This is an overhaul to the NetHack drawing mechanism.
- eliminates the need to have separate lists in drawing.c
for the things and their associated explanations by grouping
those thing together on the same inializer in a struct.
- replaces all of these options: IBMgraphics, DECgraphics, MACgraphics,
graphics, monsters, objects, boulder, traps, effects
- drawing.c contains only the set of NetHack standard symbols for
the main game and a set of NetHack standard symbols for the
roguelevel.
- introduces a symbols file that contains named sets of
symbols that can be loaded at run time making it extensible
for situations like multinational code pages like those reported
by <Someone>, without hardcoding additional sets into the game code.
- symbols file uses names for the symbols, so offsets will not break
when new things are introduced into the game, the way the older
config file uchar load routines did.
- symbols file only contains exceptions to the standard NetHack
set, not entire sets so they are much less verbose than all of
the g_FILLER() entries that were previously in drawing.c
- 'symset' and 'roguesymset' config file options for
preselecting a symbol set from the file called 'symbols'
at startup time. The name of the symbols file is not under the
users control, only the symbol set name desired from within the
symbols file is.
- 'symset' config file option loads a desired symbol set for
everything but the rogue level.
- 'roguesymset' config file option loads a desired symbol set
for the rogue level.
- 'SYMBOLS' config file option allows the user to specify replacement
symbols on a per symbol basis. You can specify as many or as few symbols
as you wish. The symbols are identified by a name:value pair, and line
continuation is supported. Multiple symbol assignments can be made on
the same line if each name:value pair is separated by a comma.
For example:
SYMBOLS = S_bars:\xf0, S_tree: \xf1, S_room:\xfa \
S_fountain:\xf4 \
S_boulder:0
- 'symbols' file has the following structure:
start: DECgraphics
Handling: DEC
S_vwall: \xf8 # meta-x, vertical rule
S_hwall: \xf1 # meta-q, horizontal rule
finish
start: IBMgraphics
Handling: IBM
S_vwall: \xb3 # meta-3, vertical rule
S_hwall: \xc4 # meta-D, horizontal rule
finish
- 'symbols' file added to the source tree in the dat directory
- Port Makefiles/scripts will need to be adjusted to move them into
HACKDIR destination
<Someone> wrote:
>>comments:
>>When you commit suicide with a potion of oil (lit), sometimes nethack
>>reports an `obj_is_local' error just after disclosing all the game
>>variables. This has been found in-game (don't ask) and reproduced in
>>wizard mode and in-game (start-scumming).
>
> 0) a neutral human wizard (the scrolls/spellbooks burning, potions boiling ;))
> 1) wish up 3 potions of oil (so that the 2 remaining will definitely kill you)
> 2) a'pply one of them
> 3) stand 1 square away from a wall, like "| @"
> 4) throw the lit potion into the wall (`h' in this case)
> 5) answer `yes' on all prompts.
The thrown potion of oil, which was extracted from any chain
during the throw, still had its timers attached when the call to
splatter_burning_oil() was made. If that killed the hero, a
panic would result during bones file creation (saving timers)
because (obj->where == OBJ_FREE) on the potion.
Remove the timer prior to splattering the oil inside.
If you get "you have summoned <demon>" after making a same-race
sacrifice, treat that as having been the "<demon> appears" message so that
the latter won't be given if it subsequently teleports to you after you've
moved out of sight of it. There are probably some more cases like this
still lurking in the code.
Give demon lords and other monsters who teleport to your location a
oneshot arrival message. Brought about by the report of the late "<demon>
appears" message delivered during its bribery demand, after the character
had already been able to see it for long enough to extract gold from a bag.
Now, if you can't see or sense a monster before it teleports to you, and
you can see or sense it after, you'll get "<monster> suddenly appears!".
The message will be given at most once for any given monster, and it won't
be shown at all if you already see/sense the monster before it teleports or
still don't see/sense it afterwards. The fixes entry is deliberately a bit
vague (and I put it in the new feature section rather than the fix section).
