[Starman01]

Hello,

even with the newest CVS-Builds from Phreak the critical Fred-error isn't resolved yet. It's continue to crash under certain conditions and damages the missionfiles while doing it, I lost a lot of work due to this   Unfortunately it is very hard to reproduce it intentionally , and I can't use any Debugbuilds for further investigation because I have no mouse-pointer, and then Fred is very hard to work with. I have no Idea if it is a code-issue or if it has something to do with our data (which certainly has some minor "this is a bad way of doing it"- things), but everything worked fine before you people changed the compatiblity at the beginning of this year. Currently we are stuck to builds from around 18.12., which have some nasty bugs to my sadness. Maybe I missed some vital change, is there some place where you sum up such vital changes ? (If not, it would be a good idea having one).

What I can tell you about my problem is that:

- It seems to happen when I place objects in large distances (of course that's necessary for a autopilot mission). I think that as the major problem because I created a good mission with the Fred from 24th february, and this one doesn't have autopilot and plays at the same area. Then I started another mission with autopilot again and the crashing began.

- After the error (however it is been created) is in the mission, I can't place anymore objects into the mission.

- At the crash, there is no strange error message appearing, neither a SCP-Code error nor something in the debug spew (it looks like it doesn't even get a chance of writing it down, the crash is immediately. Fred just crashes with the Windows error that want's to send a report (if you know what I mean).

- Never mind the white textbox for the ships-description I mentioned earlier, that seems to be a side-effect, but obviously it's just a minor memory problem, it happens to me in my old fred build too.

- After many many tries I figured out something strange. Most the times the mission is damaged that way, that I can only place one additional ship. If I place a second one, Fred crashes. However, when I save the mission after placing a single ship, I can place another one (but not always)

- In one of my many many tries I get this error message (but that's a very seldom one) :

Code: [Select]

---------------------------
Assertion Failed!
---------------------------
Assert: Player_start_shipnum >= 0 File: c:\fs2_open\code\fred2\missionsave.cpp Line: 1334
[This filename points to the location of a file on the computer that built this executable] 
 Call stack: ------------------------------------------------------------------     fred2_open_d-20060306.exe 00455426()   
  fred2_open_d-20060306.exe 0042909f()     fred2_open_d-20060306.exe 004477e6()     fred2_open_d-20060306.exe 00433e39() 
    fred2_open_d-20060306.exe 008b3fbb()     fred2_open_d-20060306.exe 008b33ee()     fred2_open_d-20060306.exe 008b0b90() 
    fred2_open_d-20060306.exe 008b106e()     USER32.dll 77d18734()     USER32.dll 77d18816()     USER32.dll 77d1c63f()     USER32.dll 77d1c665()   
 OPENGL32.dll 5f0f699b()     USER32.dll 77d18734()     USER32.dll 77d18816() ------------------------------------------------------------------
  [ This info is in the clipboard so you can paste it somewhere now ]   Use Ok to break into Debugger, Cancel to exit.
---------------------------

Usually the player is Alpha 2, and I have sometimes problems getting the player the leader in the autopilot flight, but that's another story. I can't image that being Alpha 2 is a problem, or is it ?

- Here is a mission file with the error in it, you can try for yourself. Place two ships (no matter what class) , and it will crash at the second one. If the error is within in the WCS-Data, some of you coders already have it, and who ever from the SCP wants to look for the problem can have the beta too.

http://www.scifi-3d.de/wcsaga/Starman/download/failed.fs2

- I also figured out, that sometimes (but not always) the error can be removed if I delete everything in the mission area except the player, so it must be related to objects or SEXP.

I know bugsearching is annoying and boring, but please help me so I can continue my work. Currently I can't , and that is a real problem for me.

Thanks in advance.

Starman

(edit) Two more things. I deleted all Events and Messages, Error is still there. So it has to be a problem with the objects.

Also when I launched the mission the very first time, I got an error regarding the "ai-chaise-any-except" that has been removed from the newest code. But I changed the AI-Goal in that SEXP, so it shouldn't be a problem anymore.


BTW, anyone know a fix how I can get my mousepointer back in the debug-builds ?

[karajorma]

Does the debug build report errors when you try to run it? Or does it startup normally but with no mousepointer?

[Starman01]

I have a few errors, but nothing out of the ordinary ( I mean they have been in there for a long time now).

