balrok

what you mean?

spiritguide-gossiphello is btw just casting spell 2584 on player - and i do this in the script

and the spiritguide-opcode is also sent inside mangoscode

sorry for the delay - i just wanted to write a _small_ patch to avoid that spiritguides can be seen by everyone - but then vladimir and triply told me to get rid of some old hacks and then i had to debug through the whole spellsystem to find out how to let dead creatures cast spells

so in bg2db branch now i moved spiritguides out of battleground code and into sd2 (i've included a sd2.patch in /mangos/ directory)

- it's a bit bad, that now spiritguides won't work without sd2 :-/

but also i think it isn't ok to implement such scripting functions into bg-code too.. also i could imagine to move in future every bg into sd2 (cause mostly bgs are just scripted instances/maps)

about alterac valley: cause i did much rewritings, i just wanted to wait until the dust has set cause i'm quite content with the current state now i will update the bg2db_av branch too and look forward to see triply or vladimir to talk about my patches (mostly the newest spell-things need to be talked about - i'm pretty unfamilar with spellcode )

and if you would test the patch, it will also be great - cause else i have to do hotfixes in future

so for installing - use branch bg2db

then apply sql: sql/updates/9999_* sql/not_for_master*

then use sd2 and apply inside it mangos/sd2.patch

then apply the sql ScriptDev2/scripts/battleground/bg.sql

i tested everything and couldn't find any error

edit: untested bg2db_av branch uploaded - will post again here, when i tested it

why not add this check in:

Unit.cpp

((InstanceMap *)m)->PermBindAllPlayers(creditedPlayer);

?

also it looks like you're using tabs instead of 4 spaces

and

+# Do not bind players to instances on boss kill

+# Default: 1 (true)

seems wrong - default should be false or text must be: "Bind players to instances on boss kill"

cause default we want that every player gets an id

in battleground.cpp

rewardmark()

at the beginning:

if (isArena() && isRated())

{

if (count == ITEM_WINNER_COUNT)

{

RewardItem(plr,YOUR_ITEMID,YOUR_AMOUNT_OF_ITEMS);

return;

}

}

no, sorry i pushed the wrong things

also i did many new commits now, but cause i haven't tested them all yet i won't push them to bg2db directly they can be previewed in: http://github.com/balrok/mangos/commits/bg2db9

i added many changes suggested by triply, improved some code, and then added functions for checking at db-loading if all events are there (found even a missing event with that^^) or if some eventids are wrong in database

but for the last part i have to think about a better system cause i implement many redundant functions which should get into one

so maybe tomorrow or the day after it i can again release a version which can be tested

also i will update bg2db_av then.. which should be added as own branch without bg2db (cause it actualy must have everythin from bg2db inside it to make sense)

edit put up a bg2db10 branch - there i cleaned up bg2db9 (merged commits together) and removed the pointer-thing which was slower and no good idea

and you don't get those errors with arathibasin?

anyway i updated bg2db_av to the same state like bg2db where m_activeevents got rewritten

i hope now it works (at least it compiles for me)

no - you have to revert the alterac-part and then you can merge with bg2db_av (no need to merge bg2db prior bg2db_av)

* What bug does the patch fix? What features does the patch add?

it moves (nearly) all creatures and gameobjects from the battlegroundcode to the database

* For which repository revision was the patch created?

master

* where can this patch be found

http://github.com/balrok/mangos/commits/bg2db

and if you want to use av-patch use this one:

http://github.com/balrok/mangos/commits/bg2db_av

* what needs to be done:

testing

fixing bugs:

spiritguides are also visible for living players right now

the intention of this patch is to keep the battleground-code clean.. mostly this was needed for alterac valley with several hundreds creatures/objects but for old battlegrounds it also helps (many lines of code could be removed through this)

the next intention is to keep content out of mangos - that's also why i converted all battlegrounds at once - cause the sql for the bgs will be found in future only at your database provider (but for testing this patch i have mysql in sql/not_for_master*.sql ) and not in mangos's sql updatefiles.. and cause i guess it will be a bit trouble for everyone i want to do everything at once

also with keeping content out of battlegrounds, creating custom things in battlegrounds won't be so difficult as before (or for example add those cats running around at the blacksmith in arathi is also easy possible)

so basically this patch adds no features but will allow easier future developement of bgs (hardest part was always to get the spawning/despawning working correctly - and writing down all spawns in the code :-/ )

and through that also implementation of alterac valley (see it as a first split of this alteracvalley patch)

also i could imagine using this system also for the outdoorpvp-patch..

what needs to be done?

