Having a hard cap on the number of rerolls doesn't help save CPU
usage from excessive rerolling, because if the cap is set low
enough to keep the CPU usage reasonable it isn't high enough for
players to actually use the feature.
Instead, allow capping the number of rerolls per second. (Sensible
values seem to be in the 5-10 range.) If the player attempts more
rerolls than this, show a paranoid confirmation prompt: the need to
type the answer to the prompt will slow a human user down (and if
the prompt is filled in too quickly, it will simply just be shown
again, preventing attempts to use automation to skip the prompt).
This feature was no longer fulfilling its intended purpose: the
issue is that when players are aware of it, instead of creating
interesting stories, it becomes a resource the player can rely on.
Alternative mitigation for some "unfair" early-game deaths has been
added (such as the warning shot for early-game attack wands and the
ability of iron shoes to protect against certain early traps), and
the warning shot in particular can create interesting gameplay
moments and stories of its own, especially when saving grace is
*not* present (discovering that a monster has a wand of fire and
can oneshot you is an interesting emergency situation, but less
scary if you know that its first attempt to oneshot you will fail).
The feature was also proving quite hard to code correctly, because
there are numerous cases that it "obviously" shouldn't affect
(most notably beheading due to Vorpal Blade, but also things like
purple worm digestion) which would need to be special-cased, and
because there are grey areas like wand zap bounces (which might or
might not be an intentional attempt by the player to hit themself
with the bounce) and damage from traps. Removing it saves the need
to work out, for every new damage source, whether and when saving
grace should interact with it.
Breaks save compatibility.
In earlier versions of NetHack, wands of digging generated fairly
often, allowing players to do a large amount of digging through the
maze walls in Gehennom and giving them a chance of an early escape
item.
In the current version, the need for both of these things has been
reduced: supply chests give a supply of early escape items
(including wands of digging), and Gehennom is no longer mostly made
of mazes with diggable walls. As such, there is no longer a reason
to generate wands of digging in such large numbers, even though it
was correct in the past.
This commit modifies some of the generation probabilities for wands
of digging in order to make them less abundant, in order to work
better with the code changes since the previous version.
The traps don't disappear when stepped on any more; that should be
true even when monsters step on them.
(This bug was reported privately to me via IRC, rather than via the
devteam's email address, so there isn't a ticket number to close.)
This removes or updates several true rumors that have become
misleading or inaccurate due to changes to the mechanics they were
describing, and adds several new rumors describing mechanics that
are either new, or are intended to be a larger part of the game
than they were in previous versions.
More oracularities describing the new mechanics are needed, but I
haven't managed to get into the write frame of mind to write in the
Oracle's voice yet...
Any sort of iron footwear now protects from traps, and polymorph
traps change one sort of iron footwear into a different sort of
iron footwear (removing any incentive to try to farm them, because
you will see both options very quickly).
When playtesting NetHack heavily, I observed that most of the time
it wasn't placing much strain on my wrists, but Sokoban was an
exception: travel, farmove, etc. can normally be used to avoid the
need to spam keys, but they don't work while pushing a boulder, and
the boulders often need to be pushed along precise routes, so you
have to tap out every movement. This becomes particularly straining
when pushing in the last few boulders, as you have to push them a
long way along the goal corridor.
This commit adds rolling boulder traps to Sokoban that will
automatically roll boulders along the goal corridor, meaning that
you don't have to push them there manually. This considerably
reduces the number of keystrokes needed to solve Sokoban, without
making any significant change to the difficulty of the levels.
Some of the designs had to change slightly in order to make room
for them, but not in a way that meaningfully changes the solution.
Rolling boulder traps now react to having boulders pushed onto them;
the boulder will roll until reaching the trap's launch spot
(resetting the trap if the boulder was removed), or until it hits a
wall otherwise.
Because they can now be reset, they aren't removed when stepping on
them with a misplaced boulder (although no boulder moves unless
there's one on the appropriate spot).
As of the change to allow for item probabilities that don't add up
to 1000, it's become a little difficult to figure out the exact
probabilities from the source code, which makes it hard to balance
item generation. Adding a tool to list the probabilities helps.
Part of the problem is that changing an item's probability without
balancing it elsewhere is usually wrong: doing that would in effect
take the probability equally from (or add the probability equally
to) all other items in the class, which might break the balancing
of those items due to the probability change.
As such, it is usually better to make an intentional decision about
which items should be less and which items should be more likely to
generate, then change them in a balancing way (meaning that the
probabilities of objects that weren't intentionally changed remain
unchanged). Doing a complex such change makes arithmetic errors
fairly likely, though, so it's useful to have a command that verifies
that it's been done correctly.
This command is primarily intended as a development aid, so it's
included only in debug builds and pre-release builds (the same as
other similar commands like #wizmondiff).
