[taylor]

A ton of little things went into this one and hopefully it pays off.  Has both DirectSound and OpenAL versions this time with the OGG streaming fixes.  Other things which may be of interest that are new to this build:

 - squash all bmpman bugs (I wish!  Hopefully fixed enough to calm redmenace down a little though)
 - fix crash and possible memory leak in jumpnode code
 - fix problem with previous build where OpenGL would mess up when it was minimized/restored
 - possible fix for some ATI users having a green screen in screenshots and under popups in OGL
 - cleanup of OpenGL debug info which makes it a little easier to read
 - if a OGL extension is found but a related function is not then fail right away rather than trying to use it later
 - cleanup of OGL rendering to possibly help prevent real bad shards of death (doesn't fix the main problem though, just trying to deal with it)
 - optimize glDrawRangeElements by only handling sized blocks of data that the video card is optimized for.  Don't know if this is a speed-up or slow-down but give it a try.
 - fix IBX size detection, was only checking against a smaller size than we wanted.  Only was a problem if the IBX failed to generate fully the first time.
 - prevent crash with 0 length IBX files.  It will now delete them properly rather than still trying to use it.
 - if a specular texture wasn't found during model load then don't use the normal texture as a spec.  Fixes various bmpman accounting issues.
 - IBX debug info which reports the checksum of the model and the IBX (info that's really only useful to me)

Nothing on the nebula slowdown in D3D yet.  I went through a big diff to current version from about when it started happening but didn't notice anything.  I'll go back and try to narrow down which set of changes actually caused it and work from there.

Didn't have a chance to work on the debriefing wav repeat but I did experience it.  Please check if it happens with the DirectSound build since I think that it may just be an OpenAL thing.  If it happens with both builds the report it as such and I'll give the fix a higher priority on my todo list.

http://icculus.org/~taylor/fso/willrobinson/20050613-win32.rar

I'll update this post with individual links to the release builds when I get the chance.

[Skippy]

You posted so fast that the file isn't there yet


404 Error

Edit : correct link is http://icculus.org/~taylor/fso/willrobinson/20050613-win32.rar

[Goober5000]

Link fixored.

[Starman01]

A crap, WCS-data still don't run, same problem than the last few builds, guess I have to go back to goobers december-build again

@Bobboau : Have you find anything out about mantis bugreport No. 0000423  ? This error prevent completly, that we use any newer builds. I hope that you figure out the problem soon, there are a few nifty features in the new builds, that we would like to make available in the prologue once released

[taylor]

Sorry about the link.  I usually test it before walking away from the computer but I was pretty sleepy by the time it got posted so, didn't think about it.  Good news on the shards of death front though, it might have actually been fixed.  fizz sent me a patch (he's been looking at it too) and in applying that I noticed another problem.  If it works out and we get it tested then I'll update the build with the fixes.

Originally posted by Starman01
A crap, WCS-data still don't run, same problem than the last few builds, guess I have to go back to goobers december-build again

What was the problem here?  I don't think I've heard any details of what's messing up.  PM me if you want and I'll take a look.

[Nuke]

what does...

Error: trying to set invalid bitmap as render target
File:U:\src\cvs\fs2_open.testing\code\graphics\GrD3DBmpman.cpp
Line: 652
[This filename points to the location of a file on the computer that built this executable]

Call stack:
------------------------------------------------------------------
------------------------------------------------------------------

 ...mean?

and oddly the debug build screws up even more.

the bool lists are still not working properly, did the syntax change?

[taylor]

Originally posted by Nuke
Error: trying to set invalid bitmap as render target
File:U:\src\cvs\fs2_open.testing\code\graphics\GrD3DBmpman.cpp
Line: 652

Got a full fs.log I can look at?  At what point during the game did this happen?

and oddly the debug build screws up even more.

In what respect?

the bool lists are still not working properly, did the syntax change?

WMCoolmon changed something here if I remember correctly.  Not sure what the proper syntax is though.  Supposedly he fixed whatever problem there was, or that he knew about, so you might want to ask him.

[redmenace]

Originally posted by Nuke
what does...

Error: trying to set invalid bitmap as render target
File:U:\src\cvs\fs2_open.testing\code\graphics\GrD3DBmpman.cpp
Line: 652
[This filename points to the location of a file on the computer that built this executable]

Call stack:
------------------------------------------------------------------
------------------------------------------------------------------

Looks like the second part of the BMPMAN error I was getting.

[Nuke]

i get the errors running nukemod when trying to use weapon models feature that bobboau implemented. the features have not worked in the last 5 or 6 builds. im pretty sure my tables are in order, they work in older builds and the only changes ive made were testing different possible bool list syntaxes in case it was changed

Originally posted by taylor

Got a full fs.log I can look at?  At what point during the game did this happen?

il upload a zip in a minute, it happens upon hitting commit in the loadout screen. freespace doesnt actually hault, i get the error about 3 or four times, and then i can get back in the game, but then i get the same errors again. i have to use the device manager to kill the game.

In what respect?

i get errors similar to those i recieved before the "fix" was implemented. of bool errors, and of other entries, such as $shields and $hitpoints reporting bool errors.  such as

Warning: Boolean '$Shields:' type unknown; assuming 'no/false'
File:U:\src\cvs\fs2_open.testing\code\Parse\PARSELO.CPP
Line: 1787
[This filename points to the location of a file on the computer that built this executable]

Call stack:
------------------------------------------------------------------
    stuff_bool_list()    parse_ship()    parse_shiptbl()    ship_init()    game_init()    WinMainSub()    WinMain()    WinMainCRTStartup()    kernel32.dll 7c816d4f()
------------------------------------------------------------------

WMCoolmon changed something here if I remember correctly.  Not sure what the proper syntax is though.  Supposedly he fixed whatever problem there was, or that he knew about, so you might want to ask him.

he attempted to fix the bool list parsing wich is required by the show weapon feature. hopefully he will read this and check his code.

ok, no error logs, i cant get the came to actually crash, it just pops up an infinite number of similar errors, i hit alt+n for a half an hour and still no crashie.
www.game-warden.com/nukemod-cos/tests/bugs.zip