A few minor bugs

[Manaweaver]

Some bugs....spawns now have a habit of going in the same direction they were going for quite some time after dying. Things that die aren't showing as plus signs but as the spawn(a circle still representing con color.) Hmm, there is still a bug with the scaling of the map, but turning on broken decode fixed that pretty easily. Not sure if there are any others, will edit as they come.

[Mr. Suspicious]

Some bugs....spawns now have a habit of going in the same direction they were going for quite some time after dying. Things that die aren't showing as plus signs but as the spawn(a circle still representing con color.) Hmm, there is still a bug with the scaling of the map, but turning on broken decode fixed that pretty easily. Not sure if there are any others, will edit as they come.

Non of these are bugs. Non of these are reproducable (well, they are, but only if you don't update ShowEQ correctly). Follow the steps for updating ShowEQ found in the INSTALL.newbies file * found in your ShowEQ directory * FIRST get the new libEQ.a (first delete or rename the old one) THEN get the new ShowEQ source and recompile. I made this mistake myself (no-one is infalable) and got the same things you experience now.

[speedphreak]

also had CVS tip segfault 20 mins after I exited EQ - reporting Bad Packets (Bad CRC32) on Kazaa traffic (port 1214) from the box that had EQ running on it.

[Cryonic]

That's because it is trying to decode kazaa traffic as if it were EQ traffic. One way to fix this is to use Session Tracking.

[speedphreak]

thanks for the infos. I've also firewalled off all non-EQ traffic as showEQ is all I use that box for.

[Manaweaver]

Well, I will tell you what I did, and you tell me if I didn't do it properly.
rm /usr/lib/libEQ.a
CD /usr/lib/
wget (Don't remember the FTP, but I downloaded the new libEQ.a)
md5sum libEQ.a readout was fceeee(the current libEQ.a) Cd /seq/showeq
cvs -z3 update
(I proceeded to export the qtdir and gcc's here)
make distclean
make -f Makefile.dist
make

so I think...THINK that I did it properly, but I could've been sloppy as I'm really rusty at doing this stuff. If those seem like they should be okay, I may think about doing a fresh insall of showeq...anyone else notice these bugs to help confirm? Mainly its the corpse one that I wanna resolve, considering that one can really throw ya off while trying to figure out how many mobs are around. It displays correctly if the mobs died before zoning in, but if they die while I'm in the zone, doesn't show...now I"m just rambling though. going to check the opcodes and make sure that there are no differences.

[Mr. Suspicious]

(I proceeded to export the qtdir and gcc's here)
make distclean
make -f Makefile.dist
make

you skipped 'make install'

Just follow the INSTALL.newbies, Chapter 3: Updating



3. Updating

3.1 Getting a new libEQ.a
Please consult the file README.libEQ about this issue

[]# pico /seq/showeq/README.libEQ



3.2 Updating ShowEQ from cvs
There's an update out... but you don't want
to download the whole thing again? Try this. If it gets a couple
changes, read the CHANGES file and see whats up. Then compile it
if you want.

Code:

[]# cd /seq/showeq
[]# make distclean
[]# cvs -d:pserver:[email protected]:/cvsroot/seq login
[]# cvs -z3 update
[]# cvs -d:pserver:[email protected]:/cvsroot/seq logout
[]# export CC=gcc3
[]# export CXX=g++3
[]# export QTDIR=/usr/lib/qt-2.3.2
[]# make -f Makefile.dist
[]# ./configure
[]# make
[]# make install

[z26o]

Looks like you skipped ./configure also.

Like "Mr. S" says, follow the guide (he was kind enough to post the relevent section) and you can't go wrong.

-z-

[Manaweaver]

ack, didn't miss either of those steps... usually I do most of the as a chain so I think I just forgot to mention them. Not to mention I just woke up, so my brain is only half awake =). Ever since the PoP release I've recompiled and updated 4 or 5 times so I've gotten it down pretty well. I just went through the list of commands again and I did them all in that exact order, so I guess I could just try recompiling. Maybe I'll delete the directories...either way I'm using it on the other comp right now so I don't really have time to test it. Will edit later after I've found the time to do a recompile. =/

EDIT: Taking the time right now to redo those steps just to make sure I'm good to go.

EDIT: Followed the eact steps in your excerpt from the file...still no luck corpses aren't actually showing up until I zone out and back in. They are staying as spawns still. Spawns are still going in the direction they were after they die, then stopping shortly after and reapearing where they went down. Really makes me wonder what I could've possibly have done wrong. Going to fiddle aroudn with the features and see what I can come up with.

EDIT: Yes yes, i know...crapload of edits....okay, removed all directories and libEQ.a. Downloaded a fresh libEQ.a. Downloaded a fresh install of ShowEQ....recompiling as I type. Will edit with final outcome, hopefully this will fix it.

EDIT: Well I'm not sure what the problem was, but corpses now are actually showing up as the Cyan Crosses now, thanks for the tips and the help =)

[Mr Guy]

I do have occasional "sliding" corpses.

[Mr. Suspicious]

I do have occasional "sliding" corpses.

I noticed ingame, some corpses when they drop on a hill or a ramp, do actually "slide" down =)

[sam]

I'm almost positive there is a bug in SEQ with corpse sliding. It seems to happen when you are not near the mob when it dies... like if my pet kills the mob around the corner while the mob is trying to run away in the last 5% of its life... the mob dies... showeq displays a plus sign because it evidently recieved the packet saying the mob died... the problem is, it hasn't recieved the packet yet stating where the mob(corpse) now is.

So I assume, the servers send the death packet imediately no matter how far away you are from the mob when it dies. The server does NOT send a new position packet once a mob dies, so if you have the animate spawns turned on, then you'll see the corpse sliding sometimes until the server reaches its 6 second timeout and sends a new location packet for that mob.

So the worst case would be a mob dying imediately after you recieved the position packet for it... which, if you are too far away, you will not revieve the next position packet for up to 6 seconds.

Now for the fix:
Just stop the spawn's animation wherever it is once you get the death packet. Example: you get a packet from the server saying the mob is at location 1,1,1 and is moving in a northeast direction (45 degree angle maybe). Then you get another packet 2 seconds later saying the mob has died and may or may not have also recieved a new location for the mob(corpse). Whether you recieve the new loc or not, grab the latest estimated cords from the animate spawn process and place the mob there with a null movement. Or if animate spawns is off, just place the mob at the last known location packet (1,1,1). Both of these ways will not place the corpse at the exact location, but they are so close it doesn't matter, and within the next 6 seconds, the server will have sent a new location packet to correct any mistakes in location we may have.

If one of the devs wants to tell me how wrong I am and that I dont know anything, feel free... I'm going only on what I know and understand of SEQ.

[Manaweaver]

Heh heh, yeah, I am still having corpse sliding issues, but at least they are plus signs now =). It doesn't happen often and it looks like it is only from a certain distance. I tested it out some nuking mobs from different distances and the ones that were farthest away had the bug more often. I'm killign stuff right now and haven't seen the bug once...usually I keep a pretty close proximity. Distance seems to be the key here.

EDIT: Perhaps adding something in the code about stopping movement when a mob dies rather then letting the server tell it that it isn't moving.