From a bug report, applying a polearm to make a
short-range ranged attack didn't scuff any engraving you were standing on,
unlike ordinary melee and throwing/shooting attacks. Grappling hooks had
the same omission. Fixing it led to several other minor bugs. Attempting
to target an unseen monster's spot with polearm or grapnel would yield some
permutation of "wait, there's something there" but draw the 'I' glyph at
the wrong spot. It used <u.ux+u.dx,u.uy+u.dy> instead of the actual target,
so put the 'I' one step in front of your most recent move (or throw or zap
or whatever last set u.dx and u.dy). Giving ESC when prompted for target
spot failed to use up a turn even when the polearm or grappling hook had
just been auto-wielded for use. Neither use_pole() nor use_grapple() set
`notonhead' for hmon() (called via thitmonst() in their cases; this was
academic since plain physical damage attacks don't actually care about it).
[The bad 'I' placement was a post-3.4.3 bug.]
Applying a bullwhip to attack an adjacent monster didn't have any of
those issues but did have the possibility of targetting off the edge of
the map when standing at that edge while confused or stunned.
Applying a polearm to target an 'I' would yield "nothing happens" if
the unseen monster wasn't there anymore, and it didn't bother to remove
that 'I' from the map. After changing it to do so, the phrasing no longer
made any sense. This led to a slightly bigger change than I intended:
since statues are now shown as gray monsters (does that work for tiles?)
instead of as chunks of stone, they are likely to be intentional targets
sometimes, so polearm attacks now handle them differently from other
non-monster locations. [I hope that other attack forms don't need
similar handling. Melee certainly doesn't, since walking onto the spot
is enough to distinguish statues from monsters. Having the missile pass
right through a statue's location probably suffices for ranged attacks.]
From a bug report.5 years ago? and again today:
mimics appear on the rogue level even though they're lowercase letter
monsters. <Someone> notes that during level generation, mimics pretending to
be closed doors are sometimes substituted for trapped actual doors. That
should be avoided on the rogue level both because its doorways are always
empty (enforced after the fact in extralev.c code rather than in the door
making code) and because mimics aren't uppercase letter monsters (so
should only ever appear there if they've travelled from another level).
Noticed while testing the patch for monster ranged attacks when hero
is hidden. Using #monster while in small mimic form would hide you, as
intended, but the first monster (or last monster?) who hit you hand-to-hand
would be grabbed ("It gets stuck on you."). Unlike large and giant mimics,
small ones lack such a grab attack so when you eventually returned to
normal form, u.ustuck would remain set and you would end up being stuck
to an arbitrary monster instead of releasing it from your grasp.
From a bug report, a hero who had
eaten a mimic corpse and was stuck mimicking gold was hit by an approaching
dragon's breath attack. Make monsters usually fail to notice that you're
lined up for their ranged attack when/if you're hidden from them.
The intersection of a couple of recent patches: noticed when
updating movement as a grid bug, and accentuated when fixing the attempt
to move down while levitating. If you can't reach the floor due to
levitation, don't show '>' in the list of likely candidate directions for
the prompt when digging. It was always included so that the list could
never be empty, but it's a poor suggestion to the player when levitating.
Use '<' instead in that situation; also a poor suggestion, but silly
rather than unintentional. :-)
This only affects the bracketted part of the "In what direction?"
prompt, not the actual digging (and player can pick any direction whether
it's shown in the prompt or not, so the digging code is already prepared
to handle attempts to dig up as well as down anyway).
Make the same simplification to save and restore of obj->oextra as
was done yesterday for monst->mextra: no need for a set of all 0 sizes
when the whole thing is null. Bumps EDITLEVEL again.
From a bug report, being stuck in the
floor or in lava and trying to move downwards while levitating gave "you
are floating high above the floor/lava" which contradicts being stuck.
Now you'll get "you are trapped in the floor/lava" in that situation.
I thought about letting it fall through to the chance to autodig with
wielded pick, but decided to go with just the alternate message. (Being
tethered to a buried iron ball still lets you "float high above the floor",
so this doesn't check for being trapped that way.)
