Freeciv

In case it ends up helpful, here is the back trace of the stack:

#0  invasion_funct (ait=ait@entry=0x555555b09260 <ai_types>, punit=punit@entry=0x5556123c2bd0, dest=dest@entry=true, radius=radius@entry=0, which=0) at ../../../ai/default/daiunit.c:1037

#1  find_something_to_kill (ait=ait@entry=0x555555b09260 <ai_types>, pplayer=pplayer@entry=0x55560c4871f0, 
    punit=punit@entry=0x55556137b4c0, pdest_tile=pdest_tile@entry=0x7fffffffd570, ppath=ppath@entry=0x7fffffffd578, 
    pferrymap=pferrymap@entry=0x0, pferryboat=0x7fffffffd580, pboattype=0x0, pmove_time=0x0) at ../../../ai/default/daiunit.c:1251

#2  dai_military_attack (ait=ait@entry=0x555555b09260 <ai_types>, pplayer=pplayer@entry=0x55560c4871f0, 
    punit=punit@entry=0x55556137b4c0) at ../../../ai/default/daiunit.c:1797

#3  dai_manage_military (ait=0x555555b09260 <ai_types>, nmap=<optimized out>, pplayer=0x55560c4871f0, 
    punit=0x55556137b4c0) at ../../../ai/default/daiunit.c:2590

#4  dai_manage_units (ait=ait@entry=0x555555b09260 <ai_types>, pplayer=pplayer@entry=0x55560c4871f0)
    at ../../../ai/default/daiunit.c:2917

#5  dai_do_first_activities (ait=0x555555b09260 <ai_types>, pplayer=pplayer@entry=0x55560c4871f0)
    at ../../../ai/default/daihand.c:757

#6  cai_do_first_activities (pplayer=0x55560c4871f0) at ../../../ai/classic/classicai.c:450
#7  ai_start_phase () at ../../server/srv_main.c:1144
#8  begin_phase (is_new_phase=<optimized out>) at ../../server/srv_main.c:1398
#9  srv_running () at ../../server/srv_main.c:2948
#10 srv_main () at ../../server/srv_main.c:3604
#11 main (argc=<optimized out>, argv=0x7fffffffdb88) at ../../server/srv_entrypoint.c:362

Do you have a savegame where this can be easily reproduced?

regrets, I have recently deleted it. I will change my habit and start attaching them.
(It probably has around 10K units, that I am playing around with.)

If I find this symptom in set I am working with now, I will upload it.

#1741 ends to same invasion_funct() crash, and is reproducible from a savegame.

Instead of silently (not letting anyone to know about it) ignoring such error situation, and doing not-necessarily-sane arrangements to "fix" the situation, I would handle this by adding fc_assert_ret() to invasion_funct().