Freeciv

Which kind of action enablers for the city founding the ruleset has?

In general that ai settler code seems unprepared for action enablers controlled city founding, but I have not yet analyzed it in more detail.

Reply To cazfi

Which kind of action enablers for the city founding the ruleset has? In general that ai settler code seems unprepared for action enablers controlled city founding, but I have not yet analyzed it in more detail.

I do not know the ruleset that was used, but 80% of our singleplayer games choose classic rules. Still, that leaves a chance that it's another ruleset. Our mp2 actionenablers for those are 3, and look as follows:

actionenabler_build_city_pioneer action = "Found City" actor_reqs =

target_reqs =

actionenabler_build_city_domestic action = "Found City" actor_reqs =

actionenabler_build_city_fort action = "Found City" actor_reqs =

In addition, Andreas seems to have added more debug info into stack trace and here's another one (getting several of these every day from singleplayers, even though only 0.25% of total games get it.)

debugging using libthread_db enabled Using host libthread_db library "/lib/x86_64-linux-gnu/libthread_db.so.1". Core was generated by /home/freeciv/freeciv/bin/freeciv-web --debug 1 --port 60 38 --Announce none --e'. Program terminated with signal SIGSEGV, Segmentation fault. #0 unit_transported (pcargo=pcargo@entry=0x556131b41a60)

2450 || (!is_server() && pcargo->client.transported_by != -1)) { thread is 1 (Thread 0x7f2f0fb19800 (LWP 6367)) #0 unit_transported (pcargo=pcargo@entry=0x556131b41a60)

---Type <return> to continue, or q <return> to quit--- #1 0x000055612afead4d in package_unit (packet=0x7ffe989df070,

#2 send_unit_info (dest=0x55612d291510, dest@entry=0x0,

#3 0x000055612b0b1d2f in illegal_action (pplayer=0x55612e5105a0,

#4 0x000055612b0b3a0e in unit_perform_action (pplayer=0x55612e5105a0,

#5 0x000055612b0b85bb in unit_do_action ()

#6 0x000055612b10f9c7 in dai_do_build_city (punit=0x556131b41a60,

#7 dai_auto_settler_run (ait=ait@entry=0x55612b745220 <ai_types>,

#8 0x000055612b10f8ea in dai_auto_settler_run ( ...thousands more of these dai_auto_settler_run come after....

P.S. is there a trick to make these messages have more readable formatting? Sorry for not knowing it.

Well there's a clue, in mp2 rules, we do not have any reqs caring about if the cargo is transported, so I'm guessing it's one of the bundled rulesets tripping over the actionenabler req for cargo can't be transported?

Another clue, I have so far never seen this crash in a longturn game which uses our rules. So if it's actionenabler and it's tripping in the checking for unit transported, it's all adding up to cause suspicion for checking this actionenabler req of not being transported, probably.

Yeah, basically almost any non-standard requirement for those action enablers could cause that crash. The code is not prepared for such additional restrictions.

While waiting for a proper fix, you may need to just stop the recursion (remove the recursive call). That probably leads some AI settlers to never to do anything, but it's better than crashing.

Well maybe it's opposite, and the fact we allow settling while still transported. Can I get a clue what I should do next to find it?

Reply To lexxie9952

one of the bundled rulesets tripping over the actionenabler req for cargo can't be transported?

Which ruleset is that? I don't see it in classic or civ2civ3.

Reply To lexxie9952

Well maybe it's opposite, and the fact we allow settling while still transported.

No. The recursion happens when settling is *disallowed* when the code assumes it to be allowed.

instead of stop recursion and do nothing, is there a little less ugly hack like make it decide to deboard somewhere or do something else?