11/05/03 EQLive breaks SEQ, discuss fixes here.

[Zaphod]

Hello all, and welcome to patch day...

First things first, the EQLive patch changed a bunch of opcodes and requires a new decode.cpp. I've attached the new decode.cpp to help people with the quest for opcodes...

Please respond to this thread with any opcodes and structure changes you find.

I'm hoping to get a working version of ShowEQ up later today.

Thanks and Enjoy,
Zaphod (dohpaZ)

[Talorn]

biggest improvement will be to change the following

OP_ZoneSpawns 0x0169
OP_GuildMOTD 0x01bd
OP_FormattedMessage 0x01dd
OP_AutoAttack 0x016b

[Talorn]

OP_ClientUpdate 0x0024
OP_BeginCast 0x0019

[suseuser7341]

newOpcode = oldOpcode + 2

Anyone got an idea how to automate opcode.h after you found scheme ?)

[SEQLurker]

Here's my ever so slight contribution to the seq project - I added 2 to everything in opcodes.h, except for the ones that said obsolete at the bottom, because I was tired of adding 2 by then. ;) I don't think I ever even did the math in base 10, but I make no guarantees...

Anyway, it seems to work well for me with this change.

[old_fart]

Getting seg faults after zone loads and all skittles appear using 4.3.17.1 and this opcodes.h. Seems that some spawns are well out of bounds with lots of playerSpawnPosStruct and spawnPositionUpdate errors.

[bubbahlicious]

I just whipped up a little program to do the conversion of opcodes.h. Build it and run it in the same directory as showeq/src. It will create a file called 'opcodes.h.new':

#include <fstream>
#include <stdio.h>

int main() {
using namespace std;

fstream fin, fout;

char buf[255];
char overwrite[5];
char *pos;
int val;

fin.open("opcodes.h", ios::in);
fout.open("opcodes.h.new", ios::out);

while ( fin.good() ) {

fin.getline(buf, 255, '\n');

pos = strstr(buf, "0x");

if (pos) {

sscanf(pos, "%x", &val);
sprintf(overwrite, "%4.4x\n", val + 2);
for (int i = 0; i < 4; ++i)
pos[i + 2] = overwrite[i];

}

fout << buf << endl;

}

fin.close();
fout.close();

return 0;
}

The resultant opcodes.h seems to be working for me so far when I copy zaphod's decode and just run 'make' and 'make install'. I've only tested a few seconds of playtime so I can't really vouch for stability.

[fox]

ManaDecrementCode 0x00bc

[seqmage]

From EQlive site,
All servers will be coming down at 3am PST on 11/6/03
All servers except Kael Drakkal and Sebilis will be coming down at 3am PST on November 6th 2003 Estimated Downtime 3 hours.

[BlueAdept]

OMG THEY ARE PATCHING BECAUSE YOU GUYS ARE TOO GOOD AND FIXED SEQ TOO QUICKLY!!! :) JK

It is probably to fix the occasional crashing of the client with Geforce cards when you camp out. Ive had it happen to me last night and several people on EQ's board posted messages about it.

Hopefully no opcodes will change, but it will probably have a new exe file.

Edit:seems there were a few other bugs like spawned mobs killing themselves and a vender who was selling stuff for nothing.

[Talorn]

i havent seen any changes as far as opcodes go. I know they were having problems with crashes and problems with doors in VT.

--talorn

[Alfred]

No. No client side changes. I would bet it is a quick fix to allow non-gaters a way out of jaggedpine. :)

Yes that's right.. click on the gem to get out of jaggedpine and you end up back in the new zone. :)

[who_me_use_seq]

There was also a little problem in Earth where the ring mobs spawned with a huge dot already on them. The mystical Arbitor as well. Seems like it would be a good thing but it made timing things very nasty. Also the mobs in PoEB were undergoing some sort of civil war. The various mobs were killing each other.

Yet another quality patch from SoE.

[Zaphod]

Originally posted by who_me_use_seq
Yet another quality patch from SoE.

Heh, I didn't think they'd be ready this early in the month...

Bad SOE, no biscuit.

Enjoy,
Zaphod (dohpaZ)

[monster69]

MM spawned with a nasty DoT too. We had a monk damn near SOLO him last night! :rolleyes:

Also, that neat new targeting ring doesn't work with radeon cards and some other situations.

What a GREAT patch!

Monster

[EpicCamp]

Aye, that bug with mobs killing themself got some of the EP guilds angry last night. A guild that has wiped 5 times and is clearly unable to do any rings in poearth beat it and arbiter because the mobs killed themselves and they all got poearthb flags. Drama ensued, quite funny though.

[Dark]

its not just Gefore cards playing up, my 9800pro went all strange after switching to windows mode.