Armor is now slightly more likely to generate outside Gehennom (at
the expense of gems that generate via random generation rather than
mineralisation or level rules).
Basic nonmagical armor (especially body armor) has had its generation
probabilities reduced *relative to other armor*, but outside
Gehennom, it is no less likely to generate (because armor in general
is now more likely to generate outside Gehennom). It is slightly less
likely to generate in Gehennom (but isn't typically needed in
quantity there).
Armor that has extrinsics is more likely to generate, both due to
having increased probability relative to other armor, and due to the
increased proportion of generated items being armor: this is the
primary goal of this change.
The intention behind this change is to increase the chance that
players naturally find useful armor (especially armor that they might
not have been planning to use, but that they can adapt their strategy
to make use of), rather than needing to wish for it: the chance of
finding useful armor is higher both in the Dungeons (due to the
increased probability) and in Gehennom (because it is more biased
towards armor that might be useful at that stage of the game). In
practice, in 3.6.x (prior to this change and to wishing changes), it
was quite common for players to wish up an entire set of armor at the
Castle, ignoring almost everything they'd found so far that game; I'm
hoping this change encourages more wish variety rather than spending
the majority of wishes on armor.
The old code had two main problems: a) it was very difficult for
unspoiled players to figure out how it worked (because donating too
much got you a bad result, and the exact amount you needed depended
on magic numbers that weren't stated in game, and because you had to
hide your visible gold to get a good result); b) for players who
knew the mechanics, it was somewhat exploitable and also somewhat
tedious to make use of (due to needing to hide visible gold before
donating).
This change preserves the spirit of the previous code whilst making
things more transparent for new players and less tedious for existing
players: the donation amounts for the various effects are still
roughly the same (but randomized), but the amounts you need to donate
for clairvoyance and for protection are explicitly stated (and as
before, the alignment reset is done by donating an unnecessarily
large amount and isn't explicitly stated as an option). If you have
a lot of visible gold, you still need to donate a sizeable proportion
of it to get a useful effect, but now you get a larger reward to
compensate for the larger donation (to the extent that doing this
gives comparable results to doing it as a series of small donations,
removing the incentive to hide your gold before donating).
There's also something here for those players who like to squeeze
every last point of optimality out of a game: the "obvious" donation
strategy gives decent results, but players who are really willing to
dig into the mechanics may be able to find a way to get slightly
better results on average (which if I've balanced this correctly,
will lead to a very long and complicated spoiler).
One other change is that this is now based on your peak rather than
current level, to fix an exploit in which the character was drained
down to level 1 to donate a very large amount of gold (improving by
20 AC points) and then immediately restored back to the previous
experience level using a blessed potion of restore ablity.
This breaks save compatibility, but is being pushed together with
other save-breaking changes to avoid the need for multiple bumps to
EDITLEVEL.
It takes time for an early-game monster to acclimatize itself to the
power of an attack wand: in much the same way as a nervous human may
quite possibly miss with their first attempt to use a gun in combat,
an early-game monster will always miss on its first use of an attack
wand (but from then on will understand how they work and get over
their nerves, and will hit as normal).
This is a balance change based on observed results from tournaments:
guarding against deaths to early-game attack wands requires an
unusually cautious playstyle which isn't much fun (and might not
always be possible even for the best players), so it is quite common
for them to be the cause of random deaths that it wasn't worth trying
to avoid. Although trying to dodge a monster who found an attack wand
is fun, you only actually get that fun if something makes you aware
of the danger: the monster missing with the wand is a clear way to
demonstrate the danger and let the player know that now is the right
time to take precautions.
This change could theoretically have broken saves, but probably
doesn't due to there having been a spare bit in struct monst. Just
in case, it is being pushed together with other save-breaking changes
to avoid the need for multiple bumps to EDITLEVEL.
A wand of stasis prevents teleportation (even in some cases where
it would normally not be prevented, e.g. the hero teleporting a
monster, or covetous monsters teleporting). This is intended to
provide an alternative tactic against covetous monsters (and their
AI has been adjusted to handle being under a stasis effect), but
might also be useful in other situations. It does not prevent
teleportation of objects, only the hero / monsters, and does not
at present prevent level teleportation (although I'm not sure about
this and it might well change in the future).
This breaks save compatibility, but is being pushed together with
other save-breaking changes to avoid the need for multiple bumps to
EDITLEVEL.
In previous versions of NetHack, this wasn't a problem because
teleport traps didn't *generate* on such levels, but more recent
changes have made this situation possible (e.g. a demon lord
migrates to a Gehennom filler level that already had a teleport
trap on it).
This change causes a "wrenching sensation", like stepping onto a
teleport trap with magic resistance does.
The MSVC warning assumes that any attempt to write - followed by an
unsigned number is an attempt to create a negative number (which
produces the wrong data type if you type, e.g., -2147483648).
However, it's also warning when an explicit U is given, e.g. in -1UL,
which is clearly not an attempt to express a negative value.
