TESTING: fix crash due to SZ_GetSpace console overflow

[RauliTop]

More info here: #855

Also, review this: #798

[Kasou123]

help
version
Protocol version 48
Exe version 1.1.2.7/Stdio (cstrike)
ReHLDS version: 3.11.0.776-devmpatch
Build date: 18:17:32 Apr 26 2022 (2936)
Build from: 3dc9f8c
SZ_GetSpace: overflow on gs-m <- russian
WARNING: datagram overflowed for gs-m <- russian

[stambeto2006]

Is there any solution to this problem with this message, and how is the problem fixed?

overflow on netchan->message

[stambeto2006]

@RauliTop Is this project expected to be implemented and fixed officially soon?

[RauliTop]

@RauliTop Is this project expected to be implemented and fixed officially soon?

I really didn't test it.
I'm currently doing changes on plugins, without successful results.

You can try my request downloading the binaries on Actions tab inside this repository.
Pull request has been updated now with lastest rehlds official changes.

[Alguiencomoyo]

More info here: #855

Also, review this: #798

my servers go down with that same message I discovered that it is an attacker
even then I enter the servers to ask for money in exchange so as not to continue attacking
I have everything updated rehlds, amx, etc
and even so your proposal can solve this too?
I attach some videos of how he does it

https://www.youtube.com/watch?v=tCQ0tFdUTho

https://www.youtube.com/watch?v=Ex4n9KfUiCQ

https://www.youtube.com/watch?v=WMDZvrewaNI

[mlibre2]

also affects? rehlds/engine/cmd.cpp#L122

[RauliTop]

also affects? rehlds/engine/cmd.cpp#L122

No, because after overflow it appears: 'on' %s
It's on the line I marked.

[mlibre2]

Well this problem is quite common, apparently one of the causes is the loss of the player's connection to the server. sv_timeout is a quick fix for this, but I wanted to find another one, since this cvar doesn't work quite right, so I wrote this plugin to check the communication with the player. This method has worked well for me so far, although I don't like to call a function frequently, maybe there is another more optimal way to do it, but at least it doesn't print the typical "overflow/datagram" error message to the console, which that prevents it from happening.

_{So far it has been ruled out that the false positives have been kicked without reason!}

[anzz1]

This is wrong, it does not match original code.

There is no Con_NetPrintf in SZ_GetSpace in original HLDS. If this is a ReHLDS fix, it should be wrapped in #ifdef REHLDS_FIXES appropriately.

Also I don't think Con_Printf has anything to do with the error.