> On 01/30/2012 08:20 PM, <Someone> wrote:
> The boulder from a rolling boulder trap can be generated on a
> lava pool. mkroll_launch() in trap.c, line 1584 checks only for pools
> of water.
Add dupstr() as a substitute for strdup() so that out-of-memory
handling will be consistent with the rest of nethack, and make it aware
of nethack's heap logging. It's treated like alloc() so that its caller
can be logged for NH_HEAPLOG.
I put it into use in a few places, but there are lots more candidates
besides the existing calls to strdup() that should be replaced.
While testing something, I noticed that I could eat a tin (off the
floor) while polymorphed into a bat. The code to check whether the hero
could open a tin was testing for limbs, so winged critters passed. Now
it requires hands instead of limbs, and also that the current form be big
enough to be capable of wielding something (even though you don't need to
be wielding anything to open a tin).
This means that a hero poly'd into a dog or cat will no longer be
able to serve him-/herself dinner from a tin....
Simplify many of the intrinsics macros from
#define xxx_resistance (Hxxx || Exxx || resists_xxx(&youmonst))
down to
#define xxx_resistance (Hxxx || Exxx)
by setting or clearing an extra bit in Hxxx during polymorph so that the
resists_xxx() check becomes implicit.
Unfornately there were lots of places in the code that treat Hxxx
as a timeout number--primarily for Stunned, Confused, and Hallucination;
Stunned happens to be one of the revised macros--rather than as a bit
mask, so this patch needed a lot more changes than originally antipated.
<email deleted> suggested that eating a cursed apple
give a Snow White reference. The apple eaten by Snow White is described
as poisoned, but cursed seems close enough. Fall asleep for 20..30 turns
if you eat a cursed apple when you lack sleep resistance.
While looking at fixing the mfrozen issue for monsters (there's no
way to tell whether it's been caused by sleep or paralysis, necessitating
that some messages be vague or suppressed when actions impact monsters
who can't move), I noticed a drawbridge bug for the hero. It was using
the misleadingly named Sleeping intrinsic incorrectly. When that is
nonzero, the hero is prone to falling asleep at random intervals, not
necessarily asleep right now. I've always intended to rename it to
something that's not misleading, but hadn't ever gotten around to doing
so, until now: change the SLEEPING property to SLEEPY and the Sleeping
intrinsic/attribute to Sleepy.
This may be moot for the drawbridge. I can't remember any hero ever
jumping to safety instead of being crushed by either the bridge or its
portcullis, and I'm sure sleepiness hasn't been a factor. So I haven't
included any fixes entry about misusing Sleeping when it meant u.usleep
(or better yet, unconscious(); or even better, Unaware [a post-3.4.3
pseudo-property that tests both unconscious() and fainted() when checking
whether hero is incapacitated]).
Contributed by <email deleted>, give an
alternate message when eating a corpse while hallucinating, including one
that gives homage to Tony the Tiger from old Frosted Flakes commercials
if you happen to be polymorphed into a tiger. Even players who try to
keep their characters hallucinating all the time are unlikely to ever run
into "tastes gr-r-reat!".
I rewrote the conditional expression to only test Hallucination once.
And I added the comment about omnivores, who'll never get "is delicious"
result with the current carnivore vs herbivore logic. (That behavior has
been around for quite a while, but seems somewhat suspect.)
MAC is defined for MacOS 9 but not for 10 (and it shouldn't be). Add a MACOSX
define and use it in eat.c to pick up the joke meant for all Mac systems.
Suggested by <Someone> <email deleted>, fainting from lack
of food while wielding a cockatrice corpse should fall on that corpse
and be fatal, like triggering a pit or moving down stairs while too
heavily encumbered. I put the fixes entry in the new features section
even though it probably qualifies as a bug.
I almost abandoned this when Michael beat me to it, but besides
handling the fruit rename bug it also moves `current_fruit' into the
context structure to eliminate separate save/restore for that.
A change yesterday made putting on an amulet of restful sleep avoid
clobbering the timeout from having already eaten one, only replace it if
the new timeout is shorter. This does the inverse; when eating one, if
you're already sleepy from also wearing that type of amulet, only replace
the timeout if new one is shorter. And don't clobber the other intrinsic
bits with FROMOUTSIDE, just add it to whatever ones might already be set.