Suppress the warning by rewriting the number in question as (0UL-1UL)
with a binary rather than unary minus.
These are displayed in discoveries, and a new 'price_quotes' option
allows them to be displayed for un-IDed objects in other contexts
too (the idea is that you turn on the option while identifying
objects and off for general play).
Invalidates existing save files.
Based on the principle that there should always be at least two
solutions to any given problem, this allows monster-slowing magic
to be an effective solution to fast low-monMR monsters (in
particular home-plane air elementals, who after recent commits
could reasonably be dealt with using good AC, but I want a second
good option to be available, in addition to the existing mediocre
options). This helps to make playing with bad AC (for whatever
reason) more viable.
This is one of the fairest instadeaths in the game (it gives
multiple warnings and there are reliable ways to escape it), but
is also where most of the threat of purple worms comes from. It
was also pretty much nonfunctional: the previously used formula set
the timer to somewhere around 50 turns (for a typical character in
Gehennom, which is where purple worms are normally encountered).
This new formula (which affects only purple worms, as the only
monster that engulfs and digests) is based on the same inputs, but
produces much smaller numbers, meaning that the instadeath is
relevant sometimes. A typical character at purple worm depth will
be able to kill faster than the timer if they focus on the purple
worm and put some damage on it before they get engulfed, but it is
much closer if the purple worm gets the first hit in (possibly
requiring the use of escape items, but the common wand of digging
works).
Hopefully this brings purple worms closer to their intended threat
level: previously they were somewhat more nonthreatening than they
should have been.
The previous commit caused air elementals to become almost totally
nonthreatening in the endgame (even the buffed ones on the Plane of
Air). This commit fixes that (whilst still leaving them somewhat
weaker than they were before against characters with good AC), by
doubling the damage of home-plane air elementals.
The damage of the other home-plane elementals was doubled too,
because they were mostly nonthreatening previously. On Fire, this
has no real effect as almost any character would be fire-resistant
by this point. On Earth, it makes the elementals more of a threat,
when they previously weren't.
However, it made Water too difficult (albeit more fun, because it
became important to avoid letting water elementals swam you). As
such, water elementals are made slightly slower to compensate.
(My own playtesting indicates that 6 is slightly too fast, but 4 is
too slow, so I'm hoping that 5 is the correct value.)
Players of both 3.4.3 and 3.7 have observed for a long time that
air elementals are disproportionately dangerous compared to other
endgame threats. (In particular, playtester feedback from 3.7 was
that the Plane of Air was much more dangerous than it should be,
with playtesters treating it similarly to Astral with respect to
the use of high-quality escape items.)
It turns out that this was because damage from air elementals while
engulfed was entirely ignoring AC, meaning that regardless of your
stats, you would be taking around 16.5 damage per turn while
engulfed (half physical damage helped, but nothing else did).
This commit purely fixes the bug, but does not balance around it,
which means that it causes the endgame air elementals to become
almost entirely nonthreatening. In a future commit, I plan to
balance around this change.
They still have their other current effects (aggravating monsters
and granting temporary invisibility). This is mostly meant as a way
of counteracting the "turn permanently invisible" effect of magic
traps, for players who would rather their characters remained
visible.
Thanks to paxed for helping with this commit.
This commit is intended to achieve two things: a) making Vlad's
throne a little less dangerous, b) making polymorph traps less
dangerous for players who don't have magic resistance. (At present,
uncontrolled polymorph is too dangerous in the late game for most
players to consider risking it, so they take care to avoid things
that have even a small chance of it, cutting out a lot of potential
strategies. Toning down polymorphs should make them happen more
often by increasing the chance that players will be willing to go
without precautions.)
Body armor is still destroyed (with the existing exception of
dragon armor), in order to mostly preserve the balance effect of
polymorph traps in the early game.
This removes the container-cursing effect (which is too specific
and too easy to work around) and replaces it with an effect that
greases everything (largely but not entirely positive).
This is primarily intended to help new players understand the
mechanics (the "wield the launcher, throw the arrow" sequence may
be unfamiliar to newer players), so the message is worded to
indicate the correct way to use the ammo.
Dragon corpses are capable of rotting, so it's plausible that the
skin of a dead dragon might also be possible to rot.
This is unlikely to come up in actual play at present (unless
wishing for rotten DSM, it can only end up rotting as a consequence
of brown pudding attacks): the primary motivation is to open up new
possibilities for armor-damaging attacks, which in current NetHack
aren't very relevant in the late game because everyone is wearing
dragon scale mail.
After discussing with paxed, I decided that instead of reducing the
safe enchantment limit for magical armor, it is instead better to
make the scrolls less effective on it. So this commit restores the
previous rules for the safely-enchant-from level and changes the
effectiveness of the scrolls to compensate.
