SmaugFUSS 1.9.2 is released.
This is a minor maintenance release covering new issues fixed since 1.9.1. It should take care of everything in the forums that had been reported but not yet dealt with by 1.9.1.
Commit details are available here: https://github.com/Arthmoor/SmaugFUSS
Changes For SmaugFUSS 1.9.2
Despite being specified in the code, the alignment antis were never implemented, and the class antis for Nephandi and Savage had not yet been added.
Setting all skills to 100 on a player does not properly handle adept maximization.
The mudprog_driver function to pick a random player as a target in a room had a very strong statistical bias in favor of the first chosen victim. The routine has now been corrected to do a proper random pick.
Mobs fleeing during mobile_update are not checking the ACT_STAY_AREA and can end up in areas they aren't allowed to move in, thus becoming stuck.
Due to a botched understanding of how door randomizer resets work, it was not possible to add one in the OLC.
Weather echo messages were still being delivered to people in indoor sector types.
Mobs who have the pass door effect on them never actually attempt to use it because the flags were not being checked.
It was possible to look at the sky while in an indoor sector.
is_name_prefix in handler.c did not match the definition specified in mud.h.
mpoload does not properly drop items to the ground that are loaded by object or room progs.
A memory leak was fixed in polymorph.c when reading morph data from a player file.
__FUNCTION__ is apparently not a portable method for embedding a function name. All instances have been swapped with __func__ instead.
Out of bounds array check in scramble() fixed.
Invalid bitflag check in do_oset() for trap flags fixed.
Wrong buffer size being used for IMC chess sessions in do_chess.
Memory leak in fread_memberlist() if the function drops out the bottom of the for loop for some reason.
Improper bounds checking for some loops involving static buffers.
Possible double free in load_projects if fread_project closed the file pointer before returning.
Bogus condition check for 'U' sectors (SECT_DUNNO) removed in add_new_room_to_map. 'U' is valid for SECT_UNDERWATER.
Memory leak in do_renumber if r_area is valid and check_vnums returns true.
Other minor issues pointed out by a CPPCheck analysis fixed. Not very many.