[hhh]

like all the pop mobs were killing themselves, and spawning the pp's but the pp's were not giving out flags.

The really cute thing is in the patch message they said they had 'fixed avators' when in reality, they were more broken than ever.

Thanks for wasting my guilds time the whole night :/

[LLMerc]

Can anyone confirm the validity of the fix posted here? Thanks for the code btw, compiled it and ran it. I am just wondering if offsetting OpCodes by 2 will actualy do the trick? I haven't had a chance to try it yet, so would just like a quick confirmation =)

Thanks

[Talorn]

yes. you can even grab the opcodes.h that was posted in this thread.

Works pretty good atm with these 2 updates. I am noticing a little bit that isnt decoding correctly but nothing that has been causing a problem yet.

--talorn

[karrde]

Maybe I screwed something up then. I dropped the new opcodes.h into my src dir, did a rm src/packet.o, make, make install, and had plenty of problems.

I'd have random spawns appear at extremems of the map and it even crashed on me, which I've never had do before.

I'm performing a make clean, make, make install now and will see how well it performs when I get home.

[QuerySEQ]

some of the structs are not offset by 2. I think some havent changed.. so setting all of them by 2 works for the "skittles" but other structs I think are the same as before...

anyways.. (still learning).. looking at packets and scratching head.. LOL

[guice]

Also, that neat new targeting ring doesn't work with radeon cards and some other situations.

Works with my 9700 Pro just fine.

[perlmonkey]

Originally posted by bubbahlicious
I just whipped up a little program to do the conversion of opcodes.h. Build it and run it in the same directory as showeq/src. It will create a file called 'opcodes.h.new':
[...]
The resultant opcodes.h seems to be working for me so far when I copy zaphod's decode and just run 'make' and 'make install'. I've only tested a few seconds of playtime so I can't really vouch for stability.

Hmmm do you mean:
perl -i.bak -ple 's/(0x)([a-f\d]+)/$1.sprintf("%04x",hex($1.$2)+2)/eg' src/opcodes.h ;-)

This command-line will fix the file in-place and leave the old one as "opcodes.h.bak". Working for me.

[perlmonkey]

Originally posted by perlmonkey
Working for me.

Spoke too soon. The view kept zooming out, and eventually crashed. The last messages were:


error on zone-client: size > left (size=74, left=38, opcode=0x0024)
error on zone-client: size > left (size=119, left=38, opcode=0x0024)
error on zone-client: size > left (size=74, left=38, opcode=0x0024)
error on zone-client: size > left (size=197, left=26, opcode=0x003b)
error on zone-client: size > left (size=68, left=38, opcode=0x0024)
error on zone-client: size > left (size=74, left=38, opcode=0x0024)
error on zone-client: size > left (size=97, left=38, opcode=0x0024)
error on zone-client: size > left (size=119, left=74, opcode=0x003b)
WARNING: OP_ClientUpdate (0024) (dataLen: 74 != sizeof(playerSpawnPosStruct):18 or sizeof(playerSpawnSelfStruct):30)
WARNING: OP_ClientUpdate (0024) (dataLen: 62250 != sizeof(playerSpawnPosStruct):18 or sizeof(playerSpawnSelfStruct):30)
Segmentation fault


If it's not obvious to someone why this is happening, I can probably hook up gdb and get a trace at some point....

And BTW is there a vB code that forces fixed-width, but doesn't insert the "code:" text?

[LLMerc]

Originally posted by karrde
Maybe I screwed something up then. I dropped the new opcodes.h into my src dir, did a rm src/packet.o, make, make install, and had plenty of problems.

I'd have random spawns appear at extremems of the map and it even crashed on me, which I've never had do before.

I'm performing a make clean, make, make install now and will see how well it performs when I get home.

Not sure if you have done this, but you also need to download the decode.cpp file and put it in the src folder =) It is indeed working for me, but I am not sure if it's flawless yet (hasn't crashed in an hour long session at least)

[lildr00d]

With the current opcodes and decode.cpp from above EQ seemed to work as expected for a few but as soon as I logged into PoFire problems started to happen..

Decoded fine.. saw skittles and players.. Then for no apparent reason it detected a zone change. All the sudden the coolist thing I ever saw happened.. SEQ was cycleing thru all the maps so fast I could hardly keep it.. It reminded me of Wargames when the compter was playing itself at Tic Tac Toe..

Oh and then after about 10 min of really kewl screens it crashed..

Anyways it was cool for awhile to watch it!!!



