[karajorma]

Originally posted by CP5670
However, it was giving some strange errors as well. I remember seeing this one about the sum of a ship's subsystem strengths exceeding that of the ship itself; I think I did that for gameplay balance somewhere and although it does not make sense, I wouldn't really classify it as an error, especially since it is very unlikely that anyone will actually notice unless the tables themselves are examined.

Flipside posted a few times that the same thing was causing assertion errors in the debug version of FS2_open. He was also getting crashes in the non-debug version which were probably related to the same problem so although FS2 may have been able to handle this I don't think FS2_open can.

Which makes it a good thing if it's reported as a bug

[RandomTiger]

OK, I've uploaded updates for debug and release.
It should display text now.

http://mysite.freeserve.com/thomaswhittaker/c_code/freespace/fred2_open_ogl_d.rar
http://mysite.freeserve.com/thomaswhittaker/c_code/freespace/fred2_open_r.rar

Summery of problems:

Bug 1: aldo_14, freeze on startup
Bug 2: karajorma, error on exit
Bug 3: Trivial Psychic, cain zoom distortion
Bug 4: freeze on property change

I'll put together a new debug exe to try and give some feedback on 1 and 2
Trivial Psychic, do you experience any problems like 3 in fs2_open OGL? Also could you try and replicate 4 with a debug Fred OGL

List of (realistic) wants:

1. HTL version
2. Ability to turn specular and glowpoints off

[Flipside]

I know you probs already have this in mind, but can I add .jpg and .tga support to that list please?

The new version runs fine on my setup as always

One thought, while HTL would be great, surely anyone with a card capable of supporting it will run fast enough not to need it in a mission editor?

[RandomTiger]

Originally posted by Flipside
I know you probs already have this in mind, but can I add .jpg and .tga support to that list please?

As yes, it needs detection support as well as display yes?

One thought, while HTL would be great, surely anyone with a card capable of supporting it will run fast enough not to need it in a mission editor?

HTL runs faster for everyone, it takes a different codepath were all 3D objects are batched. It shouldnt be too hard, all the code is already there. It would be nice for people with slow machines or large complex scenes. Also at some point we may want to get rid of the non htl code path meaning FRED would have to become HTL as well.

[vyper]

RT, those latest links bottom out about 1/4 of the way through download and give me an incomplete archive.

[aldo_14]

Originally posted by RandomTiger

Bug 1: aldo_14, freeze on startup

I'd gave even odds this is a system and not a code problem, though  - so don't go nuts unless someone lese has the same thing.

[vyper]

sorry, just the relase link does it

[Flipside]

Originally posted by RandomTiger


As yes, it needs detection support as well as display yes?
 

Yes, the loading errors were thrown up by the debugger, simply cos it couldn't find the PCX, but it doesn't actually see jpgs at all if they are there, even to tell me it can't read them, so I suspect it would need both.

As for the HT&L I understand now, thanks

[Goober5000]

ATTENTION EVERYONE

CP, Black Wolf, and others, this means you.

If you run the debug EXE and you get table errors, you really should fix your tables.  If you continue to play FS2 with incorrectly formatted tables, you will experience problems.  They may be noticeable (e.g. a crash) or they may be silent, but they will be there. The reason the error messages are there is to let you know that your stuff needs fixing.

Examples:
* The issue with the subsystems being more than the hull: This has the potential to cause a divide-by-0 error or to throw a bunch of negative numbers into certain mission-critical calculations.
* The issue with FRED not liking the lack of subsystems: Every ship should have a subsystem, even if it's just "Pilot".  (Cargo containers and navbuoys might be exceptions... I'd have to check.)
* The issue with the missing weapon on the loadout: At minimum, you'd have a mission where you couldn't select the Flail, which may or may not be noticed.
* And so forth.  The recent FS Port crashes were caused by assigning sound indexes that were out of range of the possible values.  Running debug builds is helpful to catch these things.

[Trivial Psychic]

RT

This new build is more stable (hasn't crashed yet), but the texture thing is still occuring.  I had a bit more time to look around and grabbed a couple more screens.  Note: I'm using the release build, not the debug build.





For the cruiser, the same effect didn't happen with any Fenris, as with this Leviathan.  Now this is the Leviathan from the Port, so there may have been some skin differences that are causing this.

later!

[karajorma]

I'm still getting the same error on exit too.