* today i've seen that the spawned spiritguides will also be visible for living units

* go for more performance (i don't think it's slow - but performance is always good )

* testing it a bit (there are soo many objects now in database and often i've written the event-ids by hand, there could be some objects wrongly spawned (but i already tried to test as much as possible on my own)

* move buffobjects to database - should only be done when poolsystem will work for instances too.. it is already possible to move them to db - but then looks more like a hack and isn't the effort worth

---

the biggest change to my old bg2db-patch was the move from one-dimensional eventids to two-dimensional ids.. through that it's quite easy to map the nodeid+state to an event (example arathi: event1=node_stables, event2=horde or alliance or neutral)

cause for one node only one state can be at the same moment (a node can't be controlled by horde and controlled by alliance at the same time) i also added in code, that only one event2 can be active in an event1

so spawning an event mostly also despawns another event

for a full list, which eventids i use please look currently in the code or sql-files.. i will create later a list of those ids

if you have questions just ask - but i mostly tried to keep code understandable

if there are still issues with this formula please write a bugreport

changing the formula is no big thing.. only finding right calculation for it

ps: i backported this patch today.. if someone likes to have this formula for tbc

sorry for the late response.. but this eventai doesn't work for me.. for me it looks like eventai is still broken there

first try to use other ports for your server (for connecting there you can edit your realmlist like: "set realmlist myserver.org:1234") i guess a ddos wouldn't attack all ports.. or?

hello - as you might have already noticed, mangos doesn't have a release-schedule, therefore no one can tell a specific date, besides "when it's ready"

so just asking wouldn't help

you need to run configure and so on, only when the makefile changes.. so you can speed up your work

another trick to improve speed is: make -j (see manpage)

hello, thank you very much.. i'm not good with eventai, so i waited until now for someone who can do this ..

i will wait until tomorrow, or the day after it and then add it to repo (until then everybody can complain or correct it ^^)

is this query also for master? - or is the eventai in master the same like in mangos-0.12 ?

also i plan to change the id of this eventai-script to something random, which wouldn't overwrite acid-scripts

----

jordanmckeag:

i also pm'ed you, but maybe other need an answer too..

answer can also be found in the thread somewhere..

the coding style isn't very good, but no-one wants to look at it..

also if you stop your server it will spam some errors about wrong guids.. i guess it's from OutdoorPvP::addobject() but it's just a guess..

try with faction_a and faction_h = 75

(don't know if this is the best solution - but those faction should make this creature friendly to the world)

if you have a (tested+working) sql-query then, you can post here or pm me, so i can add it to the repository

@andriuspel:

i think i wanted to do it since i took over the repository, thanks for reminding me

i hope my text there is understandable..

just a plain statistic for master and building only game with -O0 and -j2

12:00 without this patch and 7:30 with this patch

also again: big thanks for this patch..

with that i'll have everyday 4:30min*NumberOfMangosCompilation more time to do something else

@darkside: i pm'ed you..

but maybe now i understand with last post..

so you want to say, that the new season6 formula a team with less than 1300 rating shouldn't lose rating?

is this right? and you're sure?

@alez:

and you say maximum rating gain is 20 - but i don't understand how to determine if a team is equal-rated to set 0 as maximum loss..

or has this again todo with the magic 1300-rating

so people with less than 1300 will always gain 20 on win and always gain 0 on lose?

i'm not sure, but maybe this patch doesn't check if a creature/player has only gold as loot

cause bg-loot is:

bones->loot.gold = 123;

maybe a fix could be:

if (cVictim->GetCreatureInfo()->lootid || cVictim->loot.gold)

(but haven't looked if bones are also a Unit and if it compiles and so on.. just a guess )

edit: sorry my idea is wrong - cause cVictim is actually a creature.. so i don't know how it could affect player

the config is:

m_configs[CONFIG_ARENA_SEASON_ID] = sConfig.GetIntDefault ("Arena.ArenaSeason.ID", 1);

so you need Arena.ArenaSeason.ID = 6

what do you mean with your bug?

if a mob doesn't have loot on blizz then it won't sparkle.. you easily can check it in alterac valley in the mines.. all mobs there don't drop anything (at least with 2.4.3)

about round robin: it's long time ago that i played on blizz but i think it was set at death of a creature

oh - thank you

good point sinaka - i think this is also a bug in current code (i'm not sure if a negative rating is theoretical possible with the old 1500 system, but a check wouldn't harm too)

if noone says something against xtzgzorex's patch (espacialy the rating-formula), i would apply it together with negative-value check..

edit:

commited

i'll wait a bit - and if no problems occur, i can backport all commits from today too.. (cause we check in config - i think it's ok to have this rating-feature also in mangos-0.12 .. people who don't like it just use seasonids <6)

if(pVictim->GetTypeId() == TYPEID_PLAYER)

{

std::string ip = ((Player*)pVictim)->GetSession()->GetRemoteAddress();

}

should be the trick

griffonheart: i think only the files for it exist.. maybe trinity has already something done for it.. but i'm not sure..

fell: this patch needs to be read at least once, cause it's quite big, and making changes afterwards is more difficult than a change before..

and reviewing must be done, cause if noone understands this patch, it will be hard to implement things in future..

also i see no problem in using this branch.. it's quite up to date all the time

hello, besides that i don't realy understand the new rating system i can help to improve your code (:

"ArenaTeam::stats.rating" should be just "stats.rating" because you want to access a variable inside this class and not a static-variable from this class

and

"enemy_rating=1300.0f;" should be just "enemy_rating=1300;" cause storing a float in an integer has no effect

thats not critical i think.. but it needs a rewrite of the spawnfunctions, which i won't do in the next time