A mimic posing as a statue was displayed as a tengu statue (and
recognizeable as such now that statues are displayed as the corresponding
monster rather than rock-class back tick), but the lookat code described
it as a giant ant statue (since there was no obj->corpsenm available to
indicate the monster type, it defaulted to 0). This adds monst->mextra
field `mcorpsenm' so that mimics have a place to remember what sort of
statue or corpse they are mimicking. And it picks a random monster type
when they take such forms so that the old tengu hack becomes irrelevant.
newmextra() and newoextra() initialized pointers via memset(...,0)
which is not portable; switch to explicit assignments. The wizard mode
code to display memory used for monsters and objects added in amounts
for the miscellaneous things pointed to by monst->mextra and obj->oextra
structs but didn't include memory for those structs themselves; add it.
Simplify monster save/restore slightly; there's no need for extra zeroes
to represent monst->mextra->X sizes when monst->mextra is null.
Update the startup banner for 2009. I should have done this with a
separate patch but I'm taking a shortcut. :-]
Mentioned in the newsgroup recently: stethoscope didn't reveal
when a monster was actually a chameleon, even though that's the sort of
inside information you might expect. This adds "shapechanger" to the
other miscellaneous attributes shown. I initially had "(a chameleon)" or
"(a vampire lord)" but decided that that was too specific, particularly
for the latter where you might expect/hope that the vampire bat next to
you is just an ordinary vampire rather than a lord.
I had a pet eat a mimic corpse and take on the shape of a pyramidal
amulet. Probing described it as "mimicing a pyramidal". Many of the
names and descriptions in objects[] are only partial, so aren't suitable
for supplying feedback like this (and "mimicking" was misspelled...).
Switch to simple_typename(), although in the case of undiscovered amulet
it yields just "an amulet" rather than "a pyramidal amulet". This also
extends probing feedback for mimicry to include all monsters mimicking
something, not just those who are in the midst of eating. [That's
post-3.4.3 code and is only applicable for wands of probing. Using a
stethoscope or hitting with Magicbane will unhide a mimic before giving
extra info about it.]
From a bug report, attempting to move
diagonally when poly'd into grid bug form doesn't give any useful
feedback in the general case, and in the specific case of using 'u' to
try to move northeast with vi-style keys, it performs #untrap instead.
Diagonal directions were being classified as non-movement when in grid
bug form, so the feedback was usually just "unknown command". But 'u'
is bound to untrap as a a convenience to players who use num_pad-style
movement keys. (Move commands don't actually reach the assigned key
part of command handling, so for forms other than grid bug, !num_pad 'u'
moves NE despite the untrap function given to that key.)
Split the diagonal handling out from movement command recognition.
It now gives "you can't get there from here..." if player tries to move
diagonally as a grid bug. For direction prompts, it now gives "you can't
orient yourself that direction" instead of popping up the command assist
display. (In the prompt string showing likely candidate directions for
digging, diagonal handling for grid bugs is academic because they aren't
strong enough to wield pick-axes.)
I) When testing the shopkeeper polymorph fix a couple of days ago, I
tried to pick a form without a head (which is now rejected for shk) by
specifying "trapper" via #monpolycontrol. It worked, in the sense that
the shopkeeper didn't take that form, but it also retried with "lurker
above". After trapper was rejected, it matched "trapper or lurker above"
for monster class and picked at random from that class (and in this case,
the second try was rejected for being headless too). The code I added
(a couple of years ago?) to prevent picking invalid forms for vampire
shapeshifting was a bit sloppy with its retry attempts.
II) I also noticed that monpolycontrol would let you pick forms that
newcham subsequently rejected, like placeholder monster "giant", so that
select_newcham_form got called again, resulting in a new prompt from
monpolycontrol to choose a form without seeing any explanation why the
first try was ignored. Now you'll get the same "it can't become that"
feedback and immediate reprompt as for an invalid monster, instead of
returning newcham a form that it won't accept.
III) The change in shopkeeper polymorph was rejecting forms without hands.
That was way too restrictive; the notake() filter is sufficient.
This is all modification of post-3.4.3 stuff, so no fixes entry.
