Reported by paxed. A potion of oil, that was already in the midst of exploding,
got picked up through spot_effects(), which led to it merging with
another potion of oil and the freeing of the original obj.
The original obj pointer was still held by breakobj(), and breakobj()
proceeded to delete the obj (again).
Function nesting:
1 spelleffects()
2 -> weffects()
3 -> bhit()
4 -> bhitpile()
5 -> bhito(obj ...)
6 -> hero_breaks(obj ...)
7 -> breakobj(obj ...)
8 -> explode_oil(obj ...)
9 -> splatter_burning_oil()
10 -> explode()
11 -> zap_over_floor()
12 -> melt_ice()
13 -> spot_effects()
14 -> pickup()
15 -> pickup_object(obj ...)
16 -> pick_obj(obj ...)
17 -> addinv(obj ...)
18 -> addinv_core0(obj ...)
19 -> merged(obj ...)
20 -> obfree(obj ...)
21 -> dealloc_obj(obj ...)
8 -> delobj(obj ...)
9 -> delobj_core(obj ...)
10 -> obfree(obj ...)
11 -> dealloc_obj(obj ...)
12 -> impossible("obj already deleted)
This marks the exploding potion with LOST_EXPLODING, so that it won't
get picked up, or merged with another object during the long
sequence of functions, and that should take care of 15-21 above.
gcc has recognized various "magic comments" for white-listing
occurrences of implicit fallthrough in switch statements for
a long time:
The range and shape of "falls through" comments accepted are
contingent upon the level of the warning. (The default level is =3.)
-Wimplicit-fallthrough=0 disables the warning altogether.
-Wimplicit-fallthrough=1 treats any kind of comment as a "falls through" comment.
-Wimplicit-fallthrough=2 essentially accepts any comment that contains something
that matches (case insensitively) "falls?[ \t-]*thr(ough|u)" regular expression.
-Wimplicit-fallthrough=3 case sensitively matches a wide range of regular
expressions, listed in the GCC manual. E.g., all of these are accepted:
/* Falls through. */
/* fall-thru */
/* Else falls through. */
/* FALLTHRU */
/* ... falls through ... */
etc.
-Wimplicit-fallthrough=4 also, case sensitively matches a range of regular
expressions but is much more strict than level =3.
-Wimplicit-fallthrough=5 doesn't recognize any comments.
Plenty of other compilers did not recognize the gcc comment convention,
and up until now the compiler warning for detecting unintended
fallthrough had to be suppressed on other compilers. That's because the code
in NetHack has been relying on the gcc approach, and only the gcc approach.
The C23 standard introduces an attribute [[fallthrough]] for the
functionality, when implicit fallthrough warnings have been enabled.
Several popular compilers already support that, or a very similar attribute
style approach, today, even ahead of their C23 support:
C compiler whitelist approach
--------------------------- -------------------------------------
C23 conforming compilers [[fallthrough]]
clang versions supporting
standards prior to
C23 __attribute__((__fallthrough__))
Microsoft Visual Studio
since VS 2022 17.4.
The warning C5262 controls
whether the implict
fallthrough is detected and
warned about with
/std:clatest. [[fallthrough]]
This adds support to NetHack for the attribute approach by inserting a
macro FALLTHROUGH to the existing cases that require white-listing, so
other compilers can analyze things too.
The definition of the FALLTHROUGH macro is controlled in include/tradstdc.h.
The gcc comment approach has also been left in place at this time.
This is additional groundwork related to
https://github.com/NetHack/NetHack/issues/1320
This additional groundwork just puts some safeguards
in place to make it rather tough to end up with an
instant death from handling a cockatrice corpse in
your inventory without appropriate protection.
At this point, still no actual petrification will occur.
Wishing is powerful, so if you cannot safely handle a cockatrice
corpse, then have a wish for one result in the corpse materializing
on the floor rather than in your inventory.
Resolves#1320
The g? structs had a mix of variables that were written to
the savefile, and those that were not.
For better clarity and to distinguish those that end up in
the savefile, relocate some g? variables that get written
directly to the savefile into different structs.
This updates EDITLEVEL, although technically it probably
didn't need to, since savefile contents are not changing.
