I finally figured out why status gets updated periodically even if
none of the fields have changed. Once a temporary highlight times
out, it starts a cycle of timeouts every 'statushilites' turns. When
I worked on this before, it was convoluted but not this convoluted.
In moveloop, if 'context.botl' call bot; in bot
call evaluate_and_notify_windowport;
for each field, call evaluate_and_notify_windowport_field:
call hilite_reset_needed and set 'reset' to the result;
if 'reset' is True then do status_update
and set 'curr->chg' and 'prev->chg' to True.
Then in moveloop call status_eval_next_unhilite:
for each field
if 'curr->chg' set 'curr->time' to moves+hilite_delta;
on the call after hilite_delta ('statushilites') moves,
call hilite_reset_needed which returns True if there is any
rule for temporary highlight and set 'context.botl'.
Go back to start. If multiple fields had temporary timeouts and
they were activated on different turns so expired on different turns
you could conceivably end up with context.botl being set every turn.
My first writeup trying to explain all this was wrong. I won't
testify about the accuracy of this one in court....
This extends the highlighting data structure to track the current
rule that's in use. And for that to make sense, it eliminates the
merging of settings from multiple matching rules. So anybody with
hit-points/up/inverse
hit-points/up/green
hit-points/up/bold
will need to manually merge their rules like
hit-points/up/green&inverse+bold
or else whichever rule matches last will be the only one in effect.
There are a lot of miscellaneous changes made as I flailed about.
The three most significant ones are that there is no guesswork over
what kind of highlight rule is in effect, status_eval_next_unhilite
will only set a timeout value if the current rule is for a temporary
highlight, and hilite_reset_needed will only return True if a timeout
is for a temporary highlight (probably moot after the _next_unhilite
change).
Have the curses interface save and restore message history for use
by ^P. It doesn't spit the saved messages out into the visible
message window after restore; that's too distracting.
Noticed while testing statuslines on a small terminal window. Using
the cursor to pick locations that panned the map to view a new subset
would end up showing a new view of the regular map rather than a
different section of what was currently displayed. For farlook that
caused monsters to take on new hallucinatory forms which was fairly
inconsequential, but for #terrain and various forms of detection it
reverted to the ordinary map instead of showing the map features that
the player requested or the temporarily revealed monsters and such.
Most interfaces keep track of the whole map and just show their view
of the new subset when panning, similar to redisplay after being
covered up and then re-exposed, but tty isn't doing that. I made
same change to Amiga as to tty since the code it was using was very
similar. I haven't touched any of the other interfaces and assume
that they don't need this. I've verified that curses and X11 don't.
Implement the 'statuslines' option for tty. 2 and 3 line status are
similar to curses. Tty's version doesn't include insertion of extra
spaces for enhanced readability, or ignoring 'showexp' when space is
needed for other fields, or right justifying 'score' and suppressing
it when there isn't room for the entire number. It continues to have
abbreviated condition and encumbrance descriptions that curses lacks
which get used when the normal ones take up too much space.
'statuslines' can be set with 'O' so it is feasible to switch back
and forth between 2 and 3 lines on the fly. But only if the display
is at least 25 lines (actually ROWNO+4) or else CLIPPING is enabled
at build time.
This fixes the bug where after resorting to abbreviated condition
values it sometimes (always?) wouldn't switch back after more room
became available. Abbreviated encumbrance values had problems too
(lack of leading space and not changing value if encumbrance changed
to anything other than unencumbered) and this fixes that as well.
Fixes#179
The Valkyrie goal level has two drawbridges and one of them was set to
have 50:50 chance to be closed (raised). But 'random' for drawbridge
open/closed (or lowered/raised) state was always choosing open (lowered).
This fixes that and also changes that level to have the old settings
(southern open, northern 50:50) for 75% of the time and new settings
(both 50:50) for 25% of the time. So there's now a 12.5% chance that
both will be closed instead of both always being open (due to the bug
with handling random).
Catch up with curses and have hitpointbar work even if statushilites
is 0 to suppress other highlighting. Indirectly fixes #H8389 by
making the circumstance which triggered that bug no longer do so.
Back in December, a change was made to suppress status when u.uhp == -1.
But if the hero died with exactly that amount, the status display would
be blanked out during end of game disclosure. Force u.uhp to be 0 when
dying. That was already happening if death occurred while hero still
had positive HP, but not when damage took him/her to negative.
