How well is it working (haven't tested that much yet)?

Moonsword wrote:How well is it working (haven't tested that much yet)?

The pilot file changes or the build in general? The pilot file changes have been well tested now and shouldn't cause any problems. The filenames have changed so they don't show up in older builds that don't know how to read the new formats. Various fixes have gone in that prevent data being transferred between campagins accidentally and to prevent crashes if the actual campaign file can't be found. Although I've found and fixed a couple of bugs or things that I didn't like so much, no bugs have been reported in the last two test releases.

The build itself should work pretty well and the next one will have additional fixes and a couple of speed enhancements. There is the problem with the talking head animations crashing on level exit and I hope that problem is fixed for the most part but I will continue looking for the root cause and why it only happens under Windows.

The upcoming build will have one additional fix for the new pilot code which can crash if it can't find a campaign file in the simroom and then you escape out. It will also have the two bmpman's (one for D3D and one for OGL) merged back into one. No functionality will be lost, it will just have redundant code removed and some speed enhancements and bug fixes I made for OGL will now apply to D3D as well.

That Oct 9th Inferno build isn't well tested so if the Colossus problem doesn't happen with a regular build then it may be a specific build issue. I am getting a bit of deja vu about that though so someone has probably reported a similar firing problem.

taylor wrote:That Oct 9th Inferno build isn't well tested so if the Colossus problem doesn't happen with a regular build then it may be a specific build issue. I am getting a bit of deja vu about that though so someone has probably reported a similar firing problem.

I wasn't using your inferno build when I experienced said problem, I was using the Oct. 7th build, but its not the first one of yours to have problems. I started my current run through the FS2 campaign with your August 22nd build. The first problems I started having were during "The 6th Wonder", when the Colossus wouldn't fire at the Hawkwood. It was a corvette so I couldn't destroy it myself, so I couldn't finish the mission. Now, I restarted the mission and everything went fine. The Hawkwood was destroyed and I went on to the next mission. Then during "Feint! Parry! Riposite", the sequence seemed off and the Medusas woudn't attack the 2 Fenrises. I had to destroy them myself to jumpstart the next phase of the mission. By the time I got the problem in "Their Finest Hour", I had switched over to your Oct. 7th build (Getting some weird problems with the warp vortex in that one).

I hope you can finalize the pilot code soon so it can be introduced into CVS. I'm dying to make full use of Goober's new SEXP enhancements, but I can't do so without a pilot.

Later!

The Trivial Psychic wrote:I wasn't using your inferno build when I experienced said problem, I was using the Oct. 7th build, but its not the first one of yours to have problems.

Oh, ok. You said the Oct 9th build which is the Inferno build. The newest regular build is Oct 7th so it could have been a typo. I'll look into this in more detail. I wish I could remember who said something about the firing problem. I'll go through my IRC and IM logs to see if it was mentioned there. Feel free to file it in Mantis in case one of the other developers has any ideas.

The Trivial Psychic wrote:I had switched over to your Oct. 7th build (Getting some weird problems with the warp vortex in that one).

I've noticed that recently the warp vortex will be flipped around so that the ship seems to enter from the wrong side if that's what you are referring to. I haven't really looked into that one much other than to make sure I didn't do something weird in warpineffect.cpp. If that's not the same problem you're seeing let me know.

The Trivial Psychic wrote:I hope you can finalize the pilot code soon so it can be introduced into CVS. I'm dying to make full use of Goober's new SEXP enhancements, but I can't do so without a pilot.

I had actually planned to have the pilot code in CVS last week. The 0822 build was just to make sure nothing new broke with the pilot code but with the forums down I've just been waiting. If a few people that have used that build can report here whether they have had pilot related problems or not that would be a big help. If everything works then I can go ahead and commit to CVS tomorrow, there's nothing else holding it up.

I noticed something strange in the mission where the colossus first appears....instead of firing at the Hawkwood, it just....sits there! It doesn't fire, it doesn't do ANYTHING! I can't get past that stupid level due to that error

It's possible that Bobboau added his beam bug back in - again - and somehow it stayed around for that build. I thought I eradicated it.

Taylor, I've used your 0822 build since I first installed the SCP, and I haven't experienced any bugs at all related to the pilot files. The only ones I have seen have been the warp effect and the head ANI memory leak, which you mentioned above. (About the head ANI bug: it seems to have changed from the 0822 build to 1007 one. Before, I got stuck on "Enter the Dragon" from the port, and turning off the ANIs wouldn't work. This time, I got stuck on "The Field of Battle," but turning them off did the trick.) Otherwise, it looks great .

Top Gun wrote:About the head ANI bug: it seems to have changed from the 0822 build to 1007 one. Before, I got stuck on "Enter the Dragon" from the port, and turning off the ANIs wouldn't work. This time, I got stuck on "The Field of Battle," but turning them off did the trick.

