v90 Localhost Bugs

Initiate Mage
Joined
Jul 12, 2016
Messages
3
Reaction score
0
Hey guys, I was wondering if anyone here who is working with v90 or has worked with it in the past has come across these bugs in the localhost. I'm using the disabled damage cap localhost that can be found here http://forum.ragezone.com/f690/v90-localhost-damage-cap-834259/ because it is the only v90 localhost I can find. Any help, insight or links to the original v90 localhost would appreciated. Thanks.

Weird newline behaviour on client-side quest texts
rato - v90 Localhost Bugs - RaGEZONE Forums


Randomly occurring glitchy UI behaviour and monster disappearance/reappearance during boss battles. Even after logging out glitchy behaviour continues.
rato - v90 Localhost Bugs - RaGEZONE Forums

rato - v90 Localhost Bugs - RaGEZONE Forums

rato - v90 Localhost Bugs - RaGEZONE Forums
 
Well, I don't have any UI issues when my client is running GMS encryption game files, however when I switch them to a custom encryption the client seems to always run out of memory (that is what's happening, your client is simply running out of memory and will eventually crash).

As for your string issue, pretty sure that was a StringPool glitch, but I still need to finish coding quests before I look into that myself.
 
Upvote 0
Well, I don't have any UI issues when my client is running GMS encryption game files, however when I switch them to a custom encryption the client seems to always run out of memory (that is what's happening, your client is simply running out of memory and will eventually crash).

As for your string issue, pretty sure that was a StringPool glitch, but I still need to finish coding quests before I look into that myself.

I investigated the issue a little bit more and it seems to be related to this packet:


Code:
public static MaplePacket moveMonsterResponse(int objectid, short moveid, int currentMp, boolean useSkills, int skillId, int skillLevel) {
        MaplePacketLittleEndianWriter mplew = new MaplePacketLittleEndianWriter();
        mplew.writeShort(SendPacketOpcode.MOVE_MONSTER_RESPONSE.getValue());
        mplew.writeInt(objectid);
        mplew.writeShort(moveid);
        mplew.write(useSkills ? 1 : 0);
        mplew.writeShort(currentMp);
        mplew.write(skillId);
        mplew.write(skillLevel);
        return writer.getPacket();
    }

Basically if the client is told that a monster used a skill (if skillId != 0 and skillLevel != 0), it'll start using up a lot more memory per skill. Then if you just let it sit there, eventually the client will bug out because that memory is only freed after a delay. It's because the skill animations or something make the client use and cache so much memory for some reason that over the course of the entire fight (9 arms, 3 bodies) the client glitches out if you wait for enough time (~5 mins).

Also, if I make monsters never use skills then it seems to work fine. I can play the boss fight several times without issue.
 
Upvote 0
I investigated the issue a little bit more and it seems to be related to this packet:


Code:
public static MaplePacket moveMonsterResponse(int objectid, short moveid, int currentMp, boolean useSkills, int skillId, int skillLevel) {
        MaplePacketLittleEndianWriter mplew = new MaplePacketLittleEndianWriter();
        mplew.writeShort(SendPacketOpcode.MOVE_MONSTER_RESPONSE.getValue());
        mplew.writeInt(objectid);
        mplew.writeShort(moveid);
        mplew.write(useSkills ? 1 : 0);
        mplew.writeShort(currentMp);
        mplew.write(skillId);
        mplew.write(skillLevel);
        return writer.getPacket();
    }

Basically if the client is told that a monster used a skill (if skillId != 0 and skillLevel != 0), it'll start using up a lot more memory per skill. Then if you just let it sit there, eventually the client will bug out because that memory is only freed after a delay. It's because the skill animations or something make the client use and cache so much memory for some reason that over the course of the entire fight (9 arms, 3 bodies) the client glitches out if you wait for enough time (~5 mins).

Also, if I make monsters never use skills then it seems to work fine. I can play the boss fight several times without issue.

Well, this is just one of the many things that happens I guess.. For me, I have full mob attacks and movement for bosses (Zakum, HT tested) and I will never crash. Instead, something as simple as warping to a different map like Henesys->Perion will cripple all graphics being rendered until the client runs out of memory and just crashes the library.