Error: Unable to swap buffers
Error code: 0
File:H:\projects\Freespace\scp\fred_upgrade\code\Graphics\GrOpenGL.cpp
Line: 1489

The I get a windows error reporting window appear and only when I okay that the debug spew goes away.

The Non-debug version still requires a CTRL-ALT-DEL to get rid of.

[RandomTiger]

Originally posted by Goober5000
ATTENTION EVERYONE
If you run the debug EXE and you get table errors, you really should fix your tables.  If you continue to play FS2 with incorrectly formatted tables, you will experience problems.  They may be noticeable (e.g. a crash) or they may be silent, but they will be there. The reason the error messages are there is to let you know that your stuff needs fixing.

Goob, if i can I'm going to change fred to report these problems in release so everyone sees them no matter what. However I will set it up the user to be allowed to progress past them and get into Fred.

That seems like a good compromise.

[RandomTiger]

Originally posted by Flipside

Yes, the loading errors were thrown up by the debugger, simply cos it couldn't find the PCX, but it doesn't actually see jpgs at all if they are there, even to tell me it can't read them, so I suspect it would need both.

Crap! Im sorry Flipside, I've just realised that OGL doesnt have support for TGA and JPG. I dont want to get into doing that now because if I try to do everything all at once we'll end up with nothing.

However you can get around this by saving PCX versions of your TGA or JPG files. Thas a bit rubbish I know but hopefully someone will sort it out later. DDS could be a possible alternitive.

[Flipside]

Ok, I'll do that for now, thanks for looking

Hopefully it'll get done one day in the future, but you are right, spin too many plates and one of them'll smash

[phreak]

i think the entire texture subsystem needs torn up before i start doing anything with high color textures in OpenGL.  DDS textures may work, as i put them in a while back.

[Flipside]

DDS kinda works, all textures are displayed as white, but at least you know how big the ship is in relation to everything else


Actually, come to think of it, I may as well make PCX's LOL

Edit : This isn't so bad really, I tend to use huge UVmapped textures, so it's only a matter of converting anything in LOD0, which is going to be about 7 files at the absolute most

[CP5670]

That new version works pretty well here. Great usage of the FS font. I also got that problem kasperl mentioned, where loading up a mission and going into the briefing editor gives an illegal operation error. The briefing editor loads up properly if the first stage is blank but gives the error as soon as I try to place an icon, so it looks like it can't display briefing icons yet.

Flipside posted a few times that the same thing was causing assertion errors in the debug version of FS2_open. He was also getting crashes in the non-debug version which were probably related to the same problem so although FS2 may have been able to handle this I don't think FS2_open can.

Which makes it a good thing if it's reported as a bug

* The issue with the subsystems being more than the hull: This has the potential to cause a divide-by-0 error or to throw a bunch of negative numbers into certain mission-critical calculations.

hmm, that's interesting; I know it worked fine in the original FS2. This could be problematic since I did that for game balancing in a few places and changing it now would create balance issues in a few missions. I personally have not encountered any problems with this so I'm going to leave that stuff as it is for now, but it's something I may have to deal with later.

The frequency of those random crashes seems to be entirely dependent on the build versions and texture/effect sizes though, so it probably has to do with memory usage. I removed some stuff from the media vp and the 12_05_2003 version has been working without any crashes for a while now.

* The issue with FRED not liking the lack of subsystems: Every ship should have a subsystem, even if it's just "Pilot". (Cargo containers and navbuoys might be exceptions... I'd have to check.)

The pilot subsystem is never actually used in the original ships.tbl - it only shows up in mission files - and there are a few ships without subsystems in there. Anyway, that's not what it is complaining about, but rather the fact that the last property in a ship listing is not what it is looking for. Can $Stealth be put anywhere? I placed it directly under $Score since that is what was done with the Pegasus and Ptah, but those ships had subsystems to go after that.

[Bobboau]

it is an optional string, it will only be used if it is found in the right spot

[Goober5000]

$Stealth goes after $Score and $Thruster stuff but before $Trail and $Subsystem stuff.  And if it's a fighter, you really should have sensors, weapons, engines, etc.

You can post the subsystem vs. hull thing on Mantis and I'll take another look at it eventually, but I can't say when.

[CP5670]

I'm actually using it for those Polaris pieces made by Dark, which effectively behave like cargo containers, as I didn't want them cluttering up the radar. So $Stealth comes after $Score as it should, but since the "ships" have no trails or subsystems, the debug version is giving an error.