My first contribution... Here goes nothing...

vesuvias · #1 02-14-2004, 06:45 PM

Forgive me if I do something taboo, I have only been following this scene for a little over a week. Any way I dove head first into the code and decided to fix a few things that were bugging me (err.. yeah).

Any way here is what I did:

-Zone Bind Points (added some fields to start_zones and fixed an issue that would crash the client if you died to your bind point)
-Fixed an issue that was causing you character creation skin selections to be lost, now everyone doesn't have to look the same (Yeeaa!!)

Additions

-the start_zone table now has support for wildcarding on selection this means that DB developers don't have to put in (number of selections on map screen) *(number of races) *(number of classes)*(number of dieties) = number records in the database. putting a -1 in of these fields in the DB basically selects all of the combinations for that field. So if I only wanted to have 1 starting point for each class I would put -1 in the diety, choice and race coloumn for each record.

When specifying start_zones just remeber this SQL statement (its the one we use now to get start_zones):

Code:

SELECT x,y,z,zone_id,bind_id,bind_x,bind_y,bind_z FROM start_zones WHERE (player_choice=%i OR  player_choice=-1) and (player_class=%i OR player_class=-1) and (player_deity=%i OR player_deity=-1) and (player_race=%i OR player_race=-1) ORDER BY select_rank DESC

select_rank allows you to assign importance to the start_zone record basically the more specific the start_zone record (records with more -1 wildcards are less specific) the greater the select_rank should be. A good rule is to start at 50 and for every wildcard you use subtract 10 from this value.

This allows you to specify things like I want all characters to start in the nexus as a default. However I want all warriors to start in Halas. This would require 2 records in the DB, one for the nexus with -1 for race,choice,class and diety and a select rank of say 10. And one for the warriors that go to halas and it would have -1 for race, choice, and diety and a select_rank of say 20. If later you wanted to further specify where agnostic warriors go, you would just add one more record with a higher select_rank.

Zoneing Fixes

well after I added that code I realised that bound characters that die crash thier clients (never knew that since I was never bound before). So I decided to hunt that bug down. It turns out that this peice of code was the culperit

Code:

					if (zc->zoneID != 0 &amp;&amp; dead)
					{
#ifdef GUILDWARS
					if(animation > 65 &amp;&amp; admin&lt;80 &amp;&amp; CheckCheat()){
						if(cheater || cheatcount>0){
							Message(15,"Cheater log updated...yup your busted,its not nice to cheat.");
							char descript[50]={0};
							sprintf(descript,"%s: %i","Death zone cheat");
							database.logevents(this->AccountName(),this->AccountID(),admin,this->GetName(),"none","Death zone cheat",descript,15);
							if(cheater==false){
								worldserver.SendEmoteMessage(0,0,0,13,"&lt;Cheater Locator> We have found a cheater.  %s (Acct: %s) was just caught hacking, please show them what we think of hackers...",this->GetName(),this->AccountName());
								cheater=true;
							}
							cheatcount=0;
						}
						else
							cheatcount++;
					}
#else
					break;
#endif
					}

This section executes in the OP_ZoneChange case under handlepacket in client_process.cpp. Any way basically if you died and the zone change packet specified a zone (other than 0) then you would break??? You got a zone change requrest why wouldn't you want to finish it? Anyway taking out the break fixed the issue it seems.

I also cleaned up the code a little in OP_ZoneChange so you wouldn't do wierd things like go to -3,-3,-3 when there was a perfectly acceptable safe_zone to go too.

Character skins

The last thing I decided to tackle was character skins and this was acutually the hardest thing of all to debug. First I corrected struct that recieved info from the client on character creation as there were a few guessed at spots inside that struct. I did this through numerous tedious trial and error chracter creations (Ie create a gnome a certian way, then create a second gnome the exact same way except this time change just his beard and analysis the data we get).

After I got the data for CharCreate_Struct ok, I moved onto trying to figure out why these character specific skin changes never got back to client. It turns out that CharacterSelect_Struct had about 60 bytes of unknown data. So I then went through a fairly painful process of trying to figure out what order the data went in on its way back. Fortunatly I was able to figure it out finally. It was challenging though. Hope everyone can benefit from some of my work at least a little

.