Cleaver's ability to hit up to three adjacent targets could kill a
long worm and then try to cut it in two. When this was first reported
I was unable to reproduce it, but this time I've managed to do so.
But not reliably, so it's hard to claim that it's now fixed. However,
the new report's explanation of why it happens and suggested fix was a
big help. I had been trying to hit three tail segments, but you need
to attack a segment next to the head, then have Cleaver hit and kill
the head first (50:50 chance depending upon whether current swing is
clockwise or counter). Worm cutting would be looking at the location
of the targetted segment but there won't be any monster there when the
head dies. (Cleaver's attack itself already copes the situation where
its 2nd and/or 3rd potential targets aren't there any more by the time
it's ready to try to hit them.)
curses uses 'reversed' (LIFO) style when displaying previous messages.
Use the existing (previously tty-only) 'msg_window' option to also
support 'full' (FIFO). The actual code needed as just a couple of
lines; tweaking options parsing and the documentation was more work.
Using ^P right after resize or 'O' of align_message, align_status,
statuslines, or windowborders would result in
'curses_display_nhmenu: attempt to display empty menu'
because some memory cleanup I added several weeks back was being
executed when the curses interface tore down and recreated its
internal windows.
This fixes ^P handling by making sure that that menu (which is just
text but uses a menu to support '>'/'<'/'^'/'|' scrolling) will never
be empty and it also fixes the window deletion to not throw away
message history until it's final deletion at exit time.
^P uses a popup window to display previous messages and it was never
deleting that window, just creating a new one each time. Same with
the routine which displays an external help file. Using either or
combination of both close to 5000 times would probably make internal
window creation get stuck in an infinite loop. Delete those windows
after they're used so it'll never be put to the test.
The memory cleanup I added for map/status/messages/invent was only
being preformed at end of game, not when saving. Fix that too.
I've overhauled the status display for curses. Horizontal layout
supports both 2 lines and 3 lines which can be changed dynamically
via using 'O' to set 'statuslines'. Fields are spread out a little
more than they used to be, making it more readable--at least to me--
but the extra spaces get squeezed out when lines become too long.
If 'showexp' is on and either conditions or hunger+encumbrance go
off the right edge, experience points are suppressed (but the option
is left on, so they'll come back once there is room).
For traditional 2-line hozizontal status, if hunger+encumbrance+
conditions go off the right edge even after experience points are
knocked out, there will be a '+' in the rightmost column if there
are any conditions that are all the way off. At present it doesn't
use the tty method of switching to abbreviated condition names to
reduce their legnth. I'll probably tackle that eventually if no one
beats me to it.
For 3-line horizonal status, there was an older implementation (but
disabled via #if 0) with gold and score moving to the third line.
(I'm not sure how status conditions were handled.) This one ignored
that and modified 2-line from scratch, moving alignment from line one
to line 2 and level description, time, and conditions from line 2 to
line 3. It looks like this (view with a fixed-width font...).
Wizard the Hatamoto St:16 Dx:15 Co:18 In:8 Wi:11 Ch:7 S:25
Lawful $:21 HP:25(25) Pw:6(6) AC:4 Xp:2/21 Hungry Burdened
Dlvl:1 T:36 Blind Lev
Score is actually right aligned with the edge but I've deleted several
spaces to keep the line shorter here. The status conditions line up
with the hunger slot as that shifts due to changes in gold/HP/power/AC/
experience, and conditions prefer that column even when hunger and/or
encumbrance are blank. Howver, if the number of conditions increase to
the point where they would go off the edge, the whole list shifts left
instead of trying to stay lined up with hunger. (It's just coincidence
that the lefthand parts of lines 2 and 3 seem to line up in this sample.
In general, they don't.)
The vertical layout has reordered most of the fields and now has a few
blank lines to separate those fields into some groups for readability.
Lines have the form of
Field-name : Value
and when highlights apply, now they only affect the value portion.
Single digit characteristics are padded with a leading space so that
all six of them line up (for "18/xx", "/xx" protrudes to the right).
HP and Pw are aligned with each other. Hunger and encumbrance share a
line. When there are more than three conditions, they're shown three
per line instead of wrapping across lines. And if too many lines are
present, it will squeeze out enough blank ones to fit.
