Oolite Bulletins

I am trying to "well, just 'print' the clouds on the terrain, both are generated at the same time anyway". Currently, the result looks more like forests than clouds. I guess I have to do something about the normals, and then do some manual alpha blending. Before I put more effort into this: is it bound to fail because of something I haven't realised yet?

@Smivs
Would it be possible to share the result of the following command, please?
(replace <Oolite Home Folder> as per your installation)

$ md5sum <Oolite Home Folder>/oolite.app/Resources/Textures/oolite_asteroid_diffuse.png

_________________
Xubuntu 14.04LTS / Windows 7 Home Premium
Intel Core 2 Quad Q9450 2.66GHz - 6GB DDR2 - Palit GeForce GTX 750 Ti StormX Dual (2048MB DDR5)
"Any sufficiently advanced information is indistinguishable from noise." [Newman, Lachmann, Moore]

Sure, here you go.

$ md5sum GNUstep/Applications/Oolite-trunk/oolite.app/Resources/Textures/oolite_asteroid_diffuse.png
5ab4d87354ade6106d33a0b89c24689d  GNUstep/Applications/Oolite-trunk/oolite.app/Resources/Textures/oolite_asteroid_diffuse.png

This image is the 'OK' one copied from my Oolite 1.84 install - the troublesome original was over-written when I C&P'd this one in.
I wonder if the image just corrupted during the download.

_________________
Commander Smivs, the friendliest Gourd this side of Riedquat.

@Smivs
That was just my OCD. I confirm that this is the correct png.

_________________
Xubuntu 14.04LTS / Windows 7 Home Premium
Intel Core 2 Quad Q9450 2.66GHz - 6GB DDR2 - Palit GeForce GTX 750 Ti StormX Dual (2048MB DDR5)
"Any sufficiently advanced information is indistinguishable from noise." [Newman, Lachmann, Moore]

That's OK. I'm always grateful for the help I get here.
TBH I should have dug a bit deeper before I posted the first report. Normally with texture problems I would check the image itself and the naming etc, and had I done so I would have noticed the image was suspect as the thumb wouldn't display. Unfortunately the log entry suggested it was more of a system software problem, so I thought it worth mentioning.

_________________
Commander Smivs, the friendliest Gourd this side of Riedquat.

Oolite crashed while loading a commander - does this lot mean anything?

Problem signature:
  Problem Event Name:	APPCRASH
  Application Name:	oolite.exe
  Application Version:	0.0.0.0
  Application Timestamp:	58a7c450
  Fault Module Name:	ntdll.dll
  Fault Module Version:	6.1.7601.17514
  Fault Module Timestamp:	4ce7c8f9
  Exception Code:	c0000005
  Exception Offset:	0000000000054725
  OS Version:	6.1.7601.2.1.0.256.48
  Locale ID:	2057
  Additional Information 1:	894d
  Additional Information 2:	894d28ee7dfa51df3dffcaf4e428a1bd
  Additional Information 3:	830a
  Additional Information 4:	830ad24cc270b772f1f88abfaa68d787

_________________
The only good fnord is a dead fnord!

The only thing that this tells me is that you are running Windows 7 SP1. The million dollar question here is, is this reproducible? If you load that commander again, does it crash? Try it a few times and Shift-startup on some of them. If you can get it to crash reproducibly please let us know. I have a suspicion of what it might be (hint: issue #204 related) but can't be sure without any more information. If you can get it to crash again, I could send you an executable with a potential fix to test.

Also, was there anything about any exceptions in the log when it crashed?

Oh well - it looked informative.

No, not reproducible - it just happens every so often. And nothing in the log, I'm afraid.

_________________
The only good fnord is a dead fnord!

How often would that be? Do you want me to send you the exe with the workaround we currently have for issue 204 just in case it might be what is hitting you?

How often? <scratches head> I can't really say - don't keep notes. I'm always willing to test a workaround exe though.

_________________
The only good fnord is a dead fnord!

Here is the 64-bit version of the current trunk with issue 204 workaround applied. Feel free to test and let us know if it crashes at any point. I understand it's difficult to say if this works for you when there is no established pattern of frequency of crashes, but try for as long as you can. Let's hope for the best.

https://drive.google.com/file/d/0BwG6R5 ... sp=sharing

Thanks! <grumbles - downloads WinZip - grumbles>

_________________
The only good fnord is a dead fnord!

Which version of the workaround is that? We went through quite a few!

I'd use the one in branch issue_204_v3 - are you using the same one?

I used the retain workaround - was the simplest and fastest to compile and zip up. All proposals we have there don't go down to the root of the problem and just kind of dance around it IMO, so we really need to come up with an actual fix. Until then, any of the workarounds would do for the purposes of finding out whether the bug we are seeing here is this issue or not.

<grumbles> Can't use that without registering - stuff 'em! I'll stick with the current .exe and see if I can force a crash.

_________________
The only good fnord is a dead fnord!