Page 6 of 12

Posted: Sun Feb 15, 2009 6:54 pm
by Thargoid
Can you try running it through a fuel station and see if you get the crunch too?

Posted: Sun Feb 15, 2009 6:55 pm
by JensAyton
gorans wrote:
And this is a first compiled oolite in a long time that didn't crash on me on load player or launch from station.

Code: Select all

Exception Type:  EXC_BAD_ACCESS (SIGSEGV)
Exception Codes: KERN_INVALID_ADDRESS at 0x00000000434f4d6d
Crashed Thread:  0

Thread 0 Crashed:
0   libobjc.A.dylib               	0xfffeff18 objc_msgSend_rtp + 24
...
27  ...rcmoini.smartscrollx_sxobjc	0x007e7d7c SmartScroll_objc + 16120
28  com.apple.AppKit              	0x91f07d98 -[NSApplication run] + 736
29  com.apple.AppKit              	0x91ed879c NSApplicationMain + 440
30  org.aegidian.oolite           	0x00003904 main + 60 (main.m:13)
31  org.aegidian.oolite           	0x000022d0 _start + 336 (crt.c:272)
32  org.aegidian.oolite           	0x00002178 start + 56

CRASH to desktop...
This crash is in third-party software. I suggest you send that crash report to Marc Moini.

Code: Select all

Exception Type:  EXC_BAD_ACCESS (SIGBUS)
Exception Codes: KERN_PROTECTION_FAILURE at 0x0000000008741000
Crashed Thread:  Unknown

Error Formulating Crash Report:
*** -[NSCFDictionary setObject:forKey:]: attempt to insert nil value (key: VMUSignaturePath)
0x91bfddc0
0x91d1e4ec
0x91bfdcd0
0x91bfdd08
0x94b31554
0x00071c00
0x000804d8
0x00080614
0x00003004
0x00009ad4
0x0000b84c
0x0000b63c
0x95fdccc8
0x0000ad58
0x95faf028

Backtrace not available
Fascinating. A quick google suggests that one’s a crash in the crash reporter itself. That makes the crash report itself a bit useless, though. :-/

Posted: Sun Feb 15, 2009 6:59 pm
by goran
another_commander wrote:
I just tested with the Griff Boa and I can tell you that the turrets work with SVN 2065. Remember: Turrets need a) a target set, b) the target to be hostile, c) the target to be in range.
I KNOW this is the wrong thread but since I have no reputation to trash... ;)

Image
2065 trunk

Image
1.72.2

Maybe there's something wrong with Caduceus turret? I'm good only in copy/paste so I can't tell.
This crash is in third-party software. I suggest you send that crash report to Marc Moini.
Ohh... I have plenty of reports wihout steermouse:

Code: Select all

Exception Type:  EXC_BAD_ACCESS (SIGSEGV)
Exception Codes: KERN_INVALID_ADDRESS at 0x00000000281578bc
Crashed Thread:  0

Application Specific Information:
[script.load.world.listAll]: Loaded 54 world scripts: "ams_system" 1.00, "behemoth" 2.5, "behemoth-spacewar"....

