This one will be hard to catch.
Here is valgrind output:
==6689== Invalid free() / delete / delete[]
==6689== at 0x4023503: operator delete[](void*) (vg_replace_malloc.c:409)
==6689== by 0x828C129: cbuffer_t::~cbuffer_t() (cbuffer_t.cc:22)
==6689== by 0x8121D8C: convoi_info_t::~convoi_info_t() (convoi_info_t.cc:247)
==6689== by 0x826489A: destroy_framed_win(simwin_t*) (simwin.cc:647)
==6689== by 0x82646ED: process_kill_list() (simwin.cc:605)
==6689== by 0x8266F0E: check_pos_win(event_t*) (simwin.cc:1290)
==6689== by 0x8282293: karte_t::interactive(unsigned int) (simworld.cc:6193)
==6689== by 0x8237610: simu_main(int, char**) (simmain.cc:1274)
==6689== by 0x82B9674: main (simsys_s.cc:681)
==6689== Address 0xede7a78 is 0 bytes inside a block of size 256 free'd
==6689== at 0x4023503: operator delete[](void*) (vg_replace_malloc.c:409)
==6689== by 0x828C129: cbuffer_t::~cbuffer_t() (cbuffer_t.cc:22)
==6689== by 0x81129F4: gui_textarea_t::~gui_textarea_t() (gui_textarea.cc:31)
==6689== by 0x8121D3C: convoi_info_t::~convoi_info_t() (convoi_info_t.cc:247)
==6689== by 0x826489A: destroy_framed_win(simwin_t*) (simwin.cc:647)
==6689== by 0x82646ED: process_kill_list() (simwin.cc:605)
==6689== by 0x8266F0E: check_pos_win(event_t*) (simwin.cc:1290)
==6689== by 0x8282293: karte_t::interactive(unsigned int) (simworld.cc:6193)
==6689== by 0x8237610: simu_main(int, char**) (simmain.cc:1274)
==6689== by 0x82B9674: main (simsys_s.cc:681)
Some memory (buffer inside cbuffer_t) freed twice.
once via ~convoi_info_t() -> ~gui_textarea_t() -> ~cbuffer_t() path.
Second time ~convoi_info_t() -> ~cbuffer_t() path.
Edit. I was lucky.
Here:
convoi_info_t::convoi_info_t(convoihandle_t cnv)
: gui_frame_t(cnv->get_name(), cnv->get_besitzer()),
scrolly(&text),
text(&freight_info),
text is gui_textarea_t, freight_info is cbuffer_t
gui_textarea_t::gui_textarea_t(cbuffer_t* buf_)
{
buf = buf_;
recalc_size();
}
End result:
convoi_info_t.text.buf == &(convoi_info_t.freight_info)
Edit^2:
FIX (https://github.com/inkelyad/simutrans-experimental/commit/3efe6a265eb7d06651335ae76ced0dad6a1ecfb5)
Inkelyad,
thank you very much indeed for that fix - pushed to the 9.x branch.
9.10 incorporating this fix is now released.