Remove start_screen() and end_screen() from the
Window-port interface.
They were only ever used by tty, and there was a comment
carried to several window-ports about how they "really
should go away. They are tty-specific"
term_start_screen() and term_end_screen() are part of
terminal/NO_TERMS supporting routines now.
Add a note about NO_TERMS to include/wintty.h for clarity.
Rename tty_startup and tty_shutdown to term_startup() and
term_shutdown(). They are found in termcap.c for !NO_TERMS
like most of the other term_ routines, as well as having
versions for several of the NO_TERMS platforms. They aren't
part of the tty_interface called from the core. The tty
implementation does call and rely on them.
Remove some conditional #ifdef's around term_shutdown()
(formerly tty_shutdown()) and just ensure that all the
tty platforms have an implementation that they can link
with, even if it is just a stub presently.
Put the protype for nethack_exit in extern.h to reduce
maintenance to a single spot, and remove it from other
locations. A warning in the msdos cross-compile led to
this change.
Using #monster to make the steed use the breath weapon often
failed because the steed did not want to breathe at weak or
too strong monsters.
Make #monster force the steed use the breath, and if there is
no targets available, make the steed make some noise as feedback.
If water walking boots haven't been discovered yet and underwater
hero rises to the surface when putting a pair on, discover them.
(Sinking while removing such on water already discovers them.)
Options processing can be early, even before ttyDisplay is allocated.
If we find that TTY_PERM_INVENT initialization is happening too early,
just set a marker (iflags.perm_invent_pending) to try again a bit later.
The changes in win/share are just to be able to sucessfully
reproduce the original issue on Windows. It was easily reproduced
on Unix, just by building with TTY_PERM_INVENT in include/config.h
and setting OPTIONS=perm_invent in config file.
This originated with a bug in NerfHack in which the developer specified
an inventory for a quest nemesis, but neglected to include the Bell of
Opening in it. Since monsters' inventory contents from makemon() were
tossed out completely, this caused a situation where the Bell was
deleted and the game was unwinnable. The first part of this change is
guarding against that by adding mdrop_special_objs before discarding the
inventory. This does create a possibility where if the programmer *does*
specify a nemesis get the Bell item in their inventory, while neglecting
to remove its special case generation in makemon.c, it would generate
twice - but two Bells is better than none.
Working on that fix led me to think about a limitation of the current
sp_lev.c behavior. You could either have a monster generate with its
species-typical inventory by not specifying an inventory for it, or you
could have it generate with custom inventory but then have to use that
to clumsily reproduce the normal inventory's complex chances and
conditionals in mongets(). So the remainder of this commit implements
another flag for des.monster(), keep_default_invent, that allows for
more flexibility in two ways:
1. When des.monster() contains an inventory function and
keep_default_invent is true, the monster will retain everything it
gets from makemon() and the objects in the inventory function are in
ADDITION to those. This is useful for augmenting a monster's default
kit with something to make them more threatening, or just more loot.
2. When des.monster contains no inventory function and
keep_default_invent is false, the monster will get NO inventory even
if its species is normally supposed to. I'm not sure where exactly
this would be used, but it doesn't hurt to have it available.
When keep_default_invent is not specified at all, the behavior remains
the same as it is now - if inventory is provided, default items are
discarded, and if not, they are kept.
This helps avoid a potential chicken-and-egg scenario
with the system configuration file (sysconf).
If sysconf wasn't accessible at the expected location, it
caused an immediate exit, without relaying any helpful
information. That happened even when using:
'nethack --showpaths'
That's particularly unhelpful, because the --showpaths
output might have been useful towards understanding where
NetHack was looking for such things.
That left you without an easy recourse to identify where
the game is looking for the sysconf file. That might be
especially troublesome if you didn't build the game
yourself.
There was a transcription error in the comments in cstd.h for
the standard list of header files, where only the description
remained for <stdlib.h>, not the name of the file itself.