[Not 100% true. newcham() doesn't explicitly reject humans anymore,
relying on polyok() test instead. Kops are the only humans who pass it.
If they're not supposed too, they need M2_NOPOLY in their definition.]
From a bug report, a shopkeeper who's been
polymorphed into a limbless creature shouldn't be able to come take your
possessions when you die since you can't pick up items when in that shape.
Rather than add new special case handling for shopkeepers, prevent them
from taking on shapes that render them unable to behave like shopkeepers
so that they can continue to catch thrown pick-axes and assorted other
things that would otherwise need a lot of extra checking introduced.
Ditto for temple priests, vault guards, and quest leaders. Restriction
also applies to wizard mode use of #monpolycontrol, the only place where
players can actually notice this.
It could--perhaps should--restrict them even further, forcing forms
that retain speech capability, but I didn't go that far.
This started out as a one line change. After I saw someone in the
newsgroup mention that Sunsword's light was inferior to that of a lamp,
I decided to make it work better (than in 3.4.3, that is, becoming the
same brightness as a lamp) when blessed and worse when cursed (useless to
hero but still visible if wielded by a monster). But then it needed to
change light radius when its curse/bless state changed, and it needed
message feedback when doing so, and that got kind of complicated. I
wouldn't have bothered if I'd known what I was getting into, but I don't
want to throw it away now that I've done all this work....
Sunsword now gives a light radius of 3 when blessed (same as a lit
lamp), radius of 2 when uncursed (same as a lit candle and as it has been
providing since added in 3.4.0), and a radius of 1 when cursed (nearly
but not completely useless, as mentioned above). Also, it now "shines"
rather than "glows" since we usually use the latter for temporary effects.
Allow potions thrown straight down while hero is mounted to target
steed instead of always hitting the floor. There's already a fixes35.0
entry for potions hitting worn saddle instead of the creature wearing it,
but that only happened when thrown at a nearby saddled monster, not when
the hero was stuck mounted on a steed with cursed saddle.
Eight separate "the <obj> glows <color>" messages were too many...
(four for dipping an item in holy/unholy water, four more for throwing
either of those potions at a saddled monster). Replace the repetitive
code, leaving just two such messages. Setting of bknown flag for the
dipped object is the only intentional change here. That used to be done
unconditionally but now requires that the hero see the glow color.
From the newsgroup:
As you read the scroll, it disappears.
The scroll erupts in a tower of flame!
Your spellbook of force bolt catches fire and burns!
[...]
Your potion of paralysis boils and explodes!
Something seems to be holding you.
Do you want your possessions idenified?
This character survived the burning and boiling objects then succumbed
to the initial flame explosion.
A fatal explosion calls done() without any explanation. Or rather,
it dated to the time where it would immediately follow "the scroll erupts
in flame" (or "you are caught in <some kind of> explosion") without the
chance for intervening messages. Then when item destruction was moved
sooner (for bones file purposes), message sequencing was left dangling.
This patch adds a new "it was fatal" message in front of done().
Post-3.4.3 problem noticed by Michael: some code which was supposed
to be inside the end of a loop got placed in front of the wrong closing
brace, ending up beyond that loop. As a result, decay of old attribute
exercise wasn't working as intended, and out of bounds array modification
was occurring (clobbering u.abon.a[0], which probably messed up resetting
strength when taking off rings of gain strength which were worn during an
exercise check).
From the newsgroup: hero's steed can become untame if killed while
it is wearing an amulet of life saving, leaving the hero still mounted
and resulting in repeated "placing steed onto map?" warnings when the
steed tries to move. Force the hero to be thrown off the mount in that
situation.
From a bug report, the feedback
you get when a monster summons insects or snakes is the same when blind
as when you can see. A comment in the code stated as much, but fixing
it is relatively straightforward. (Or not; there are actually a lot of
cases to be handled; this covers enough of them, I hope.)
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.
From a bug report, code that always
produced plural feedback (used only for values of N greater than 3) was
unnecessarily handling plural vs singular. No change in game behavior;
just one less function call in the code.
