[chief1983]

If data doesn't make it over to the new pilot file during the conversion, that's not corruption.  That's just a broken pilot converter.  That converter hasn't been tested as much as other parts of the pilot code, so I'm not surprised it's not working flawlessly yet.

Just to clarify, when you run an Ant8 build, it converts your old player files to a new file, not modifying the original file (or it should not anyway).  That's why using a non-Ant8 build should still work fine, and not have noticed any changes made by the Ant8 builds.

If after the conversion, whatever made it over still has other problems after using just an Antipodes 8 build, like things still won't save, or settings disappear, or the game crashes because of the new pilot file, then yeah, that's probably corruption.

[Zacam]

No Antipodes specific bug-fixes this time around yet, but now updated to match to Trunk 6945 since 6912 was getting a little long of tooth.

[Talon 1024]

I'm still getting the same bug that I was a while back with the GTC Duke having four secondary banks.  I have no idea what is causing it, but I've corrected my modified setup like General Battuta asked me to a while ago when I first experienced this bug, and that didn't help.

I do have a feeling that it's something to do with the persistent variables, like Karajorma pointed out a while ago.

[attachment deleted by ninja]

[General Battuta]

Buuuh that's weird. What's up with the normal and .3612.vp versions of your MVP files? Is it supposed to be like that?

[Talon 1024]

It is supposed to be like that.  See reply #1 in the 3.6.12 MediaVPs release thread.  It's an update to the 3.6.12 MediaVPs.

[General Battuta]

Good to know. I got nothin'. 

[Parias]

Heh. Co-incidentally I just started a fresh Blue Planet: AoA playthrough and I think I'm hitting exactly the same problem as Talon:

Code: [Select]

(...etc)
Got event GS_EVENT_ENTER_GAME (2) in state GS_STATE_RED_ALERT (44)
Entering game at time =  22.417
ASSERTION: "pm->missile_banks != NULL" at object.cpp:783
 Ship GTC Hyperion has 4 secondary banks, but no missile banks could be found.
Full log attached. I was actually doing a playthrough of AoA with the stock 3.6.12 "Final" build earlier today when I hit an unrelated crash a bit later on in the campaign. I figured my pilot file was hosed since it got the point where I couldn't even select him from the barracks (game would crash immediately) so I just blew him away completely, decided to grab this A8 r6946 build on a whim to check it out, and was rolling along in AoA nicely until I hit the Red Alert mission for the GTC Duke "Ross 128" segment and now I can't advance. Only bringing it up here since Talon made mention of exactly the same symptom too, so either we both have the same underlying configuration problem or something's buggered.

[attachment deleted by ninja]

[karajorma]

I do have a feeling that it's something to do with the persistent variables, like Karajorma pointed out a while ago.[/color]

I don't think it is. I fixed that bug.

This is some kind of red alert bug. The whole red alert system really needs to be refactored to prevent bugs like this one.

[The E]

Okay, let's bring a bit of life back into this. Here's a build based off of the current antipodes revision (7090). It has a bunch of error checking and fault recovery added to it that should reduce crashyness. Please test this.

http://blueplanet.fsmods.net/E/fs2_open_3_6_13d_SSE2_Antipodes.7z
http://blueplanet.fsmods.net/E/fs2_open_3_6_13d_Antipodes.7z

[SypheDMar]

After testing build with no mod and WiH  back to campaign switch to mediavps, it seems to work as intended. Noticed that it defaults to multiplayer instead of single. No crashed occurred and campaigns seem to remember where it last left off. Didn't do extensive testing nor did I test multi, but I did go to the multi screen and nothing bad happened there, either.

[Zacam]

Changelog Post for Antipodes Build r7108

Commit gaps can be safely assumed as Trunk sync's or code sync's against trunk. Listed here are changes specifically to the code base for the new Pilot code or related issues pertaining to it.

Code: [Select]

------------------------------------------------------------------------
r7018 | The_E | 2011-02-18 07:22:28 -0800 (Mon, 18 Feb 2011) | 1 line
Changed paths:
   M /branches/antipodes/code/bmpman/bmpman.cpp
   M /branches/antipodes/code/graphics/gropengltexture.cpp
   M /branches/antipodes/code/playerman/managepilot.cpp

Fix for bug encountered in BP, From now on "0" is no longer a valid handle for bmpman slots
------------------------------------------------------------------------
r7033 | The_E | 2011-02-26 00:53:09 -0800 (Mon, 26 Feb 2011) | 1 line
Changed paths:
   M /branches/antipodes/code/fred2/campaigntreewnd.cpp
   M /branches/antipodes/code/fred2/fredview.cpp
   M /branches/antipodes/code/fred2/messageeditordlg.cpp
   M /branches/antipodes/code/fred2/missionsave.cpp
   M /branches/antipodes/code/misison/missionparse.cpp