Thread 0 Crashed:
0   com.apple.CoreFoundation      	0x953c8648 __removeHandler2 + 400
1   com.apple.CoreFoundation      	0x953c8c4c _CFDoExceptionOperation + 1284
2   org.aegidian.oolite           	0x00023fcc -[Universe newShipWithName:] + 368 (Universe.m:2940)
3   org.aegidian.oolite           	0x00023db0 -[Universe newShipWithRole:] + 64 (Universe.m:2898)
4   org.aegidian.oolite           	0x0002e690 -[Universe populateSpaceFromHyperPoint:toPlanetPosition:andSunPosition:] + 2800 (Universe.m:1377)
5   org.aegidian.oolite           	0x0002d7b0 -[Universe setUpSpace] + 3116 (Universe.m:1064)
6   org.aegidian.oolite           	0x0003aef0 -[Universe initWithGameView:] + 2612 (Universe.m:287)
7   org.aegidian.oolite           	0x0000dbf8 -[GameController applicationDidFinishLaunching:] + 500 (GameController.m:228)
8   com.apple.Foundation          	0x928f4114 _nsnote_callback + 372
9   com.apple.CoreFoundation      	0x9533c7c8 _CFXNotificationPostNotification + 920
10  com.apple.Foundation          	0x928f17bc -[NSNotificationCenter postNotificationName:object:userInfo:] + 88
11  com.apple.AppKit              	0x9355ed34 -[NSApplication _postDidFinishNotification] + 108
12  com.apple.AppKit              	0x9355ec4c -[NSApplication _sendFinishLaunchingNotification] + 80
13  com.apple.AppKit              	0x934e65f0 -[NSApplication(NSAppleEventHandling) _handleAEOpen:] + 260
14  com.apple.AppKit              	0x934e5e28 -[NSApplication(NSAppleEventHandling) _handleCoreEvent:withReplyEvent:] + 88
15  com.apple.Foundation          	0x92914988 -[NSAppleEventManager dispatchRawAppleEvent:withRawReply:handlerRefCon:] + 480
16  com.apple.Foundation          	0x9291475c _NSAppleEventManagerGenericHandler + 236
17  com.apple.AE                  	0x90d24ce0 aeDispatchAppleEvent(AEDesc const*, AEDesc*, unsigned long, unsigned char*) + 164
18  com.apple.AE                  	0x90d24be8 dispatchEventAndSendReply(AEDesc const*, AEDesc*) + 40
19  com.apple.AE                  	0x90d249ec aeProcessAppleEvent + 212
20  com.apple.HIToolbox           	0x9564b4ac AEProcessAppleEvent + 52
21  com.apple.AppKit              	0x934e3954 _DPSNextEvent + 1156
22  com.apple.AppKit              	0x934e30dc -[NSApplication nextEventMatchingMask:untilDate:inMode:dequeue:] + 112
23  com.apple.AppKit              	0x934dcd98 -[NSApplication run] + 736
24  com.apple.AppKit              	0x934ad79c NSApplicationMain + 440
25  org.aegidian.oolite           	0x00002190 _start + 336 (crt.c:272)
26  org.aegidian.oolite           	0x00002038 start + 56

Posted: Sun Feb 15, 2009 7:36 pm
by Screet
another_commander wrote:
I just tested with the Griff Boa and I can tell you that the turrets work with SVN 2065. Remember: Turrets need a) a target set, b) the target to be hostile, c) the target to be in range.
...and sometimes the turrets do fire, but when I look from the outside, it appears as if they don't. Maybe that's also reason for a good amount of "turrets don't work" reports like mine in the past! I noticed that the turrets do work even if they don't fire visibly is that the ship I targeted was destroyed and I got the bounty. I don't know why I sometimes could not see shots - maybe I would have seen them further in the distance, but that the viewpoint and angles to the enemy just didn't work out well in those cases.

However...this also leaves the question of what did change between 1.72.2 and SVN2065 that such a bug was removed, possibly without even noticing that a bug has been removed and when the next test release could be expected, as I'm surely not the only one who uses ships with turrets. At least, I have all tools on this machine and just would need to use SVN and spend some time to get everything done. I also ask because I once took that test version that was made for ATI users...and switched back to the original version when my logfile was spammed by deprecated-warnings.

Screet

Posted: Sun Feb 15, 2009 11:39 pm
by JensAyton
Screet wrote:
...and switched back to the original version when my logfile was spammed by deprecated-warnings.
You can turn those off using logcontrol.plist, y’know. :-)

Posted: Sun Feb 15, 2009 11:51 pm
by Screet
Ahruman wrote:
Screet wrote:
...and switched back to the original version when my logfile was spammed by deprecated-warnings.
You can turn those off using logcontrol.plist, y’know. :-)
I didn't have a look at that...I knew that some part is configurable, but I thought: "Hell, did they change everything? Will it still run OK or should I wait until they fixed all OXPs for that release?". Maybe I was too afraid, but then I knew that the 1.72.2 runs...and that it's just a graphics bug I've been playing with since the beginning. The 9.1 drivers fix the garbled display problems very well, but not always, and it still reports 0MB of texture memory for OpenGL, slowing things down.

In the AMD forum someone said that there are multiple ways to request texture buffer size, and that things would run if the only way would be used that is implemented correctly by the drivers. Someone else suggested buying a new motherboard with new RAM in order to provide fast enough memory to the graphics card that way...and I've got a 4870x2 with 2GiB...ARGH!

