3.8.0 beta1, 64bit archlinux: units stop when too many are on the map

[Baŝto]

This occurs when too many units are on the map.
New movement commands are given, but not obeyed.
I’m not sure if they start walking when enough units got killed, but they don’t start walking when a movement command is fulfilled.

this behavior can be observed in this scenario: http://www.file-upload.net/download-7987022/pigdefense.zip.html

It’s a towerdefense with just pig waves and just 10sec between new waves, every wave will be 130% of the wave before. Since the pigs are slow and CPUs need some time to die, there is a huge amount of pigs on the map. The CPUs are still alive when the first units stop moving.


Edit (tomreyn): Removed [bug] tag, not needed on the bug reports forum.

[titi]

I tried it and I have no such problems :-(  . But I played and killed the pigs on my square.

Update: Strange in a second try I got the behaviour you talked about. I will look at it.

[Baŝto]

I just tested it again.
I uploaded my --verbose output

That’s a screenshot before (wave 13)…



and after (wave 15) the occurance of that bug.

[softcoder]

When running this scenario i get lua errors:

Code: [Select]

[] *ERROR* In [/home/softcoder/Code/megaglest/trunk/source/glest_game/game/script_manager.cpp::init 468]
error [Error calling lua function [global] error: Runtime error: [string "global"]:54: attempt to call field 'getn' (a nil value)
Stack Trace:
../mk/linux/megaglest:Shared::Platform::megaglest_runtime_error::megaglest_runtime_error(std::string const&, bool)address [0xb01bd2] line: 237
../mk/linux/megaglest:Shared::Lua::LuaScript::endCall()address [0xa9cbdd] line: 596

in the xml file i see a bunch of: x.getn() calls

Are you sure this scenario works?