It's complaining about a few things in def_species.tbl, two weapon-explosions doesn't have an LOD0 anim, that for the weapon "empty" (whatever it is) 3 of 4 Lods are missing and a few times that it couldn't load the hud animation for the file "fadeicont-node"

Those are the startup error messages. The debug spew itself has something interesting in it, it seems to have a problem with some of our capship-subsystem, though everything shows up fine in fred and is working ingame. For example :

Warning: Ignoring unrecognized subsystem comm, believed to be in ship KIS_Kamrani-class.pof
Warning: Ignoring unrecognized subsystem nav, believed to be in ship KIS_Kamrani-class.pof

Those are subsystems I use on every capships, they are modelled destroyable systems. The pof contains a submodel called "comm" for example. That one is then labeled as "$special=subsystem" and "$name=communications". A normal "communications" (in the subsystem editor) isn't included then. The ships.tbl refers to the "comm", and ingame it shows up as "communication" like in retail. Never had any problems with that, so I think it shouldn't be that bad.

Some of our models still have that problem: "Serious problem loading model XXX, 36 normals have been capped to zero". I have no Idea where this error come from, since there are only a few ships that have it, and those seems to be fine modelled. The models are working perfectly ingame, and only Fred-HTL complains about that. Older version that I don't run in Fredhtl don't care and mention that thing.

But all these things are in there from the start. So either you guys changed now something that does care about that, or the problem is somewhere else.

What about you Kara, you have our beta. Is the error appearing for you too ? Should be in every Fredbuild past 20060224.

[karajorma]

I got quite a few errors running the debug build but I was hoping you'd have fixed them. See when the debug build is giving you an error it really means that it should be fixed as soon as possible. When you run a retail build you won't get those errors which means that instead of them being trapped and warned about instead they simply wander off into other parts of the code.

Just running the retail build because debug complains is sort of like removing the bulb from the engine hot light on your car's dashboard and thinking that you've fixed the problem cause the light doesn't come on any more.

Bottom line is basically that you need to fix your tables and ships so that you don't get any assertions or warnings first. Otherwise it's really hard to know what the cause of your problems could be.

[Starman01]

I'll try, but that's certainly not the problem ! We have dozend of peoples playing the game, and we have no errors and no random chrashes, on mutliple machines with multiple configurations. IMO those are just nitpicking errors with no bigger effect. What about the mission, can you reproduce the error on your side ?

[Starman01]

Two errors remains, that I can't identify because I don't know about what tables the code *****es. Could please someone with code-experience tell me the error-location ?

1) There is no weapon called "empty" in the weapons.tbl, and no ship with a "empty" turret-entry.

Code: [Select]


Warning: For 'empty', 3 of 4 LODs are missing!
File:c:\fs2_open\code\weapon\weapons.cpp
Line: 1200
[This filename points to the location of a file on the computer that built this executable]

Call stack:
------------------------------------------------------------------
    fred2_open_d-20060306.exe 0057279e()
    fred2_open_d-20060306.exe 005743a1()
    fred2_open_d-20060306.exe 005751f5()
    fred2_open_d-20060306.exe 004474c7()
    fred2_open_d-20060306.exe 0043fc67()
    fred2_open_d-20060306.exe 008b3b57()
    fred2_open_d-20060306.exe 008b33ee()
    fred2_open_d-20060306.exe 008b0b90()
    fred2_open_d-20060306.exe 008b106e()
    USER32.dll 77d18734()
    USER32.dll 77d18816()
    USER32.dll 77d1b4c0()
    USER32.dll 77d1fd29()
    ntdll.dll 7c91eae3()
    USER32.dll 77d201f7()
------------------------------------------------------------------

2) This one comes up 3 times. I can't find the Icon, nor a table-entry referring to it. We have replaced the jumpnodes, maybe that's related ?

Code: [Select]


---------------------------
Warning!
---------------------------
Warning: Couldn't load hud animation for file 'Fadeicont-node'

File:c:\fs2_open\code\hud\hud.cpp

Line: 2364

[This filename points to the location of a file on the computer that built this executable]



Call stack:

------------------------------------------------------------------

    fred2_open_d-20060306.exe 0044724e()

    fred2_open_d-20060306.exe 004474fe()

    fred2_open_d-20060306.exe 0043fc67()

    fred2_open_d-20060306.exe 008b3b57()

    fred2_open_d-20060306.exe 008b33ee()

    fred2_open_d-20060306.exe 008b0b90()

    fred2_open_d-20060306.exe 008b106e()

    USER32.dll 77d18734()

    USER32.dll 77d18816()

    USER32.dll 77d1b4c0()

    USER32.dll 77d1fd29()

    ntdll.dll 7c91eae3()

    USER32.dll 77d201f7()

    USER32.dll 77d20291()

    fred2_open_d-20060306.exe 008b191a()

------------------------------------------------------------------


[taylor]

1) There is no weapon called "empty" in the weapons.tbl, and no ship with a "empty" turret-entry.