To see the vertical status, you need a display size of at least 106
columns with 'windowborders' explicitly off, or 110 with them on; also
set option 'align_status' to 'right' or 'left'. (With borders on,
including the default 'auto' setting, the vertical status appears at
width of 108 columns, but does so by hiding 2 columns of the map; using
110 columns avoids that.) Resizing from outside the game or changing
align_status via 'O' both cause dynamic reconfiguration of the layout;
there's no need to save, make config changes, then restore.
I've noticed (with curses interface) that there are a lot of
update_status(BL_FLUSH) calls when nothing on the status lines
has changed. It happens while just wondering around, not due to
deliberate botlx update. This eliminates one of the causes.
Updating hero's experience points or score (via u.urexp) was
requesting a bottom lines refresh even if the relevant options to
see those changed values was off. The botl code would send a flush
directive even though nothing visible was modified. (I have a fix
for that too, but am holding back while hoping to find some of the
other causes of unnecessary botl requests.)
It took me a while to track this down: if you use 'O' to create
hilite_status rule(s) for status condition(s) and you specify multiple
attributes in the rule creation menu, it accepts them but it was
parsing the new rule(s) incorrectly and only supported one attribute.
if (mask & bit)
something;
else if (mask & other_bit)
something_else;
else if (mask & yet_another_bit) ...
effectively stops at the first bit matched. (At the time that that
was written, the menu leading to it only accepted a single attribute.)
More groundwork for overhauling the status display for curses, plus
a few functional changes. It was doing a full status update for
every changed field (except conditions), instead of waiting for a
flush directive after gathering multiple changes at a time. Since
it already does gather every change, the fix to wait is trivial.
This decouples 'hitpointbar' from 'statushilites'. When highlighting
is off, it uses inverse video only. When on, it behaves as before:
using inverse video plus the most recent color used to highlight HP
(which can vary if that has rules to highlight changes or percentage
thresholds) but ignoring any HP attribute(s). This also enables the
latent 'statuslines' option and changes 'windowborders' option from
being settable at startup only to changeable during play.
'statuslines' can have a value of 2 (the default) or 3 and applies to
'align_status:bottom' or 'top'; it's ignored for 'left' and 'right'.
At the moment, setting it to 3 only allows status condition overflow
to wrap from the end of line to 2 to the beginning of line 3, and if
window borders are drawn they'll clobber the last character on line 2
and first one on line 3. There's no point in trying to fix that
because it will go away when the main status overhaul changes go in.
Condition wrapping for vertical orientation (left or right placement)
was already subject to the same phenomenon and will be superseded too.
This also changes the meaning of the 'windowborders' value so could
impact players using source from git (or possibly beta binaries for
Windows, but not for OSX where curses interface wasn't included).
Old:
0 = unspecified, 1 = On, 2 = Off, 3 = Auto (On if display is big
enough, Off otherwise; reevaluated after dynamic resizing);
Unspecified got changed to 3 during curses windowing initialization.
New:
0 = Off, 1 = On, 2 = Auto;
0 gets changed to 2 for default value at start of options processing.
So old value of 2 is changing meaning and explicit old value of 3 is
becoming invalid. Implicit 3 changes to default 2. Explicit 3 could
be the subject of a fixup but there isn't much point since 2 can't
have a similar fix. Users who are using old 2 or explicit 3 will need
to update their run-time config files.
This adds 'statuslines' to the Guidebook and moves some other recently
added documentation of curses options from among the general options
(section 9.4) to "Window Port Customization options" (section 9.5).
None of them have been added to dat/opthelp which seems to be missing
all the wincap options.
Originally I made a lot of changes (mostly moving C99 declarations to
start of their blocks) to the old '#if 0' code at end of cursstat.c,
but have tossed those, except for one subtle bug that assumed 'int'
and 'long' are the same size.
Change the generic status line location "End Game" to relevant element
name "Earth", "Air", &c. ("Plane of <element>" might be too long if
hungry and encumbered and afflicted by conditions. "Astral Plane" is
already specific so not affected.)
Honor hilite_status rules specifying color even if curses-specific
option 'guicolor' is off.
Update status from scratch when 'O' is used to manipulate hilite_status
rules.
The curses interface was ignoring video attributes (bold, inverse, &c)
when color is toggled off or if built with TEXTCOLOR disabled. Honor
attributes regardless of whether color is displayed.