The change from long to unsigned long for monst.mstrategy may bring
some lint complaints along with it. The various constants (STRAT_xxx) used
to populate it are still signed. I didn't increment EDITLEVEL for this;
existing data should still work ok.
Allow config file entries to adjust win32 console colours.
The following entries in a config file are examples:
OPTIONS=palette:black-0-0-0
OPTIONS=palette:red-210-0-0
OPTIONS=palette:green-80-200-0
OPTIONS=palette:brown-180-100-0
OPTIONS=palette:blue-0-0-200
OPTIONS=palette:magenta-128-0-128
OPTIONS=palette:cyan-50-180-180
OPTIONS=palette:gray-192-192-192
OPTIONS=palette:dark gray-100-100-100
OPTIONS=palette:orange-255-128-0
OPTIONS=palette:bright green-0-255-0
OPTIONS=palette:yellow-255-255-0
OPTIONS=palette:bright blue-100-100-240
OPTIONS=palette:bright magenta-255-0-255
OPTIONS=palette:bright cyan-0-255-255
OPTIONS=palette:white-255-255-255
This uses an undocumented way to adjust the console
colours in a win32 console application. The method and
code snippet used comes from www.catch22.net by James Brown.
This page:
http://www.catch22.net/about.asp
states the following:
"you do not have to pay anything to use the software, and there are no
licencing terms for any sourcecode that you may download from this site.
This means you can freely use any sourcecode or portions of code in
your applications, whether they be free software or professional, retail
products."
Turn being unconscious (via several reasons, including fainted from
hunger) into a pseudo-property named `Unaware' and use it in several
places where only being asleep was checked. #H202 was about a stunned
character who got the recovery message when it timed out while fainted.
This suppresses messages for several difficulties when they begin or end
while hero is Unaware. Messages about fatal illness, sliming, or
petrification aren't suppressed; they're too important to hide from the
player. "You feel ..." messages come out as "You dream that you feel ..."
when Unaware; fairly lame but hopefully adequate.
From a bug report, being hit with a
yellow light explosion while fainted from lack of food caused blindness
but being hit while sleeping did not. Make being in fainted state become
a protection against light-induced blindness.
From a bug report, the fallback selection criteria
(used when everything is extinct?) in rndmonnum() was excluding hell-only
monsters when outside of Gehennom, but failed to exclude never-in-hell ones
when inside. [Some of the never-in-hell monsters are Angels, but the rest
are all cold based creatures. That must date to when fire resistance was
required for the hero, which is no longer the case. Should those cold
monsters retain their never-in-hell setting?]
This also fixes a latent copy/paste bug in the unused mons[] definition
of Cerberus (it was the only unique monster which failed to specify G_NOGEN).
<Someone> says he reported this five years ago, but I don't think I ever
saw it. Receiving an artifact as the result of #offer would add it into
the discoveries even if the hero was blind and never saw it. It either
needed to have the dknown bit set, as if it had been seen, or else should
not be added to the list. I've opted for the latter.
Mentioned by <Someone>, who suspected that the fix for C343-114 dealt with
this. When a giant picks up a boulder from a location where more than one
is present, line-of-sight would be inappropriately cleared, as if the
remaining boulders were transparent. (He said that he got no pickup
message when the giant picked up the boulder. I do, and have not tried to
figure out whether this is a post-3.4.3 difference.) Pushing a boulder
onto a landmine had the same bug, although you wouldn't notice unless there
were at least three boulders to be pushed (mysterious ability to push more
than one boulder in a turn strikes again: first one triggers the trap and
is destroyed, allowing hero to see through any others; second fills in the
resulting pit; third gets moved to the former trap spot and [probably--I
haven't checked] re-blocks vision; fourth, if any, stays put; if only one
or two are present, the change in vision is expected and the fact that it
happens too soon in the two boulder case would probably go unnoticed.
Polymorph of a boulder also opened up vision without checking whether any
others are present.
An existing, evidently overly optimistic, fixes34.4 entry covers this.