Neither should have any observable effect on game play, so no fixes entry.
Implement something <Someone> suggested a long time ago: eating a
disenchanter corpse has a chance to remove an intrinsic. Uses the same
routine as nighttime gremlin attacks, which chooses an intrinsic randomly
and attempts to remove it, so has no effect if it chooses one the hero
lacks. This can be used to remove "aggravate monster" but is much more
likely to target something the player wants to keep. [By the way, a lot
of potential candidates are missing: sleep, shock, and disintegration
resistance and teleport control come immediately to mind.]
Also, it has been bugging me that you can get both strength and
fire/cold/shock resistance from the same fire/frost/storm giant corpse.
The code prevents mind flayer corpses from conferring both intelligence
and telepathy, so strength handling was inconsistent (even though it
predated mind flayers...). This causes strength boosting to be treated
as an extra candidate when selecting an intrinsic to confer, so you'll
either get strength or resistance (which might be a no-op) but not both
from those giants. And it special cases the other giants to have the
same 50% chance for boosting strength, even though the alternative in
their case is to do nothing instead of trying to confer something else.
Lastly, it now gives a message when you succeed in gaining strength.
From the newsgroup:
The tin opens like magic!
<some interruption occurs>
You stop opening the tin.
Either it opens immediately or it doesn't, so the "opens like magic"
message is inaccurate. Rather than simply changing the phrasing, this
gives blessed tins a 50% chance to really open immediately so that their
contents are available for eating on the same turn, and 50% to behave as
before but with a message which is suitable for the single turn delay.
Hero poly'd into a metalivore always gets the same-turn case when eating
any tin. Use of a tin opener has a chance to do so (always when blessed,
50/50 for same-turn vs 1 turn delay when uncursed, 33/33/33 same-turn or
1 or 2 turn delay when cursed).
Overall, blessed tins are better than they used to be, since half
of the time you'll save a turn, but they're still not reliable to eat in
the midst of combat since sometimes you'll need another turn and will be
likely to get interrupted in that situation. Uncursed tin openers still
give the same behavior as opening blessed tins, so are also better than
they used to be. Blessed tin openers are now superior, and cursed ones
are slightly inferior in addition to being welded to hero's weapon hand.
For GOLDOBJ configuration, relax the 52 object limit for inventory
when gold uses the special $ slot instead of a letter. Takes care of an
old buglist entry from the beta testers. [It will need to be revisited
if we ever implement multiple coin types that can't all fit in one slot.]
Also for GOLDOBJ, prevents nymphs and monkeys from stealing coins,
since allowing that made their steal-item attack be a complete superset
of leprechaun's steal-gold attack.
New macro slimeproof() to decide whether something is susceptible
to turning into green slime. Most of this is just making use of existing
cached permonst values in damageum() and mdamagem() and shouldn't affect
anything. I wanted to avoid mixing that in with the actual slime changes
which are coming.
A couple of things noticed when looking at the death-by-brainlessness
code. The 3.4.3 code ran a loop while life-saving was keeping the hero
alive, which would work if someone added other sources of life-saving than
the amulet but not if they added some form which didn't get used up when
it kicked in. Post-3.4.3, the dev code eliminated the loop but was no
longer guarding against additional forms of life-saving. This attempts to
clear the relevant field once any form of life-saving takes effect (for
brainlessness). It's not perfect since someone could change `Lifesaved'
to look at something other than the hero's properties, but at least it
still avoids the risk of getting stuck in a loop if someone makes a really
bad customization.
Also, make life-saving use min(2*level,10) instead of flat 10 for
amulet or 8*level for explore/wizard survival when it saves someone whose
max HP have been clobbered somehow. Other places assume that 1 HP per
level is the lowest the hero will have; saved-life gives a modest bit more.
Lastly, some post-3.4.3 code to make ghosts/shades immune to brain
sucking was using mon_nam() to start a sentence; Monnam() is needed there.
Make polymorphing or changing alignment perform a touch check (as is
done when catching lycanthropy) on wielded weapon(s) to see whether the
hero can still use them in his new form. Part [2 of 2] will update
retouch_equipment() to check all items in use rather than just weapon(s).
(A comment or two in part 1 already refers to expected behavior of part 2.)
