Hard Light Productions Forums
General FreeSpace => FreeSpace & FreeSpace Open Support => Topic started by: Shevek on January 11, 2010, 07:09:31 pm
-
I've never posted here before, but I've been lurking here for ages. 3.6.10 looks so amazing; I haven't wanted to play FreeSpace this much since the original retail version was in print. So I'm very frustrated that when I try to start the campaign, the game crashes after the first training mission loads, by freezing with the sound looping. My hardware is a MacBook Pro 2.1, so Core 2 Duo processor and ATI X1600 graphics card. I'm running the game on a fresh install of Ubuntu 8.10 with flgrx installed (annoyingly, I had to make a separate Linux install just for FreeSpace, because newer kernels don't support the X1600 with 3D acceleration). My FSO install is more or less fresh, and I typically play with most eyecandy turned on. To find out if it was a mission-specific problem, I tried skipping the training missions, and going straight to the first actual mission. The game crashed to desktop when I clicked into the ship selection screen, with the error "ASSERTION FAILED: "!resize" at graphics/gropengltexture.cpp:944". Log of that is attached. I'm thinking the problem may have to do with my graphics driver, but I'm not sure what to do about it. One thing I could try is installing the version of the driver from ATI's website, instead of the one from the Ubuntu repository, but I'd like some input first. Suggestions?
Thanks!
[attachment deleted by admin]
-
Put this: http://www.mediafire.com/?dyejkjlqmm3 into mediavps/data/effects
-
I did, but the game still crashes, though now at unpredictable spots, including in the tech room (I have 3D tech room models turned on).
I don't know if this matters, but I also get weird artifacts above and below the opening movie.
Thanks for the prompt response.
-
You need to use -no_gls as command line argument to disable OpenGL shader support, which is not properly supported by lower end Radeon X1k series.
-
If it's not supported, why do the normal maps display perfectly? It seems like if theres a fix for it in WinXP by using a different .dll file, there ought to be something I can do in Linux.
-
The problem is that that particular series of cards needs a few driver-level hacks to implement the shaders correctly. Later cards implemented the hardware correctly, thus the driver hacks were removed. As a result, the latest available drivers for that card are somewhat broken.
The Windows workaround consists of using an atioglxx.dll from an older driver that was known to work correctly. I don't know how possible a similar solution for Linux is.
-
That, and whatever else you do, please remove -height from the command line.
And the Assert on !Resize is indeed fixed by the linked to cubemap.dds file The_E linked to.
Try this package as well: FSU_and-PostProc_Shaders.7z (http://zacam.ueuo.com/fsu/FSU_and-PostProc_Shaders.7z)
You don't necessarily need the post-processing portion, but make sure main-v and main-f SDR files are in mediavps\data\effects, that may help.
If/when it does crash at another spot, another log would be helpful. In fact, run debug until it crashes, copy out the fs2_open log file to another location, and then run it again until it crashes and see if it's the same thing (but at a different time) that it's crashing on.
Additionally, temporarily removing MV_Advanced or using the MV_Abolish.vp could go a long way to providing some additional information as well.