Details:
gb.bases -> svb.bases
gb.bbubbles -> svb.bbubbles
gb.branches -> svb.branches
gc.context -> svc.context
gd.disco -> svd.disco
gd.dndest -> svd.dndest
gd.doors -> svd.doors
gd.doors_alloc -> svd.doors_alloc
gd.dungeon_topology -> svd.dungeon_topology
gd.dungeons -> svd.dungeons
ge.exclusion_zones -> sve.exclusion_zones
gh.hackpid -> svh.hackpid
gi.inv_pos -> svi.inv_pos
gk.killer -> svk.killer
gl.lastseentyp -> svl.lastseentyp
gl.level -> svl.level
gl.level_info -> svl.level_info
gm.mapseenchn -> svm.mapseenchn
gm.moves -> svm.moves
gm.mvitals -> svm.mvitals
gn.n_dgns -> svn.n_dgns
gn.n_regions -> svn.n_regions
gn.nroom -> svn.nroom
go.oracle_cnt -> svo.oracle_cnt
gp.pl_character -> svp.pl_character
gp.pl_fruit -> svp.pl_fruit
gp.plname -> svp.plname
gp.program_state -> svp.program_state
gq.quest_status -> svq.quest_status
gr.rooms -> svr.rooms
gs.sp_levchn -> svs.sp_levchn
gs.spl_book -> svs.spl_book
gt.timer_id -> svt.timer_id
gt.tune -> svt.tune
gu.updest -> svu.updest
gx.xmax -> svx.xmax
gx.xmin -> svx.xmin
gy.ymax -> svy.ymax
gy.ymin -> svy.ymin
Related note:
There are some pointer variables that are heads of chains that were not
moved from 'g?' to 'sv?', because they are not actually written to the
savefile directly, but the objects/monst/trap/lightsource/timer in the
chains they point to are. That can be changed, if desired.
Examples: gi.invent, gm.migrating_objs, gb.billobjs, gm.migrating_mons,
gf.ftrap, gl.light_base, gt.timer_base
When using #loot to put items into a shop-owned container on a shop's
floor, you are asked "Sell it? [ynaq] (n)" for each item, but the 'a'
and 'q' choices only worked as y or n for the current item. By the
next one, the preferred answer had been reset to default and ynaq was
asked again.
Set a flag in use_container() to have in_container() set the sell vs
don't sell state for the first item but not for any others. Reset
the state at the end of use_container() instead of after each item in
in_container().
This bug was present in 3.6.x, also in 3.4.3, and probably earlier.
If gold is moved into throne's coffer chest (or added to exchequer
monster's minvent) while quivered by the hero, it wasn't being unworn
to remove it from quiver slot. That could lead to a crash. Example
was segfault during 'f' command.
freeinv() expects caller to unwear item being removed from inventory;
reverse_loot() wasn't doing so.
This reverts commit 378648bd9c.
The problem was triggered by marking the 'objlist' argument in
merge_choice() prototype with __attribute__((nonnull)) when it
shouldn't have been, then a followup which relied on that. The
'objlist' argument might be Null. Instead of passing its address to
force it to be non-Null, remove the attribute.
If your inventory is full and you aren't already carrying a loadstone,
you can pick one up into the overflow slot. But if you are already
carrying one and the one you're trying to pick up won't merge with it
(only criterium that matters would be BUC state, I think), you can't
pick it up and get a message saying so. If loadstone isn't known
yet, the message always referred to it as "gray stone" rather than
"stone called <whatever-you-called-it>".
PR #1140 added checking the thrown, stolen, and dropped flags of an
item when testing whether it would merge (at my suggestion...) with
a stack in the target list (hero's invent). That interferred with
picking it back up--whether via autopickup or explicit pickup--while
inventory was full even when the item would otherwise be mergable.
There was some trial and error involved when trying to figure where
to put the fix but things seem to be working.
This replaces a static analyzer workaround and could possibly bring
its unwarranted complaint back.
Pull request from entrez: in the class filtering menu for multi-drop
and for loot in-or-out of a container, make choosing 'A' without any
other filter choices (such as all, specific class(es), cursed, unpaid,
just-picked-up, &c) become a no-op.
I started with the pull request and then undid much of it. It would
have been simpler to start from scratch. If you don't have option
paranoid_confirmation:AutoAll set, when choosing 'A' for all-without-
prompting as the only selection, operate as the PR has made things
work: effectively, 'A' by itself is ignored and the operation ends
with nothing happening.
However, if you do have paranoid_confirm:A set, then continue treating
'A' by itself as if 'A'+'a': everything. Since paranoid confirmation
is specified, that will be followed by a confirmation prompt.
This also adds a context-sensitive hint to the menu about how the 'A'
entry works, shown every time when 'cmdassist' is On or just once (per
session) when that is Off.
The documentation probably needs some updating.