From a bug report, having hallucination time out while
mimicking an orange (instead of gold, after eating a mimic corpse), you'd
still get the hallucinatory end-of-mimicking message about not wanting to
be peeled. If hallucination state is toggled, update the pending message
and change the hero's appearance. In practice, only the orange-to-gold
case can occur. Anything which might trigger gold-to-orange will have
terminated the hero's mimickery befort that happens.
<Someone> reported being swallowed by his pet purple worm during
Conflict, then being stuck inside once Conflict ended. I'm not entirely
sure what dog_move() intended by the "swallowed case handled above" comment.
It returns without letting the pet move when the distance between pet and
hero is 0; that wasn't much in the way of "handling" being swallowed.
Grabbing pets did let go, but peaceful monsters didn't until you actually
attempted to move away from them. Now all four combinations (grabbed or
swallowed by tame or peaceful monster) are handled the same: the monster
will let the hero go next time it gets a chance to try to move, using up
its move in the process.
When the hero arrives on a level and a monster at his destination can't
be relocated to make room for him, goto_level() would place the hero on top
of the monster. From a bug report, who said that the
game panicked (without providing specifics, at least so far). I wasn't able
to reproduce a panic but get a pair of impossible warnings and can confirm
that there is a monster on the same spot as the hero, which could easily
lead to strangeness depending upon what actions the monster attempts to
perform. This fix causes a non-relocateable monster in that situation to
be moved to the migrating monsters list for later arrival back on that same
level. That's inconsistant with the migrating monster arrival routine,
which kills off any monster it can't place; I'm not sure which action is
more reasonable, deferred arrival or outright removal.
There are three or four dozen ``(void) rloc(mon)'' calls which don't
do anything special when rloc fails to move the monster. Those need to be
reviewed and the ones where it's making a space for some other monster have
to do something about failure. (Failed teleport attempts can simply leave
the monster in place, so most of those calls won't need any extra handling.)
Forwarded from the newsgroup by <Someone>: temple priest might
abandon his post via teleport if conditions are obscure enough. Change
rloc_pos_ok() to only accept spots inside the same shop or temple when a
shopkeeper or temple priest is teleported to a random destination. rloc()
tries rloc_pos_ok() 500 times before reverting to goodpos(), so this will
usually succeed for a large room; it may fail for a small one (reverting
to the current behavior, more or less). Shopkeepers or priests who get
polymorphed into a critter which teleports to the stairs when in need of
healing will still leave their shop or temple if wounded (no change).
Priests resist if the player tries to teleport them, but shopkeepers
don't. So for direct attack by the player, this only affects shopkeeper
destination. But it affects both types as far as being hit by quantum
mechanics (probably caused by player's use of conflict) or if polymorphed
into monsters which steal and then flee (again, probably caused by the
player since those strong monsters won't voluntarily polymorph).
From a bug report...): pushing a boulder onto a level teleporter trap
could repeat the
You push the boulder and suddenly it disappears!
message. That would happen whenever the teleport destination was the same
as the current level (20% chance). The boulder wasn't being moved onto the
trap location so was still present when the pushing code tried to handle
the next one in the pile. I've changed so that pushing stops whenever a
pushed boulder is affected by a trap, and also so that the boulder gets
moved as usual when a level teleporter fails to send it somewhere.
I've always thought it's pretty strange that pushing ever operates
on more than one boulder in the same turn in any situation, but I haven't
changed that for the non-trap cases. (Usually the first boulder pushed
ends up blocking the second one, so you get a "you try to move it, but
in vain" message which seems odd since you just moved one. But if there's
a pool of water or lava in the path, you can actually push multiple
boulders successfully.)
From <Someone>:
Pet picks up 8 spears.
Pet wields 8 spears.
Pet thrusts its spear at Foe.
The routine to handle a monster attacking the hero are already uses "mon
thrusts one of its spears" in this case, so make monster against monster
messages do the same. Also, it's no longer necessary to save one monster
name before formatting another when using two monster names in the same
message, so switch to the more straightforward usage here. (The Blind
check is needed in the mhitu case but not in the mhitm one, where it's
redundant because the caller has already verified that both monsters'
locations are visible, but I left it in.)