Remove several extraneous inclusions of the standard C99 headers.
Tested on the following afterwards:
Linux (using hints/linux.370) including tty, curses, qt6, and X11
macOS (using hints/macOS.370) including tty, curses, qt5, and X11
Windows MSYS2 using sys/windows/GNUmakefile
Windows Visual Studio using sys/windows/Makefile.nmake
msdos cross-compile on Ubuntu using djgpp cross-compiler
Issue reported by k21971: winter wolf cub and hellhound pup were
defined with alignment -5 (chaotic) while winter wolf and hellhound
were 0 (neutral).
K2 suggested that winter wolf plus cub both be neutral and hellhound
plus pup both be chaotic but I've gone another way: both cub and
pup are now 0 and both adults are -5.
Fixes#1340
Various bits I had in progress before Michael's commit.
Mainly forget engravings when bones are saved instead of leaving them
flagged as seen for the next hero who gets the level.
Since EDITLEVEL is being incremented for the previous patch anyway,
add the "name_from" bit to the obj struct now, as groundwork for
the code change mentioned in a TODO comment in bones.c:
/* strip user-supplied names */
/* Statue and some corpse names are left intact,
presumably in case they came from score file.
[TODO: this ought to be done differently--names
which came from such a source or came from any
stoned or killed monster should be flagged in
some manner; then we could just check the flag
here and keep "real" names (dead pets, &c) while
discarding player notes attached to statues.] */
if (has_oname(otmp)
&& !(otmp->oartifact || otmp->otyp == STATUE
|| otmp->otyp == SPE_NOVEL
|| (otmp->otyp == CORPSE
&& otmp->corpsenm >= SPECIAL_PM))) {
free_oname(otmp);
}
Also, the bitfield per-byte groupings identified in the struct
weren't accurately reflected, so rearrange the Bitfields,
and correct the per-byte groupings.
This invalidates existing save files and bones.
GitHub issue reported by ars3niy:
https://github.com/NetHack/NetHack/issues/1303
@ars3niy commented on Oct 27:
Stepping on a square with a dust or "graffiti" engraving while blind
produces no message because presumably you can't read them by swiping
the floor with your hands, however the engraving glyph still shows up on
the map afterwards. While this helps zen players, it looks like a bug.
@ville-v commented 3 days ago:
Searching while blind also reveals the engravings. Here is a save file
demonstrating the issue.
[...]
This adds an erevealed bit to engravings, to accompany the the eread
bit that is already there.
eread: refers to the text of the engraving
erevealed: refers to the engraving map symbol
Hopefully, this resolves issue 1303 without creating additional bugs.
This invalidates existing save files and bones.
Fixes#1303
After finding a trap on a chest or a large box, remember it
as trapped: "You see here a trapped large box."
Randomly generated chests and boxes can be obviously trapped.
Allow defining obviously trapped containers via lua.
Invalidates saves and bones.
ENHANCED_SYMBOLS is defined by default in config.h.
The msdos build tried to #undef ENHANCED_SYMBOLS
in tilemap.c, but doing it in there created a mismatch
between the data struct definition for glyph_map in wintype.h
and the initializers generated in tilemap.c
Move the msdos build catch for ENHANCED_SYMBOLS to
one single place in config1.h so that the code and data agree.
Issue reported by loggersviii: attempting #untrap from an adjacent
doorway can move the hero diagonally out of the doorway.
A followup comment by elunna pointed out that a monster's attack that
results in knockback can produce similar result.
Fixes#1305
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).
I don't think this solves the recent light source reports,
but it changes a couple of things in an attempt to get more
information.
1. Having gy.youmonst.m_id field always be zero makes it tough
to distinguish it from uninitialized memory, or a random memory
value. This changes the m_id for the hero's gy.youmonst.m_id
to always hold the identifier 1, instead of 0.
