- gcc warned that the "anything" initializer needed more braces
- gcc also warned of a couple unused variables
- WIN_STATUS, when STATUS_VIA_WINDOWPORT is not defined, had no type
It was possible for status_finish to get called twice, therefore free() could end up called twice.
Add a macro symbol for the argument to status_initialize().
Adjust some field widths so that there is a little bit of breathing space.
Add the ability to select the windowtype on the command line on Unix using
a new -wwindowtype option. I had thought the proposed patch could core
dump, but the default windowtype selection occurs earlier and ensures
that raw_print will always work. So, the only problem with the proposed
patch was it didn't move the linux and sco special-case code until after
the selection was made. That special-case code really should be moved to
to wintty.c, IMO since it doesn't affect other windowtypes.
The default symbol for lit and unlit corridors are the same. This makes
the lit_corridor option a no-op where the defaults are used and also means
that using a light scroll/spell a corridor has no obvious effect. To
address that, I special-cased the lit corridor symbol and change its color
to bright white when its the same symbol as the unlit corridor symbol (I
didn't change the default color since I thought that made the lit corridor
look strange using the windows console interface).
DEC C warned about a problem and that caused make to quit.
| idx = (++idx % 2);
|........^
|%CC-W-UNDEFVARMOD, In this statement, the expression "idx=(++idx%2)"
|modifies the variable "idx" more than once without an intervening
|sequence point. This behavior is undefined.
Introduction of a new set of window port status display
routines. The new routines are conditional on
STATUS_VIA_WINDOWPORT
being defined in config.h. See the experimental section,
where the #define resides for the time being.
Introduction of a new set of window port status display
routines. The new routines are conditional on
STATUS_VIA_WINDOWPORT
being defined in config.h. See the experimental section,
where the #define resides for the time being.
Introduction of a new set of window port status display
routines. The new routines are conditional on
STATUS_VIA_WINDOWPORT
being defined in config.h. See the experimental section,
where the #define resides for the time being.
Just From a bug report: getting interrupted and then
resuming would sometimes produce two instances of the "You finish" message
(and evidently consumed an extra turn in the process). I think this is
an old problem and that it's just coincidence that it showed up right after
the patch dealing with avoidance of stale context for 'A'; the interruption
has to occur when there is just one turn left in removing the final item
so doesn't happen very often.
Prompted by the report that it was possible to check a corpse at an
adjacent spot while levitating: substantially revamp the handling of
applying a stethoscope to corpses and/or statues. Aside from the missing
reachability check, the old code suffered from grammar problems when
multiple corpses were present and it didn't try to figure out the gender
of the monster who left the corpse.
This now lets non-healers check corpses and statues at any time
(previously that only worked for them while hallucinating). For healers
it gives some new feedback: when listening to a corpse, they'll be given
a hint if it (actually any corpse in the pile at that location) is going
to revive, and when listening to a statue they'll be given a hint if it
contains any items. The existing hint when a statue is actually a trap
is still only given to healers.
Groundwork for forthcoming stethoscope fix: add an object list
traveral routine (there must be something like this already, but I
couldn't find it...) and a couple of timer access routines.
For now, the code is conditional on BARGETHROUGH
being defined, while it gets tested further. While behavior is
different with and without BARGETHROUGH defined, savefiles
are the same either way.
After this patch is applied, only the riders have the M3_DISPLACES
bit set, but the Wizard and Vlad probably should too. Any others?
Oops, I spotted a typo that I made in the new text.
For the branch, there is no need to re-synch, as the
the generated .txt version is also patched with this patch.
An optimization.
Split the patterns into two chains, one for things you want "always picked up", and one for things you want "never picked up". That way, the
entire exception list isn't being checked twice, once for each type of
check. The check for things to "always pick up" only checks the
chain at the AP_GRAB index, and the check for things to "never
pick up" only checks the chain at the AP_LEAVE index.
- Also modifies the toggle autopickup '@' message
appropriately for the AUTOPICKUP_EXCEPTIONS build.
(it just tacks on ", with some exceptions" if there are some)
Guidebook used both "behavior" and "behaviour". I think this was the
only word not using the US-English spelling. Also fixed a spacing bit
in Michael's latest rev to the .mn version.