Here are the diffs:

First SQL changes

Code:

ALTER TABLE start_zones ADD bind_x FLOAT DEFAULT "0" NOT NULL
ALTER TABLE start_zones ADD bind_y FLOAT DEFAULT "0" NOT NULL
ALTER TABLE start_zones ADD bind_z FLOAT DEFAULT "0" NOT NULL
ALTER TABLE start_zones ADD select_rank TINYINT UNSIGNED DEFAULT "50" NOT NULL

Now the database.cpp diff:

Code:

Index: database.cpp
===================================================================
RCS file: /cvsroot/eqemu/eqemu/eqemu/Source/common/database.cpp,v
retrieving revision 1.1.1.11
diff -r1.1.1.11 database.cpp
1164,1165c1164
&lt; 				cs->gender[char_num]	= pp->gender;
&lt; 				cs->face[char_num]		= pp->face;
---
> 				cs->gender[char_num]	= pp->gender;				
1167a1167,1174
> 				cs->face[char_num]		= pp->face;
> 				cs->haircolor[char_num] = pp->haircolor;
> 				cs->beardcolor[char_num]= pp->beardcolor;
> 				cs->eyecolor2[char_num] = pp->eyecolor2;
> 				cs->eyecolor1[char_num] = pp->eyecolor1;
> 				cs->hair[char_num]		= pp->hairstyle;
> 				cs->beard[char_num]		= pp->beard;
> 				
6551c6558
&lt; 	cout&lt;&lt;"Choice:"&lt;&lt;in_cc->start_zone&lt;&lt;endl;
---
> 	//cout&lt;&lt;"Choice:"&lt;&lt;in_cc->start_zone&lt;&lt;" Class:"&lt;&lt;in_cc->class_&lt;&lt;" Race:"&lt;&lt;in_cc->race&lt;&lt;" Diety:"&lt;&lt;in_cc->deity&lt;&lt;endl;
6556c6563,6564
&lt; if (RunQuery(query, MakeAnyLenString(&amp;query, "SELECT x,y,z,zone_id,bind_id FROM start_zones WHERE player_choice=%i and player_class=%i and player_deity=%i and player_race=%i", in_cc->start_zone, in_cc->class_, in_cc->deity, in_cc->race), errbuf, &amp;result)) {         
---
> 	int qLen = MakeAnyLenString(&amp;query, "SELECT x,y,z,zone_id,bind_id,bind_x,bind_y,bind_z FROM start_zones WHERE (player_choice=%i OR  player_choice=-1) and (player_class=%i OR player_class=-1) and (player_deity=%i OR player_deity=-1) and (player_race=%i OR player_race=-1) ORDER BY select_rank DESC", in_cc->start_zone, in_cc->class_, in_cc->deity, in_cc->race);
> if (RunQuery(query,qLen,errbuf,&amp;result)) {         
6564c6572,6597
&lt;          in_pp->bind_zone_id = atoi(row[4]); 
---
>          in_pp->bind_zone_id = atoi(row[4]);
> 		 //if we haven't set a bind_zone_id lets set it to the starting zone
> 		 if (in_pp->bind_zone_id ==0) in_pp->bind_zone_id = in_pp->zone_id; 
> 		 in_pp->bind_x = atof(row[5]);
> 		 in_pp->bind_y = atof(row[6]);
> 		 in_pp->bind_z = atof(row[7]);
> 		 //if bind x,y,z are all 0 then we need as a default to use the safe point
> 		 //for the bound zone
> 		 if (in_pp->bind_x ==0 &amp;&amp; in_pp->bind_y ==0 &amp;&amp; in_pp->bind_z ==0){
> 			char *query2 = 0;
> 			MYSQL_RES *result2;
> 			MYSQL_ROW row2;
> 			if (RunQuery(query2, MakeAnyLenString(&amp;query2, "SELECT safe_x,safe_y,safe_z FROM zone WHERE zoneidnumber=%i",in_pp->bind_zone_id), errbuf, &amp;result2)) {
> 				if (mysql_num_rows(result2) != 0) {
> 					row2 = mysql_fetch_row(result2); 
> 					in_pp->bind_x = atof(row[0]);
> 					in_pp->bind_y = atof(row[1]);
> 					in_pp->bind_z = atof(row[2]);
> 				}
> 				mysql_free_result(result2);
> 			} else {
> 				safe_delete_array(query2);
> 				LogFile->write(EQEMuLog::Error, "Database: could not find a zone entry in database for the starting bound zone.");	
> 			}
> 
> 		 }
6569c6602,6603
&lt; 			LogFile->write(EQEMuLog::Error, "Database: could not find start_zones entry in database. Using Defaults..");
---
> 		  LogFile->write(EQEMuLog::Error, "Database: could not find a start_zones entry in database for this choice,class,race,diety. Using Defaults..");
> 		  
6574c6608,6609
&lt; 		LogFile->write(EQEMuLog::Error, "Database: could not find start_zones table in database. Using Defaults..");
---
> 		LogFile->write(EQEMuLog::Error, errbuf);
> 		LogFile->write(EQEMuLog::Error, "Database: SQL Error. Using Defaults..");

eq_packet_structs.h Diff

Code:

Index: eq_packet_structs.h
===================================================================
RCS file: /cvsroot/eqemu/eqemu/eqemu/Source/common/eq_packet_structs.h,v
retrieving revision 1.1.1.10
diff -r1.1.1.10 eq_packet_structs.h
151c151,156
&lt; /*1600*/	int8	unknown1600[60];	// ***Placeholder
---
> /*1600*/	int8	haircolor[10];	
> /*1610*/	int8    beardcolor[10];
> /*1620*/	int8	eyecolor2[10];	
> /*1630*/	int8    eyecolor1[10];
> /*1640*/	int8	hair[10];	
> /*1650*/	int8    beard[10];
604,606c609,611
&lt; 	/*0068*/	int32	haircolor; //guess
&lt; 	/*0072*/	int32	eyecolor1; //guess
&lt; 	/*0076*/	int32	eyecolor2; //guess
---
> 	/*0068*/	int32	haircolor;
> 	/*0072*/	int32	beard;
> 	/*0076*/	int32	beardcolor;
638,640c643,645
&lt; /*0128*/	int32	beard;//guess
&lt; /*0132*/	int32	beardcolor;//guess
&lt; /*0136*/	int32	face;
---
> /*0128*/	int32   face;
> /*0132*/	int32	eyecolor1;//its possiable we could have these switched
> /*0136*/	int32	eyecolor2;//since setting one sets the other we really can't check

and finally the client_process.cpp diff

Code:

Index: client_process.cpp
===================================================================
RCS file: /cvsroot/eqemu/eqemu/eqemu/Source/zone/client_process.cpp,v
retrieving revision 1.1.1.12
diff -r1.1.1.12 client_process.cpp
98,99c98
&lt; 	#endif
&lt; 	
---
> 	#endif	
1098c1097
&lt; 				case OP_Death: {
---
> 				case OP_Death: {					
1100,1101c1099
&lt; 						break;
&lt; 					
---
> 						break;					
1585c1583
&lt; 				case OP_ZoneChange: {
---
> 				case OP_ZoneChange: {					
1606a1605,1606
> 					
> #ifdef GUILDWARS
1609,1618c1609,1619
&lt; #ifdef GUILDWARS
&lt; 					if(animation > 65 &amp;&amp; admin&lt;80 &amp;&amp; CheckCheat()){
&lt; 						if(cheater || cheatcount>0){
&lt; 							Message(15,"Cheater log updated...yup your busted,its not nice to cheat.");
&lt; 							char descript[50]={0};
&lt; 							sprintf(descript,"%s: %i","Death zone cheat");
&lt; 							database.logevents(this->AccountName(),this->AccountID(),admin,this->GetName(),"none","Death zone cheat",descript,15);
&lt; 							if(cheater==false){
&lt; 								worldserver.SendEmoteMessage(0,0,0,13,"&lt;Cheater Locator> We have found a cheater.  %s (Acct: %s) was just caught hacking, please show them what we think of hackers...",this->GetName(),this->AccountName());
&lt; 								cheater=true;
---
> 						if(animation > 65 &amp;&amp; admin&lt;80 &amp;&amp; CheckCheat()){
> 							if(cheater || cheatcount>0){
> 								Message(15,"Cheater log updated...yup your busted,its not nice to cheat.");
> 								char descript[50]={0};
> 								sprintf(descript,"%s: %i","Death zone cheat");
> 								database.logevents(this->AccountName(),this->AccountID(),admin,this->GetName(),"none","Death zone cheat",descript,15);
> 								if(cheater==false){
> 									worldserver.SendEmoteMessage(0,0,0,13,"&lt;Cheater Locator> We have found a cheater.  %s (Acct: %s) was just caught hacking, please show them what we think of hackers...",this->GetName(),this->AccountName());
> 									cheater=true;
> 								}
> 								cheatcount=0;
1620c1621,1622
&lt; 							cheatcount=0;
---
> 							else
> 								cheatcount++;
1622,1623d1623
&lt; 						else
&lt; 							cheatcount++;
1625,1626d1624
&lt; #else
&lt; 					break;
1628c1626
&lt; 					}
---
> 
1652,1655c1650,1651
&lt; 		
&lt; 					tarx=zonesummon_x;
&lt; 					tary=zonesummon_y;
&lt; 					tarz=zonesummon_z;
---
> 					//zone debugging
> 					ZonePoint* zp = zone_point;										
1669,1673c1665,1673
&lt; 					else if (zonesummon_x == -3 &amp;&amp; zonesummon_y == -3 &amp;&amp; (zonesummon_z == -3 || zonesummon_z == -30) &amp;&amp; database.GetZoneName(m_pp.bind_zone_id)) {
&lt; 						strcpy(target_zone, database.GetZoneName(m_pp.bind_zone_id));
&lt; 						tarx = m_pp.bind_x;
&lt; 						tary = m_pp.bind_y;
&lt; 						tarz = m_pp.bind_z;
---
> 					else if (zonesummon_x == -3 &amp;&amp; zonesummon_y == -3 &amp;&amp; (zonesummon_z == -3 || zonesummon_z == -30)) {						
> 						if (database.GetZoneName(m_pp.bind_zone_id)){
> 							//zoneing to bind point
> 							strcpy(target_zone, database.GetZoneName(m_pp.bind_zone_id));
> 							tarx = m_pp.bind_x;
> 							tary = m_pp.bind_y;
> 							tarz = m_pp.bind_z;
> 							
> 						} //else bind point isn't set and we will zone to the zone safe point
1693c1693
&lt; 						tarheading = zone_point->target_heading;
---
> 						tarheading = zone_point->target_heading;						
1696a1697
> 						
1699a1701
> 						
1704c1706
&lt; 					else {
---
> 					else {						
1707,1709c1709,1711
&lt; 						tarx=-1;
&lt; 						tary=-1;
&lt; 						tarz=-1;
---
> 						//tarx=-1;
> 						//tary=-1;
> 						//tarz=-1;
1719c1721
&lt; 					if (target_zone[0] != 0 &amp;&amp; admin >= minstatus &amp;&amp; GetLevel() >= minlevel) {
---
> 					if (target_zone[0] != 0 &amp;&amp; admin >= minstatus &amp;&amp; GetLevel() >= minlevel) {						
1737c1739
&lt; 						if (m_pp.zone_id == zone->GetZoneID()) {
---
> 						if (m_pp.zone_id == zone->GetZoneID()) {							
1749a1752,1753
> 							//zoneing to another zone so we need to the let the world server
> 							//handle things with the client for a while
1763,1764c1767,1768
&lt; 					else {
&lt; 						LogFile->write(EQEMuLog::Error, "Zone %i is not available", zc->zoneID);
---
> 					else {						
> 						LogFile->write(EQEMuLog::Error, "Zone %i is not available because target wasn't found or character insufficent level", zc->zoneID);

