[Kazan]

I located a bug in the _vm_* functions in Windebug.cpp related to the code enabled by defining NEW_MALLOC.

I think i have it all cleaned up and using malloc,realloc,free again - but for the moment i think memory usage tracking is broke in relation to realloc.. i'll fix it tonight if it is.

[Gregster2k]

Yanno, I don't know why, but I get intermittent NTDLL.DLL error messages in FSO these days. -_-;; --- i be darned if i figure out what is causing it.

I use OpenGL on a Radeon 9800 Pro. My PC is using the 2.6.42 version of Radeon Omega Drivers, has display driver version...

Display Driver version "8.14-050512m-023864C-ATI-OMEGA".

When I installed those on my laptop its Mobility X600 started "snow-freezing" in GTA: San Andreas periodically. Returning to the original ASUS-supplied ATI spec drivers for my ASUS V6V laptop fixed it.

If I ever choose to run FSO on the laptop in future, and I probably will, I'm going to have a problem reporting bugs to you guys ... the ASUS drivers provided aren't Catalysts, but they have the Control Panel, which reports this version:

Display Driver version "8.11-050212a-021287C".

I have no idea which Catalyst driver corresponds to that...anyone know? If it's too old a Catalyst i may have a problem -_-;;

[karajorma]

Originally posted by Gregster2k
Yanno, I don't know why, but I get intermittent NTDLL.DLL error messages in FSO these days. -_-;; --- i be darned if i figure out what is causing it.

mv_music.vp

Remove it and all will be fine.

(If you're using mv_zPack from 3.6.5 then you should simply extract the music.tbl from this link and stick it in your data\tables folder or upgrade to the 3.6.6 VP files and delete the old version of zPack).

[Mav]

Ok, first as I stated in the other topics, this build worked very well for (though still having the white squares problem and briefing icons/debriefing text misalignment) before  I tried to get some old mods of mine to work.
Currently the trouble isn't that serious (the rest of the screen getting black if some dialogue box pops up, showing again as soon as the box is closed; and yes, even if I don't use said mods...), but I'm starting to fear for my windows install, as the trouble with my old PC started in this manner...

Now, what I actually wanted to ask here is, when I tried to debug one of these mods with Fred-debug, I got several errors that I could solve, but also one that I have no idea what it refers to (it's only there with that specific mod). The error message I get is:

"Assert: count < max_ints
File: C:\Languages\Visual Studio Projects\Visual C++\fs2_open-CVS\code\Parse\PARSELO.CPP
Line: 1914
[This filename points to the location of a file on the computer that built this executable]

Call stack:
------------------------------------------------------------------
    fred2_open_G-20050715-dbg.exe 004fef34()
    fred2_open_G-20050715-dbg.exe 00501d62()
    fred2_open_G-20050715-dbg.exe 005020ae()
    fred2_open_G-20050715-dbg.exe 0044f092()
    fred2_open_G-20050715-dbg.exe 00447581()
    fred2_open_G-20050715-dbg.exe 00905fa2()
    fred2_open_G-20050715-dbg.exe 00905904()
    fred2_open_G-20050715-dbg.exe 009034a9()
    fred2_open_G-20050715-dbg.exe 00903945()
    USER32.dll 77d18709()
    USER32.dll 77d1d297()
    USER32.dll 77d1b368()
    USER32.dll 77d1e840()
    ntdll.dll 7c91eae3()
    USER32.dll 77d218a4()
------------------------------------------------------------------"

Now, why the hell is it complaining about a value being below a max value?  Could it actually mean the opposite? And really "count" is a damn explanative name for a variable - WHAT actually does it count ???
Please nobody get insulted, I'm just a bit stretched on non-explanative error messages.  And the other messages did a good job in helping me with debugging.



Oh, and while I'm at it - another mod seemed to be more or less working, but there were several models where fred complaind about their geometry and that it had to cap several normals or so (but it still did load those models). Any idea what might be the cause for this?