More tweaking brought about by the report of falling through a
trapdoor above the quest locate level and landing below it (which won't
happen randomly any more, but can still be achieved via controlled level
teleport). You don't get any message for the locate level when arriving
from below, and giving the full initial message if you manage to make a
return visit from above produces silly results. This adjusts the logic
for delivering those messages; once you've been on the level, you won't
get the full initial message later even if it wasn't given the first time.
You will still get the shorter secondary message, up until the nemesis has
been killed. (In some cases it might not make much sense since its wording
is based on the assumption that you've already seen the full message.)
The handling for these messages will never be completely correct
unless the messages themselves are rewritten (and the result would
probably end up with really wishy-washy phrasing so there's not much point).
They assume that you'll be arriving via the stairs, even for the case where
you do come from above rather than unexpectedly from below, and they can
be misleading or confusing if you arrive someplace else. We could improve
things by having alternate FIRSTLOCATE and NEXTLOCATE messages for use when
not on--or next to, for the case where a pet displaces your arrival--the
stairs. But I'm not going to attempt to compose those.
From a bug report, it was possible
to fall from above the quest locate level to below it even though it has
nondiggable floors (hence no trapdoors or holes to pass through). Since
the game can't verify that it is nondiggable (which could vary on a role
by role basis depending upon their quest descriptions) when it is not the
current level, restrict falling past it only when it hasn't been visited
yet. Impose the same restriction for random level teleport. This enforces
proper sequencing of the quest feedback, which was the more significant
thing that went wrong in the reported case (player finally got the full
locate level message on a return visit, when descending from above, after
having already cleared out that level on his way back up from falling).
Once the locate level has been reached or passed, it is no longer a
barrier to falling or random teleport. (When it is eventually visited,
there's no attempt to remember whether it allows holes, since that
information and the corresponding fall check would need to be extended to
every level in the dungeon. Also, controlled teleport is still allowed to
bypass it even when it hasn't yet been visited, so the "enforces proper
sequencing" claim above is an exaggeration.)
From a bug report, applying a pick-axe will
wield it in place of the primary weapon but wouldn't update the alternate
weapon. Make objects which become wielded when applied (pick, lamp, whip,
grappling hook, pole-arm) honor the pushweapon option (where the previously
wielded weapon becomes the alternate weapon without use of 'x' command) the
same as when explicitly wielded. Old behavior matched the documentation,
but that seemed overly restrictive.
The devteam feedback was to place casts in the code
in question.
This puts explicit casts on some code that was being
compiled into 'int64' then stuffed into smaller types with
VC2005.
I'm pretty sure that some pre-standard compilers don't know how to
apply an initializer to a variable of type union. Unfortunately, I don't
have access to one to check. Fortunately, there's no need to explicitly
initialize `zeroany' since the default value is what we want--the first
field will be set to zero or null as appropriate (null in this case).
Strictly speaking, this isn't adequate; what if long is wider than a
pointer rather than narrower? Using `= {DUMMY}' didn't handle that case
either; the ordering of the union's fields controls which bits get stored.
As a practical matter, it should make no difference. As long as the code
reading a union accesses the same field as the code writing that union set
up in it, anything in extraneous bits should be irrelevant--except perhaps
when a debugger tries to format things. The whole issue has always been
implicitly based on the assumption that null pointers have all bits zero
in the first place; that's typical but not guaranteed.
botl.c conversions. All the ports seem to be using genl_status_update(),
rather than a window port specific version, so botl.c was the only place
this had to be adjusted.
Also a uudecode cast for the result of strlen, since it isn't using
config.h
Remove some more code that forced pointers into a long int, and
vice versa where information could be lost (P64 platforms such as
WIN64 have a 64 bit pointer size, but a 32 bit long size.)
This 3rd part deals with region functions switching
some arguments from type genericptr_t to 'anything'.
Like the previous 2 parts, this needs to increment
EDITLEVEL in patchlevel.h.
Remove some more code that forced pointers into a long int, and
vice versa where information could be lost (P64 platforms such as
WIN64 have a 64 bit pointer size, but a 32 bit long size.)
This 2nd part deals with timeout functions switching
some arguments from type genericptr_t to 'anything'.