[ Caveat: compiles ok on branch code but only play tested on trunk code;
the do_wear.c diff is a lot different between the two variants and the
trunk one includes some whitespace cleanup. ]
<email deleted> reported that having a spellcasting monster
destroy some armor while you're in the midst of using 'A' to take that
armor off would result in a crash. The problem was actually more
widespread than that: having a nymph steal worn items (accessories as
well as armor), or a succubus remove them, or being interrupted by monster
activity and then reading a scroll of destroy armor prior to resuming 'A'
could all produce a similar crash. 'A' relied on stale context and could
attempt to manipulate an equipment slot which had become empty, ultimately
leading to an attempt to dereference a null pointer.
The 'R' command didn't have this problem since any accessory gets
removed immediately. The 'T' command already had handling for this:
there's only one item to deal with and multi-turn take off only applies
to some of the slots; the donning() check followed by cancel_don() took
care of those. Only 'A' was vulnerable to the problem and it wouldn't
necessarily need to be interrupted and resumed; loss of the current
multi-turn item or any pending item would be enough--but I'm not sure
whether such item loss could occur without also interrupting the current
activity, so resumption of previous 'A' was probably a requirement for
triggering the crash.
This makes shield and shirt handling be similar to other types of
armor instead of relying on the fact that none of them need to have any
attribute adjustments when put on or taken off. However, there are
still assumptions (the `cancelled_don' stuff) that some slots don't have
any eligible items requiring more than a single turn to use; that should
probably be changed.
External names longer than 31 characters trigger a compiler warning
for me about truncation, and that causes make to quit. So shorten the
two long names. Also, call the cleanup routine for the FREE_ALL_MEMORY
configuration.
An enhancement to the original to allow
two types of exceptions:
always pickup
never pickup
To specify "never pickup", begin the pattern string with '>'
which is NetHack's down direction symbol (for "leave it down")
To specify "always pickup", begin the pattern string with '<'
which is NetHack's up direction symbol.
For example, to pick up all arrows, regardless of the pickup_
types setting:
autopickup_exception = "<*arrows"
Add config.h experimental option AUTOPICKUP_EXCEPTIONS.
It's an interface-only change which allows you to add lines to your
config file to selectively avoid autopickup of items based on their
text description that is displayed when you pick them up. It does
it by matching a pattern against the xname singular return value.
For example:
autopickup_exception = "*corpse" will avoid picking up corpses, even if
food (%) is in your pickup_types.
autopickup_exception = "*brown*"
will avoid picking up any brown items (why, I do not know)
autopickup_exception = "*loadstone"
will NOT avoid picking up loadstones, unless they are already
identified, because the xname string will be "gray stone", so no
match there.
The matching has no knowledge of in-game objects, it is just
a text pattern match, thus it is an interface change, not a gameplay
change, and it is meant as a convenience for players.
There was code higher up that could destroy the obj,
so this eliminates any risk of trying to dereference a
bad pointer in the deferred silver message by saving
a copy of the name earlier on.
From the newsgroup:
> <email deleted> (<Someone>)
> Newsgroups: rec.games.roguelike.nethack
> Subject: Question: Why don't silver wands give silver damage?
> Date: 9 Nov 2003 09:18:50 -0800
> Organization: http://groups.google.com
> Lines: 7
> <email deleted>
>
> I had a character cornered by a werejackal the other day. I'm not too
> bothered by the death but why didn't the silver wand he desperately
> wielded in his final moments do extra damage? I mean, silver rings do
> so why not wands? I realise this is a pretty minor problem since not
> that many people will be going around wielding wands, but still.
>
> ~<Someone>
There was a code path for objects such as wands that avoided
all the silver checks. Now fixed.
Acknowledge Schroedinger's cat at end of game.
- determine cat's status:
- give points for the animal which accompanied
you.
or
- include dead cat in the box contents.
Schroedinger's cat is the only ordinary creature
that could actually ascend with you.
This patch doesn't deal with any supernatural creatures
including:
- djinn in bottles
- ghosts in bottles
