Issue reported by mkuoppal: drum of earthquake triggers a sanity
check warning which the fuzzer escalated to panic.
Analysis by entrez. If a pit gets created next to a pool or moat or
lava, liquid might flow there and replace the pit. But the drum of
earthquake code assumed that the pit was still there. If there was a
monster there and it wasn't levitating or flying and it wasn't killed
by the liquid, it got marked as trapped even though the pit was gone.
'sanity_check' noticed.
With difficulty I was able to reproduce the impossible warning before
the fix, but the are a bunch of random factors at play. After the
fix I can't reproduce it again, but that's not a guarantee that it's
actually fixed. The analysis seems correct though and the fix is
based on dealing with that.
Closes#1170
Allow 'm p' to pay via menu when menustyle is traditional and to pay
via the old sequence when it's combination, full, or partial. Also
revise the "Itemized billing?" prompt to accept 'm' as well as 'ynq'.
Answering 'm' will switch from the old sequence to the menu (whether
you got to that prompt via m-less 'p' for traditional or 'm p' for
other styles).
... and have more than 1 billed item, and using non-traditional
menustyle.
I opted to add an extra field to the bill struct, because
that made the code cleaner.
Breaks saves and bones.
During very early startup, Windows may not have loaded
the tty window procs yet, and it is running with safeprocs.
It will eventually load the tty stuff. If the currently
operating window port fails in can_set_perm_invent(),
try the check for WC_PERM_INVENT again explicitly on
the tty windowport.
There is a slight change in behavior here when collecting build-time
option information. If MAXOPT is too small, collecting will now save
MAXOPT lines and then skip anything that should come after. It used
to always keep the very last line, each time replacing whatever the
previous last line was, but would leak memory if that ever happened.
And after yesterday's "fix" it would try to keep that very last line
beyond the array bounds.
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.
"Grand poohbah" is a common term while "high poohbah" is not, and it
seems appropriate for hallucinogen-distorted high priests. (It seems to
most commonly be spelled "grand poobah" but I left that alone; "poohbah"
seems like a decent compromise between that and the spelling of the
character's name in the Mikado).
Add pickup_stolen option to autopick items stolen from you by a nymph or
monkey, even if they don't match your normal autopickup settings.
Replace was_dropped, was_thrown with a 2-bit bitfield that can contain
values LOST_DROPPED, LOST_THROWN, and LOST_STOLEN (or 0), since they
should all be mutually exclusive anyway as they track the most recent
way the item left the hero's inventory.
[Rebase/merge conflict fixed up. PR]
Previously, if you threw a dagger into a pile of daggers, you'd pick up
the entire pile with pickup_thrown on. Since pickup_thrown and
dropped_nopick options are supposed to apply specifically to items
you've handled, don't merge items with different values in those fields.
This is based on a feature in UnNetHack (and I think some other variants
as well). If the hero intentionally drops an item with 'pickup_dropped'
disabled, don't autopick it back up when walking over that square again.
Typically when the player drops an item, it's because she doesn't want
it in her inventory any more, and this option stops autopickup from
defeating that goal (especially useful for tasks like stash management
without a container). Players have come up with workarounds to this
problem like toggling autopickup when approaching their stash pile or
adding name-based autopickup exceptions to allow them to exclude
individual items from autopickup, but this behavior should reduce the
need for those things.
I think 'pickup_dropped' is a little unfortunate because it suggests
equivalence to 'pickup_thrown' (i.e. any dropped items will be
automatically picked up regardless of autopickup exceptions). Calling
it something like 'nopick_dropped' might be better, but as far as I can
tell options cannot start with the word 'no' because it's interpreted as
a negation of the rest of the option name.
When known discoveries was displayed on a window of type NHW_MENU,
header lines sent to it via menu_add_heading() disappeared into the
bit bucket. Switching back to putstr() made them show up. But I also
changed the type to NHW_TEXT. A menu_add_heading() in artifact.c was
overlooked, and after the window type change that started triggering
a panic when '\' or '`a' was used while any artifacts were discovered.