Yeah I made a couple of changes. It's been crashing when it tries to access data from the ani from within missionmessage.cpp. The data it tried to access was the reference count, so it would know how many times it has to pass through anim_free(). Now it just goes through anim_free() until it's been completely unloaded, there is an error, or if it tries more than 25 times. Not very pretty but I thought it was worth a try. The other change was to make sure that it never tries to free an anim (just like the original code) if they are turned off. As long as you turn the guage off before loading a mission then it won't leak memory.

There is some data corruption somewhere that's causing this. The same code works flawlessly under Linux and OSX but craps out under Windows for some reason. I'll track it down one of these days though.

taylor wrote:

The Trivial Psychic wrote:I had switched over to your Oct. 7th build (Getting some weird problems with the warp vortex in that one).

I've noticed that recently the warp vortex will be flipped around so that the ship seems to enter from the wrong side if that's what you are referring to. I haven't really looked into that one much other than to make sure I didn't do something weird in warpineffect.cpp. If that's not the same problem you're seeing let me know.

I actually posted the problem in that build's original release thread, but HLP went down right away so I doubt you got the chance to see it. Basically, the warp ani is often missing a 90 degree chunk. Sometimes, the warp effect is all there, sometimes said chunk is missing, sometimes its only half there, sometimes its only a quarter there, and somtimes its not there at all. This bug was not present in the 0822 build.

Also, I noticed that 3D warp is no longer default, which Goober has confirmed. I would like to request that future builds include a flag in the launcher to reactivate the 3D warp, instead of having to extract and open up every mission in FRED and click the box to get it to work. The return to the 2D warp may be partially responsable for the above problem of the warp ani.

Speaking of FRED, any word on your attempts to repair FRED's rendering? You should take a look at Goober's SEXP enhancements . They're worth incorporating.

Later!

For the 3D warp thing can't we simply make the game check if the 3D hole is there and use it if it is?

At the moment AFAIK the game checks for warp.pof if it wants a 3D hole and warphole1 for 2D holes.

Can't we simply have it use 3D holes if the extra pof is there?

TP is correct that it would be hugly ennoying to have to extract and resave every mission just to make the 3D warp work again.

Goober5000 wrote:It's possible that Bobboau added his beam bug back in - again - and somehow it stayed around for that build. I thought I eradicated it.

I've barely looked at this but going over a couple of files that I know were updated just before I made that build I noticed something in aicode.cpp version 2.69. In ai_fire_from_turret() it checks dist_to_target to make sure it's not outside the firing range. The change (on or about line 12052) in 2.69 adds a minimum distance as well which is 0 by default. The problem I see with that code is that this dist_to_target can be a negative number. I don't know if that's part of the problem in this case or not but at least some weapons fire is getting cut off because of it.

The Trivial Psychic wrote:Speaking of FRED, any word on your attempts to repair FRED's rendering?

It's better but still has too many positioning problems in HTL mode. I put FRED coding on hold while I try and finish up my current code and get everything in CVS. Everyone keeps asking for the IBX and pilot code as well as some OGL texture fixes so I needed to get all of that finalized. I heard that RandomTiger was starting to work on FRED again but I haven't heard from him directly yet. I think he's a lot more familiar with the code and could probably track down the remaining problems faster than I can.

karajorma wrote:TP is correct that it would be hugly ennoying to have to extract and resave every mission just to make the 3D warp work again.

The problem is that it would be even more hugely annoying for people to compile their own version of the media vp that doesn't include it, if they wanted to disable it.

A command-line thing is easy enough, though. Consider it done, once this latest mess of schoolwork is over.

taylor wrote:That Oct 9th Inferno build isn't well tested so if the Colossus problem doesn't happen with a regular build then it may be a specific build issue. I am getting a bit of deja vu about that though so someone has probably reported a similar firing problem.

Well cool - until today I didn;t know you guys were still making Inferfno builds. TBH, I'm not sure how we haven't hit the ships limit in TI already, so we're going to need to make the switch to an Inferno build pretty soon.

Goober5000 wrote:A command-line thing is easy enough, though. Consider it done, once this latest mess of schoolwork is over.

I think this would be a better use of the SCP options screen actually. I've got this partially done already with glow/spec maps and ambient_factor in there already. I plan to run through the current cmdline options and pull out a few more that would work better in the options screen later today and it would be easy enough to add a warp option as well. I'm using some filler graphics for the interface right now but the code should be done this weekend and then I'll know what new graphics are needed and can start pestering some of the art people.

Black Wolf wrote:Well cool - until today I didn;t know you guys were still making Inferfno builds.

Technically we aren't since it's a pain to maintain two builds and since the Inferno builds didn't work well with other builds or networking. It's not officially supported in other words. In that new Inferno build I made some changes to address various concerns and hopefully make it work in tandem with regular SCP builds. First thing is that networking is disabled, period. The upped table values breaks it anyway but having a nice "This feature is disabled" popup will hopefully help avoid bug reports. The second thing is that pilot files are stored in a different directory (data/players/single/inferno/) since they aren't compatible with normal SCP builds. That Oct 9th build contains everything that the regular Oct 7th build contains: IBX code, new pilot formats and all of the various bug fixes.

Can I ask a question on one of those features:

What is the IBX code and what does it do?