Loading, Please Wait... (Zone: 'unk_zone_30326')
Zone: Zoning, Please Wait... (Zone: 'unk_zone_30326')
Loading Filter File: /usr/local/share/showeq/filters_unk_zone_30326.conf
>> No such file: '/usr/local/share/showeq/filters_unk_zone_30326.conf', creating new one...
Zone: ChangeCode: Server, Zone:unk_zone_30326
Error opening map file '/usr/local/share/showeq/unk_zone_24084.map'!
Can't find spawn point file /usr/local/share/showeq/unk_zone_24084.sp
Loading, Please Wait... (Zone: 'unk_zone_24084')
Zone: Zoning, Please Wait... (Zone: 'unk_zone_24084')
Loading Filter File: /usr/local/share/showeq/filters_unk_zone_24084.conf
>> No such file: '/usr/local/share/showeq/filters_unk_zone_24084.conf', creating new one...
Zone: ChangeCode: Server, Zone:unk_zone_24084



WARNING: OP_ZoneChange (0134) (dataLen:4 != sizeof(zoneChangeStruct):76)!
WARNING: OP_ZoneChange (0134) (dataLen:4 != sizeof(zoneChangeStruct):76)!
WARNING: OP_ZoneChange (0134) (dataLen:4 != sizeof(zoneChangeStruct):76)!
WARNING: OP_ZoneChange (0134) (dataLen:4 != sizeof(zoneChangeStruct):76)!
WARNING: OP_ZoneChange (0134) (dataLen:4 != sizeof(zoneChangeStruct):76)!
WARNING: OP_ZoneChange (0134) (dataLen:4 != sizeof(zoneChangeStruct):76)!
WARNING: OP_ZoneChange (0134) (dataLen:4 != sizeof(zoneChangeStruct):76)!
WARNING: OP_ZoneChange (0134) (dataLen:4 != sizeof(zoneChangeStruct):76)!
WARNING: OP_ZoneChange (0134) (dataLen:4 != sizeof(zoneChangeStruct):76)!

[Nstalkerga]

Doenst seem to be related specifically to PoF.

Played there for about 2-3 hours last night part of which being some serious raiding..


Worked find for me their.

All in all its doing great.

Keep up the good work :)

[perlmonkey]

Originally posted by perlmonkey
Spoke too soon. The view kept zooming out, and eventually crashed.

I realized I had patched opcodes.h, but not picked up the new decode.cpp. I did that, and all was well. I have not had the zooming-out problem again, but that happened in upper Guk, and I have not been back there. Perhaps only certain zones are seeing this problem?

[suseuser7341]

@lildr00d: OP_ZoneChange is 0x136. Your Log shows that you used still the old value which was 0x134.

[Zaphod]

Originally posted by perlmonkey
I realized I had patched opcodes.h, but not picked up the new decode.cpp. I did that, and all was well. I have not had the zooming-out problem again, but that happened in upper Guk, and I have not been back there. Perhaps only certain zones are seeing this problem?

No, from your errors it was obvious that you forgot the decode.cpp.



error on zone-client: size > left (size=74, left=38, opcode=0x0024)
error on zone-client: size > left (size=119, left=38, opcode=0x0024)


These errors are almost exclusively related to having a bad decode.cpp.

Enjoy,
Zaphod (dohpaZ)

[suseuser7341]

And thanks perlmonkey, I knew perl would be great to do it, just I am too rarely using it to figure it out how to do it :)

[Leetlewon]

Procedure I used to get ShowEQ working again -

Downloaded newest version. Applied patch.

Dropped new decode.cpp and new opcodes.h into src.

Compiled & installed. Works great!

[cbreaker]

It works for me too but seq has crashed twice on me now, whereas before the EQ patch it was rock solid.

I'll run it from a terminal this time and post the details if it crashes again.

Thanks for the patch.

---
Okay here's the last screen of messages:


WARNING: OP_WearChange (0125) (dataLen:0 != sizeof(SpawnUpdateStruct):9)!
error: size > left (size=138, left=74, opcode=0x0024)
error: size > left (size=172, left=154, opcode=0x003b)
error: size > left (size=27399, left=198, opcode=0x0024)
WARNING: OP_WearChange (0125) (dataLen:0 != sizeof(SpawnUpdateStruct):9)!
WARNING: OP_WearChange (0125) (dataLen:0 != sizeof(SpawnUpdateStruct):9)!
error: size > left (size=15111, left=270, opcode=0x0024)
WARNING: OP_WearChange (0125) (dataLen:0 != sizeof(SpawnUpdateStruct):9)!
WARNING: OP_WearChange (0125) (dataLen:0 != sizeof(SpawnUpdateStruct):9)!
error: size > left (size=138, left=92, opcode=0x0024)
error: size > left (size=96, left=84, opcode=0x0024)
error: size > left (size=241, left=46, opcode=0x0024)
error: size > left (size=216, left=83, opcode=0x0024)
error: size > left (size=241, left=77, opcode=0x0024)
error: size > left (size=239, left=62, opcode=0x0024)
error: size > left (size=16650, left=144, opcode=0x0024)
error: size > left (size=239, left=8, opcode=0x5a03)
error: size > left (size=138, left=128, opcode=0x0024)
error: size > left (size=102, left=75, opcode=0x0024)
error: size > left (size=252, left=52, opcode=0x0024)
WARNING: OP_ClientUpdate (0024) (dataLen: 171 != sizeof(playerSpawnPosStruct):18 or sizeof(playerSpawnSelfStruct):30)
Segmentation fault

