Forums

I've been seeing some posts about old KR clients and I can tell that the community is still interested in this language and old files. What I couldn't find anywhere is Korean serverfiles or server-side translations such as the equivalent of translate.lua or oxquiz.lua. locale_string.txt already contains the Korean translations so that's the easy part. Item and mob names can be found in the latest GF proto dump for anyone looking for them Skill names can be obtained by any skilltable.txt read in EUC-KR encoding Client side translations are obviously obtained from one of the Korean client downloads available out there. So all that's really missing here is quest and quiz strings, if anyone knows how to get them please drop a link down in the comments, if the link contains translations for Dragonsoul and Belt system that would be additionally awesome! If you are looking for these resources as well drop a reaction to this post so more people see it

How to fix the Aura of the Sword/Stealth/Feather Walk/Swiftness and similar effects being attached to the mesh after death: This is about what @WhiteBambo asked specifically, I've seen a lot of clients having this and this is not relevant to togglable slots and skills. The reason this is happening is because of the server's 3-second stun-death (classic mechanism from YMIR) and this should be an easy fix by this guide: Please adjust changes for your own source!!! Normally, when a monster or player kills you, the server first sends a stun packet and only 3 seconds later the real death with the affect clear. The client handles that stun packet by calling Die() at once for your own character, and Die() wipes the affect flags without running each affect's off path. When the server's "flags off" update arrives 3 seconds later, the client sees no change and never detaches the visuals that live outside the affect slot array, such as the Aura of the Sword weapon glow, the reach scale and the stealth alpha. That is why only your own corpse shows it, never other players' screens, and why /kill and one-shot deaths are clean: those skip the stun packet. The client fix below makes the early clear run every off path properly. Making deaths instant server-side also avoids it, but that is a gameplay decision, not a fix. Some client-side fixes: Swiftness and Feather Walk can survive this fix in some clients, as well as overwrite eachother, or show to observer clients even if the character's movement has stopped! Test whether you have this before applying the following fix: Catching false-positive flags and disabling them (the actual client-side fix for this issue):

hi where i can download?

Sadly I have to say goodbye to this site. Sadly I cannot change my email address to anything not Google. I don't use Google anymore so I don't even have access to the email inbox linked to this account. This is unfortunate but support here is basically non existent. Saying farewell to this forum. You can't ban me because I'm closing my account myself. The project will stay available over at Github and maybe some other forums.

Mob Death Dissolve System When a mob/NPC dies, its model dissolves away in irregular patches (disintegration look) over a few seconds, instead of the corpse just popping out. [*]Noise-based patch dissolve (not a flat fade), triggered on death [*]Mobs/NPCs only — players excluded [*]Configurable speed & patch size, and toggled by one define (easy revert) Download

hello

Looking forward to test it

Dear friends, I found a SQL-side bug in the "Great Shop" system that was causing the `player.item` table to bloat. I have fixed it—the implementation is shown below—but I would like to ask for your opinion: do you consider this an optimized and correct fix? Thank you in advance for your comments, and I wish you success in life. void ITEM_MANAGER::SaveSingleItem(LPITEM item) { #ifdef ENABLE_FIX_SKIPSAVE_ITEM_LEAK if (item->GetSkipSave()) return; #endif if (!item->GetOwner()) { DWORD dwID = item->GetID(); DWORD dwOwnerID = item->GetLastOwnerPID(); db_clientdesc->DBPacketHeader(HEADER_GD_ITEM_DESTROY, 0, sizeof(DWORD) + sizeof(DWORD)); db_clientdesc->Packet(&dwID, sizeof(DWORD)); db_clientdesc->Packet(&dwOwnerID, sizeof(DWORD)); sys_log(1, "ITEM_DELETE %s:%u", item->GetName(), dwID); return; } sys_log(1, "ITEM_SAVE %s:%d in %s window %d", item->GetName(), item->GetID(), item->GetOwner()->GetName(), item->GetWindow()); TPlayerItem t; t.id = item->GetID(); t.window = item->GetWindow(); switch (t.window) { case EQUIPMENT: t.pos = item->GetCell() - INVENTORY_MAX_NUM; break; #ifdef ENABLE_BELT_INVENTORY_EX case INVENTORY: if (BELT_INVENTORY_SLOT_START <= item->GetCell() && BELT_INVENTORY_SLOT_END > item->GetCell()) { t.window = BELT_INVENTORY; t.pos = item->GetCell() - BELT_INVENTORY_SLOT_START; break; } #endif default: t.pos = item->GetCell(); break; } t.count = item->GetCount(); t.vnum = item->GetOriginalVnum(); switch (t.window) { case SAFEBOX: case MALL: t.owner = item->GetOwner()->GetDesc()->GetAccountTable().id; break; default: t.owner = item->GetOwner()->GetPlayerID(); break; } thecore_memcpy(t.alSockets, item->GetSockets(), sizeof(t.alSockets)); thecore_memcpy(t.aAttr, item->GetAttributes(), sizeof(t.aAttr)); #ifdef ENABLE_SOULBIND_SYSTEM t.sealbind = item->GetSealBindTime(); #endif db_clientdesc->DBPacketHeader(HEADER_GD_ITEM_SAVE, 0, sizeof(TPlayerItem)); db_clientdesc->Packet(&t, sizeof(TPlayerItem)); } LPITEM ITEM_MANAGER::CreateItem(DWORD vnum, DWORD count, DWORD id, bool bTryMagic, int iRarePct, bool bSkipSave) --inside #ifdef ENABLE_FIX_SKIPSAVE_ITEM_LEAK item->SetSkipSave(bSkipSave); #else item->SetSkipSave(false); #endif @ Amun @ martysama0134 @ Intel @ Karbust @ Geoffrey @ Mali @ Owsap @ Distraught

Hello, I added a reaction to the post, but when I try to add another one, a message pops up saying I can only add one reaction today—yet I still don't see the download links Please help.

eline sağlık başkan

You should look for core syslog/syserr. In there, look for the timestamp and check what happened. Usually DCs go in there. Look in all cores, start with the 3.

The problem Server time calculation in Metin2 client uses only one packet to synchronize it. Server sends current timestamp once, client adjusts time according to it. If packet retreival is late so is calculated offset. This results in desynchronized in game interface, ex. affects are visible in the affect shower 2 seconds longer than they are actually applied. The fix Solution is fairly simple: by periodically measuring the Round Trip Time (RTT) and factoring in half of that delay to calculate the exact network offset, the client can continuously synchronize its clock with the server.

niceee

Thamnks

There were a lot of omissions and errors in the tutorial, but if you put some thought into it, it works. If there's anything I can do to help with the implementation, let me know.