SEQ Borked with 8/15 patch.

[fransick]

Looking at it real quick it looks like playerProfileStruct may have been shortened just a little. BUT playerProfileStruct is part of charProfileStruct and unless I am mistaken, my eyes are telling me that ALL charProfileStruct should have is the playerProfileStruct in it now (nothing from 19960 onwards). This is from comparing the same character from a working zone.log and a non-working zone.log. I have no idea where they are putting the rest of the charProfileStruct or if they simply are no longer passing it to the client. I don't think I could test that adequately though without first figuring out what small changes they made to playerProfileStruct.

The good news is that it doesn't look like many opcodes changed.

A quick check on a few known values and they are definitely passing some data from charProfileStruct. I checked on radiant and ebon crystals for example as well as ldon points and those are still coming across so, sadly, it isn't as easy as eliminating charProfileStruct data I don't think.

[rogues]

Ok, I'll take another look. I'm only trying to compare what I see in the log files, I think you have a better way of comparing than I do, right?

Ok, I think I see where it is, IS data in playerProfileStruct. the potion belt had been starting at 19642 and now it's starting at 13699.

Why were they allowing for such a huge difference in the BandolierStruct (offset 13108) to 19508 for the InLineItem? MAX_BANDOLIERS is only 20, so how do you get from 13108 to 19508?

[fransick]

Ok, I'll take another look. I'm only trying to compare what I see in the log files, I think you have a better way of comparing than I do, right?

Not really... I have a few reference logs from past working patches that I compare to and I usually look for know values in the new log to kind of see where things are. As a non-programmer I was pretty proud of myself when I figured out how the hex values were being expressed in the packets. For instance, when I looked up Radiant Crystals earlier, I converted 1464 (number of Radiant Crystals toon has) to hex (5b8) then searched the log for "b8 05" and looked at the data around it and saw that the value for Ebon Crystals was also nearby so deduced I was in the right place and they are still sending that data over.

No voodoo or skill here, just old fashioned scut work since I have little to no code experience, haha.

[rogues]

Ok, so we're doing the same thing. Where I had found AAs to be at 13200 I'm finding them at 12917 now. That's right before the bandolier structures. But where before the potion data was at 19642, it is now at 13699. So the removed data is somewhere in there (or at least a good chunk of it).

[fransick]

Ok, I'll take another look. I'm only trying to compare what I see in the log files, I think you have a better way of comparing than I do, right?

Ok, I think I see where it is, IS data in playerProfileStruct. the potion belt had been starting at 19642 and now it's starting at 13699.

Why were they allowing for such a huge difference in the BandolierStruct (offset 13108) to 19508 for the InLineItem? MAX_BANDOLIERS is only 20, so how do you get from 13108 to 19508?

Bandolier and potion data changed a lot and look to be variable now. In the old struct, bandolier data was fixed length and lots of room was left for each piecce of data. If you didn't use all the bytes reserved for, say, the name of bandolier, it would just fill in the remaining space with zeros. Now it is just sending over as much as it needs and no "fluff" or wasted bytes. It looks like this:

Old Log:

13552 | 31 48 53 00 00 00 00 00 00 00 00 00 00 00 00 00  | 1HS............. 
13568 | 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00  | ................ 
13584 | 16 da 01 00 6f 0d 00 00 52 69 74 75 61 6c 20 49  | ....o...Ritual I 
13600 | 6d 70 61 6c 65 72 00 00 00 00 00 00 00 00 00 00  | mpaler.......... 
13616 | 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00  | ................ 
13632 | 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00  | ................ 
13648 | 00 00 00 00 00 00 00 00 cb d9 01 00 05 0e 00 00  | ................ 
13664 | 4b 6e 69 67 68 74 20 43 61 70 74 61 69 6e 27 73  | Knight Captain's 
13680 | 20 53 68 69 65 6c 64 00 00 00 00 00 00 00 00 00  |  Shield......... 
13696 | 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00  | ................ 
13712 | 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00  | ................ 
13728 | cf da 01 00 ba 0d 00 00 53 68 61 72 72 61 27 73  | ........Sharra's 
13744 | 20 52 69 74 75 61 6c 69 73 74 69 63 20 44 6f 6c  |  Ritualistic Dol 
13760 | 6c 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00  | l............... 
13776 | 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00  | ................ 
13792 | 00 00 00 00 00 00 00 00 f2 8b 01 00 2d 09 00 00  | ............-... 
13808 | 48 75 6d 61 6e 6f 69 64 20 52 65 64 75 63 74 69  | Humanoid Reducti 
13824 | 6f 6e 69 7a 65 72 00 00 00 00 00 00 00 00 00 00  | onizer.......... 
13840 | 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00  | ................ 
13856 | 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00  | ................