Non-elven magical armor now only gets +1 from blessed scrolls of
enchant armor when enchanted from +3, coming to a total of +4. But
scrolls of enchant armor are now more effective on nonmagical and
elven and previously unenchanted armor, giving more of an incentive
to use them in the early game.
Cursed and uncursed scrolls of enchant armor are now also more
powerful than they were (but less powerful than blessed scrolls),
hopefully making it a more interesting decision as to whether to use
scrolls of enchant armor even if you don't have the means to bless
them.
Running low on HP in monster form isn't the same sort of critical
problem that running low on HP while not polymorphed is, because the
character changes back when the monster form HP runs out rather than
dying. (Indeed, running out of HP in monster form is often
intentional.)
The exception is when wearing unchanging (which implies both that the
monster form is intentional and that running out of monster HP would
be fatal), so low monster-form HP is treated as a major trouble in
that case.
Inspired by <https://nethack-yanis.github.io/yanis/4724.html>.
Counting wizkits as starting inventory causes side effects (giving
you skills from them, making the wizkit items formally identified,
etc.). Some of these side effects are undesirable (because, e.g.,
it can give you skill in a weapon you're restricted in), and the
others are mostly neutral (because wizard-mode players can identify
anything they want to). As such, it's better to treat the wizkit as
something you obtain immediately after entering the dungeon, rather
than something you had all along.
Prior to this commit, Archeologists had an incredibly difficult
start, worse than was intended. This is intended to make it a bit
easier (whilst making the role more different from other roles) via
allowing them to identify scroll types earlier in the game than the
other roles are able to (they do it using a knowledge of ancient
languages that lets them read scroll labels that other roles
couldn't).
These are often an important part of a character's build. There's
no purpose in listing them in disclose because the player generally
already knows what spells and skills they had and doesn't need them
identified, but they're useful when looking at someone else's game
or reminiscing over a past game.
The character doesn't know that an unlabeled scroll is blank until
they look at the inside of the scroll, but that could be done
either by reading or by writing.
These two enhancements were interacting with each other in weird
ways (paupers would start with no auto-IDs but force bolt, but
gain the level 1 auto-IDs upon training any skill). Change the
interaction so that paupers don't get the level 1 auto-IDs (which
wizards get to start with), but do get the level 3 auto-IDs in
skills that they manage to advance.
This adds a "reroll" option that lets players reroll their
character's attributes and starting inventory. Although I generally
think doing this makes the game worse, a) some players are going to
do it regardless and b) if a player is going for a challenge game,
rather than to win, it may be required. So in the absence of an
option like this, players repeatedly start and quit games instead,
creating a large number of junk logfile entries and generally
causing problems for other players on the same shared machine
(because repeatedly reloading the game is very CPU-intensive).
This should in theory be windowport-agnostic (although in practice
it may not be). Tested on tty, X11 and curses; on tty and X11 it
works fine (although X11 treats the change in attributes as
something that needs a status highlight), on curses it is slightly
jankier in terms of what other windows are drawn in the background
(but still plays correctly and I suspect this is a pre-existing
bug).
To form a complete implementation, we will need to consider the
following:
- Should there be a delay on a) starting the game and/or b)
rerolling? If so, what should it be (maybe configurable via
sysconf?)
- Should we take more steps to discourage players from rerolling?
It would be bad if players see the option exists and turn it on
just because it exists, or (worse) treat it as condoning the
particular style of play.
- Should we take steps to detect that players are rerolling
manually and a) tell them to use the option instead, b) tell them
that this is not an intended way to play (and may make the game
less enjoyable and/or prevent them getting the practice they need
to eventually win)?
Breaks save and bones files.
This doesn't implement inventory rerolling, just adds the
infrastructure: it's now possible to call u_init_inventory_attrs
multiple times and the starting inventory/attributes replace those
from the previous call rather than compounding.
This is helpful for if we ever allow playing multiple games in the
same session: because the arrays are not modified, they can now be
used more than once.
The important part of the "don't offer while impaired" change was to
prevent offering while confused. However, it was also extended to
other status conditions: stunning seems fine, but hallucination was
problematic (both because it makes a large number of messages
inaccessible, and because hallucination is more of a long-term status
effect than the other two and players may sometimes choose to play
with it for a large portion of the game). So make the change trigger
only on stunning and confusion, not hallucination.
This also updates the changelog for the change, because while
connected to the erinys changes, it's technically separate and is
relevant even in games where erinyes are never summoned.
For example, being hit by the bounce of a wand of fire means that
the main character could take damage twice in a turn, which would
kill even through saving grace; and scrolls and potions could burn
up after that and finish off the last HP, even if the wand only hit
once. This commit changes it to track all damage done during the
turn, and prevent HP dropping below 1 from damage until the next
player action or the next turn boundary, whichever comes first.