Closes#1143
The 'A' option in the #loot or 'D'rop menu selecting every selectable
item when used on its own has been the cause of many bag of holding
explosions and other typo-related frustration. Now that it operates on
other filters rather than overriding them, actually require some other
filter be selected for it to have any effect. This means that 'A' on
its own will do nothing, but 'A'+'a' will still act like 'A' alone
previously did. I think this will reduce the rate of serious typo
accidents in games, without being intrusive and while still maintaining
'A' as a useful option (which I think it is, in it's 3.7 incarnation --
I use it all the time in combination with other filters, especially
justpicked).
Moving over at item that's resting on ice gives a message about there
being ice present and then about the item, whether mention_decor is On
or Off. With it On, you'll get a message about being back on solid
ground as soon as you leave the ice. With it Off you wouldn't get
that at all if not levitating; that's the basic no-mention_decor
behavior for ice. However, if you were levitating, you would get a
delayed "back on solid ground" message when moving over some other
object, which might occur quite a bit later. Autopickup handling is
calling describe_decor() when the hero is levitating and some of that
wasn't appropriate for no-mention_decor.
This issue has been present since I first implemented mention_decor,
not introduced by recent back_on_ground() changes.
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]
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 picking up multiple items (or looting multiple items) in one
operation, only show encumbrance for the first item that causes that
to change instead of complaining about non-zero encumbrance on every
item. The very first item is treated as if it caused a change, so
you get "you have a little trouble lifting <obj>" even if you were
already burdened, same as before.
Only lightly tested.
Simplify suppression of highlighting for menu header lines during end
of game disclosure. Didn't actually affect as many things as I was
expecting.
Plus a bit left out of the optfn_dogname() parsing commit.
Instead of just accepting an attribute, it's now possible to
use a color, or both color and attribute, for example:
OPTIONS=menu_headings:inverse
OPTIONS=menu_headings:red
OPTIONS=menu_headings:red&underline
Default is still just inverse.
This lets the player change the menu heading color without
needing to use menu colors for them.
Also makes it so the core uses NO_COLOR instead of 0, for all
the menu lines which don't have any prefedefined color.
Tested for tty, curses, x11, qt, and win32
Put everything through a single function that can handle all the
complicated parts of using the correct proposition for different terrain
types, and will not just call things "solid ground" indiscriminately.
This got complicated but I'm not sure if it's possible to do it much
simpler while still using the distinct names for each type of terrain
(unless you are OK with the sentences sounding sort of wonky).
When moving off a body of water with mention_decor on, describe_decor()
would produce messages like "You are back on cloud", "You are back on
wall", and "You are back on air bubble." For those types of terrain
that were producing odd sentences like this, change the format to "You
are back in a cloud", "You are back in a wall", "You are back in an air
bubble", and "You are back in the air."
Modifying an() [actually just_an()] to treat "<thickness> ice" and
"frozen <hallucinatory liquid>" as special cases which shouldn't be
prefixed with "a" or "an" affected using something like "shaved ice"
or "frozen yogurt" as named fruit.
|a) shaved ice
|b) frozen yogurt (weapon in hand)
now have article "a" preceding them:
|a) a shaved ice
|b) a frozen yogurt (weapon in hand)
However, the existing cases
|c) iron bars
|d) an iron bars (weapon in hand)
still get item 'c' wrong. 'd' is slightly odd but that's because the
fruit name is ambiguous as to whether it's singular or plural.
Classify nearby ice as "solid" (no melt timer), "sturdy" (more than
1000 turns left), "steady" (101 to 1000 turns left), "unsteady" (51
to 100 turns left), "thin" (15 to 50 turns left), or "slushy" (1 to
14 turns left, matching walking on ice with the Warning attribute).
[I'm not thrilled with "steady" and particularly "unsteady".]
I was originally going to do this just for probing downward, but ended
up also doing it for look-here and getpos's autodescribe. It nearly
got out of hand and touched more files than anticipated.
'mention_decor' ought to treat moving from ice firmer than thin to
thin or slushy, from thin to slushy, from slushy to any other, and
from thin to firmer as if moving onto different terrain but I haven't
attempted to tackle that.
The melt timer could work more like a candle's burn timer, triggering
at intermediate stages and resetting itself, so that ice which changes
to a weaker state under the hero could be reported to the player. But
this doesn't implement that.
For Autoselect-all confirmation, accept yes|n|q (or yes|no|quit)
rather than just yes|n (or yes|no). An extra query routine is needed
to support it, but the existing paranoid_query() can just call the
new paranoid_ynq() with an extra argument, keeping things painless.
