[Bug 37705] Heroes of Might and Magic VI Shades of Darkness Textures Fail to Display in Most Campaign Maps

[wine-bugs at winehq.org]

https://bugs.winehq.org/show_bug.cgi?id=37705

Paul Gofman <gofmanp at gmail.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |gofmanp at gmail.com

--- Comment #24 from Paul Gofman <gofmanp at gmail.com> ---
Created attachment 63529
  --> https://bugs.winehq.org/attachment.cgi?id=63529
wined3d: Do not treat  _SAMP_SRGB_TEXTURE state as a boolean value.

I tested the game from Ubisoft website (started by UPlay; basic package, some
campaigns look missing). The game worked right away with d3dcompiler_47 native
override only. I could reproduce the problem by starting "Dungeon Campaign"
("The Other Elves"). The screen gets dark during the intro animation, and after
that the map is dark on load. If I save the game, restart and load, or reset
device somehow (e. g. by changing resolution), the map is displayed nicely.

The problem is that at the start of intro animation the game sets some big
nonzero value (0x7e41882a) for D3DSAMP_SRGBTEXTURE (sampler 0). It does it
somewhere close to the end of frame rendering, and goes rendering next frame
with this value set. It sometimes switches D3DSAMP_SRGBTEXTURE to 0 for a short
while, but sets it back to the same value after, and most of the textures are
accessed with SRGB correction under Wine since that. Sampler state setting
looks exactly the same under Windows. The same sequence recorded and displayed
by apitrace displays without dark textures under Windows, and with those under
Wine. Ultimately disabling D3DSAMP_SRGBTEXTURE state handling makes it display
like on Windows.

Weird enough, Windows seem to not turn SRGB correction mode on when this value
(0x7e41882a) is used. I am attaching the patch which includes the test (which
succeeds for me on a Windows machine) showing that setting 2, 100 or that value
from the game trace does not make native d3d9 to perform SRGB correction. It
does that if the value of 1 is set.

The patch attached fixes the issue for me (tested on the campaign start
referenced above). The match introduces a comparioson of D3DSAMP_SRGBTEXTURE
value with 1 instead of treating it as a boolean value.

-- 
Do not reply to this email, post in Bugzilla using the
above URL to reply.
You are receiving this mail because:
You are watching all bug changes.