Complaint about stability

[10053r]

I am complaining here so that the devs will see it. Apologies if this is off topic for the development section. However, I am getting very frustrated about the stability issues in spring. It seems that more than half of my games end in crashes of one sort or another. I have no specific bugs to report, because diagnosing this type of random crashing is very difficult, and I can't figure out if anything happening in the game is causing it.

Basically, if a game goes on for more than 30 minutes, chances are pretty good it will end in a crash. Sometimes everything goes in a straight line through other objects until either I quit after a minute or so, or it quits by itself. Sometimes I actually get the crash reporter. Either way, all I want is to finish a game! Please? I can't code spring myself, so I'm dependant on you guys to step up and fix these bugs.

[Masse]

i havent crashed like in months...

[Min3mat]

same

[jcnossen]

As you can understand, I can't really do anything about "random crashes", but just be sure to send in the crashreport when you can.
They really work well for debugging.

[SwiftSpear]

I've gotten the blue screen crash a few times in really long games, and I've ping out crashed before (usually where units just drift off is a ping out crash) but I can't say crashing has been a major problem for a while now.

[10053r]

Interesting. Do you guys play long games on big maps? I typically play AA on maps of at least 15 x 15.

And I understand, Zaphod. I'm mostly just venting...

[PauloMorfeo]

I've been having many issues too.
One of them, just «disapeared» literally, without change of version, lobby server or reinstall. Just stopped happening...

Now, i have other issues. But they seem to have «disapeared» a few days ago, too.

Funny thing how something like software, which should always work or never work (have at all times similar behavior), has «moods». I think i'll convert to the software religion and learn new prayers.

[AF]

The innards of spring arent correct. This can most blatantly be seen by how ti effects the AI itnerface, and how some people get lots of bugs while some peopl dont, and they all seem somewhat random.

[Tired]

Playing only under hosts that you know as reliable might help somewhat - .info and watch their cpus and synch. Maybe won't help you, but usually doesn't hurt.

[Betalord]

These "random" crashes are probably not caused by spring itself. It could be your RAM, graphic card, mother board, pretty much anything. Could also be by overheating (usually cpu or graphic card). Anyway, some bugs have been fixed with next version, perhaps your problem will be gone with it.

[AlienDNA]

Indeed, Betalord. I have been playing with one friend, for 5 long hours on a 20x20 map, trying to destroy eachothers mega-defence, attacking with enormous amount of units, and there was no crashing. Another friend of mine did have stability problems, but fixed them by doing a complete reinstall. Then there are indeed hardware related problems, like bad memory or corrupted clusters. Not to mention the mind-numbing misery video-drivers can cause. Sometimes it can even be a combination, a videocard not working right with a certain motherboard, or processor.

TA spring has been running VERY stable here for quite a long time here. So I wouldn't just blame the devs for stability-issues, nor spring itself, for the problem might be your machine just as well.

[AF]

You should all check your system settings, to see if your computer displays bluescreens of death, ro if it simply reboots when they happen so you never see them.

Very important to read and google for things written on blue screens of death, and fi you fall into the group of people whose machines come with bluescreen of deaths set to reboot the instant they occur with no logging fo them or messages, then you need to change that.

For all yuo know it's because you've got somethign plugged itno your usb hub it doesnt like that uses a faulty driver...

Causes of blue screens of death:
Bad RAM
Bad drivers
Bad programs
A computer that's gonna go bang in the next few weeks.
Heavy static damage

BLSOD only occur when soemthing very bad happens at the core fo widnows that should never happen on a perfectly good computer.

[Caydr]

Spring has never crashed for me (I've only been playing since 65b2 tho). Seems like maybe it works a lot better on certain hardware or something.

Irritating when people send me bug reports - AA is unstable, etc. Well hey, guess what, AA never crashes OTA. Never. Like, ever. Period. Never-never-never. I'm very proud of the stability I've achieved. If Spring crashes while AA is running, it's coincidental and would've happened even if you were playing XTA.

I really wish stability could be improved though, for these other people. Although I've never had a crash, I know in average at least 1 player and up to a quarter of the players in a large game will crash out.

[AF]

I hate how the engine doesnt give NTAI the correct information sometimes, or how it makes a single instance of NTAI manage all the players it's been assigned to yet only be able to give commands and get info from one of those players.

And how it sometimes discards commands as impossible but executes them at other times in the same situation.

[Dwarden]

spring is often crashing when some unit dies or there is way too many explosions etc ... very often not where i'm looking or can see at all ... i keep submitting crashlogs in that cases ...

[el_muchacho]

I guess there may still have bugs in the memory management as there is no reference counter nor any sort of automatic memory management scheme AFAIK, so it's fairly complex (i.e much more than the average OO program), with lots of small object allocations. Using such a scheme probably increases the stability, but there is always a non-negligible performance penalty. As it isn it's not impossible that some obscure cases arise from time to time, that are very difficult to reproduce and therefore to debug without some very sophisticated static code analysis tools (like Coverity). And there was at least one such bug solved in the last release.

Now, with a version of Spring running on Linux, maybe it is possible to use Valgrind, which could help detecting such bugs.

But for some reasons taht I don't understand completely, Spring is more stressing for my hardware than all other games I have and I've had a couple of reboots (not crashes, reboots) during game in the past.

[Zenka]

I sometimes (say once a 20 games) got a reboot.
Always when a unit dies.

Spring also reboots my Pc when a home-made s3o unit dies.
(I think that I've set the dead-model to a LLT).

Didn't saw Spring crash in a long-long time. (Reboots are worse though)