(I thought this might be the right thread for this, but on the other hand, I possibly should have posted it in the modding thread? If this is an issue, some admin move this post over there, please. Or just someone point me to post it there anew.)

[taylor]

Originally posted by Mav
Now, why the hell is it complaining about a value being below a max value?  Could it actually mean the opposite? And really "count" is a damn explanative name for a variable - WHAT actually does it count ???

What mission or mod are you trying to use?

Basically this is just saying that it's trying to load in a series of numbers, up to a particular amount ("max_ints"), and for each one that it finds it increments "count".  So this error is that, in a list of numbers, there are more numbers than it's expected to find.  This can cause a memory overwrite problem if it's only got "max_ints" worth of storage space and it attempts to load more than "max_ints" into that memory location.  Basically, it will either crash or have bugs if the mission or table in question continues to have this error in it.

[Mav]

Hmmm... I didn't load any mission, I just started fred-dbg and got that message. The mod I'm using is my personal shuffling-together of things, it's quite possible that there are errors in it. It just would be nice to know where I have to search for them.  One thing I though in-between is that I might have too many weapons defined in my weapons.tbl - could it be this?

[taylor]

Originally posted by Mav
It just would be nice to know where I have to search for them.  One thing I though in-between is that I might have too many weapons defined in my weapons.tbl - could it be this?

It wouldn't be too many weapons, just too many things in a list.  So, say that the problem was in ships.tbl, with "$Default SBanks:".  The maximum number of secondary banks is 4, it's hard coded in the source.  So if you specified more than 4 items in that list then it would give you the error you received.  A list is anything inside of parenthesis (unless it has XSTR in front of it) so you can just search for those in your tables and that will help figure out where the problem is.

And since you are running a debug build it should produce a fs.log which would have parsing errors in it.  You may have to create an empty txt file named "debug_filter.cfg" in your freespace2 directory to get all of the parsing errors to be reported but that will help you figure out where the problem is hiding.

[Mav]

Thanks, I'll see if I can find anything (including the fs.log you mentioned).
I know about secondary/primary bank limits more or less though, so I'm rather sure this isn't the point. On the other hand, quite some of the bugs FRED found for me are rather stupid - I really wondered about me making these...

[Mav]

A new thing. It possibly is not that specific to this build, but since I got it with this one, I'll post it here...


What I'm talking about is capship shieldmeshs like on the SSD Diablo that can be downloaded at F2S.
I tested them pretty often when I was still using SCP v.3.5.5 and below, and there they did work almost flawless . (no yelling intended )

I always wondered people stating over and over that they didn't work. Now I tried again with this build (which, as I already stated, works fine for me elsewise), and guess what - FS crashed upon some (beam-)bombers firing at the Diablo!
It showed exactly the same symptoms as a faulty shieldmesh (too far away from the ships' hull) would have in 3.5.5 or earlier and retail (trust me, I made enough of them ).
Only point is that the Diablo's shieldmesh wasn't faulty...


So it seems, somewhere in-between it was broken...
I jumped from 3.5.5 over a short stop in 3.6.0 directly to this build, so I can't say when it actually started.

I'd just like to have them working again...

Does anyone know what's wrong and/or with what version it started?
And is anything planned for solving it? (please?)

[Trivial Psychic]

Have you tried a debug build?

[Mav]

I tried a few days ago (after seeing your reply...) . The debug build gives me the same count-error as fred-debug, and thus refuses to start. So I'd first need to solve that error - problem is, I don't have any idea where to search for it...
I don't have any ships with more than three primaries and four secondaries. Could too many ship flags or too many allowed weapons for a ship-type cause it?

Well anyway, no insult intended (the SCP-team is really doing a great job ), but when I get an error-message, it would be quite helpful if it points me to the point that caused the error (even if it only says "error is in line xyz of file ###.tbl" ) . Granted, most error messages already do so (and that was really helpful to me), but it would be nice if this could be extended to all possibly mod-specific error-messages, like this "count"-thing (maybe it's already being worked on and I just don't know - I apologize if this should be the case).

