE_CHECK fail, CheckId 261 account Id 5

2011-06-30 03:19:02 RESULT PAGE_CHECK fail, CheckId 799 account Id 5

2011-06-30 03:53:15 RESULT PAGE_CHECK fail, CheckId 134 account Id 5

2011-06-30 04:36:18 RESULT PAGE_CHECK fail, CheckId 261 account Id 11

2011-06-30 17:12:57 RESULT PAGE_CHECK fail, CheckId 782 account Id 11

2011-06-30 17:58:36 RESULT PAGE_CHECK fail, CheckId 88 account Id 11

Thanks for any reply man

There are false positives indeed. A friend of mine told me warden is banning for having Ati Tray Tools opened, which only changes Anti Aliasing and other graphic stuff.

This quite screws up everything if not fixed. Checks that found Ati Tray Tools as cheat are page ckecks with id 88, 261 and 783 (with `address` = 174688 AND `length` = 37)

This one looks to be false positive as well (TOM_RUS version, MEM_CHECK)

(438, 243, '', '', 11287980, 8, '04000000903C9F00', NULL);

I use this but i have many false detections how i can create my database ? i launch 3.3.5a game

what do you mean by creating database?

i am having this crash

http://pastebin.com/TCrjus73

wd->str.c_str() is NULL and wd->str.size() is 24 ? :-o

There are some dupes in SQL data (TOM_RUS version)

SELECT MAX( id ) 
FROM`warden_data_result` 
GROUP BY address, length, data, result
HAVING count(*)>1
ORDER BY MAX( id )ASC; 

DELETE FROM `warden_data_result` WHERE id IN 
(385,450,473,476,477,482,483,484,488,497,554,569,617,626,652,655,667,694,747,758,763,765,791);

And what do you think MEM_CHECK is for? I'm not sure what you're definition of "modify the client" is, but if it's screwing with WoW's data I would most certainly call that modifying the client.

It may check only the code, not data (as data that are dynamic probably can't be digitally signed) - like run speed that's just a variable in memory, that may change every now and then.

But I don't know - that's why I was asking ;-)

Can someone explain what the "checks" check for?

If I understand it correctly (which I may not), this check types

PROC_CHECK = 0x7E, // uint Seed + byte[20] SHA1 + byte moluleNameIndex + byte procNameIndex + uint Offset + byte Len (check to ensure proc isn't detoured)

MEM_CHECK = 0xF3, // byte moduleNameIndex + uint Offset + byte Len (check to ensure memory isn't modified)

MPQ_CHECK = 0x98, // byte fileNameIndex (check to ensure MPQ file isn't modified)

LUA_STR_CHECK = 0x8B, // byte luaNameIndex (check to ensure LUA string isn't used)

TIMING_CHECK = 0x57, // empty (check to ensure GetTickCount() isn't detoured)

check for modification of the client itself and theese

DRIVER_CHECK = 0x71, // uint Seed + byte[20] SHA1 + byte driverNameIndex (check to ensure driver isn't loaded)

PAGE_CHECK_A = 0xB2, // uint Seed + byte[20] SHA1 + uint Addr + byte Len (scans all pages for specified hash)

PAGE_CHECK_B = 0xBF, // uint Seed + byte[20] SHA1 + uint Addr + byte Len (scans only pages starts with MZ+PE headers for specified hash)

MODULE_CHECK = 0xD9, // uint Seed + byte[20] SHA1 (check to ensure module isn't injected)

look for malicious programs/drivers, ok?

And let's say I have the source code of of some well-known cheat program, like WoWEmuhacker (I don't, but _someone_ does). And I rename it, do some really minor modifications in the code and recompile it, so its checksums are different. Now it can't be detected by any of the last four checks and probably naither can it be by the rest, because it does not modify the client, only it's data in memory (like movement speed). Is this correct?

I merged master into blink1 & blink3. It should work, but is not tested (I am still on 3.22a client). Blink3 most likely cannot be applied on preWotlk branches due some changes in teleport code.

As for the development: I chose to make various patches (v1, v2, v3, and wanted even v4) mainly because I thought that slighter modifications would have bigger chance of being accepted than complete (and maybe incorrect or hacky) rewrite. Patches 1-3 only modify and correct the clearly incorrect pre-cast check ;-) I was afraid that modification of the actual "blink" code could cause performance issues or likely be too hacky. Maybe I should have post it in Under Review

Anyway I never finished v4 probably of 2 reasons -

1. few people posted better patches (at least they said they were ) and

2. I (playng a mage) was perfectly happy with results of blink3. Both meaning that further develompent is not that necessary and I better iinvest my time elsewhere

The server where I play switched to TC, where blink works fine (similarly to blink3 ;-) but in TC it teleports you a little higher in the air) so I have no further plans with this patch (exept for maybe creating a 0.12-compatible blink3 version).

2sidsukana if it does not teleport in instances, then you probably don't have vmaps. Without them blink cannot function properly (but my patch will still improve it ).

It's a pitty that previous versions with not removed icon check are not aviable

Dodge and parry are not diminished.

Gains from dodge rating, parry rating, agility and defense (and maybe from strength for DKs?) is what is subject to DR ... and that should be calculated during "rating->%" and "stat->%" conversion

For miss you need to separate base miss and miss from defense rating, however I am unsure what if you do not have maxed "base" defense skill ... ?