With today's patch, that same data now looks like:

13088 | 31 48 53 00 52 69 74 75 61 6c 20 49 6d 70 61 6c  | 1HS.Ritual Impal 
13104 | 65 72 00 16 da 01 00 6f 0d 00 00 4b 6e 69 67 68  | er.....o...Knigh 
13120 | 74 20 43 61 70 74 61 69 6e 27 73 20 53 68 69 65  | t Captain's Shie 
13136 | 6c 64 00 cb d9 01 00 05 0e 00 00 53 68 61 72 72  | ld.........Sharr 
13152 | 61 27 73 20 52 69 74 75 61 6c 69 73 74 69 63 20  | a's Ritualistic  
13168 | 44 6f 6c 6c 00 cf da 01 00 ba 0d 00 00 48 75 6d  | Doll.........Hum 
13184 | 61 6e 6f 69 64 20 52 65 64 75 63 74 69 6f 6e 69  | anoid Reductioni 
13200 | 7a 65 72 00 f2 8b 01 00 2d 09 00 00 44 41 00 44  | zer.....-...DA.D

It looks like it's sending over only what is needed and is variable? Same thing with the potion belt. I am guessing similar changes will be found elsewhere, these were just the first two I stumbled across when looking at the packets.

[rogues]

So basically we're at the point of needing to do what purple said to do, which is a little beyond me at this point. Is NetClient like wireshark?

[Razzle]

Those structs are very simple. It is a null (00) term string of variable length, and then another. The two 32bit unsigned ints. That looks to be one bandolier slot. Previously those were fixed length character arrays. Now they are only as long as needed.

Look for patterns like a string with (00) at end. Spaces in string show as (20).

[rogues]

Ok, so where BandolierStruct was of length 320 before and there were allowances made for 20 (ie MAX_BANDOLIERS), that accounts for the 6400, which is how there is such a big jump from 13108 to 19508. You are saying the type changes from BandolierStruct to.... something that can be variable in length, which is where my c fails me because it's been 20 years.

[Razzle]

There are lots of places they could save space and thus save bandwidth. Saving bandwidth equates to saving money.

[fransick]

Ok, so where BandolierStruct was of length 320 before and there were allowances made for 20 (ie MAX_BANDOLIERS), that accounts for the 6400, which is how there is such a big jump from 13108 to 19508. You are saying the type changes from BandolierStruct to.... something that can be variable in length, which is where my c fails me because it's been 20 years.

I think that is exactly what Razzle is saying. Whereby the old bandolier struct was 320 the one I posted above from today's patch is only approximately 115 bytes. One of my toons that has several bandoliers setup has about 1319 bytes between the start of his bandolier data and potion data. Another toon that doesn't have any bandoliers setup has only 744 bytes before getting to potion belt data. When I set up a single bandolier the 744 jumped to 800 bytes. The only consistent thing I could see was that on both characters the bandolier data started at byte 12955.

Not sure if that's help or not Rogues

[Razzle]

What is opcode for player profile? Not gonna have time to go through packets. I am decompiling client so i can check for bit fields that turn on/off char arrays. If i get a few hour i should be able to track down big gotchas.

[fransick]

what is opcode for player profile? Not gonna have time to go through packets. I am decompiling client so i can check for bit fields that turn on/off char arrays. If i get a few hour i should be able to track down big gotchas.

0x078e

[rogues]

@fransick - right, I was really just answering my earlier question about how the current everquest.h could have so much space for the bandolier, it was just because I didn't understand how much they had been making allowance for. Now I do, but unfortunately what I don't understand is how you build the structure out to handle variable lengths, and how to get it to understand where the bandolier structure has ended and the potion belt starts if it can very depending on what the toon has.

[fransick]

@fransick - right, I was really just answering my earlier question about how the current everquest.h could have so much space for the bandolier, it was just because I didn't understand how much they had been making allowance for. Now I do, but unfortunately what I don't understand is how you build the structure out to handle variable lengths, and how to get it to understand where the bandolier structure has ended and the potion belt starts if it can very depending on what the toon has.

Looks like we'll both be learning as we go from here on out! Good times...

[rogues]

Indeed!

spawnStruct is variable in length so I tried to learn from there, but it seems to be variable based on the union of different types of options (if NPC, if PC, etc) which I don't think will work in this case. I'll poke around at it more tonight.