Feel free to critique...

Ves

Monrezz · #2 02-15-2004, 01:10 AM

:o

Wow, great work! All this after only a week as well

Don't suppoose you can post the whole file? Unles you can tell me how to use the diff :P

Monrezz

Scorpious2k · #3 02-15-2004, 02:25 AM

Good to see another code tinkerer join the ranks

Nice work.

vesuvias · #4 02-15-2004, 03:51 AM

Quote:

Originally Posted by Monrezz

:o

Wow, great work! All this after only a week as well

Don't suppoose you can post the whole file? Unles you can tell me how to use the diff :P

Monrezz

The 3 source files changed can be gotten here.
client_process.cpp goes in the zone directory and the other two go in the common directory.

Hope that helps...

Don't forget about the database changes (the 4 alter table SQL statements)...

How do I submit these changes so they can be added to the next release?? Does posting it here do that for me? Should I PM someone specific?

Ves

Monrezz · #5 02-15-2004, 09:14 AM

Thanks, got it all compiled and it runs smoothly. Didn't crash on death either

As for submitting - when a dev reads this (trump) I'm sure he'll merge after a little testing.

Monrezz

samandhi · #6 02-15-2004, 09:36 AM

Or you can catch them on IRC chat and let them know.

m0oni9 · #7 02-15-2004, 10:11 AM

Not to knock on anyone, but if you're responsible for merging files you really should be familiar with diff/patch. If you're in the Windows world it can be a bit foreign, but there are some implementations out there that you can use. I haven't used it, but WinMerge (http://winmerge.sourceforge.net/) looks like it would work for this purpose.

Keep up the good work vesuvias

farce · #8 02-15-2004, 05:16 PM

only one guy merges the code into eqemu source... Trumpcard.

the rest of us just like to test out other people's tweaks.

Scorpious2k · #9 02-16-2004, 12:28 PM

Quote:

Originally Posted by m0oni9

I haven't used it, but WinMerge (http://winmerge.sourceforge.net/) looks like it would work for this purpose.

I use it all the time to merge the Scorpious2k custom code into new releases. It works well.

