[redmenace]

I got another loading crash in Exodus.

[redmenace]

I got a crash during Dunkerque while attacking nehema bombers.

[redmenace]

I got another crash during "Dunkerque."

Involved "Assert: aigp != NULL"

[redmenace]

I have a crash in "Their Finest Hour"

the fs.log is also saying that I do not have a debug_filter.cfg, but as asked I have put an empty file in c:\games\freespace2\data and c:\games\freespace2\SCP\data.

[taylor]

Originally posted by redmenace
Involved "Assert: aigp != NULL"

Fixed last night.  Same bug as your previous support ship docking problem.  The code that was supposed to fix it never got executed but now that's fixed.

the fs.log is also saying that I do not have a debug_filter.cfg, but as asked I have put an empty file in c:\games\freespace2\data and c:\games\freespace2\SCP\data.

[/B]
If you created the file with notepad then make sure it doesn't have an extra .txt extension on it which notepad always tacks on.  If it gets listed as a text file in Explorer then that's what happened but you wouldn't see the .txt if you have hide extenstions turned on.

[redmenace]

Here is the file I created.

[redmenace]

I hope the rest of the information is helping.

[taylor]

Bah, stupid Windows version.  It goes in ${HOME}/.fs2_open/data under Linux and OSX but the Windows version just has it in the same directory as the binary.  Just move it and it should work fine.

The bug reports are deffinitely a help, though I am starting to wonder if you're just padding your post count.

The first Dunkerque crash where you are attacking the bombers looks to be hitting one of the first bugs you posted about, the model_get() crash.  I added the extra NULL check there which is what it hit.  That helps figure out where it's coming from but I still can't say that I know exactly why it happened.  It looks to just be a problem with WMC's updated jumpnode code as it doesn't appear to handle model_load() errors very well but the rest of the code should have picked up the slack, at least in a debug build.  I'll add an extra debug printout at the problem area and see if that can't help figure out what's going on.  I'll try and reproduce as well but haven't been able to so far.

[redmenace]

Originally posted by taylor
Bah, stupid Windows version.  It goes in ${HOME}/.fs2_open/data under Linux and OSX but the Windows version just has it in the same directory as the binary.  Just move it and it should work fine.

alrighty then

Originally posted by taylor
The bug reports are deffinitely a help, though I am starting to wonder if you're just padding your post count.

Well if I am padding my post count, it is better than spamming in hidden and abandoned forums. And more constructive I might add. Honestly though, as I said in the internal, I am testing all of the retail missions. So I figure this is the best way to communicate severe and bothersome crashes in order to have them fixed for the next release

Originally posted by taylor
The first Dunkerque crash where you are attacking the bombers looks to be hitting one of the first bugs you posted about, the model_get() crash.  I added the extra NULL check there which is what it hit.  That helps figure out where it's coming from but I still can't say that I know exactly why it happened.  It looks to just be a problem with WMC's updated jumpnode code as it doesn't appear to handle model_load() errors very well but the rest of the code should have picked up the slack, at least in a debug build.  I'll add an extra debug printout at the problem area and see if that can't help figure out what's going on.  I'll try and reproduce as well but haven't been able to so far.

Good to hear. Hopefully the more thorough debug spew will help get all these fixored.

Also here is some more debug spew from the bogus hitpos error.
http://dynamic.gamespy.com/~freespace/forums/attachment.php?s=&postid=680554

Can you also post an updated build tonight?

[taylor]

Originally posted by redmenace
Well if I am padding my post count, it is better than spamming in hidden and abandoned forums. And more constructive I might add. Honestly though, as I said in the internal, I am testing all of the retail missions. So I figure this is the best way to communicate severe and bothersome crashes in order to have them fixed for the next release

I know I certainly don't mind it.  They are good bug reports which makes it easier to find out what's going on.  I never mind that.

Can you also post an updated build tonight?

I'll try.  I'm working on a audio releated problem now and am testing a pilot file corruption fix.  If I don't feel good enough about a public build tonight then I'll just PM you a link and let you know what to stay away from.  Everything looks good so far though.

[redmenace]

Thats good. As soon as I get done with the BMPMAN error and get you a decent debug spew, I will start working on Open GL so we start making progress towards our next release.

[WMCoolmon]

I went over the jumpnode code; it's pretty simple and doesn't look like there should be any problem with it unless the model fails to load. I made a small change that should help speed things along and keep it from checking a nonexistent model when looking for which jumpnode something is in.

[redmenace]

In "Apocalypse" I got a BMPMAN related error. I am not sure if it is the one you are looking.

[Trivial Psychic]

Originally posted by taylor to redmenace
If I don't feel good enough about a public build tonight then I'll just PM you a link and let you know what to stay away from.