[SirKnightly]

Here is the nightly for Windows on 17 Aug 2010 - Revision 6377

Group: SSE2
fso-WIN-SSE2-20100817_r6377.7z
MD5Sum

Group: Standard
fso-WIN-Standard-20100817_r6377.7z
MD5Sum

Group: Inferno_SSE
fso-WIN-Inferno_SSE-20100817_r6377.7z
MD5Sum

Group: Inferno_SSE2
fso-WIN-Inferno_SSE2-20100817_r6377.7z
MD5Sum

Code: [Select]

------------------------------------------------------------------------
r6351 | Sushi | 2010-08-05 09:17:20 -0500 (Thu, 05 Aug 2010) | 1 line
Changed paths:
   M /trunk/fs2_open/code/weapon/shockwave.cpp
   M /trunk/fs2_open/code/weapon/weapon.h
   M /trunk/fs2_open/code/weapon/weapons.cpp

Support for allowing bombs to be destroyed by weapon blast and/or shockwaves. Weapon flags are "takes blast damage" and "takes shockwave damage"
------------------------------------------------------------------------
r6353 | The_E | 2010-08-10 13:41:12 -0500 (Tue, 10 Aug 2010) | 1 line
Changed paths:
   M /trunk/fs2_open/code/graphics/gropenglextension.cpp

Remove requirement for SM3. While this will not allow cards that are only capable of SM2 to use the standard shaders, now shaders can be written for them.
------------------------------------------------------------------------
r6354 | The_E | 2010-08-11 08:08:51 -0500 (Wed, 11 Aug 2010) | 1 line
Changed paths:
   M /trunk/fs2_open/code/fireball/fireballs.cpp
   M /trunk/fs2_open/code/parse/sexp.cpp

Fix for assertion encountered with the warp-effect sexp. Also updated sexp help to indicate correct usage.
------------------------------------------------------------------------
r6374 | The_E | 2010-08-16 17:04:23 -0500 (Mon, 16 Aug 2010) | 1 line
Changed paths:
   M /trunk/fs2_open/code/ai/aiturret.cpp
   M /trunk/fs2_open/code/parse/lua.cpp
   M /trunk/fs2_open/code/ship/ship.cpp
   M /trunk/fs2_open/code/ship/ship.h

Allow scripting to override turret AI target selection
------------------------------------------------------------------------
r6375 | The_E | 2010-08-16 18:18:10 -0500 (Mon, 16 Aug 2010) | 1 line
Changed paths:
   M /trunk/fs2_open/code/parse/lua.cpp

And of course, I screwed it up a bit. Reading the documentation is A1 SUPAR.
------------------------------------------------------------------------
r6376 | IssMneur | 2010-08-17 00:37:38 -0500 (Tue, 17 Aug 2010) | 1 line
Changed paths:
   M /trunk/fs2_open/code/parse/sexp.cpp

Clarify the help text for set-object-speed-x, set-object-speed-y, and set-object-speed-z
------------------------------------------------------------------------
r6377 | IssMneur | 2010-08-17 00:42:55 -0500 (Tue, 17 Aug 2010) | 1 line
Changed paths:
   M /trunk/fs2_open/code/parse/sexp.cpp

Fixes mantis 2275: Improving remove-background-bitmap and remove-sun-bitmap so that they issue an Error if an invalid index is passed via the sexp.
------------------------------------------------------------------------



[The E]

Code: [Select]

------------------------------------------------------------------------
r6353 | The_E | 2010-08-10 13:41:12 -0500 (Tue, 10 Aug 2010) | 1 line
Changed paths:
   M /trunk/fs2_open/code/graphics/gropenglextension.cpp

Remove requirement for SM3. While this will not allow cards that are only capable of SM2 to use the standard shaders, now shaders can be written for them.


Okay, here's something for you to try, IF you have a GPU that until now couldn't support the FSO shaders. Download this file, and place it in mediavps_3612\data\effects. Then start the game with the -normal commandline option (or without the -no_glsl option). Please post a debug log, even if it works OK.

[Commander Zane]

SM2 support! You're a ****ing hero!
The bomb thing sounds gnarly.

[The E]

Please note that I can't make guarantees for post-processing (as it doesn't work on my end), and that if you are using this new shader, I'd love to see a debug build, regardless of whether or not they work.

[MatthTheGeek]

Well, still doesn't work for me. We have already talked on IRC about this, and btw I notify that I'm using the most recent drivers that can run FSO - any driver more recent make it freeze.

Log attached.

[attachment deleted by ninja]

[swashmebuckle]

Mac report here.  Seemed to work fine for normals (assuming I followed the instructions correctly), not sure if bloom was working or not--probably not.  It did spend a lot of time loading up models I already should have had cached versions of, like the Hatty (looked at it in the ships lab).

[attachment deleted by ninja]

[The E]

Load delay was due to cache files being regenerated.

Okay, another question: What kind of performance can you see? Better or worse than fixed-pipeline rendering? Any visual differences (aside from normal maps, of course) ?

[swashmebuckle]

Actually normals worked fine for me before as well--are they being treated differently with this shader?  I'll turn up my detail settings to see if I can notice any differences.  Is there something I'm supposed to be looking for in particular?  And is there a key command to use fixed pipeline rendering so that I can compare things side by side in a mission?  Max fps is unfortunately locked at 60 for me so the ship lab isn't much use in seeing what is faster...

Edit: Oh, this might have fixed the environment mapping issues I had on OSX where random files would be reflected rather than the skybox.  Will have to crosscheck things further with FotG...

[The E]

Unfortunately, there's no way to switch renderers mid-mission, you'd have to start the game with -no_glsl. And no, you shouldn't really see any differences, as this shader is supposed to be functionally equivalent to the old one, but you never know.

[swashmebuckle]

OK, Massive Battle 1 with maxed detail settings and me just sitting still looking at the action hit a low of 6-7 fps with the initial clash of forces.  The same point with the fixed render pipeline hit a low of 9-10 fps.

[chief1983]

Swash, I think the skybox reflection stuff was fixed by OS X 10.6.3 or 10.6.4 and their respective updates to the OpenGL drivers.  But if this fixed it that'd be awesome too as that's been a longstanding OS X issue.

[swashmebuckle]

Yeah nvm there's still some map weirdness there on 10.5.8

Tried the same mission with the old shaders and got a low of 8-9, but considering how it plays out differently every time I'd say they're pretty darned close.

[chief1983]

10.5.8 is not likely to be fixed unless they patch their OpenGL drivers in a 10.5.x updated the same way they fixed them on 10.6.  Is there another big 10.5 patch in the pipeline?

[Aardwolf]

Ooo, bombs that can set off other bombs! Now there could be a reason not to fire too many swarm bombs

[Echelon9]

Swash, I think the skybox reflection stuff was fixed by OS X 10.6.3 or 10.6.4 and their respective updates to the OpenGL drivers.  But if this fixed it that'd be awesome too as that's been a longstanding OS X issue.

Even with this week's OpenGL patch for 10.6.x there are still issues with Apple's OpenGL / GLSL drivers not following the spec properly. For instance, one bug that we have to work around in the MediaVP shaders is still present.

[chief1983]