Split: Crash and fullscreen problem

[Harmless]

I can start a new thread if that's desirable.
--------------------------------------------------
Re: crashing (uncommanded program termination without obvious residual)

I have the built-in AMD Radeon graphics in my laptop. The crashes are routine occurrences that are very situation specific and are sometimes repeatable from the most recent save game file. Lending credence to the graphics driver theory, in one case the crash was perfectly repeatable 3 times in a row from the saved game file, but once I restarted my system, the crash was no longer repeatable from the same sequence of events.

The bug reports were sent to [email protected] on 5/21 and 5/23. These submissions followed the Oolite guidance for bug reports, including attached save game files, PC HW/SW configuration (via Windows System Info xml file), the most recent saved game file, latest.log and previous.log, and a description of the issue, along with different sets of steps that could replicate the crash, as well as sets of steps that seemed to interrupt the problem and prevented a crash, all starting from the moment of restoring the attached saved game file.

Because the crashes occur in the station interface, with a habit pattern of saving immediately after each docking, it's just a minor inconvenience.
--------------------------------------------------
Re: time compression

Thanks -- I realize that I was working of old information, which as I recall said that only in full screen mode, while paused, left-right arrows changed rate of time flow. Thanks for the update -- CTRL-F, pause, left-right arrow works great!
--------------------------------------------------
Re: full screen mode
I'll be happy to send a complete bug report with screenshots via [email protected]. Somehow I wound up with the silly idea that I should only report serious bugs, such as the uncommanded program termination/crash bug.
*(after some research) I now realize that the bug report guidelines say to work through GitHub issue tracker for less-critical bugs, so I'll report the full screen issue via GitHub.
*(edit) I just loaded an ran Oolite on a different laptop, in full screen mode, without issue. Seems unique to my own ancient laptop.
--------------------------------------------------
Re: auto-documention
I'll explore Doxyfile/Doxygen and Objective C, thanks for the heads up! It's always good to expand our horizons, up to our assimilation limits
--------------------------------------------------
Again, thanks for the warm welcome and the very responsive follow up!

[another_commander]

I have moved the discussion to Testing and Bug Reports so that we don't clutter the introuctions topic too much.

Thanks for the feedback FWIW, I have not seen any reports from oolite.bug.reports - not sure if anyone from the rest of the team has received something. In any case, I suggest that you report any kind of bug, regardless of its severity, to github. That is where active development takes place. If you have the emails of the report(s) you have already submitted, feel free to open corresponding issue(s) for them and copy the text of those emails there. It's the only sure way to track them.

Regarding crashes, we did fix a very sneaky bug involving loading of saveiles post-1.84. It usually didn't show up, but it was there and it could be that this is what was triggering the crash in your case. Alternatively, try to see if there is an option in the Radeon control panel that resembles threading optimization and disable that if you find it active. If you want to try a binary that fixes the known crash, in case that was indeed the problem, try downloading the latest nightly build from the github nightly builds repository. The current nightly build seems more stable than 1.84 and contains also a bunch of new features that you might like. You can install it and run it in parallel with 1.84. Of course, being nightly build, cutting edge etc etc, some caution is recommended. It is fine and stable today, but it is not guaranteed to continue being so tomorrow, if a new build is committed.

[Cody]

Alternatively, try to see if there is an option in the Radeon control panel that resembles threading optimization and disable that if you find it active.

If memory serves, the Catalyst AI is the nearest thing to it.

[Harmless]

Great, that was the fastest response to a bug report that I've ever seen! Honestly, I should not have imposed the full screen issue on you, since it appears to be unique to my set up. Thanks for the suggestion -- I updated the Radeon driver to 15.200. Also, I downloaded, installed, and have just begun using oolite-trunk-1.85.0.412349f-dev.x64.exe.

Although this hasn't resolved the full screen issue, I have learned a few more things about it.

As you suggested, moving the Oolite window onto the secondary monitor provides a faultless full screen mode on the secondary monitor. Surprisingly, within Windows 7 screen resolution window, setting the secondary monitor as primary (which relegates the laptop to secondary status) also provides a faultless full screen mode on either monitor. I think we can safely say that this one is one of those quirks that is very specific to my setup, and now, thanks to your assistance, I know how to work around the problem locally.
------------------------------------------------------------------
Regarding the crashes under 1.84 x64, since the nightly build may fix the crash problem, I'll make a full report on it only if it crops up again.

So, thanks again for a really active and helpful community and for this great gaming experience!

Very respectfully,
Harmless

[Harmless]

*** Disregard this comment, I'm becoming confused about which version I have running ***

I noticed that the CTRL-F key, which worked great in the 1.84 x64 build, does not seem to have any effect in the latest nightly build.

*(edit) Although CTRL-F doesn't seem to have any effect, CTRL-SHIFT-F shows FPS. Then, when I pause and use left-right arrow keys, it doesn't seem to affect the flow of time.

Very respectfully,
Harmless

[another_commander]

Note that the key combination for FPS display is Shift+F, not Ctrl+f.

Good to know that you have a workaround for the full screen thing. I recall from past reports that Radeons seem to have a generic issue with keeping track of game window coordinate changes, to which the command line parameter -nosplash comes in very handy. If you would like to give it a final try, set the monitor setup to the configuration that shows the problem, then launch the game with the -nosplash parameter. Then try to switch to fullscrenn. My spidersenses tell me it might just work.

[Harmless]

You're spot on, "oolite.exe -nosplash" prevents full screen from crossing multiple screens on my AMD Radeon dual monitor setup!

Thanks!

[ashtiboy]

and it might be that oolite is only trying to use 500MB's of ram instead of the normal 2-4 GBs that all new sistems requre at the mimum for proper GC collection