From a bug report, trying
to invoke a wielded artifact after changing alignment resulted in "the
<artifact> evades your grasp" but it remained wielded, contradicting the
message. This adjusts the message in touch_artifact() if the object is
already in inventory, and adds retouch_object() to handle cases where
failing to be able to touch ought to force unwearing/unwielding.
There's newsgroup discussion about a mind flayer sucking the brains
of a ghost, plus speculation that the fix for C343-121 has already taken
care of it. I agree with them that ghost's brains are too insubstantial
to be hit by brainsuck damage from mind flayers' tentacles. C343-121 is
vague enough to cover this; the corresponding fixes entry was more specific
and is augmented here.
From a bug report, but
just received: if you choked while eating a cockatrice egg and survived,
the turn-to-stone sequence wouldn't be initiated. For such eggs, turning
to stone starts when the food is finished rather than when it's started,
and fpostfx() wasn't called for food that triggered choking so the egg
went away without actually being finished. Fortune cookies, lumps of
royal jelly, and a few other things suffered from the same situation.
From a bug report: you could vomit when polymorphed into a rat but real life
rats can't/don't vomit. The latter was confirmed by <Someone> and <Someone>.
While testing a fix for this, I discovered a couple of other problems.
Healing magic which cured sickness failed to heal Vomiting (potion or
spell; unicorn horn deals with them separately). Enlightenment failed to
report Vomiting (it's not shown on the status line). Most significant was
that vomiting_dialogue() called vomit() twice (also make_confused() and
make_stunned() three times for every once intended). It was dividing the
remaining turns by 3 and then using that value to decide what to do, but
only message display took into account that the same divided value would
occur on 3 consecutive turns (or just 2 for the final countdown to 0,
because dialog routine gets called before timed-property decrement).
From the newsgroup: if polymorphed into a metallivore, you could eat
artifacts that you couldn't touch with your hands. Now you can't eat ones
which evade your grasp, and you take some damage from other touchy items
on--actually, prior to--the first bite. (Those still "taste delicious" if
you survive; I'm not sure whether this ought to be changed.)
There was code to give feedback if you attempted to offer the Amulet
on a regular altar instead of the final high altar, but that code was
unreachable; getobj() yielded "that's a silly thing" whenever you picked
an amulet while not on the Astral (or recently changed, Sanctum) level.
This allows you to try to offer the real or fake Amulet of Yendor on any
altar, but they'll only be listed as likely candidates when on the Astral
level. Conversely, it no longer lists carried corpses as likely candidates
at the Astral high altars; they're still acceptable but not what the hero
is supposed to be fiddling with there. Also, allow corpses on the floor
to be offered on high altars, fixing a complaint we've gotten a few times
over the years. (Unfortunately there's no way to suppress them as likely
candidates on the high altars while still allowing them to be sacrified.)
Change safe_qbuf() so that instead of picking one of three strings
for sprintf() to plug into a prompt string, it actually constructs the
full prompt string itself. Also pass in the unformatted object and a pair
of formatting functions instead of performing dual formatting in advance.
The actual formatting is done via new routine short_oname() which also
takes an object and a pair of formatting routines plus a target length.
It uses the first routine, typically xname() or doname(), and formats the
object, then if the result is too long it makes some transformations, and
tries again. If truncating "called foo" and "named bar" down to 12 chars
and omitting "uncursed, rustproof, thoroughly corroded" attributes still
result in a string that's too long, it uses the other formatting routine.
The latter calls one of several jacket routines around simple_typename()
to produce a short result.
This has been through about four incarnations now and has gotten a
bit less testing each time, but I need to get it in place before I end up
running out of gas and abandoning it. I've got some changes to shk.c
(where safe_qbuf is needed but not currently used) that now need to be
redone and will come eventually.
From a bug report: if you
attempted to eat a Rider corpse and got the 1/7 chance that non-yet-rotten
food will be treated as rotten, then also got "the world spins and goes
dark" result for rotten food, you would both survive the eating attempt
and also end up with a partly eaten Rider corpse. This patch treats Rider
corpses like lizard and lichen corpses; they'll never yield rotten food
effects. That way, they'll always be fatal to eat. They'll still end up
being partly eaten if you are life-saved, but since they'll immediately
revive, the only way you'll know that is to use probing or stethoscope to
discover that they've revived at less than full health.
