[taylor]

Because of one particular bug fix this is OpenAL only.  If anyone using OGGs for music or (de)briefing voices has any problems here please let me know.

Fixes:
 - extra bmpman debugging info (for redmenace)

 - little cleanup of OpenGL code and a slight speedup for OGL page flips

 - fix OGG streaming error that would cause audio to repeat the first half second or so

 - debug builds in 1024x768 will show the right-hand memory info further on the right rather in the middle of the screen.  OAL builds will not show "S-HRAM" since it's never going to be used.

 - fix save_screens (used with popups) and print_screens (screenshots) in D3D mode while in a window (for Raa)

 - properly (as much as possible) account for D3D texture usage

 - make gamma ramp inversion fix for OGL that Sticks made for D3D

 - fix possible crash issue when you hit pause in a multiplayer mission


http://icculus.org/~taylor/fso/willrobinson/20050606-win32-OAL.rar (5.6meg)

Any of the previously listed (my last build thread) Mantis bugs that have not already been updated as fixed or not need to be with this build.  If they are not confirmed as still broken they will soon be resolved as fixed.

Report any new bugs, blah, blah, blah.  I get tired of always saying that.

[Hippo]

Whats the difference between OGL and OAL?

[redmenace]

Not sure if you are aware of it but the first second of a debriefing wav file gets repeated. Also the Shards of death are inflicting the Hecate, but not the Leviathon.

I also played 4 or 5 missions in a row and recieved one of the bmpman  errors.

Also Nebula Lightning Sound is not working. It could be related to when Open GL minimizes for an error. But I am not sure in any way.

[Trivial Psychic]

Originally posted by Hippo
Whats the difference between OGL and OAL?

OGL = Open Graphical Language (?), as an alternative to D3D
OAL = Open Audio Language, as an alternative to... um... what you normally use.

[WMCoolmon]

OpenAL is a wrapper for whatever sound system is actually used by the OS. It also adds some nifty features like surround sound and doppler shifting.

In the case of Windows, it adds an extra layer of abstraction on top of DirectSound. As a result, the same code can be used on Linux and Windows.

[taylor]

Originally posted by redmenace
Not sure if you are aware of it but the first second of a debriefing wav file gets repeated.

Of a WAV?  Damn, didn't check those after the OGG fix.  Should be simple to track down though.

Also the Shards of death are inflicting the Hecate, but not the Leviathon.

Aarrrgg!  Forgot to add the IBX debug patch.  And regarding that second set of cruiser01 files that you posted, they were still wrong.  Different data than the previous set that you posted but two subobjects had extra verts than what I'm able to recreate.  So much for 1 in a 100 I guess since you had two wrong in a row.  I'm going to get you all of my IBX files so that you don't have to deal with this but it is a problem that needs to be addressed.

I also played 4 or 5 missions in a row and recieved one of the bmpman  errors.

That blasted Tercoc01.tga again.  I had trouble with that one file before but thought that I had finally fixed it.  It's used and released properly in an earlier mission and next time it gets released after the second try which looks right on the surface.  Problem is that it's trying to get used again after that.  I'm not sure if it's a loading or an unloading error but I'd bet it's a loading problem or something in the model code.

At the very least I should have enough info to build a test case now though and replicate this on my end.  In case it was needed I made this very easy to fix by design, just change a "true" to a "false".  The problem with that though is you'll only be able to play 6 or 7 missions in a row before you run out of bitmap slots and start getting graphics corruption.  Another last resort fix is to just never release that file and that's easy to do as well, just a bad hack, but will avoid running out of bitmap slots.

Either way this won't happen for 3.6.7 and I'll try and get it fixed as soon as possible.  If you keep getting more bmpman errors please post them so I can make sure they are all from the same file.

Also Nebula Lightning Sound is not working. It could be related to when Open GL minimizes for an error. But I am not sure in any way.

Yeah I did make a change here (OGL minimizing) but didn't intend to ever put it in CVS.  I'll assume that's the problem but if it's still messed up in the next build that anyone releases then let me know.

[Nuke]

did the bool array parse thingie that was preventing vwep from working get fixed?

[redmenace]

Ok, since we have new bmpman debug spew, I will have to go through D3D and search for the error in there as well. Since that is now on hold for the time being I will have to start concentrating on documentation for you know what

[Trivial Psychic]

Originally posted by taylor

Originally posted by redmenace
Not sure if you are aware of it but the first second of a debriefing wav file gets repeated.

Of a WAV?  Damn, didn't check those after the OGG fix.  Should be simple to track down though.
[/B]

I get this too, but not every time.  I was testing out the Port again, and noticed this.  Not every debriefing stage is effected by this.  Sometimes its the first stage, sometimes its the second, sometimes it doesn't happen at all.  Also, if this bug occurred, and then I selected the mission statistics screen, then went back to the debriefing, the same bug won't occur.  It's weird like that.  I have tested out TBP a bit more and can report that the original briefing & command briefing repeat bug isn't occuring for ogg, but I haven't tried it in ogg debriefings yet.

[taylor]

Originally posted by redmenace
Ok, since we have new bmpman debug spew, I will have to go through D3D and search for the error in there as well. Since that is now on hold for the time being I will have to start concentrating on documentation for you know what

Squash'ed!  I think so anyway.  I pulled a couple of extra functions out of my experimental tree which should work well to solve the problem that I noticed.  The problem was just how model_page_out_textures() was used and the way I was trying to save bitmap slots.  One_function + two_different_methods_of_working = big_trouble.  Not exactly sure how I got it to work so well in the first place given the problem but now it should work well regardless of how it's used and still get the proper results.

A new build with this and other bug fixes should be released at some point later today.  Also fixed that jumpnode related crash which you had gotten before.  A few OGL changes may help prevent the shards of death.  Won't help the fact that the generated data is still wrong but hopefully OGL may handle it better now.  Plus vert rendering will be cut up and restricted to what the video card says that it's best at handling at a time.  Don't know how well that translates to speed (up or down) but the more complex the model the more chance of a big speed drop with the old method so hopefully this is better.

[redmenace]

how about nebuleas?

[Trivial Psychic]

Originally posted by taylor
A new build with this and other bug fixes should be released at some point later today.

I await with baited threath.

[taylor]

Originally posted by redmenace
how about nebuleas?

Haven't found anything yet but I'll keep working on that today and hopefully figure it out before posting the new build.