For menustyle:full plus paranoid_confirmation:Autoall, if you include
'A' in the class filtering choices then you're prompted for whether
you really meant it. (Same behavior as in the past few weeks.) Yes
auto-selects items matching all other chosen classes, or from all
classes if 'A' is the only choice. (Again, same behavior.) n|no
moves on to menu:full's second menu. If anything else was chosen
along with 'A', that's what the second menu will offer. (Same as
past few weeks but revised from initial implementation.) If 'A' was
the only choice, it will now use 'a' and offer a choice of everything
in the second menu. (That's a change; it used to cancel if declining
to honor 'A' and nothing else was present.) Now <return> without
answering or q|quit or ESC skips the second menu, whether or not the
rejected 'A' was the only choice made in the first menu. (New.)
Other paranoid confirmations still just accept yes and n|no responses,
treating ESC as n|no.
When paranoid confirmation for menustyle:full's "Autoselect all" is
enabled, it is handling yes vs no backwards.
The initial implementation had it right, then a negation got dropped
when it was expanded to support yes|no via paranoid_confirm:Confirm.
ESC is being treated as 'n' rather than as cancel. Fixing that will
take some more work. This just adds the missing negation to make 'y'
and 'n' behave properly.
An issue from nearly three years ago, reported by Anerag: asking
player whether to really pray isn't paranoid enough because it
accepts 'y' rather than requring "yes".
This changes it to require "yes" followed by <return> or <enter> if
paranoid_confirm:Confirm is also set. (A side-effect of that is
explicit "no<return|enter>" is required instead of just 'n' to
decline--for all the paranoid confirmations, not just for prayer.)
This extension of paranoid:Confirm applies to paranoid:AutoAll too.
A comment asks why paranoid_confirm:pray is different from the other
paranoid questions in the first place. The answer is that when it
isn't set, no confirmation prompt is issued at all. The others all
have y/n confirmation prompts when the corresponding paranoid option
isn't set.
Once upon a time there was a boolean option called 'prayconfirm' that
issued "really pray?" prompt when True. It was added after players
whinged about typing Alt+p when they meant to type Alt+o. When the
more advanced 'paranoid_confirmation' was introduced, paranoid:pray
superseded prayconfirm, but it still only issues a confirmation
prompt where there normally wouldn't be one rather than change an
existing one to require "yes<return|enter>" instead of 'y'.
Closes#303
If confirmation for menustyle:Full 'A' choice is on and player
chooses 'A' but answers no at the "Really autoselect All?" prompt,
just remove 'A' from the list of choices and act on the rest. If it
was the only choice, the menu operation will behave as if no choices
had been made. (That was the previous behavior even when other
choices were also present.) But when other choices have been made
along with 'A', use them without 'A': the normal second menu will
be put up, listing objects matching the specified classes and not
autoselecting anything.
Issue most recently reported by Xdminsy (previously reported by
others): it is too easy to accidentally pick choice 'A' in object
class selection menus for menustyle:Full. Previous change relevant
to this was to exclude choices 'A' and 'a' from being set by '.'
(choose all entries) and '~' (toggle all entries). That was an
improvement but doesn't help with pressing shift when meaning to
type 'a' by those who type faster than they cogitate.
This implements a suggestion by janne-hmp: add new choice for the
paranoid_confirmation option, 'Autoall' (synonym 'Autoselect-all').
If the player sets this and includes 'A' among the choices for
class selection, prompt to confirm whether to honor it. Like
confirmation for praying, it adds an extra y/n prompt rather than
change an existing y/n prompt into a yes/n or yes/no one. If the
player declines, then nothing is selected and the operation is
cancelled rather than putting the menu back up to choose again.
OPTIONS=paranoid_confirm:autoall requires at least two letters
('au') even if the 'a' is capitalized. paranoid_confirm:a means
confirm attacking peaceful monsters. And it should be
OPTIONS=paranoid_confirm:autoall pray swim
if someone just wants to add autoall to the default paranoid bits.
The Guidebook hasn't been updated to describe the new choice since
it seems likely that it might undergo adjustments.
Closes#1065
Use of select-all, select-page, toggle-all, and toggle-page in the
object class selection menu for menustyle:full excluded A (auto-
select all), a (all classes), B, C, U, X (bless/curse states), and
P (recently picked up) but allowed u and x (unpaid and used-up shop
goods) to be set. Treat u and x like the other pseudo-classes;
they have to be chosen explicitly rather than be set with . and ~.