The problem you run into with diff is that the source changes very dynamically. Several times a day in some cases at the developer level. So you end up with a diff that doesn't match the source you are working with.

Quote:

Keep up the good work vesuvias

Much agreed. We need a lot more people like vesuvias.

Scorpious2k · #10 02-16-2004, 01:09 PM

I'm going to merge this into our server's code right away. And into cvs as soon after that as I can.

We may have a problem. In 5.5 (eqlive patch compatible) the player profile has changed... it may/may not affect your fixes on player appearence.

vesuvias · #11 02-16-2004, 02:37 PM

Quote:

Originally Posted by Scorpious2k

I'm going to merge this into our server's code right away. And into cvs as soon after that as I can.

We may have a problem. In 5.5 (eqlive patch compatible) the player profile has changed... it may/may not affect your fixes on player appearence.

Great!! I am glad people are finding some of these fixes useful. I fixed actual apperences inside the game today (Now that you can configure a character to look like you want you will notice that no one else in game sees you like that). It was a similar problem with the packet structs not being fully known or wrong in some cases (Probably evolved to thier current state overtime). I will try and get those diffs out within the next day or two (I have to run through and remove all my debuggin code first

).

Currently I am trying to analyze the packet sniffs from both a successful spell cast and bugged one (the spell bug is really pissing me off, so i am going to fix it). If I can get that fixed I will probably turn my attention toward why zoning is so unstable sometimes.

I figured 5.5 might change everything completely and if thats the case it's no big deal. Everything I am doing now I chalk up to learning experiences. I have never had the need to analyze packets before so this project is providing me with some interesting challenges. I have more fun tinkering than playing for some reason (guess what race I favor when I do play online :lol: ).

Ves

Rojan · #12 02-17-2004, 09:23 AM

Quote:

Originally Posted by vesuvias

Quote:

Originally Posted by Scorpious2k

I'm going to merge this into our server's code right away. And into cvs as soon after that as I can.

We may have a problem. In 5.5 (eqlive patch compatible) the player profile has changed... it may/may not affect your fixes on player appearence.

Great!! I am glad people are finding some of these fixes useful. I fixed actual apperences inside the game today (Now that you can configure a character to look like you want you will notice that no one else in game sees you like that). It was a similar problem with the packet structs not being fully known or wrong in some cases (Probably evolved to thier current state overtime). I will try and get those diffs out within the next day or two (I have to run through and remove all my debuggin code first

).

Currently I am trying to analyze the packet sniffs from both a successful spell cast and bugged one (the spell bug is really pissing me off, so i am going to fix it). If I can get that fixed I will probably turn my attention toward why zoning is so unstable sometimes.

I figured 5.5 might change everything completely and if thats the case it's no big deal. Everything I am doing now I chalk up to learning experiences. I have never had the need to analyze packets before so this project is providing me with some interesting challenges. I have more fun tinkering than playing for some reason (guess what race I favor when I do play online :lol: ).

Ves

The zoning seems to be improper zone coordinates followed by incorrect safe points in the db (may not be the db creators fault though see below).

I have been messing around with it and once I set the correct points I can zone back and forth #zone in and out without crashing the client to desktop.

I saw another thread here that shows a fix to the steamfont zone crash, check that one out and I bet it will fix many of the zone crash problems, basically there is an 8bit variable where there should be a 16bit variable so zone points are not getting either sent or received correctly.

Here is the title of that thread -- Got akanon to steamfont zoneline to work (code & db)

Great work btw.

wize_one · #13 02-17-2004, 09:34 AM

Rojan.
in the worldbuilders forum i posted up a fixed zone table. has everything except GoD zones.
here is the link to it
http://www.eqemulator.net/forums/viewtopic.php?t=12672

all #zone are working right. no underworld or bad locs.

Rojan · #14 02-17-2004, 10:05 AM

Quote:

Originally Posted by wize_one

Rojan.
in the worldbuilders forum i posted up a fixed zone table. has everything except GoD zones.
here is the link to it
http://www.eqemulator.net/forums/viewtopic.php?t=12672

all #zone are working right. no underworld or bad locs.

Cool thanks, that fixes the safe points which #zone uses and should stop the client from crashing when it gets a bad or out of range zone in point.

Scorpious2k · #15 02-17-2004, 11:34 AM

Quote:

Originally Posted by Rojan

I saw another thread here that shows a fix to the steamfont zone crash, check that one out and I bet it will fix many of the zone crash problems, basically there is an 8bit variable where there should be a 16bit variable so zone points are not getting either sent or received correctly.

That's already in the distro, should be in cvs... if it isn't it will be in the next release