2. write_ls was taking the stashed pointer in the light source,
and using it to immediately extract the m_id field and search
for that m_id. This changes the approach slightly, to actually
try and locate the stashed pointer itself in one of the monster
chains. Only if the monster pointer is located, do we dereference
it to obtain the m_id field.
3. For the interim, mark the saved ls with another set bit when
there has been a failure to locate the monst. At this time,
no code is acting on that bit, but it can be seen in a debug
session.
Hopefully, the next report will provide enough information to
understand the scenario a little better.
Using 'f', if hero is wielding a polearm, and a monster is in range,
don't switch away even if we do have ammo in quiver and a launcher
in the inventory.
Experience equivalent to killing a monster is gained when starting a turn
adjacent to and being able to see the monster.
Breaks saves.
Idea and parts of code via dNetHack
GNU make looks first for a file called GNUmakefile, ahead of
looking for Makefile and then makefile.
Renaming sys/windows/Makefile.mingw32 to sys/windows/GNUmakefile
allows:
o src/GNUmakefile (for use by GNU make) and src/Makefile (for use
Microsoft nmake) to both reside in the src folder during build.
o src/GNUmakefile will be used by GNU make, without having to
explicitly specify "-f GNUmakefile" on the GNU make command line.
o src/Makefile will be used by Microsoft nmake, without having to
explicitly specify "-f Makefile" on the Microsoft nmake command line.
For the gcc build, the movemement of sys/windows/GNUmakefile needs
to be copied to src/GNUmakefile as part of the build process (see
sys/windows/build-msys2.txt).
For the Microsoft Visual Studio command line build with nmake,
sys/windows/Makefile.nmake needs to be copied to src/Makefile as
part of the build process (see sys/windows/build-nmake.txt).
They are both copied to the src folder from their respective
repository source file names when the nhsetup.bat file is used.
When a mimic in door form is hit by a wand of locking or wand of
opening or corresponding spell, bring it out of concealment like
was recently done for being zapped while in chest form. And give
some feedback rather than just changing the mimic's form to 'm'.
Give more detailed feedback when bumping into a mimic while moving.
Because newsym() would be called only on the head position of the worm,
if hilite_pet was on, the segments and head of a long worm would have
mismatched highlighting in the immediate aftermath of taming or
untaming.
There was an issue with Windows mingw build because the function
prototypes were not available. Place them into a distinct
header file nhregex.h and include it from extern.h, and
available for cppregex.cpp to include without the rest of
extern.h (which can give some problems with c++).
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.
Add enlightenment feedback for Sunsword's blocking of becoming blind
from light flashes. It uses an extra property so that wizard mode
can report the reason.
EDITLEVEL is being incremented, so existing save and bones files are
invalidated.
In sys_early_init(), the values for sysopt.gdbpath and
sysopt.greppath were being assigned by calling dupstr()
without ensuring that a value previously assigned by
dupstr() were free()'d.
Also, the sysopt struct definition is changed to sysopt_s,
not because there's anything wrong with the original
struct sysopt sysopt;
but because I couldn't convince the debugger to use the
correct thing when trying to track down the leak.
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
Related to #1309https://github.com/NetHack/NetHack/issues/1309
K2 commented: "This might help - k21971/EvilHack@afed641"
A comment in there states:
"Fix: sections of wall being visible when they shouldn't yet.
This has been a long-standing bug for as long as I can remember, and qt
appears to have figured it out. What was happening: the player would all
of the sudden see a section of wall in an area that they hadn't explored
yet. It was discovered that this was only occurring if that section of
wall had any type of tree up against it."
The fix there attempts to leave trees out of the check_pos non-zero return,
so give that a shot.
I didn't attempt to reproduce the situation myself,
and therefore cannot confirm that this does resolve it.
Feedback on effectiveness or side-effects are welcomed. If someone is
able to confirm that this resolves the issue without creating new
issues, we can close it, otherwise this can be reverted.