[JamTune]

LoL wow thought it was just something I did wrong. I zoned into VT and it didn't load the map but all the mobs where there. I loaded the VT map and was ok I can work with this.. Then WOW I was in CS,GE and lots more in a matter of sec. Now I know that wizards can travel fast but man I had no idea I could zone like that;)

[lildr00d]

Seems stable now.. I downloaded the opcode.h -- I figured I fubared something adding 2 to all the codes myself which is why I was haveing issues.

Everything that I use SEQ work.. skittles, where I am, where my guild is.. and more importantly where my corpse is.. har har

[cbreaker]

Weird. I did the same and it's really unstable. Crashes about every 5 minutes. Oh well I'll wait for the official release.

[CiscoKid]

Are the attached opcodes.h and decode.cpp files that are in this thread now in CVS or we still have to patch 'em in before compiling?

[mirulo]

First thx a lot for the great job !!!!! Thx at Developpers and contributing members for this !!!!


Are the attached opcodes.h and decode.cpp files that are in this thread now in CVS or we still have to patch 'em in before compiling?

These 2 updates appear in cvs -z3 update :

#make distclean
# cvs [...] login
# cvs -z3 update
# cvs [...] logout
# export [...] (if you need to do recomp qt, gcc...)
# make -f Makefile.dist && ./configure && make && make install

For me, with these 2 updates, the errors aren't no more in terminal window but Seq will crash (freez) all times... :/
(i've ATI Radeon 7000 familly's on my Seq computer, PIII 600, 512 Ram, RedHat 9 under Gnome Desktop)

(So sorry for my poor english, but i'm french :) )


Also i have one question: When i run Seq with root account, no error message appear in terminal but when i start Seq with my user account (with SU command of course) i have this message at begin: "Session Management error: Authentification Rejected, reason: None of the authentification protocoles specified are supported and host-based authentification failed "

It's bad Doc ??? :) You think i must prefer run Seq in root or this message aren't important ??

Thanks in advance !!

[BlueAdept]

Originally posted by mirulo
For me, with these 2 updates, the errors aren't no more in terminal window but Seq will crash (freez) all times... :/


When you say crash, do you mean that it will just sit there for a while (with the hard drive flashing), but eventually will decode?

When I upgraded to .17 and added the 2 files, I had to revert to using --itemdb-disable added onto the command line. Without it, it would just make my hd flash for about 1-2 minutes every time I zoned. .16 I didnt have this problem (but earlier versions I did).

[mirulo]

it would just make my hd flash for about 1-2 minutes every time I zoned

WOOOOOT Thx BA !!!!! In fact it's that: Seq freez and take about 30 sec/1min for decode each zone !!!! But i've all and he work very good !!!

Happy face !!! That's work !!!! You're the best !!! :D


So, for my second problem ... !!!??? :rolleyes: LOL


Bybye =)

[quackrabbit]

A patch to take bring CVS up to date with EQ as of 11/05/03.

NOTE: DO NOT USE THIS IF YOUR SEQ IS WORKING!

This is just a one stop patch that includes all of the fixes mentioned in this thread.

To install, download CVS then apply this patch to your showeq directory. (patch -p0 < cvstonov05.patch)

EDIT: File removed due to it now being obsolete. Thanks Zaphod!!

[PhoenixZorn]

Originally posted by BlueAdept

It is probably to fix the occasional crashing of the client with Geforce cards when you camp out. Ive had it happen to me last night and several people on EQ's board posted messages about it.

Hopefully no opcodes will change, but it will probably have a new exe file.

Edit:seems there were a few other bugs like spawned mobs killing themselves and a vender who was selling stuff for nothing.


The campout bug was experienced by me for a total of 24 hours, in which time I had dumped my install of Win2k AS for a copy of Win XP Pro. Needless to say, as soon as I realized what was going on, I switched back to 2k AS... with 2 accounts and one windows box... it's a pain in the ass to crash every time you camp.

By the way, those campout crashes that I experienced happened on wednesday... and I reinstalled 2k thursday morning at about 3am...