From a bug report, getting "your
clumsy kick does no damage" when attempting to kick the location of a
concealed mimic or hidden monster didn't bring the target out of hiding.
Mentioned by <email deleted> in his report about
sound anomalies, it was possible for a sleeping hero to "gladly start
removing armor" when attacked by a nymph or succubus. The code
explicitly wakes up the hero if he has fainted from lack of food because
"can't charm without waking you" (according to the existing comment),
but it didn't handle other forms of sleep and paralysis. You could get
|You fall asleep.
|The wood nymph charms you. You gladly start removing your armor.
|The combat awakens you.
Now in the same situation you'll get
|You fall asleep.
|You wake up.
|The wood nymph charms you. You gladly start removing your armor.
From a bug report, the text change applied
when you try to give an artifact's name to am item of that artifact's type
would choose a letter from 'a' through 'y' when replacing the randomly
selected target letter. Rather than fixing the off by one bug which
prevented 'z' from being chosen, this switches to the existing routine
used for mangling engravings. (Unfortunately the fix is not as simple as
first expected, because wipeout_text() doesn't guarantee to change text
which has spaces in it and all the quest artifact names have those.)
From a bug report, pets able to eat
acidic and poisonous corpses (black naga was the case cited) would eat
green slime corpses without turning into green slime, unlike the hero.
This prevents such monsters from eating green slime unless they're
starving, implements transformation into green slime for the case where
it does get eaten, and prevents non-pet gelatinous cubes from devouring
such corpses. meatobj() is reorganized to hopefully become clearer, and
it removes the assumption that the object eater is a g.cube in case we
ever adopt slash'em's "tasmanian devil" monster.
Monsters with digestion attacks who swallow green slime monsters
are turned into green slime, but ones who swallow hero poly'd into green
slime are not. This doesn't address that.
When dragon scales or dragon scale mail becomes embedded in poly'd
hero's skin, the '[' command would report "you are not wearing any armor"
but the '*' command showed "a set of <color> dragon scales (embedded in
your skin)". The '[' feedback is more accurate but the '*' feedback is
probably more useful. This changes '[' to be more specific when lack of
worn armor is accompanied by embedded dragon scales/scale mail.
I put the fixes entry in the new features section since the old
behavior wasn't a bug.
From a bug report, polymorph of self due to
breaking a wand also polymorphed various items that were dropped in the
process, unlike the case of zapping polymorph at monsters which excludes
dropped items from being poly'd. This polymorphs the pile at the hero's
feet before polymorphing the hero. I first tried to handle it using
obj->bypass like with monsters, but that didn't work. Post-3.4.3, the
bypass handling is also used for polyself (by retouch_equipment()) and
it was getting reset at an inconvenient time.
He also complained that he failed to get "you feel shuddering
vibrations" when some polymorphed objects got destroyed. That message
is issued by weffects() which do_break_wand() doesn't call. It ought to
be fixed, but this patch doesn't address it.
Lastly, add code to prevent objects guarded against polymorph via
obj->bypass from getting used up when creating polypile golems.
The report that killing a Rider on an altar allowed you to sacrifice
its corpse was a wizard mode-only phenomenon in 3.4.3 (because you needed
to use an altar at some location other than the Astral Plane, hence also
needed to use ^G to get the Rider there), so not really a bug. But a
post-3.4.3 change has made it possible to offer corpses from the floor on
the astral level, unintentionally making it possible to sacrifice Rider
corpses. This makdes #offer check specifically for them and trigger their
revival if the attempt is made, same as done by pickup.
Some post-3.4.3 code to clean up thrown or kicked objects which were
in transit at the end of a game didn't work correctly for kicked objects,
leading to an "obj not free" panic if you kicked an object at some point
and didn't kick anything else before the game was over. Unlike thrownobj,
kickobj wasn't being cleared after use.
Add an entry to the set of randomly placed engravings. The revival
of Doctor Who, which seems to be referred to as Series 1 through Series 4
ignoring umpteen years of earlier versions, showed graphiti of "BAD WOLF"
intermittently throughout season two and brought it back for the finale
of season four. As far as I noticed, it was always written in all caps.