This one is a weapon_expl.tbl thing.  Unless you tell it that empty has only 3 LODs it will try to load 4.  Just add:

Code: [Select]

$Name:  empty
$LOD:       3to your weapon_expl.tbl and you should be good to go.

[Starman01]

Hm, did that and now it complains that 2 of 3 are missing. I now entered

Code: [Select]

$Name:  empty
$LOD:        1

and the error is gone. Though I have to admit, that I have no idea what I just did I don't know of any weapons, explosion or other stuff in our data called "empty".

[taylor]

Now I feel stupid.  It was complaining that 3 of the 4 LODs were missing, which means it can only find 1 file called empty.  I should have paid more attention. 

Though, if you don't have a file called "empty" in the data anywhere then I'm not quite sure why it's finding even that one. 

[Starman01]

Now I found it, there is a file called empty.ani in the interface folder, and guess what, it's empty No Idea what is was originally meant for, but there has certainly be a reason. Though the minor problem is now fixed, the big one still remains.  And like I said, it is very unlikely that these minor issues are leading to the crash with the new fredbuilds, when the old ones never complained about anything

[karajorma]

Yeah but now that you've fixed the tables you can send them to me and I can run the debug build through VC6 and find out exactly what is causing the bug

[Starman01]

I will do that asap . However, the "fadeicont-node" is still unsolved. I have to wait for Tolwyn, maybe he knows what this is used for.

[Trivial Psychic]

By the name of it, I'd expect it to be part of the Icons for briefings, but the error message suggests that its something referenced by a HUD entry.  I suggest you go through every one of your ships (even the non-player ones) to see what hudicons are listed.

[Starman01]

Yesterday night (while lying in Bed) I had a shining It can only be a entry in icon.tbl missing it's icon.ani  and so it was
Now all errors from the freddebug-build are fixed. However, I can't run the game with a debugbuild, it's just not starting (though no
error or such). I can't remember however when I was able to run the game last with a debug build, guess those are more problems related
by my machine.

[Starman01]

@Karajorma : Could you please PM your Email-Adress ? I have the tables ready, but I can't upload them because our FTP is on strike

[Backslash]

After some crazy tests regarding the first issue, I am starting to be able to reproduce this error.  But the really weird thing is, it seems to involve waypoints.

I downloaded the mission in this thread, replaced all the ships with FS2 equivalents, and then tried it.  It crashes upon adding two more ships.  This confirms it is not a problem in WCS.

Then I removed all the waypoints, saved it as a new mission, restarted Fred, opened that new mission, and placed a whole bunch (30+) of new ships.  No problems!

So, we've got it narrowed down some at least...

[taylor]

After some crazy tests regarding the first issue, I am starting to be able to reproduce this error.  But the really weird thing is, it seems to involve waypoints.

http://lore.maxgaming.net/~scp/mantis/bug_view_advanced_page.php?bug_id=0000628

[Starman01]

So, the child has finally a name I'm so glad it's not a problem with our data. But if this workaround is safe and will not harm the mission files, I can at least switch to a newer build. However it will not be that easy always placing waypoints the last. Sometimes I need to add new ships for better balancing (which means, pls fix asap )Is the "place one ship, save mission" reported to be save too ?

Thanks to everyone helping me search the error

Starman

[Trivial Psychic]

Yesterday night (while lying in Bed) I had a shining It can only be a entry in icon.tbl missing it's icon.ani  and so it was
Now all errors from the freddebug-build are fixed.

Be advised, Debug FRED won't warn you about ALL the errors when loading up the program.  I'd suggest that you also try placing (one at a time) every ship on the grid in FRED debug.  If any errors are skipped by the openning debug check, it'll find them when you place them.

[Tolwyn]

1) There is no weapon called "empty" in the weapons.tbl, and no ship with a "empty" turret-entry.

I am using empty.ani for following entries

Code: [Select]

$Anim: empty
and in some cases

Code: [Select]

$Icon:                        empty
or

Code: [Select]

+Tech Anim:                     empty
or

Code: [Select]

$Impact Explosion:            empty
Mostly it is used as a placeholder.