3e597b6882
Changes to be committed: modified: doc/fixes36.1 modified: src/ball.c I looked up the original bug report that led to bug page C343-20 "When dying immediately on entering a level, the map may show you dying on the previous level." It was received public report U891: > When one is being punished and goes down a staircase and dies because the > ball and chain fell on their head, one gets the message about their death > while the old level is still being displayed. I wasn't sure whether this > was a bug or not because on one hand it wouldn't make much sense to > generate a new level if the character is going to die anyway. However, > that being said it does make a difference if the character is about to go > down into a level where one cannot leave bones files, ie medusa or the > first level of the mines (if i remember correctly). So, if your character > dies from this does the bones file get left on the level you were on > (which is still displayed at the time of death) or the level you died as > soon as you got to (but was never displayed)? Thanks! Pat had remarked in response: "So this is just a display issue; game play works as intended (for the program; I imagine you weren't planning to get killed." A debug trace in wizard mode 3.6.1 beta shows that the relevant code path is this: NetHack.exe!done(int how) Line 908 NetHack.exe!losehp(int n, const char * knam, char k_format) Line 2678 NetHack.exe!drag_down(...) Line 823 NetHack.exe!goto_level(d_level * newlevel, char at_stairs, char falling, char portal) Line 1316 NetHack.exe!next_level(char at_stairs) Line 1157 NetHack.exe!dodown(...) Line 954 NetHack.exe!rhack(char * cmd) Line 3416 NetHack.exe!moveloop(char resuming) Line 464 NetHack.exe!main(int argc, char * * argv) Line 104 This patch clears the display for the situation in drag_down(), so the old level is not shown.
75 lines
3.7 KiB
Groff
75 lines
3.7 KiB
Groff
$NHDT-Branch$:$NHDT-Revision$ $NHDT-Date$
|
|
|
|
General Fixes and Modified Features
|
|
-----------------------------------
|
|
doc/*.6 man pages and corresponding doc/*.txt text copies were out of date
|
|
data.base entry for "lava" had wrong first name for Don Woods' attribution
|
|
cursed genocide of "none" sent in monsters, but "that's enough tries" didn't
|
|
update MAXPLAYERS documentation in sysconf file and allow 0 for it
|
|
wizard mode: don't include feedback about named fruit for ^X and enlightenment
|
|
looking at distant objects while wearing the Eyes of the Overworld made their
|
|
up-close descriptions known when not intended
|
|
message when cursed wand zapped by a monster happens to explode was suppressed
|
|
if hero was deaf, even though that message has no audible component
|
|
support explicit 'symset:default' and 'symset:Default symbols' in options
|
|
crash during startup if player name set as 'player' in defaults
|
|
any existing vampire shape-shifted into critter (fog cloud, bat, wolf) became
|
|
an unkillable critter if vampires were genocided
|
|
unlike in previous versions, an uncursed scroll of enchant armor failed to
|
|
uncurse the piece of armor being enchanted (change was unintentional)
|
|
slice of {pizza,cake,&} pluralized as "slouse of ..." due to false match
|
|
with "lice" (discovered pre-3.6.0-release)
|
|
change dipping prompt to not ignore 'called' and 'named' attributes of item
|
|
to be dipped
|
|
avoid 'the unlabeled {scroll,spellbook} fades' when blank item is hit by water
|
|
wrong message given when high priest on astral plane rejects being assigned a
|
|
name (got the one for unique monsters instead of the one for priests)
|
|
negative intrinsic protection shouldn't confer MC=1, "you are warded" (not
|
|
possible from divine protection but is possible from eating rings)
|
|
make a slight adjustment to the quickmimic() sense wording
|
|
fix typo in passage 1 of The Colour of Magic
|
|
falling asleep when reading dull spellbook ignored sleep resistance
|
|
getpos() complaint about invalid movement keystroke didn't describe meta-chars
|
|
accurately
|
|
'realtime' value in xlogfile was incorrect if 'checkpoint' option was active
|
|
make a previously-discovered scroll written with marker while blind have its
|
|
label known so it can be read while blind
|
|
#name or C for discoveries list that spanned multiple pages would exit on
|
|
space instead of advancing to next page (workaround: use '>' instead)
|
|
don't show the old level when you die going down the stairs because of an
|
|
iron ball
|
|
|
|
|
|
Platform- and/or Interface-Specific Fixes
|
|
-----------------------------------------
|
|
tty: M-N gave "Unknown command 'M-" with "'." finishing the sentence on the
|
|
line below it, leaving bogus '.' displayed on the top row of the map
|
|
tty: specifying all four of role, race, gender, and alignment still prompted
|
|
for confirmation with "Is this ok?" before starting play
|
|
unix/X11: in top level Makefile, some commented out definitions of VARDATND
|
|
misspelled pilemark.xbm (as pilemark.xpm)
|
|
win32gui: getversionstring() was overflowing the provided Help About buffer
|
|
win32gui: guard against buffer overflow in in mswin_getlin()
|
|
MacOSX: initial binary release was built from out of date source code that
|
|
had 'BETA' and 'DEBUG' inappropriately enabled
|
|
X11: core bug for '`' (backtick) command was only noticed by X11 interface,
|
|
which issued impossible message "add_menu: called before start_menu"
|
|
|
|
|
|
General New Features
|
|
--------------------
|
|
naming Sting or Orcrist now breaks illiterate conduct
|
|
different feedback for reading a scroll of mail created by writing with marker
|
|
|
|
|
|
Platform- and/or Interface-Specific New Features
|
|
------------------------------------------------
|
|
|
|
|
|
NetHack Community Patches (or Variation) Included
|
|
-------------------------------------------------
|
|
|
|
|
|
Code Cleanup and Reorganization
|
|
-------------------------------
|