Server crash when a player goes higher then Y255 about cuberite HOT 19 CLOSED

Yep, it's ravenscroftj's new drowning code - I think it could be fixed with a constraint to player height < 255.

from cuberite.

Should now be fixed, closing.

from cuberite.

Not fixed...

from cuberite.

It doesn't fully crash now, but freaks out and throws an error:
"Verification failed: a_Chunk.UnboundedRelGetBlockType(RelX, RelY+1, RelZ, BlockIn), file ..\source\Player.cpp, line 1352".

from cuberite.

That's the same "crash". The server fails an assert and therefore commits suicide.

from cuberite.

Ah, ok. Just forked to take a look at it.

from cuberite.

Is there a programmatic way for the server to call the suicide helpline? Mortality rates in this department will get quite high if the server keeps on doing this.

from cuberite.

Or when the player goes a little into the void.

from cuberite.

I submitted a patch that should temporarily solve it. It just skips running the offending method if you're outside of the 0-255 block range.

from cuberite.

The suicide code is there on purpose, so that the devs will inspect whenever the code is invoked in invalid situations. Because when the code is called, the caller assumes that it will receive the correct block value, it doesn't anticipate that there's no block at those coords, so usually the caller has to be fixed.

from cuberite.

Okay :)

What does this method do anyway, and what does it do to have to run when there are no blocks?

Tiger

[Apple iPad Mail App]

On 11 Aug 2013, at 10:00 AM, "Mattes D" [email protected] wrote:

The suicide code is there on purpose, so that the devs will inspect whenever the code is invoked in invalid situations. Because when the code is called, the caller assumes that it will receive the correct block value, it doesn't anticipate that there's no block at those coords, so usually the caller has to be fixed.

—
Reply to this email directly or view it on GitHub.

from cuberite.

The patch is okay, mostly, just some cosmetics to fix and it'll be merged.

from cuberite.

Fixed?

from cuberite.

Should be. Please re-test, and close if okay.

from cuberite.

I still get the error.

from cuberite.

:(

from cuberite.

Can you verify that the error is still relevant to the cPlayer::SetSwimState() function? It might be an altogether different failure.

from cuberite.

I'm not getting any crash when above Y=256, so whatever crash you're getting is probably not caused directly by this. Can you provide any more info? stacktrace...

from cuberite.

Oh, the crash was very picky, it triggerred only between Y 255 and 256, nowhere else.

from cuberite.