Crash after entering in orbit

[GoogleCodeExporter]

What steps will reproduce the problem?
1. Load save game
2. Leave planet, go to next planet II on same solar system (Beta Giclas)
3. It's a water world. Select Scan
4. Either wait a few seconds or select autoscan and then wait

Attached: capture, log, save game.

Original issue reported on code.google.com by Awakened...@gmail.com on 8 Dec 2012 at 4:56

Attachments:

• starcon2.08

[GoogleCodeExporter]

I confirm that the crash happens even if you do nothing (not even a scan).
It may be related to the display of the planet rotation (it crashes always at 
the same point in the rotation).
Attached save game right before you get to the bugged planet, to save time for 
the developer. 

Original comment by Awakened...@gmail.com on 8 Dec 2012 at 4:59

Attachments:

• starcon2.09

[GoogleCodeExporter]

Here is another one that crashes upon rotation.
I notice that it happens when the planet is rotating horizontally from right to 
left?
Alpha Giclas IV

Original comment by Awakened...@gmail.com on 8 Dec 2012 at 5:18

Attachments:

• starcon2.11

[GoogleCodeExporter]

[deleted comment]

[GoogleCodeExporter]

Alpha Giclas VII-C (yellow) crashes also, same save game as above

Original comment by Awakened...@gmail.com on 8 Dec 2012 at 5:33

[GoogleCodeExporter]

Here is another buggy planet, same problem as the others: Alpha Ptolemae IV
Load attached save game

Original comment by Awakened...@gmail.com on 8 Dec 2012 at 7:57

Attachments:

• starcon2.18

[GoogleCodeExporter]

I'll try to take a swing at this one, too.

Original comment by Jaakko.M...@gmail.com on 8 Dec 2012 at 2:23

[GoogleCodeExporter]

Revision 608 I just uploaded might have solved this. Does it work now?

Original comment by Jaakko.M...@gmail.com on 8 Dec 2012 at 3:51

[GoogleCodeExporter]

I am now using Revision 610 and it's still crashing at the exact same places

Original comment by Awakened...@gmail.com on 8 Dec 2012 at 8:09

[GoogleCodeExporter]

Issue 10 has been merged into this issue.

Original comment by Jaakko.M...@gmail.com on 8 Dec 2012 at 9:10

[GoogleCodeExporter]

Issue 12 has been merged into this issue.

Original comment by Jaakko.M...@gmail.com on 8 Dec 2012 at 9:10

[GoogleCodeExporter]

Same crash at Alpha Giclas VII-C (yellow moon). 
Tilt of the planet: 6 degrees. Also seems to be rotating horizontally. Seems to 
be the common elements of all crash. See attached save game, right next to 
crashing planet.

Original comment by Awakened...@gmail.com on 8 Dec 2012 at 10:06

Attachments:

• starcon2.13

[GoogleCodeExporter]

Latest interesting remark:
so far, all my planet crashing happened for planet with a tilt of 6degres in 
the scan statistics. Something related to the math calculation probably ends up 
with a division by zero or a fail approximation that cause some floating point 
problem... something.
I'll keep an eye to see if future crashes continue to be for 6 degree planets.

Original comment by Awakened...@gmail.com on 8 Dec 2012 at 10:22

[GoogleCodeExporter]

Thanks for the detailed info, that's great!

This indeed is a strange one. It seems that (at least) at these planets, the 
code tries to access the planet graphic data array beyond the array's limits.

You see, the planet is laid out in an [x*y] sized array, each array elementing 
being a pixel of the planet. For some reason, on these problematic planets the 
code wants to access x*(y+1) elements, running past the boundary. 

The fix I've made just reserves some extra memory for each plane, giving them 
each x*(y+1) sized chunk of memory, thus preventing these problem planets from 
reading/writing into wrong place. Hope you get the newest .exe soon so you can 
test this fix too.

Original comment by Jaakko.M...@gmail.com on 8 Dec 2012 at 10:24

[GoogleCodeExporter]

Great!  Reminds me of my good old days of programming. Still do quite a bit of 
scripting in my job in VB from time to time, but 95% of my work now is in 
system analysis and business analysis... ;-)
Have you checked the first element (y=0) for these problematic planets? If the 
softwares attempts to access array y+1, it may be because it's starting off by 
one and the first element may be skipped or empty. 
Thanks for working on it!

Original comment by Awakened...@gmail.com on 8 Dec 2012 at 10:30

[GoogleCodeExporter]

Bug is fixed on revision 619 YAY!!! I had to test using the debug exec rather 
than the normal exec however. I don't know why it worked this time; last time I 
used the debug exec it was asking for msvcr110d.dll which I couldn't find 
anywhere on the web. This time it asked me for msvcr100d.dll, a file I had 
already found and downloaded, so this time it worked...

Original comment by Awakened...@gmail.com on 9 Dec 2012 at 4:27

[GoogleCodeExporter]

Hmm... strange stuff :-s But most important thing is that you got the debug 
exec working! I'll close this happily resolved issue.

Original comment by Jaakko.M...@gmail.com on 9 Dec 2012 at 5:44

• Changed state: Fixed

[GoogleCodeExporter]

Confirmed solved on revision 625

Original comment by Awakened...@gmail.com on 9 Dec 2012 at 8:37

• Changed state: Verified