This will allow for longer sexps.
------------------------------------------------------------------------
r7034 | The_E | 2011-02-27 12:06:44 -0800 (Mon, 27 Feb 2011) | 1 line
Changed paths:
   M /branches/antipodes/code/pilotfile/csg.cpp

Adding a few CLAMPs and log warnings to ensure validity
------------------------------------------------------------------------
r7062 | The_E | 2011-03-13 01:07:43 -0800 (Mon, 13 Mar 2011) | 1 line
Changed paths:
   M /branches/antipodes/code/bmpman/bmpman.cpp

A few changes to bm_get_next_handle().
------------------------------------------------------------------------
r7076 | The_E | 2011-03-18 13:24:07 -0800 (Wen, 18 Mar 2011) | 1 line
Changed paths:
   M /branches/antipodes/code/bmpman/bmpman.cpp

Fix for bug noted by Zacam. Note to self: read the code before toying with forces you don't understand.
------------------------------------------------------------------------
r7083 | The_E | 2011-03-20 11:00:18 -0800 (Thu, 20 Mar 2011) | 1 line
Changed paths:
   M /branches/antipodes/code/pilotfile/csg.cpp

More guarding against reading invalid values.
------------------------------------------------------------------------
r7091 | The_E | 2011-03-26 06:29:35 -0800 (Thu, 26 Mar 2011) | 1 line
Changed paths:
   M /branches/antipodes/code/misison/missioncampaign.cpp
   M /branches/antipodes/code/playerman/managepilot.cpp
   M /branches/antipodes/code/playerman/managepilot.h

Minor patch to address issue with custom mainhalls not loading for fresh pilots. I hate this one.
------------------------------------------------------------------------
r7094 | Zacam | 2011-03-29 06:54:07 -0800 (Thu, 29 Mar 2011) | 4 lines
Changed paths:
   M /branches/antipodes/code/Makefile.am
   M /branches/antipodes/code/bmpman/bmpman.cpp

Reported by rsaxvc:
<limits.h> in bmpman (Probably due to MAX_INT being used)
Pull the "demo" dir out of the makefile
http://www.hard-light.net/forums/index.php?topic=75332.0
------------------------------------------------------------------------
r7098 | Zacam | 2011-03-29 12:49:05 -0800 (Thu, 29 Mar 2011) | 1 line
Changed paths:
   M /branches/antipodes/code/bmpman/bmpman.cpp

Removing platform #ifdef from the limits.h include as un-necessary.
------------------------------------------------------------------------



[General Battuta]

SEXP LENGTH BOOST GET

  ;7those are eyebrows of excitement

[mjn.mixael]

The latest Ant8 build I have won't load the next mission if its an 'Incoming Transmission' (continued) mission. I was playing with FSU testing, but also tried on Port/mediavps3.2 which I know worked before. Here's a log.

http://pastebin.com/4QGxADRN

This was with Port 3.2 (pretty sure the important part is the same)

http://pastebin.com/e9FFCmtm

[Pyro3d]

I was able to fix that problem in the attached .diff, mjn.mixael. Is there any reason that the red alert code is saving ship statuses from all (?) previous red alert missions?


[attachment deleted by ninja]

[Zacam]

Your diff (comment attributed to you) was committed to the Antipodes Branch on r7191.

Sadly, I just could not get the reported situation to manifest at all. We'll have some more testing done of course.

[chief1983]

I was wondering about the first part of that diff, as it didn't seem to relate to the second part.  Were those two different bug fixes or something?

[Pyro3d]

The first part of the diff was the fix to mjn.mixael's (and my) problem.

The second part was fixing a assertion that I found upon running the fixed build. It looks like it occurred when a ship object had old data loaded into before the function got to the new data, causing an assertion in the change_ship_class() function to fail when one of the subsystem hitpoints exceeded the expected maximum hitpoints of the subsystem (I hope that makes sense.) As said in the comment, nuking old red alert data before loading new data was probably not the best solution to it, but it did clear out the old data and prevented the assertion from asserting itself.

EDIT:
Added fix for mantis 2425

[attachment deleted by ninja]

[SypheDMar]

This is probably really annoying given that it's a campaign-specific bug, but it might affect others too.

Using this build (didn't try latest nightlies because testing is extremely hard), when I tried to order a Pegasus wing to disarm a Fenris cruiser, they fail to disarm correctly for minutes, so I ignored the cruiser. This happened on Windmills. When I tried it with 3.6.12, it disaled the cruiser within seconds.

[chief1983]

Well nightlies and this are completely different.  The new pilot code has specifically had a few campaign-related bugs I think.  But knowing if it happens in trunk would be critical to fixing it.

[Zacam]

I'll test that before I upload updated builds. Right now I'm trying to resolve some issues that affect both Antipodes and Trunk before providing updated builds for Antipodes.

Updates -are- on the way though, and it will include both of pyro3d's patches.