I wonder how long it takes until I get all those problems solved that result from the malfunctioning ATI drivers. Since there's currently no OpenGL acceleration, I threatened them with german law...if all germans would demand their money back and sue ATI for advertising a missing property, it would be surely more costly for them than to fix their drivers :twisted:

Screet

Posted: Mon Feb 16, 2009 8:38 am
by Eric Walch
Thargoid wrote:
Sorry, I meant that the boundary box of the caddy seems (at least visibly) to fit inside the hole in the station.
When you look in the code the first level of collision detection compares collision radii. So it is in fact a sphere and no box. And the sphere peeks out of the box in the mid of the faces and cuts its corners of.
e.g. with the buoy repair station the bounding box has a size of 3000 meters. When flying along the sides, 100 meter outside, my ships very often collided near the centre of the face were the sphere comes out the most. I now let them fly at 200 meters outside the blue representation of the box and every thing is fine.
So the box is only a rough representation of the size. Also with rotated substructures, the box around that substructure becomes very large.

Posted: Mon Feb 16, 2009 7:12 pm
by Thargoid
As some people are having problems, I've updated the OXP to introduce also a fuel satellite. This appears in much the same place as the fuel station might (at a 1 in 3 chance of replacing a station), and works by simple proximity (within about 100m of the satellite, just don't crash into it or try and dock with it) rather than fly-through.

It's available via Box.Net from the same URL as normal (you can get it here).

It's tested with a Caduceus, and works fine. Doesn't work with very big ships (cos they're too big), but if you can afford one of them then you can afford to go to the main station ;)

And I've done some more testing with the fuel station. It is actually possible to fly through with an unmodified Caddy. But you have to be dead-centre and minimum speed (single bar speed, player.ship.speed=6). Anything higher you go splat, if off-line you bounce.

Tried it with some debugging going, will sort through the logs soon.

Posted: Mon Feb 16, 2009 7:19 pm
by Screet
Thargoid wrote:
As some people are having problems, I've updated the OXP to introduce also a fuel satellite.
...
Tried it with some debugging going, will sort through the logs soon.
Thank you very much!!!

Screet

Posted: Mon Feb 16, 2009 7:56 pm
by Thargoid
Sad to report that enabling the octatree's as per Ahruman's console command doesn't make anything visibly collide, there's still plenty of gap.

And nothing at all appears in the logs :(

So still no further idea as to what turrets and large entities are doing to each other, so unless someone clever can look into it I'd suggest to use the new satellites instead ;)

Posted: Thu Feb 19, 2009 10:39 pm
by Thargoid
It's been brought to my attention (thanks Lestradae) that in a very specific set of circumstances (witchspace jumping into a system without a fuel satellite from one with one when you're near it) causes a system crash.

So I've implemented a fix and uploaded it to Box. You can grab Fuel Station v1.11 here (same URL as normal).

Posted: Tue May 12, 2009 9:32 pm
by DaddyHoggy
Thargoid - you are the wizard of nifty gadgets and I've been thinking about the Fuel Stations.

We've got a lot of Fuel-like companies in YAH (Texacoo, citi-goo etc) could an additional gadget be fitted in the hold of the ship - a lean burn injector system - that burns witchfuel at a slightly reduced rate when you hit the injectors - perhaps even giving you a slightly increased rate of speed and/or acceleration over the standard fuel (like Shell's Optimax claims).

Filling up with normal fuel at Stations or sunskimming does not activate the system - it must be the special fuel bought from the Fuel Stations - perhaps the Stations could also be sub-sponsored by some of the Racing Teams?

A reason (other than convenience) to visit these wonderful structures.

Just trying to add a little potential back-story richness to the Ooniverse - feel free to dismiss as you see fit.

Posted: Tue May 12, 2009 9:36 pm
by Thargoid
You could potentially have it burn less by occasionally adding 0.1ly of fuel periodically when you're running under injectors (speed above max speed). The required variable set by the fuel station, and unset when you buy fuel elsewhere (via the equipment screen).

Would be an intricate script to code up, but it should be possible to do with some thought. Speed & acceleration won't be possible as those are read-only parameters to JS.

But not for for me for the moment, as I'm busy with one or two other projects (Aquatics v2.0, coming RSN).

Posted: Tue May 12, 2009 10:45 pm
by ClymAngus
I've done some redesigning the caddy should go through a little better now.
www.crimsonforge.co.uk/cloister/neocaduceus.zip

Smaller, sleeker better gun accuracy. You know all the things that titted people off with the original.

Posted: Wed May 13, 2009 10:34 am
by Screet
ClymAngus wrote:
I've done some redesigning the caddy should go through a little better now.
Smaller, sleeker better gun accuracy. You know all the things that titted people off with the original.
...and even further improved look ;)

Screet