It's really just the WZ files along with your graphics card I guess. This isn't anything new and has been a common issue even in v62 :/
 
Upvote 0
Yeah the graphic glitches is a known thing happening especially on NVIDIA graphic cards and in Royals and Legends (v62) the usual guide during bosses is to GFX during bosses right before a boss dies and during so

Basically if your client turn into a 'glitch mode' things like your inventory start to disappear half and such, but a thing also 100% happening is that when a 'revive' happens (aka Zakum 1 dies into Zakum 2) the client immediately crashes without an error. This also happens a lot of times if you simply move your mouse, or just for no reason. An other thing I've noticed is that when you are in 'glitch mode' Fraps (recording program) refuses to screenshot or capture video until you GFX, which is kind of interesting. An other effect for Vista users is that your desktop background will change to black for whatever reason. Taking screenshots with Scroll Lock while in glitch state also immediately crashes your client.

As described earlier people are advised to GFX a lot which basically is changing the graphic settings (using hotkeys, or else you'll crash) so the graphics is 'refreshed' if you change your graphic settings (low to high or high to low, etc) it fixes the issue but the same thing can trigger it again as expected, so people heavily having this issue GFX like every 20 seconds using hotkeys.

EDIT: Honestly I want to believe the GFX issue was introduced in v56 when they changed the client to support Window Mode (which also completely changed how the client runs afaik). I never remember GFX being a thing back in v55.
 
Last edited:
Upvote 0
Pretty sure #1 is a server (source) related issue and not client, as I've seen it in v83 servers and v62 servers before too.

No it's client sided. It usually occurs when editing the client's strings with Stredit or anything like that. (I guess that's stringpool related like eric said).
It usually doesn't occur on a 100% clean localhost, but you can also fix it on a edited one by removing the extra n's in their respective lines (probably by hand). That's how a friend of mine did it at least.

(Pretty sure I posted about this before but w/e)
 
Upvote 0
The n **** is from a client edit.. Usually from stredit. Or some dumbass that doesn't know how to edit the client properly.

I know a fix for the UI glitch but the friend who helped me find it wants to keep that private, it's an issue only low versions have(prob pre-bb)
 
Upvote 0
I know a fix for the UI glitch but the friend who helped me find it wants to keep that private, it's an issue only low versions have(prob pre-bb)

too bad this fix probably doesn't fix foreign encryptions loaded into the client.. seems all the libraries i've tried that import all image data from scratch causes corruption in some areas and then throws out of memory as soon as your new client gets ingame :(
 
Upvote 0
too bad this fix probably doesn't fix foreign encryptions loaded into the client.. seems all the libraries i've tried that import all image data from scratch causes corruption in some areas and then throws out of memory as soon as your new client gets ingame :(
Most likely not, but I'm not sure as I find custom encryptions to be useless.
 
Upvote 0
Most likely not, but I'm not sure as I find custom encryptions to be useless.

Well, changing a simple header is useless (or, even the read-only trick). But when it comes to having to stop idiots from stealing all the custom-made sprites we use, changing the wz/client-side encryptions entirely is quite useful. Also, if you were to use custom encryption and modify a lot more, you could easily block all your famous PE exploits and packet sniffers.
 
Upvote 0
Well, changing a simple header is useless (or, even the read-only trick). But when it comes to having to stop idiots from stealing all the custom-made sprites we use, changing the wz/client-side encryptions entirely is quite useful. Also, if you were to use custom encryption and modify a lot more, you could easily block all your famous PE exploits and packet sniffers.
Till someone releases your encryption and it's all pointless
 
Upvote 0
Do you know what grammar is? You act all big but spell like a noob.

But if your protection gets cracked(which can be done in 2 days automatically) you can't change or fix it.

You dumb as **** talking big game with no game. (one hunna percent)
 
Upvote 0
Do you know what grammar is? You act all big but spell like a noob.

But if your protection gets cracked(which can be done in 2 days automatically) you can't change or fix it.

please unpack and unvirtualize my dll's in 2 days or less.

WZ Ciphers when parsing the files is never sent over a network unlike the client ciphers, so you can't just find a UserKey/Sequence easily. The actual encryption functions are additionally modified so you'd have to figure them out as well. I didn't think unpacking and unvirtualizing a VMProtected DLL was easy to crack, but hey, if you have the skill and tools (and not to mention, time) to be able to reverse everything, then go right ahead. In my defense, I'll just change everything up and try to protect it better.
 
Upvote 0
Back