If you write the scroll by description, you obviously know what the
label is because you specified the label (even if you didn't know
what the scroll was). When writing an unidentified scroll by type,
though, and getting lucky, you don't necessarily know the label of
the resulting scroll.
This fixes a couple of bugs: a long-standing bug in which writing a
scroll by label could fail even if you've already seen a scroll with
that label (due to the game not tracking whether or not you've seen a
scroll if it doesn't have a name); and a somewhat newer bug in which
spellbooks auto-identified by Wizard knowledge were marked as having
been encountered (rather than as known but not encountered).
Breaks save file compatibility, but not bones files.
This is partially for the pun, and partially because the "wish for
scrolls of charging to identify them" strategy has been nerfed in
previous commits and this offers an opportunity to discover what
scrolls of charging are without randomly encountering one.
There was only one point in the code at which this caching was
being done, and it was incorrect: it's possible for the result of
near_capacity to change during a monster turn because monster
actions can change either inventory weight or carry capacity.
The bug was particularly relevant in cases where a character
polymorphed into a slow weak monster gets attacked by a monster
that moves at normal speed: due to the polyform being slow, the
normal-speed monster gets in a lot of attacks and causes a
rehumanization, but due to the polyform being weak, it was
burdened at the start of the monster turn, and so when that
penalty is (due to the bug) applied to the next turn it can
mean that the character misses the next turn too, and may end up
dying as a result.
The existing messages made sense for brief dips into water, but
didn't make sense when using an oilskin sack for an extended
period underwater (and also assumed that the player was able to
see the sack). This commit changes the message to make sense
(and to be less spamy) if the hero enters water and remains there,
and prevents oilskin sacks self-IDing if the hero is blind and
thus can't see the water.
It makes sense that a vegetable would do less damage than a hard
object would, as they're generally fairly soft, so it seems like a
likely thing for players to try if they're *intentionally* trying
to hit for zero damage (which could be useful in certain niche
cases, e.g. to wake up a sleeping monster without damaging it).
Part of the reasoning behind thrown objects not hitting mimics is
that a character who doesn't know there's a mimic there wouldn't
aim at it. But if you know there's a monster there (e.g. via
telepathy or monster detection), you would aim at it. Pushing a
boulder at it and hearing a monster works too (which is important
in cases where a mimic is trapped behind a boulder in the Sokoban
corridors).
Part 4 of implementing wish spreading. (This is now a complete
implementation, although the details are likely to change - but it
makes sense to commit something with the right balance properties,
and then tweak it based on feedback from playtesting.)
This helps to make the Amulet of Yendor feel special, and restores
approximately the same average number of wishes per game as existed
prior to the nerf to wands of wishing.
Placing the wish in allmain helps to avoid the wish happening at an
awkward place in the game's control flow, and is simpler than
testing every possible mechanism for gaining items for bugs (message
order is a common issue when trying to place it in addinv-related
functions, and this also avoids issues with the wished-for item
immediately invalidating an assumption that was made by the calling
code).
It is possible that this would be better as an invoke effect,
although I like the impact of picking up the Amulet and immediately
being given a wish.
This is partly for balance reasons (so that clearing the Castle
helps towards the level 14 Quest unlock) and partly as a clue to
players spoiled on previous versions that the Castle wand now works
differently.
Part 3 of implementing wish spreading. These items are each worth
most of a wish for non-illiterate games (most games wish for a
magic marker, and a magic lamp gives a wish 80% of the time).
The placement of the random item could be better (currently it is
purely random, which is occasionally interesting but often boring),
but this will serve as a base for experimenting with the balance
properties of the moved wishes.
Part 1 of implementing wish spreading. Vlad's throne is now
guaranteed to eventually give a wish, but has a range of other
powerful (and mostly bad) effects, like removing intrinsics, that
can be much harder to deal with than typical throne effects.
This causes the difficulties to get out of sequence, but we agreed
that this is a less important rule than keeping monster IDs stable.
Breaks save and bones files, because it changes monster IDs.
This is both a bugfix (the hero would be unlikely to aim their
throw at what appeared to be an object) and a balance fix (it was
possible to, somewhat tediously, defend yourself from mimics by
throwing gold pieces at them, and for many players this became
standard strategy in shops, negating the threat from mimics).
These two types of monster were extreme outliers in terms of where
they appeared versus how lethal they were (3-4 times as deadly as
other monsters that appeared at similar depth, based on statistics
from actual play), so their generation depth has been manually
modified.
Breaks save and bones files (because monster type IDs have to be
sorted numerically by difficulty, and changing difficulties thus
changes the IDs, but the IDs are used to identify the monsters in
save files).
These were unbalancing the game a) in the Castle and b) if they
woke up unique monsters (most notably the Wizard of Yendor).
I considered adding a difficulty check, but this commit instead
just directly fixes the symptoms. (It doesn't make sense for the
Castle to contain a monster that would kill or be killed by its
inhabitants: they should have died long before the hero arrived.