Like part 1, this needs to increment EDITLEVEL in patchlevel.h.
[the problem in the earlier rev was tracked to cleanup_burn(),
where arg was holding a (genericptr_t) timer id, and
passed directly to del_light_source() as is.]
P64 (Win64) has a 64 bit pointer size, but a 32 bit long size.
Remove some code that forced pointers into a long int, and
vice versa where information could be lost.
This part deals with light source functions and their
arguments mostly, and switches some arguments
from type genericptr_t to 'anything'.
P64 (Win64) has a 64 bit pointer size, but a 32 bit long size.
Remove some code that forced pointers into a long int, and
vice versa where information could be lost.
This part deals with light source functions and their
arguments mostly, and switches some arguments
from type genericptr_t to 'anything'.
Hide pointer formatting in alloc.c by eliminating the need for callers
to know how big a buffer is required. I generally prefer the caller to
pass in its own buffer for this sort of thing, but in this case the usage
is almost entirely for debugging so using static buffers results in less
clutter in the rest of the code.
> Michael Allison wrote:
> There are unresolved functions in the trunk if you build without WIZARD
> defined and attempt to link:
Also to a lesser extent, in the 3.4.4 branch.
<Someone> wrote on Tuesday, July 27, 2004 at 06:46:15
> In the region.c function rest_regions allocates storage for the possible
> enter_msg and leave_msg strings. But the function free_region does not relese
> this storage.
Also ensures that some code that is currently ifdef'd out
makes copies of the strings into memory from alloc()
to ensure that no problems with free() result if the function
gets passed a literal string.
<Someone> wrote on Tuesday, July 27, 2004 at 06:46:15
> In the region.c function rest_regions allocates storage for the possible
> enter_msg and leave_msg strings. But the function free_region does not relese
> this storage.
Also ensures that some code that is currently ifdef'd out
makes copies of the strings into memory from alloc()
to ensure that no problems with free() result if the function
gets passed a literal string.
> NetHack feedback form submitted by
> <email deleted> on Friday, June 30, 2006 at 17:31:12
> ---------------------------------------------------------------------------
>
> mailversion:1.35
>
> nhversion:3.4.3
>
> nhfrom:Our 3.4.3 source release, unmodified
> comments:
> telnet nethack.alt.org with the terminal set to 21 rows.
> Choose to pick a char, not accept pot luck, and game segfaults.
> (same happens from linux console)
I was able to reproduce something similar in win32 by setting
the console to 21 rows. As he stated, don't let the game pick you
character for you to reproduce the problem. As soon as I chose
Archeologist the problem occurred:
Where:
In hack.c, weight_cap()
if (Levitation || Is_airlevel(&u.uz) /* <email deleted> */
#ifdef STEED
|| (u.usteed && strongmonst(u.usteed->data))
#endif
)
Variables:
carrcap 200
u.usteed 0x00000000
&u.uz 0x005e54aa
youmonst.data 0x00000000
Examination of the preprocessor output of that section
of code reveals that
"Levitation" becomes:
(u.uprops[47].intrinsic || u.uprops[47].extrinsic ||
((youmonst.data)->mlet == 5))
so it is the is_floater(youmonst.data) causing the crash.
Call stack:
weight_cap() line 2300 + 24 bytes
inv_weight() line 2342 + 5 bytes
calc_capacity(int 0) line 2354 + 5 bytes
near_capacity() line 2365 + 7 bytes
bot() line 607 + 5 bytes
docorner(int 47, int 19) line 2378
erase_menu_or_text(int 5, WinDesc * 0x00a22550, char 0) line 994 + 25 bytes
tty_dismiss_nhwindow(int 5) line 1664 + 15 bytes
tty_select_menu(int 5, int 1, mi * * 0x0006fc40) line 2248 + 9 bytes
tty_player_selection() line 442 + 16 bytes
pcmain(int 3, char * * 0x00a20eb0) line 457
main(int 3, char * * 0x00a20eb0) line 91 + 13 bytes
This adds a check for a valid youmonst.data in
bot().