Other idea... could all those limits be comprehensively listed somewhere?

[WMCoolmon]

I've tried to give the relevant limit in the error messages I've made.

However, there's a problem with making a lot of 'em. Right now there are only Warnings (which only show up in debug builds, but don't kill Freespace 2) and Errors (which DO show up in all builds, but kill the app)

There's no room for a middle-ground problem AFAIK, which is a problem because either won't show up to people and they'll forget about them or they will stop execution of the program (even if they aren't critical).

EG, right now, if you have two ships with the same name the second will be excluded. It's not a problem that'll kill missions, unless the name was changed from one used in another mission accidentally, but it should be something the person running the app should know about.

[StratComm]

I was particularly thrilled to see the error message for "web cursor not found" with an explanation that the executable was in the wrong directory (ok, so I double-clicked an executable in WinRAR) since that has come up more than once.  There are still a few loading/parding errors that I wish people saw and could fix, but the situation is pretty well-handled ATM I think.

[Goober5000]

Originally posted by WMCoolmon
and Errors (which DO show up in all builds, but kill the app)

It doesn't, actually.  If you click OK the game will keep on running, as long as the code recovers gracefully from the Error.

We might have to make this more explicit in the Error message.[/B][/quote]

[ImmortalZ]

The thrusters on the cruiser is locked to on. Is this a known bug?

I'm running the build in this thread with the following vps:

(all downloaded today off SCP)
core
effects
textures
models

My commandline params:

-spec -glow -pcx32 -jpgtga -2d_poof -rlm -ballistic_gauge -snd_preload -env  -ambient_factor 75

Any ideas?

I also notice the white box problem. But I saw another (older) build which had fixed this problem and the coder said that the fix was commited to CVS. If so, why is it still here?

P.S: My vid card is an X800XT running Catalyst 5.5.

[Cobra]

i think it's supposed to be that way to make a little realism. no flare would look unrealistic, IMO.

by the way.

[ImmortalZ]

Ok, I can take that.

One more thing, there is no environment mapping or specular maps visible in D3D mode; only in OGL. (The Perseus looks da bomb!). Is this also supposed to be so? I've been reading the forums for 2 days now and I got the impression that OGL lags behind D3D in getting features implemented.

Edit : Thanks for the welcome!

I'm an ardent fan of FS2, used to FRED back in the day. Maybe when I get some time off college, I can join up with the Academy and refresh some of my (pathetic ) skills. This is some major @#$% you've done with the source:eek:  My congratulations to team/contributors!

Edit 2 : Corrected above, thanks kara.

[WMCoolmon]

Originally posted by Cobra
i think it's supposed to be that way to make a little realism. no flare would look unrealistic, IMO.

by the way.

Actually it'd look better if there was no engine trail, but the glow was still there.

I dunno if it's possible to gradually fade the flare, though.

[karajorma]

You won't see bump mapping in any mode. They've not been implemented yet. The spec maps may be designed to look like bumps on some ships though.

The lack of Shine maps in d3d is also a known error with ATI cards and will be fixed after 3.6.7 is released (hopefully soon). At this point we'll be getting a whole host of graphical improvements

[ImmortalZ]

Bump mapping : I meant environment mapping. I saw somewhere in the Wiki that it was implemented and is supposed to look beautiful.

Also, there is one more thing I discovered.

I changed over to OGL and immedietely noticed the colours being off (I could see green and red shades in the hangar interface). It looked like a gamma problem, but I couldn't reduce it from the game options. The gamma settings in my graphics control panel is not changed in anyway.

Then suddenly enough I started getting "Error reading from registry. Stopping" from the loader (gave me one hell of jump). I went in regedit and found the key and lo and behold, there were two gamma entries. One at 1.00 and one at 1.80. Changed the latter to 1.00 and I was in graphics heaven with OGL.