What does not
Game is semi-unstable without native msvcrt.dll from VC++ 6.0 runtime. Crash might happen right after "loading" screens when entering or leaving "battle mode", when traveling from one part of the world map to another, e.t.c., but it is not a strict rule. With current Wine version these crashes seem to be pretty rare - I had to play the game for about 30-40 minutes to finally hit this bug.
With native dll override (use "winetricks vcrun6sp6" or "winetricks vcrun6") the game works perfectly, I had played it straight for several hours in a row without an issue.
Game require a native DLL from DirectX 9 redist to start. Actual DLL required is d3dx9_33.dll, but there are issues with some shaders (mostly water surfaces) when using that one. Installing native d3dx9_36.dll instead does the trick - most of the time game behaves perfectly rendering smooth glitch-free picture. There still might be problems with water surfaces sometimes, but usually it is enough to save at the problematic place, exit the game, start it up again and load the save to "fix" the problem. Observation show that this pixel shaders bug is an "atomic" thing - you either hit it or not during single continuous gameplay session. In other words, would you hit this bug or not is determined by some unknown factors at the moment you start up the game. If you happen to be lucky and hadn't hit it at the game startup - game would behave perfectly until you exit it.
What was not tested
Everything relevant was tested (exploring the game world, interacting with NCPs, traveling from one part of the game world to another, fighting battles, entering castles and towns, e.t.c.). Hadn't played the game for an extended period of time (say, for more than 4-5 hours) as this playback session was mainly done to write this APPDB report, so there still might be some uncovered bugs that happen too infrequently to be caught during few testing hours.
Game was tested on computer with following specs: AMD FX 8120 CPU @3.1GHz, 8GB DDR3 RAM, GeForce GTX 550 Ti GPU with 1GB VRAM, nVIDIA Binary Driver v.275.09.07. Specifically for this test I had used a fresh clean Wine prefix in which I had installed d3dx9_36 and vcrun6sp6 using Winetricks (fixes game refusing to start problem and infrequent game crashes at "loading" screens). There's a possibility of hitting another bug with this game if case more recent version of nVIDIA Linux GPU driver is being used (285.x and newer). This bug manifests itself as an incorrect in-game rendering of character models. Player model itself and models of other non-player characters disappear for the most part (i.e. most of the model surfaces are not rendered at all). I believe this bug to be a bug in nVIDIA GPU drivers which somehow get into a "corrupted OpenGL state machine" state, most likely as a result of incorrect vdpau usage by notorious Adobe Flash Player . I'm still trying to investigate this case further and hope that it'd be possible to pinpoint the root cause of this bug in the future.
|Operating system||Test date||Wine version||Installs?||Runs?||Used|
|Show||Ubuntu 16.04 "Xenial" amd64 (+ variants like Kubuntu)||Jan 20 2017||2.0-rc2||Yes||Yes||Platinum||ottopower|
|Current||Fedora 14||Apr 17 2012||1.5.2||Yes||Yes||Gold||Alexey Loukianov|
|Show||Fedora 14||Oct 03 2011||1.3.29||Yes||Yes||Gold||Alexey Loukianov|
|Show||Ubuntu 10.10 "Maverick" i386 (+ variants like Kubuntu)||Dec 13 2010||1.3.9||Yes||Yes||Gold||dr. Hannibal Lecter|
|Show||Fedora 12||May 22 2010||1.1.44||Yes||Yes||Gold||an anonymous user|