Nearly two years ago, <email deleted>
suggested that lembas wafers and cram rations be treated like fortune
cookies and never yield the rotten food result. I'm guessing that cookies
are handled that way so that rotten food feedback doesn't override false
rumor delivery when they're cursed, rather than because they're considered
to be rot-proof. This implements <Someone>'s suggestion, except that cursed
lembas and cram will still behave like rotten food.
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.
The latest Micrsoft compilers complain when a function is
assigned to a function pointer, and the function's argument
list does not match the prototype precisely.
It was evem complaining about the difference between this:
int x()
{
[...]
}
and a prototype of
int x(void);
when assigning that function's address to a function pointer.
This quiets those warnings, without suppressing the mismatch
check altogether for more serious mismatches.
It seemed inappropriate to allocate and tack on an oextra
structure just to mark it as unmergable, and the oextra
struct itself wouldn't be released until a save/restore took
place.
This uses one of the freed up oattached bits to prevent the merge.
move oattached and oname and other things that vary
the size of the obj structure into a separate
non-adjacent oextra structure, similar to what has
already been done for mextra. The obj structure
itself becomes a fixed size.
New macros:
#define ONAME(o) ((o)->oextra->oname)
#define OMID(o) ((o)->oextra->omid)
#define OMONST(o) ((o)->oextra->omonst)
#define OLONG(o) ((o)->oextra->olong)
#define OMAILCMD(o) ((o)->oextra->omailcmd)
#define has_oname(o) ((o)->oextra && ONAME(o))
#define has_omid(o) ((o)->oextra && OMID(o))
#define has_omonst(o) ((o)->oextra && OMONST(o))
#define has_olong(o) ((o)->oextra && OLONG(o))
#define has_omailcmd(o) ((o)->oextra && OMAILCMD(o))
changed macros:
has_name(mon) becomes has_mname(mon) to correspond.
The CVS repository was tagged with
NETHACK_PRE_OEXTRA
before commiting these, and
tagged with
NETHACK_POST_OEXTRA
immediately after. The diff
between those two tags is this oextra patch.
The associated mail daemon changes to use an oextra
structure instead of a hidden command located in the
name after the terminating NUL, have not been tried
or tested.
<email deleted> wrote:
> Eating gold in a vault (or polymorphing a pile of gold into 1 gold piece)
> doesn't anger the guard.
This addresses the eating part of that report, but the hero
has to get caught doing it.
<email deleted> wrote:
> * MAIL Abuse: one can polymorph into a gelatinous cube, send themselves lots of
> scrolls of mail, and eat them, gaining free nutrition.
Change the description of leather spellbook to leathery. Eating it
does not violate vegan or vegetarian conduct. [By shear coincidence, I got
``Blecch! Rotten paper!'' for it while testing this.]
<Someone> wrote:
> * Eating a leather spellbook doesn't violate
> vegetarian conduct. I suggest you rename it to
> something else. (May I suggest "bark"?) Also, I refuse
> to consider "leather" to be a colour or texture.
> * While as a water nymph in water, seducing an unseen
> monster on dry land sometimes gives away its identity.
> * When submerged iron golem monsters residually
> "rust", and the rust damage kills them, it doesn't
> report their death.
> * Drum of Earthquake says "the entire dungeon shakes
> around you" even when you're in the Endgame or the
> Quest's top level.
> * Wishing for a statue of a guard, then destoning it,
> then chatting to it, makes it say "drop that gold"
> without context. This could probably be extended to
> other cases, too (like Croesus?)
This corrects only the first complaint in the list above.
This patch alters wiz_identify so that it displays an
inventory menu with all items shown identified
without actually identifying them.
You can just press ^I (or whatever the wiz_identify command
is) a second time, while the menu is displayed to actually identify
them (tested with TTY menus only).
Reveal more tin details at end of game disclosure.
b - an uncursed carrot
c - an uncursed tin of broiled kobold meat
d - an uncursed tin of stir fried fox meat
e - 4 uncursed tripe rations
f - 4 uncursed food rations
g - an uncursed cream pie
n - 4 uncursed tins of spinach
o - 3 uncursed tins of pureed newt meat
p - an uncursed homemade tin of fox meat
Note that in the case of homemade and rotten
it sounded better to have the term before the
word tin, rather than after:
homemade tins of newt meat
rather than:
tins of homemade newt meat
The wishing code should probably be
changed to reflect this so people can wish for
a "homemade tin of newt". As it stands, they must
wish for a "tin of homemade newt" which differs
from the final display. That is not included.