Or is it simple 5% base miss and everything else is diminished (ie. there's no difference between "skilldiff" from level difference and from defense rating) :confused:

bump ;-)

what about keeping the icon check as last resort check if removing it causes problems?

maybe just remove all auras with

AURA_INTERRUPT_FLAG_UNK15 = 0x00008000, // 15 removed by casting a spell?

?

but some creatures can jump without any spells

Then maybe replace

+            const_cast<CreatureDataAddon*>(addon)->move_flags &= ~(MONSTER_MOVE_JUMP|MONSTER_MOVE_UNK4);

with

+            const_cast<CreatureDataAddon*>(addon)->move_flags &= ~(MONSTER_MOVE_UNK4);

?

Well, it's incorrect anyway ...

I think that EffectMiscValue determines the horizontal speed and vertical speed (or angle?) should be calculated from distance and horizontal speed so target lands (+-) right in front of caster (damage can probably be used to override it)

For example http://www.wowhead.com/?spell=37370 (EffectMiscValue=300) should pull players very fast, with very low vertical speed ... and to the caster, not far behind him like on mangos

And http://www.wowhead.com/?spell=32265 (EffectMiscValue=125) pulls players at much smaller horizontal speed, but higher vertical. And comments says "If you are far away, you are also pulled significantly upwards" ...

spell healing and spel damage can be in different rate affected by spell power.

I don't think so - last time when I looked into spell_bonus_data table there was only "direct_bonus" column, not "direct_bonus_damage" and "direct_bonus_heal" and because healing and damage is done by the same spell, it gets same bonus for damage and than again for healing.

Well, we could include the new formula, but in my eyes it would be better, to do it in another patch, since

1) its some different issue

2) its a version issue: someon, who is running an older server may want to stick to the old formula but wants do get rid of the problem that is adressed by the actual patch.

Well, yes, I just thought that there is (although a very little) chance that this bug gets magically fixed by switching to new formula (and using max level instead of spell level)

I found a little bug related to this fix - according to wowhead comments, if Drain Soul is casted on a target below or at 25% HP, the damage should be still 4 times higher even if the target gets during the channneling over 25% of its HP (f.e. it is healed by somebody...).

Currently, the damage is 4x higher if at the time of the periodic tick is target below or at 25% HP.

That's quite common amongst all HoTs/DoTs and delayed spells - their bonus damage, crit chance, multiplier ets. should be taken from the moment the spell is cast, but Mangos calculates that when they deal damage.

Thank you for explanation

But I still do not understand one thing - what is the difference between damage done a taken? Aren't those two supposed to be equal? I.e. the damage that caster does to target is equal to the damage that target takes from caster...Or do I interpret it wrongly?

Damage done does not include effects that increase/reduce damage taken - eg. dampen/amplify magic, curse of elements. And yes, theese should be calculated on hit.

Unsure how crit chance should be calculated, but most likely chance to critically hit should be calculated on cast and chance to BE critically hit (ie. modiifers like Molten armor or resilience) should be calculated on hit.

sorry if im wrong but I think this patch missed the DD spell bonus form Unstable Affliction in this part :confused:

according to http://www.wowwiki.com/Spell_power_coefficient

the DD spell coefficient for unstable affliction is 180%

maybe this should not be covered by default calculation and inserted in the table as

if not, at least someone more experienced could look at the DD spell coeficient of UA

sorry about bad english.

The triggered spell is 31117

EDIT: Strange that I wrote this post in reply to the following one, but it is shown before it :-o

Just note, talent Warbringer applies 3 auras. I think that 262 only enables charge in combat. Making charge usable in any stance should most likely be handled by aura 275 (ignore_shapeshift) and making charge remove movement impairing effect may be aura Override Class Script (so instead for aura 57499, you should be looking for SPELL_AURA_OVERRIDE_CLASS_SCRIPTS with misc value 6953 .... and maybe move it to separate patch )

And triggering juggernaut spell (crit bonus) is IMO not related to aura 262 as well

Note that, at some point beween 2.2 and wotlk, the downranking formula has been changed. AFAIK Mangos still uses old formula.

Ghostcrawler made the following statement about the downranking formula:

If caster level is less than max caster level, then the penalty = 1.0.

If caster level is at or greater than max caster level, then the penalty = (22 + max level - caster level) / 20.

The penalty is capped at 0. Max caster level per spell is also pretty inconsistent, but it's around 4 levels higher than the level at which you get the spell.

Basically, once the caster is 3 levels higher than the max cast level of the spell, the standard coefficient of the spell is multiplied by that penalty (losing 5% of the coefficient per level), with the coefficient reaching zero when the player is 22 levels higher than the max cast level.

Arena UnitID is NYI in mangos and addons for 3.x use it

"arenaN"

Opposing arena member with index N (1,2,3,4,5).

we need to check only HOLY school spell damage, in db we can check only normal spell damage.

Are you serious? The fact that db data are used only for "normal" spell damage and not spell specific (judgement is holy isn't it?) would render the table useless for most spells

better use delete/insert, because afaik REPLACE works only in Mysql

Just for note: base at nos4r2zod and my tests patch have real problem with some auras stacking. So generic rules must be rewrited.

Which auras?

Man, aura 262 is allso used here. http://www.wowhead.com/?spell=52437 You can't rename an antire aura only for one spell.

lol ... that's already mentioned above :-)

Thanks, now I understand how do you mean it, I wrongly considered it for a bug

If you mean 30% of stamina from master, it is also stated in the Character Info tab on the official, not only on wowwiki, the real HP difference between pets on the offi and mangos must be caused by something else than wrong % of stamina shared.

Then maybe incorrect (outdated) base HP ?