When testing singularizing of fruit names I noticed that "bunches of
grapes" became "bunche of grapes". makesingular() had a comment about
not recognizing "es" and suggesting that recursion could solve it. But
makeplural() already handled things like that without resorting to
recursion, and it also recognized more compounds than just "foo of bar"
and "*man-at-arms" (such as "pie a la mode" and "soup du jour"). This
moves the compound phrase recognition into a separate routine so that
both makesingular() and makeplural() will handle the same stuff, and it
modifies makesingular() to do as well as makeplural() when processing the
front half of compound phrases (the "foo" part in "foo of bar").
Also, a minor plurization tweak: algae was recognized as already
plural but larvae and several similar words weren't.
Inventory formatting for a single slime mold object would be strange
if the user entered a plural word or phrase for fruit name. Forcing the
user-specified value to be singular as it's being set up as current fruit
avoids that. [Something like <Someone>'s "bunch of grapes" is unaffected;
it's already singular and correctly handled as such by makesingular().]
This may have a side-effect of limiting the creativity of some players
who try to trick others via bones files gimickery, but I think the extra
consistency in object naming during ordinary play is worth it.
According to the cvs log info, this issue was actually mentioned in
a patch I made ("fix M203...") in October, 2005. I have no recollection
of that at all....
Each time save and restore is performed, the ffruit list gets
reversed. Since additions can be made when it is in backwards order or
in original order depending upon whether an odd or even number of save/
restore cycles have taken place, the numeric sequence of its entries is
ultimately arbitrary. So there's no point using extra code to force new
ones to be added at the end of the list; just put them at the beginning.
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.
From a bug report:
The following steps do not yield the expected fruit:
1) start nethack in explore mode (with a wand of wishing)
2) change fruit name to "tomato"
3) save/restore
4) change fruit name back to "slime mold"
5) save/restore
6) wish for a fruit; you get a tomato
7) check options; fruit name is set to "slime mold"
If you specified a fruit name that already existed in the list,
fruitadd() always set current_fruit to the fruit with
the highest fid encountered in the list to that point, instead
of the fid of the matching entry.
From the newsgroup: if no co-aligned artifacts are available when
attempting to give the first divine gift for an offered corpse, nothing
would be given. Since the gift counter stayed at 0, subsequent attempts
to select one would still treat it as first gift and always fail again.
The first divine gift from offering corpses must match the hero's
alignment; after that, nonaligned artifacts (Frost Brand, Giant Slayer, &c)
are added to the pool of choices for further divine gifts. It's pretty
easy for a chaotic character to use up the co-aligned artifacts before
getting any divine gift. There aren't many chaotic ones and some are
inelgible due to being race-specific items. Wishing for Stormbringer--or
loading a bones file which contains it--and creating Sting and Orcrist
via naming will do the trick for an elven hero.
This patch expands the pool of candidates to include nonaligned
artifacts during first gift selection if no co-aligned ones are available.
use makedefs --grep in Makefile.doc
call make clean in doc from make clean in top
add commented out rule to produce mdgrep.h from mdgrep.pl
macosx1.5: don't chown/chgrp for single-user install
unixmain.c: work around C90 warning for Mac-specific code, fix last fix
makedefs.c: temporarily disallow blank after control introducer until docs
catch up
mdgrep.pl: add ALLDOCS, clean up generated file's header
Finally found a flag combination that will complain about declarations mixed in
with other code: -ansi -pedantic.
Clean up the violations of that I just introduced and add that flag to the
Mac 10.5 hints file. (Note that there is one warning left in unixmain.c -
it's in Mac-specific code.)
Add SHELLERS - people allowed to use ! command with same syntax as WIZARDS.
Add new hints file for 10.5, since the rules and commands for groups changed
(new commands introduced in 10.4, old ones removed in 10.5; creating a new
user under 10.4 gave you a matching group, in 10.5 it doesn't). Also move
shared build into roughly right place in file system when being installed
for root - don't use ~root.
Makefile.top - don't remove ./-p unless it exists (that's always annoyed me).
fix error invoking macosx.sh
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.