Note: The CVS repository was tagged with NETHACK_PRE_MEXTRA
prior to application of this patch to allow easy withdrawal if necessary.
Adds a new mextra structure type that has a set
of pointers to various types of monster structures
including:
mname, egd, epri, eshk, emin, edog
Replaces the mextra bits in the monst structure
with a single pointer called mtmp->mextra of type
(struct mextra *).
The pointer can be null if there are no additional
structures attached. The mextra structure is not
adjacent to the monst structure.
Reduces the in-memory footprint of the monst that
has no other structures attached, at the cost
of adding 6 extra long ints per monster to
the save file
The new mextra structure has the mextra fields
independent of each other, not overlapping as was
the case with previous NetHack versions.
This patch doesn't do anything to capitalize on
that difference however.
Consolidates vault.h, epri.h, eshk.h, emin.h and edog.h
into mextra.h
Adds a macro for checking for whether a monster has
a name:
has_name(monst)
This fixes the magic trap panic
expels() -> spoteffects() -> dotrap() ->
domagictrap() -> tamedog()
because the monst no longer varies in size so no
replacement is required.
Extend the capabilities of corpse_xname() so that various callers can
be simplified. It can how handle an article prefix, effectively turning it
into corpse_doname() (not quite; still need doname() to see a count when
quantity is more than one, or to see bless/curse state). It can also handle
inclusion of adjectives like "partly eaten" or "bite-covered". For unique
monsters those come out in the form
the Chromatic Dragon's partly eaten corpse
instead of the old
partly eaten Chromatic Dragon corpse
[so wishing probably needs to be taught about potentially finding a monster
name before assorted adjectives such as blessed; also, name_to_mon() needs
to learn how to cope with the possessive suffix].
A sizeable chunk of this patch deals with consolidating some of the
redundant "petrified by a cockatrice corpse" handling. It may be possible
to consolidate all remaining instances together since they're quite similar,
but I didn't think about that until just now and I want to get this patch
over with.
From a bug report, bribeable demons will demand money when
hero has fainted from lack of food and hero can pay while unconscious. I
decided to just borrow from vault guard behavior and have the hero regain
consciousness. It turns out that reset_faint() has been broken since a
long ago (before my time...) change to nomul() [nomul(0) is a no-op while
fainted since multi is negative then]. Now fixed; both bribe-demanding
demons and vault guards will cause fainted hero to wake up when they arrive.
If hero can't move for some reason other than fainting, demons will skip
the bribe demand and immediately become hostile (vault guard in that case
goes away after saying that he'll return). There is no deafness handling;
perhaps the bribe demand is accompanied by sufficient pantomiming for the
hero to figure it out? ;-)
Also fix an unintended potential alignment hit against the player if
bribeable demon is killed after becoming hostile due to misjudging displaced
hero's location.
Fix a couple of problems From a bug report. Eating a Rider corpse is fatal, but eating a live Rider's
brain was not--now it will be, both for monster mind flayers and for player
poly'd into one. Also, there was no check for cannibalism when poly'd hero
eats brains--now there is. Not mentioned in the report: eating Medusa's
brains will now be fatal just like eating her corpse. And pet mind flayers
who eat the hero's brains will gain some nutrition like they do when eating
monster brains.
Creating a common eat_brains() routine turned out to be something of
a mistake; there is only a tiny amount of overlap among the u-vs-m, m-vs-u,
and m-vs-m cases.
Makefiles need a dependency update to add edog.h for eat.c.
Make petrification initiation or termination go through a new routine,
make_stoned(), instead of manipulating its countdown timer and delayed
killer directly. No change in behavior.
There's no reason in terms of bug risk or game play or saved data why
this shouldn't be done in the branch too, but so much of the surrounding
context has already diverged between trunk and branch that it's trunk only.