So for liches/zombies to exist in the Castle at all, there must be
a truce.)
+15 wasn't very impactful in the late game and late mid-game, but
was much too lethal in the early game (wearing the ring for a while
near the start of the game would make the game unwinnable as very
out-of-depth monsters spawned, and they would still be there even
after removing the ring and usually capable of one-shotting an
early-game character). This commit changes it to a doubling of
level difficulty rather than a flat increase: that makes it more
relevant in the late-game where a +25 or +50 might potentially have
an impact, and more survivable in the early game (although it still
spawns monsters that are difficult for the point in the game, there
is now a chance that you might survive long enough to be able to
take the ring off and clear off all the out-of-depth monsters that
spawned).
Previously, the code for monster healing was repeated every time it
was needed; this commit sends it all through a common function, which
will make it easier to make changes to how monster healing works in
the future.
This is just a code reorganisation and won't have any gameplay
effect unless I made a mistake.
Luck from sacrificing is now limited by the value of the sacrifice.
This fixes two exploits, both of which rely on getting luck up to
maximum as soon as you have an altar, a luckstone, and a few
rations, via altar-camping until you accumulate enough luck. One of
them is to use the resulting luck to throw off the balance of combat
via using it to make hit chance calculations irrelevant. The other
is to use it to get crowned early in the game; in particular,
getting crowned pre-Sokoban is often viable and, especially for
chaotic characters, solves most of the game's difficulty at that
point (because the intrinisics and weapon are enough to carry a
character to the Castle given even mediocre luck with finding armor).
After this commit, becoming crowned very early in the game is more
difficult (likely requiring unicorns and identified gems), and the
hit chance gain from luck becomes a more gradual gain over the
course of the game rather than all happening immediately upon
finding the altar and luckstone.
In addition to making the game more balanced, this also discourages
grinding by reducing the incentive for altar-camping, so it will
hopefully make it more fun as well.
In 3.6, artifact gifts are often either a) entirely useless or
b) gamebreaking, neither of which is really ideal from a balance
perspective.
This commit aims to make artifact gifts more useful in the early
game by greatly increasing the chance for situational artifacts to
generate positively enchanted. However, the most powerful
artifacts will now only be gifted if you offer a high-value corpse,
meaning that they are only likely to be accessible later in the
game. The selection of which artifact to gift has become more
complicated in order to a) increase the chance that it fits the
character and b) reduce cheese strategies (e.g. it is no longer
possible for elves to force the gifting of Stormbringer as the
first sacrifice gift).
gnome-terminal was recently fixed to work correctly with
TTY_TILES_ESCCODES (and other similar private-use terminal code
sequences), and thus this isn't an incompatible combination any
more.
The new "riding blocks stealth" code was reading u.usteed at an
early point in the restore process, before its value was reliable
(dorecover()->restgamestate()->set_uasmon()->steed_vs_stealth());
because restgamestate() happens prior to restlevelstate(), the
value of u.usteed will be a stale pointer from some previous game,
and attempting to determine whether the steed is flying will crash
the game.
steed_vs_stealth() doesn't actually need to be called during the
restore process (because BStealth is saved in the save file), so
this can be easily fixed by omitting the call to it during the
restore.
The new formula is: (xlevel + Con)% chance of regenerating 1 hp
each turn.
This formula has been extensively playtested throughout the whole
game (including two ascensions). The intention is to make late-
game combat more interesting: early game the HP regeneration rate
is potentially slightly faster but not significantly changed, but
in the midgame and lategame is substantially slower because there
is no longer a big regeneration boost once the character's xlevel
is in the double digits.
With the new formula, I'm finding that my characters have to heal
with potions (rather than by waiting) in places that they never
had to before (e.g. lower Dungeons, and upper Gehennom), which in
turn means that fighting efficiently is now more important than it
was before. (In fact, in one of the games I wished for potions of
full healing on Astral for safety, although I think I would still
have won without.) It's also generally the case that you can no
longer regenerate "mid-fight": you need to disengage in order to
heal up. This made the game more fun as it meant that escape items
became more relevant, and I was using a greater range of items
throughout the game than I normally would.
The ring of regeneration has also been slightly buffed: it now
heals an extra 1hp per turn unconditionally (rather than becoming
less effective as the character levels). In both my test
ascensions, I found a ring of regeneration, but intentionally
refrained from using it in order to ensure that the new HP
regeneration rate would be tolerable even without one.
The recent commit to interpret walking into a shopkeeper as a "pay"
command was triggering in too many circumstances. Check to ensure
that the monster that we're walking into is a known shopkeeper
before activating the special case.
Prior to this commit, there was no good way to deal with swarms of
weak, good-AC enemies using magic; trying to play a wizard as a
pure spellcaster would make bees and ants very difficult to deal
with (because they would usually dodge force bolts).