Not sure how other interfaces were dealing with this.
Thinko in the reformatting, not triggered during testing because I was
using !CRASHREPORT. With it enabled, setting a watchpoint in gdb on
something unrelated resulted in nethack going into crashreport_init()
and never caming out. (I imagine that it would have eventually but 10
or 15 minutes with nothing noticeable taking place was unbearable, and
the program hadn't even really started yet.)
Fixup some of the inconsistently formatted code that has been
introduced recently or been building up for a while. Done manually.
I wasn't systematic except for looking for lines ending in '&' or '|'
(which wouldn't find such things if they're followed by a comment)
so there might be lots more. I changed a bunch of C++-style //...
comments to old style C /*...*/ so that they'll match the rest of
the core's code rather than because they shouldn't be used.
A couple of things I noticed while running umpteen tests for tty
perm_invent. Remove the update_inventory() from unmul(), and limit
the one that deals with seeing inventory when recovering from blindness.
Just a drop in the bucket overall, and the screen updates nearly
instantly for update_inventory() except when debugging perm_invent so
players aren't likely to notice this.
As part of the tty resize handling revision, code dealing with the
perm_invent window was moved out of tty_destroy_nhwindow() was moved
into a separate routine. The new routine would have been called for
a window of NHW_PERMINVENT, but WIN_INVENT doesn't have that type,
just ordinary NHW_MENU, so the cleanup wasn't happening, resulting in
a memory leak.
Silence a bunch of complaints from heaputil about freeing Null
pointers. The C standard allows that but it makes the data collected
about mallocs and frees not match up when nethack has MONITOR_HEAP
enabled.
opttext[] was given 120 elements but the MAXOPT macro used to limit
populating it was only 40. There happen to be exactly 40 lines of
runtime info--at least for my config--and the last slot was going to
waste, so the final dupstr("") at the end of build_options() wasn't
being tracked and freed.
It's interesting the whatever other memory checker is being used
failed to find something caught by old heaputil. Maybe the configs
running whatever it is are generating smaller options text?
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.
A few add_menu_heading() calls were added to the '\' and '`' commands
which use a text window rather than a menu. For some reason they were
using create_nhwindow(NHW_MENU) but only populating it with text so
the heading lines sent as menu entries were lost. (Might panic with
Qt; I didn't check.)
If riding over ice, check whether the steed, rather than the hero, is
cold-resistant or floating to determine whether it should slip, since it
is the monster which would actually be in contact with the ice.
Fumbling makes the hero fall from the saddle, but the justification was
weak if the only fumbling source is riding over ice (the messages were
things like "you drop the reins" which made more sense from magical
fumbling). Make all fumbling from ice alone go into the ice-specific
"slip on the ice" block and add a chance to fall from your mount there.
If fumbling from another source while riding on ice, the hero will
always fall from his steed, since that's what happens on normal floor --
ice had actually been reducing this chance.
This gets rid of a FIXME and K4056 internal bug report.
Allow the comma to be quoted as follows:
SYMBOLS=S_ice:\,
or
SYMBOLS=S_ice:','
Disclaimer:
The use of the comma on the map could conflict with future
use of that currently unused symbol for other intended purposes.
Fumbling was apparently meant to make it harder to keep your footing
when an earthquake created a pit under you, requiring a 1/5 roll to
stay upright, but because it was added as an additional OR it actually
just gave the hero an additional (albeit unlikely) chance to retain her
footing. Make it actually have a negative impact on the hero's ability
to retain his footing rather than a minor boost.
The tracks left by hero were cleared when player saved and
restored the game, or changed levels. Now the tracks are
saved in the dungeon level, so changing levels keeps the tracks
left by hero in that level.
Also increased the length of tracks from 50 to 100, and
simplify the tracking function.
Thing not done: fade out old tracks when returning to a level.
Breaks saves and bones.