Because the description seen by the player is "you have never changed
form" rather than "you have never polymorphed yourself", make death by fully
turning into green slime (even if that death is avoided via lifesaving)
violate the conduct. Suggested by <Someone> 9-Dec-2004. Likewise, eating
a mimic corpse and temporarily turning yourself into a pile of gold also
violates that conduct. Mentioned by someone--probably <Someone>, or possibly in
the newsgroup--a long time ago.
A post-3.4.3 change dealing with reaching into pits resulted in "you
sit on the air" if you used the #sit command after escaping a pit trap.
Change can_reach_floor() so that caller explicitly controls whether being
on the brink of a pit is a condition that prevents reaching the floor.
This also splits a fairly common message about not being able to reach the
floor into a separate routine.
There is still oddness here: if you're polymorphed into a flyer,
#sit yields "you sit down" followed by "you fly over a pit" (latter occurs
when escaping trap activation). A ceiling hider behaves similarly, but
the second message is "you escape a pit" and doesn't sound quite as silly.
Perhaps #sit should pass TOOKPLUNGE to dotrap(), or maybe there's some
better way to handle this?
A few weeks ago someone in the newsgroup posted how he got an intrinsic
value for increased damage up to something like +74 via eating rings (on his
way to killing the Riders a zillion times for maximum score--he wanted to be
sure to kill them in a single hit each time so that his use of keystroke
macros wouldn't get out of sync by extended combat). Make it moderately
difficult to get beyond +20 and impossible to get past +40 for corresponding
attribute when eating rings of increase damage, increase accuracy, and
protection. Since you could also wear a pair of +7 rings you can still get
an awfully high total, but it won't be unlimited any more and most people
willing and able to go to such extreme lengths would undoubtedly prefer to
be wearing other types of rings.
[This ought to be suitable for the branch version too but I'm not going to
spend the effort to migrate it there.]
Recently From a bug report, reducing
the value of a shop object via cursed enchantment was ignored by shopkeeper.
This replaces the existing costly_cancel() routine with costly_alteration()
which performs a similar task: bill for any item whose value has been made
less. The hero owns the resulting object but must pay for the original one
before being allowed to leave the shop.
This covers the majority of cases where bill_dummy_object() was already
being used: cancelling a charged or enchanted item, casting drain life at
same, diluting potions or blanking scrolls or books by dipping them into a
potion of water, dulling a weapon by engraving with it, eating unpaid food
or opening unpaid tins, applying a cream pie to hit yourself with it in the
face, applying a wand to break it, burning something by dipping it into lit
potion of oil, and clearing potions by dipping a unicorn horn into them.
The shop billing behavior for those actions hasn't been changed, just
consolidated into one place which delivers a common message for them.
This also covers many cases which weren't being handled: stripping
wand or magic tool charges via cursed scroll of charging, reducing a charged
ring's enchantment via same, reducing weapon or armor enchantment via cursed
scroll of enchant weapon or armor, stripping an item's rustproofing via
confused enchantment, making a crysknife revert to a worm tooth, unblessing
potions of holy water or uncursing potions of unholy water. (That last one
won't be billed if it's the result of prayer rather scroll, spell, or #dip.)
And this tries to handle the reverse situation more thoroughly too:
many actions which improve the value of an unpaid item now also cause the
shop bill to be updated to reflect its new higher price. Aside from the
basic enchanting and charging magic, it covers converting dragon scales into
dragon scale mail and worm tooth into crysknife. Some things which might be
expected to inflate shop prices, like rustproofing or increasing the number
of charges in a wand, don't actually affect the price. And there are bound
to be cases where the price is affected but I've overlooked.
While auditing nomul() I noticed unconscious() treats (multi < 0 && !nomovemsg)
as unconscious. This explains the behavior in M29 (unconscious message
while performing #turn). I checked all the places with this combination,
and found a few that did not appear to fall under the "unconscious" category.
Most I changed to use You_can_move_again to ensure the same display w/o the
unconscious behavior. Also:
- found another string that unconscious() should have considered
- vomit() now sets nomovemsg, one caller was also doing this redundantly
- vomiting_dialogue() was calling stop_occupation() after vomit(), which can
reset multi. I reversed the order and removed a doubly-redundant nomul call.
tele() still has a problem: some cases where multi < 0 should probably take
a branch like the unconscious() branch but with a different message.
doturn()'s behavior - turn then wait - is also less than perfect, but I
think this is a known problem.
