[SirKnightly]

Here is the nightly for 22 October 2022 - Revision 11b37bbea



Group: Win64-SSE2
nightly_20221022_11b37bbea-builds-x64-SSE2.zip (Mirror)


Group: Linux
nightly_20221022_11b37bbea-builds-Linux.tar.gz (Mirror)


Group: MacOSX
nightly_20221022_11b37bbea-builds-Mac.tar.gz (Mirror)


Group: Win32-SSE2
nightly_20221022_11b37bbea-builds-Win32-SSE2.zip (Mirror)

Code: [Select]

------------------------------------------------------------------------
commit 41074956a
Author: Goober5000
Commit: Goober5000

    more fixes
 code/scripting/api/libs/mission.cpp |  2 +-
 code/scripting/lua/LuaConvert.cpp   | 15 ++++++---------
 2 files changed, 7 insertions(+), 10 deletions(-)

------------------------------------------------------------------------
commit c63ce5de8
Author: MjnMixael
Commit: GitHub

    Allow writing to player mission specific stats (#4801)
 code/scripting/api/objs/player.cpp | 115 ++++++++++++++++++++++++++++++++++---
 1 file changed, 106 insertions(+), 9 deletions(-)

------------------------------------------------------------------------
commit 17e42e92c
Author: MjnMixael
Commit: GitHub

    cockpitinactive gauges should render in chase view (#4804)
 code/hud/hud.cpp | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

------------------------------------------------------------------------
commit debedf26e
Author: Goober5000
Commit: Goober5000

    partial fix
 code/scripting/api/libs/mission.cpp |  7 +++----
 code/scripting/lua/LuaConvert.cpp   | 10 +++++++++-
 2 files changed, 12 insertions(+), 5 deletions(-)

------------------------------------------------------------------------
commit 49fb2f07c
Author: Goober5000
Commit: Goober5000

    refactor per feedback, tweak to fix circular #include, use accessor, reorder, fix namespaces, fix API documentation
 code/scripting/api/libs/mission.cpp | 96 ++++++++++++++++++++++++++++---------
 code/scripting/lua/LuaConvert.cpp   | 25 ++++++----
 code/scripting/lua/LuaConvert.h     | 13 +++--
 3 files changed, 100 insertions(+), 34 deletions(-)

------------------------------------------------------------------------
commit f24277c39
Author: Goober5000
Commit: Goober5000

    make clang happy
 code/scripting/api/libs/mission.cpp | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

------------------------------------------------------------------------
commit 452c8370d
Author: Goober5000
Commit: Goober5000

    model instances are not needed for debris in many cases
 code/debris/debris.cpp     | 28 +++++++++++++++++-----------
 code/hud/hudtargetbox.cpp  |  4 ++--
 code/model/modelrender.cpp |  6 +++---
 3 files changed, 22 insertions(+), 16 deletions(-)

------------------------------------------------------------------------
commit 77179dc18
Author: Goober5000
Commit: Goober5000

    add mn.createDebris
 code/debris/debris.cpp              | 271 +++++++++++++++++++++++-------------
 code/debris/debris.h                |  14 +-
 code/scripting/api/libs/mission.cpp |  83 +++++++++++
 code/scripting/api/objs/enums.cpp   |   4 +
 code/scripting/api/objs/enums.h     |   4 +
 5 files changed, 276 insertions(+), 100 deletions(-)


[Trivial Psychic]

I'm getting a CTD on this build when starting one of my ITB missions.  It doesn't do this when using the Oct. 21st build.  The mission begins with the player's wing launching from a fighterbay hanger.  When I play with a debug build, I get this from the log:

Assert: "model_instance_num >= 0"
File: modelread.cpp
Line: 3782

I believe that it may have something to do with the Nova#shadow model, as I just tested out each mission that includes this ship, and I get a CTD at the start of each one.  The lab doesn't complain about the model, though.

Just a bit earlier on in the debug log, I get this:

WARNING: Failed to find uniform 'sEnvmap'.
WARNING: Failed to find uniform 'sIrrmap'.
WARNING: Failed to find uniform 'sEnvmap'.
WARNING: Failed to find uniform 'sIrrmap'.
WARNING: Failed to find uniform 'shadow_map'.
WARNING: Failed to find uniform 'shadow_map'.
WARNING: Failed to find uniform 'shadow_map'.
WARNING: Failed to find uniform 'shadow_map'.
WARNING: Failed to find uniform 'tex'.
WARNING: Failed to find uniform 'depth_tex'.

Running the debug build from the 21st gives those same warnings, but doesn't cause a crash.

Is this an oops with regards to new code additions, or is the new code cracking down on a model problem that has up to this point, not been a concern?  If it is the latter, I'd like to know what's wrong with the model so I can fix it.

Thanks.

[Goober5000]

There isn't enough information to be sure.  What mod are you running, and what mission produces the crash?

[Goober5000]

Addendum: On the assumption that this is TBP, I downloaded the mod, created a basic mission with the Nova#shadow, tested it, and did not experience a crash on the 10/22 nightly.  So I'll need the debug log and one of the missions that produces the crash.

[Trivial Psychic]

OK.  I've done some further testing, and have determined that the presence of the Nova#shadow is coincidence.  It seems to be a collision thing.  I started up a mission, turned, and rammed anything nearby, and got a CTD.  In most instances, the start=CTD happens when I'm starting in a fighterbay.  For one of the missions, I changed it to starting just in front of a ship, and had no problems.  Then I turned around sharply and it CTDed.  I thought that perhaps it had something to do with certain textures coming on screen, so I tried viewing the model in other missions with no problems.  Then I retried the same mission, put some distance between me and the launching ship, then came back around and things went smoothly.  My guess is that my sharp turn put my wingmen (who had 'guard Alpha 1' as their orders) were closer to the ship and hit it, causing the crash.  Flew a few more missions that were giving no problems, then rammed into the Nova#guns, and had a CTD.  I thought that maybe it was with THAT model, but then I tried another random mission and rammed an EA transport, and got a CTD.  Either its the collision itself, or something about playing the sound.  Either way, it's something that was introduced for this build.  Oct. 21st runs smoothly.

[Goober5000]

What mission do you get the crash in?  Either attach it here or tell me the file name.

Multiple people have said they are getting crashes on this build, but that doesn't help unless I can reproduce the crash.  Narrowing it down to collisions helps, but I tried crashing into a few things and no CTD occurred.

[Goober5000]

Ok, by combining information from your report and MjnMixael's report, I figured out this was being caused by shadows being enabled.  I was able to reproduce the crash and I now have a confirmed fix.

[Trivial Psychic]

Awesome!  Thanks for sticking it out with me.