Also, toggling 'hilite_pet' On during play wouldn't do anything if the
curses-specific 'petattr' option had been left as None. (It worked as
intended if set in starting options.)
Change the test for whether fonts.dir exists (added to the script
in 3.6.0, for automatically setting up possible use of the NH10 font
under X11) from 'test -e file' to 'test -f file' since the latter
seems to be more universally available. When present, fonts.dir is
plain text, so a test for "exists and is a regular file" rather than
one for general existance is appropriate.
If nethack is built to use graphical tombstone but file rip.xpm is
missing from the playground, there would be a crash if the rip output
was shown. My first attempt to fix it prevented the crash but didn't
display any tombstone, just the last couple of lines of output which
follow the tombstone. This keeps that in case of some other Xpm
failure, but checks for rip.xpm via stdio and reverts to genl_outrip
for text tombstone if it can't be opened.
Extend the earlier support for Delete/Rubout in getline() to the
text entry for extended commands. In other words, treat <delete>
and <backspace> as synonyms in both places.
Some reformatting too, but only in a couple of the files.
Kicking a stack splits off one item (except for gold coins) and
propels it, but the range for how far it would move was calculated
before the split using the entire stack's weight. So a large stack of
small items might fail with "thump" (which the report suggested hurt
the hero, but it doesn't) and none of the stack would move. Splitting
sooner looked complicated because of several potential early returns
between the range calculation and the eventual kick, so this hacks the
stack's quantity to get the intended weight instead.
showed non-empty containers in inventory (including the one being
applied) with a 'for sale' suffix during put-in operations, as if the
shop was trying to sell it to the hero. Amount shown was cumulative
value of its contents. (Using /menustyle:T doesn't show the container
being applied so this wasn't visible with it unless other non-empty
containers were being carried.)
Two or three fix attempts solved one problem but introduced another.
This one seems to finally get things right but considering that there
was trial and error along the way, my confidence isn't great.
If steed ate a mimic corpse and started mimicking an object or dungeon
furniture, the hero was able to keep riding. Force a dismount when
that happens, even if steed takes on monster shape rather than object
or furniture. After that, #ride to remount non-monster will fail
unless using wizard mode's "force mount to succeed" action, in which
case steed's eating finishes immediately and it returns to normal.
This doesn't address the older report that mounted hero can continue
to move around while the steed is eating.
number_pad got removed from Guidebook.tex when mouse_support was added
by cc0e7a8750fd4e07d3a44592c38f5912d4e603de; put it back. Evidently
a cut-and-paste mishap when copy-and-paste was intended, probably
provoked by '\numbox{}'.
Remove \numbox{} from mouse_support; it isn't needed there since the
choices aren't trying to line up with '-1'.
Both Guidebook.mn and Guidebook.tex:
Remove number_pad's "for backward compatibility" from mouse_support
since the latter is brand new.
New petattr list of uppercase letters seemed intrusive when rendered
with a bold font, so switch to lowercase. (Option processing already
supports either case.)
Mention the '=' prefix for hilite_status 'absolute' threshold.
[As usual, the Guidebook.tex changes haven't been tested.]
Report forwarded from spam filter so not included in bugzilla list.
Make corpse revival feedback be more consistent. Some of the healer-
specific flavor is still there.
Generally, fish should lay their eggs in the water and
not on land, but the game was only allowing the opposite.
Eels are catadromous and lay their eggs in the Sargasso Sea,
not in the dungeon.
Something else noticed while testing #H8271: toggling perm_invent on
with 'O' didn't show anything (at least with curses) until some later
action caused it to be updated. Make updating persistent inventory be
included with full redraw and set the need_redraw flag when toggling
perm_invent.
Realized while fixing #H8271: if persistent inventory got an update
while wearing or taking off was in progress (not within user's control
since hero is busy) the item in question was flagged as "(being worn)"
even though it wouldn't be worn if putting on got interrupted. Update
doname() to show "(being donned)" or "(being doffed)" instead of
"(being worn)" when corresponding operation is in progress. (During
testing, I was able to observe "being doffed" but never managed to see
"being donned".)
A relatively recent change moved 'obj->known = 1' when wearing armor
from before setworn(), which issues an update_inventory() call, to
afterwards. There wasn't any particular update then, so observing
the enchantment of armor by wearing it wasn't being reflected in the
persistent inventory window if that was enabled.