This commit adds a new spell designed to be very good against
swarms of weak enemies: "chain lightning", which does 2d6 lightning
damage to every monster around you. It has an initially short range,
but can chain from monster to monster to cover potentially large
distances (as long as none of the monsters en route are shock
resistant or tame/peaceful; the spell can't chain past shock
resistant monsters and avoids peacefuls). Monsters within one
space of the visible lightning bolts are affected. Unlike other
lightning effects, this one does only 2d6 damage, not enough to
blind affected monsters.
This commit breaks existing save and bones files (thus the
EDITLEVEL increase).
The change to Sokoban difficulty calculation introduced a balance
issue: previously the Sokoban prize was obtainable before item-
destruction attacks became common and would protect from them,
whereas now they commonly appear in Sokoban itself, before the
prize is accessible. This makes scrolls and potions much less
usable as escape items, and potions of healing less usable for
HP recovery, because they either get stashed or destroyed (and
in either case aren't usable by the character in time-critical
situations).
This commit attempts to alleviate the problem by adding a way to
protect items from cold and fire in the early game. It's a
little unreliable, and requires equipping an item that has bad
stats and generally isn't otherwise used, but which is generally
easy to find early on.
Not included in this commit, but probably necessary for a
"finished" version of the feature: some way of letting players
know the new mechanic. It shows up in enlightenment, but there
should probably be at least rumors (and maybe even a major Oracle
consultation) hinting at the mechanic. We might also want to
change the unIDed description of the dwarvish cloak as a hint,
although that would require, e.g., new database entries.
The same checks were being repeated for every damage type; this
sends them through two centralised functions (one for checking
whether an extrinsic blocks a specific instance of item destruction
and one for the enlightenment message), so that new mechanisms of
item destruction prevention will need to change only one point in
the code.
Wizards now start out recognising all level 1 spellbooks, making
it possible to write a low-level spellbook in order to start
training a spell skill of choice.
As such, it shouldn't be affected by the heat of the ground in
Gehennom (which is hot enough to boil but not to burn). The oil
still heats but won't break its bottle.
In Gehennom and on the Plane of Fire, the ground is hot enough to
boil potions (although not hot enough to, e.g., burn scrolls).
The potions sometimes survive this (almost always if it's your
potion, it's blessed, and you have maxed Luck), but often don't.
In addition to making a lot of flavour sense, this serves a
gameplay purpose in that it reduces the number of potions that
are deathdropped by monsters in the late game. In Gehennom,
monsters often generate with potions to use defensively, but then
get killed before they have a chance to use them: this produces a
surfeit of potions that players tend to convert into holy water or
potions of full healing (and in general it doesn't make much sense
that the basic potion of healing primarily generates in Gehennom).
This commit approximately halves the number of useful potion
deathdrops, whilst still allowing monsters access to their
potions; when the monster dies, it drops the potion and this has a
chance of destroying it.
The previous commit makes things somewhat harder for vegetarian
characters. Normally, making a conduct harder isn't a huge issue,
but vegetarian Monk is an intended/encouraged combination. As such,
compensate by adding some additional food rations to Monk games
specifically; the Quest home level is early enough to provide the
extra food, and as the level is Monk-specific, it will provide the
food only to Monks specifically. (Although the Quest generally
requires level 14 to enter, the Quest home level specifically
doesn't.)
Playtesting has shown that there is too much permafood in the game
at present: in the late-game the only food-related problem is how
much to carry in order to avoid burdening yourself. In the early
game, food could previously have been a problem prior to
Minetown/Sokoban (thus the recent commit to add a guaranteed ration
in the upper dungeons), but past that point, there is easily enough
food generated on the ground. Additionally, the recent commits to
make healing sources more available in the early game reduce the
amount of time that needs to be spent waiting to heal, thus
further reducing food requirements.
The main purposes of food as a mechanic are to given an incentive
to press onwards and to discourage grinding. However, if monsters
are deathdropping non-corpse permafood, then beyond the very
early game, grinding actually generates more food than it uses up,
so the nutrition mechanic doesn't do its job properly.
Playtesting (including a full ascension!) has shown that there is
still plenty of food available even without deathdrops available
(my test game had 8 spare non-deathdrop food rations upon reaching
the Castle, at which point nutrition is no longer an issue due to
the Castle food stores and the huge numbers of C- and K-rations
dropped by the soldiers). I will address the potential problems
this causes for vegetarian Monks in a future commit.
With potions of healing becoming much more common, the multi-step
alchemy recipe from potion of healing to potion of extra healing to
potion of full healing is likely to become even more overpowered
(and it was somewhat unbalancing even beforehand).
This change restricts alchemy involving diluted dipped potions to
alchemize only two potions at a time. This means that potions of
healing can stil be alchemized into potions of extra healing as
efficiently as before this commit, and so can potions of extra
healing into potions of full healing; but the multi-step recipe
is now limited by requiring a lot of potions of gain level or gain
energy. As such, this is intended to make potions of healing into
an item primarily useful in the early game, and discourage hoarding
them for the late game.
