Instead of outright destroying the armor, the spell will instead
first erode the armor. The spell hits 2-4 times, so if it hits
the same armor 4 times, it will get destroyed. This does not
hit erodeproof armor.
Also change the scroll of destroy armor, so that blessed one will
destroy a cursed armor, if hero is only wearing that.
dmgval is also used for monster attacks; move the healer knife
bonus to actual hero-specific code, and make sure it only
applies to hand-to-hand combat.
Styled as anatomy knowledge, a healer attacking a monster in melee
with a knife gets +1 damage per 6 dead monsters of that type,
capped at +3 damage.
Handwaving away the fact we're not keeping track who actually
killed the monsters...
Idea from aosdict.
The lua des.reset_level() call did not reset the special level
coder, so some values were kept and couldn't be changed.
Adjust the movement tests for this change.
Add a new debug flag prevent_pline, which prevents all messages
from going out to the UI. This prevents the tests from stopping
for -more-.
Add rudimentary tests for applying whistles, camera, and stethoscope.
Test generation of every object, both via des.object and obj.new.
Expose FIRST_OBJECT and LAST_OBJECT numbers to lua.
Add lua nh.int_to_objname, a function to convert integer value to
object base name and class.
Allow creating new nethack lua object by specifying id and class.
Expose constants NUMMONS, LOW_PM, and HIGH_PM to lua.
Allow converting an int to monster type name.
Create one of each type of monster in the lua tests.
Potion of object detection and then looking at a generic object glyph
on the map would try to "discover" the generic object, and the loop
in discover_object would go too far, overwriting the exclusion_zones
pointer.
Take the easy route and just prevent discovery of generic objects;
there are probably other places where the generic objects should
be handled too, but the fuzzer hasn't hit them.
Luck has a massive effect on the to-hit chance; maximum luck alone
(which almost everyone has past the midpoint of the game) gives
10 points to to-hit, so accounts for 50% chance alone, excluding all
other effects.
Multiple variants do something similar to this, so it is well tested.
This version comes from xNetHack by copperwater <aosdict@gmail.com>,
and allows the +1 or -1 luck adjustments of early game, such as full moon,
to have an effect.
The safe armor enchantment limit is lowered by one, if the armor
is innately magical. This takes off 3-7 points of AC from
a typical ascension kit, but should not really have any effect
for early game.
Also clean up the relevant code a bit.
Add some basic spellcasting stuff to the tutorial: read a spellbook,
cast a spell. If the hero doesn't have enough energy, just adds
a note saying so.
Remove/restore the known spells when entering/leaving the tutorial.
Seems like several new players did not find the "hidden" stairs.
Add iron bars to show a glimpse of them - the hero still needs
to find the secret door in to the room.
Noticed a strange oversight in knockback: Hero hitting a monster
did not cause knockback, unless polymorphed into a monster.
Add knockback chance if we're using a weapon, not twoweaponing,
and dealing some damage.
Weapons that can do knockback are lucern hammer, bec de corbin,
dwarvish mattock, (silver) mace, morning star, war hammer,
club, quarterstaff, aklys, flail, pick-axe, and grappling hook.
Iterating over a large set of locations in a selection caused a memory leak.
Lua couldn't do garbage collection in the middle of the iterator function,
so it eventually ran out of space, and just quietly dropped stuff.
When I reworked amnesia to not forget levels or objects, I removed
the forgetting from the mind flayer attacks. I intended to add
something to replace it, but forgot ...
Grimtooth is now permanently poisoned, protects the wielder from
poison, and can be invoked to throw poison.
Permapoison code comes from xNetHack by copperwater <aosdict@gmail.com>.