Take another crack at describing yesterday's do-again fix. Having
'autoquiver' enabled wasn't necessary to encounter the problem.
Also, 'in_doagain' is an int rather than a boolean.
It's possible to get a rolling boulder trap which doesn't have any
boulder. That isn't invalid, but if/when it happens on a shallow
level it shouldn't be covered by the corpse of a fake adventurer
since such a trap won't kill anyone.
One-line fix is much shorter than attempting to describe the problem.
^A could misuse previous input if 'f'<direction> needed to fill the
quiver and there was nothing suitable, so that the sequence became
'f'<what to throw>. If previous <direction> was an inventory letter
that was occupied, and the item it that slot wasn't already worn in
some other slot, it would be put in quiver slot. Then player would
be asked for direction rather than immediately throwing it since the
what-to-throw prompt had just used up the last of the ^A queue.
Miscellaneous formatting included....
The fuzzer likes to set options randomly; the combination of
DECgraphics symbol set (on a display capable of rendering it) plus
eight_bit_tty produces a bizarre map display. Make DECgraphics
override eight_bit_tty rather than the other way around.
Leather jacket doesn't take multiple turns to wear, so wearing it
wasn't calling Armor_on() and recently moved 'uarm->known = 1' didn't
get executed. Not reported yet but had the same issue: fedora and
dented pot wouldn't call Helmet_on().
Closing nethack's window sets 'program_state.stopprint' to inhibit
disclosure interaction, but shopkeeper claiming hero's stuff or vault
guard claiming hero's gold didn't honor that and just issued normal
pline messages. For win32, they got delivered in a popup even though
nethack's window had gone away.
Make those two end-of-game situations honor 'program_state.stopprint'.
[Fix not tested on win32...]
about the armor. Wearing armor sets obj->known, making its enchantment
be shown when it gets formatted, because the AC value on the status line
lets the player deduce what that is. It was being set at the beginning
of the wear operation. If the armor got stolen before it became fully
worn, the enchantment was still shown. Defer that until the end of the
operation. An attentive player can still deduce the enchantment if the
item is stolen (because its protection starts immediately) but the hero
won't learn that enchantment unless the donning completes.
This might be suboptimal but it isn't qualitatively different from
watching a pet walk/not-walk over items whose bless/curse state isn't
known or dropping unidentified items in a shop to check their price.
The player can deduce something that the hero doesn't know yet.
Support <delete> (aka <rubout>) during getline(). It doesn't actually
honor the current erase_char value set up for the terminal, just
treats DEL the same as ^H. (The previous lack of support had nothing
to do with terminfo specifying ^H; the handling is hard-coded.)
tty treats escape while there's already some input as kill_char (erase
the input but get more from scratch) and returns ESC if there isn't.
curses was doing the first half but not the second, so not providing
any way to communicate "cancel" back to the core. Fix is simple.
Other getline() bug fixes:
1] there was a wprintw("%*something") which was passing the value from
strlen (type 'size_t') to the "%*" argument (type 'int'). That's
always wrong (size_t is guaranteed to be unsigned) and could be severe
(if size_t is different width than int--as on current OSX systems--
depending upon the internals of argument passing).
2] strncpy() only supplies a terminating '\0' if the input is shorter
than the number of characters specified.
A lot of reformatting is warranted but I only did the getline routine
(manually, so might have missed stuff).
There was no provision for malloc() potentially returning Null and it
wasn't integrated with nethack's MONITOR_HEAP. 'heaputil' shows that
the curses interface is leaking like a sieve. If some things are
actually being allocated separately and then freed from within curses,
those need to be thoroughly documented and maybe switched back to
malloc().
The curses interface already has a hack to keep 'Count: 12', 'Count:
123' intermediate multi-digit counts out of its message recall history
for ^P, but it was flushing real messages when getpos()'s 'autodescribe'
reported what the cursor moved over. Overload the count hack to support
putstr(WIN_MESSAGE, ATR_NOHISTORY, text)
(which is what custompline(SUPPRESS_HISTORY, ...) eventually calls).
The conditional logic for when to create the 'count_window' was pretty
convoluted. This simplification has the same semantics but I don't
have PDCURSES to actually verify that.