At present, potions of healing generate primarily in Gehennom,
which causes significant balance issues (e.g. you don't have them
to heal in the early game, or even if you do, you can't identify
them in time to use them). In this series of commits, I'm aiming to
make potions of healing a more viable early-game healing source,
which means making them both (much) more likely to generate, and
easier to identify so that they are actually usable in the early
game.
This commit radically increases the generation chance of potions
of healing (reducing the chance of most other potions slightly to
compensate), and gives them a unique base price. This should make
them fairly easy to identify either by price-ID or quantity-ID
(and the unique base price is chosen to be fairly easy to figure
out even for unspoiled players).
These are primarily in chests (apart from one guaranteed good food
item on the Mines branch level), and are quite likely to be potions
of healing, although other items that are useful for early-game
survivability are also possibilities.
This is part of a series of commits that aim to make the early game
less about waiting to heal up and more about pressing forwards. In
particular, this means that characters need likely access to
healing sources other than waiting/backtracking/hiding in closets.
In a future commit, I plan to make permafood generate primarily
through exploration (rather than drops from monsters) in order to
deter waiting around or grinding; the early guaranteed food drop is
present to give the more nutrition-intensive characters (e.g. orc
wizard or vegetarian Monk) a fair chance to reach the more abundant
food sources in Minetown or Sokoban.
In particular, Wizards now get a magic marker guaranteed (but with
fewer charges than magic markers naturally generate with). This is
intended to improve the game in two ways: it reduces the incentive
for startscumming, and it gives Wizards a method to gain an
additional low-level spell early if they wish (but the marker does
not have enough charges to get higher-level spells quickly using
this method).
Previously, Wizards got a boost to the chance of writing unknown
spellbooks based purely on being a Wizard (with the chance still
luck-based), leading to a very large power spike when the Wizard
gained access to a luckstone and the ability to max out luck.
This had two main issues: this power spike came *after* the major
early-game difficulty spike, often leaving Wizards forced to deal
with it without having appropriate spells; and it promotes
grinding (for Luck and for Magicbane) at an early point in the
game, meaning that the Wizard early game effectively followed a
sequence of extreme difficulty -> grinding -> minimal difficulty,
which isn't very good balance-wise.
With this commit, Wizards lose their advantage to writing unknown
spellbooks by guessing, and instead learn spellbook IDs based on
their spell skills (advancing a skill gives knowledge of higher-
level spellbooks). This means that writing unknown spellbooks
becomes guaranteed with sufficient skill, but has no advantage
over non-Wizards in schools where the Wixard does not have
sufficient skill.
Due to Wizards' skill caps, there are two spells which they can't
ever write guaranteed: create familiar and charm monster. Create
familiar is a fairly niche spell (that doesn't match the Wizard
playstyle that well) and being unable to write it is not a major
problem. The inability to easily write charm monster is
intentional.
NetHack has historically had problems (both in terms of interface
and in terms of gameplay exploits) caused by unidentified items
not stacking with identified items. The 3.7 change to cause items
picked up by monsters to become unidentified has exacerbated this,
e.g. there's an undesirable strategy in which players give
ammunition to hostile monsters in order to unidentify it, allowing
it to stack with other unidentified ammunition so that the whole
stack can be enchanted with a single scroll.
This commit makes it possible to stack items with different
identification statuses (unles Blind), in effect causing the
unidentified items to be identified by comparing them to the
identified items (this is consistent with the mechanic via which
two stacks of unidentified items can be determined to be the same
as each other). The hope is that this will prevent any gain from
exploits involving intentionally unidentifying (or intentionally
failing to identify) items, that it will prevent interface clutter
caused by (e.g.) a Ranger's main projectiles becoming partially
unidentified as a nymph picks them up and subsequently failing to
stack, and that it will reduce the incentive to backtrack to an
altar to merge, e.g., a stack of formally-known-uncursed and a
stack of informally-known-uncursed food rations to stack.
If the game ended while the player was in the middle of a repeated
command (that was not cancelled), this would spuriously cause a
"while helpless" to appear in the death message and logfile entry.
This bug is hard to observe because most things that can kill the
character also cause repeated commands to stop repeating, but is
easy to reproduce by command-repeating the #offer command that's
used to ascend (as "n20#offer" or "20#offer" depending on the
control scheme): doing so produced "ascended, while helpless"
prior to this patch.
This is a well-defined operation, so bwrite() should be able to
handle it. However, when running under glibc, fwrite() produces an
unexpected return value for 0-byte writes, which makes bwrite()
think that the write failed (causing a panic).
This change implements 0-byte writes by not calling into libc at
all, so that we don't have to worry about how to decode the return
value of fwrite().
Alex Smith