The persistent inventory window wasn't being updated if you toggled
the 'sortpack' option interactively. (The new setting was honored
once something else triggered a 'perm_invent' update.)
The logic for toggling 'fixinv' seemed backwards. I hope this "fix"
for that hasn't actually broken it.
Make some progress on a couple of next minor release checklist
items, hopefully without introducing too many new bugs. This
is just the initial commit, and work continues.
Checklist items:
Savefiles compatible between Windows versions, whether 64-bit
or 32-bit in little-endian field format.
Selection of file formats:
historical (structlevel saves),
lendian (little-endian, fieldlevel saves),
and just for proof-of-concept, ascii fieldlevel saves
(the ascii is huge! 10x bigger than little-endian).
For the fieldlevel save, all complex data structures recursively
get broken down until until it is one of the simple types that
can't be broken down any further, and that gets when it gets
written to the output file.
New files needed for this build:
hand-coded:
include/sfprocs.h
src/sfbase.c - really a dispatcher to one of the
output/input format routines.
src/sflendian.c - little-endian output writer/reader.
src/sfascii.c - ascii text output writer/reader.
auto-coded (generated):
include/sfproto.h
src/sfdata.c
This is just one approach. I'm sure there are countless others
and they have different pros and cons.
For producing the auto-coded files a utility called
universal-ctags, that is actively maintained and evolving,
was used to do all the heavy-lifting of parsing the
NetHack C sources to tabulate the data fields, and store
them in an intermediate file called util/nethack.tags
(not required for building NetHack if you already have a
generated include/sfproto.h and src/sfdata.c)
util/readtags (also not required for building NetHack
itself) will decipher the nethack.tags file and produce
the functions that can deal with the NetHack struct data
fields.
You can obtain the source for universal-ctags by cloning it
from here:
https://github.com/universal-ctags/ctags.git
The combination universal-ctags + util/readtags has been
tried and tested under both Windows and Linux, so it is
not tied to a particular platform.
Note: util/readtags will work only with universal-ctags
output, so other ctags are unlikely to work as-is.
Universal-ctags can be build from source very easily
under Linux, or under Windows using visual studio.
Prevent the fuzzer from randomly toggling the 'silent' option. If
you use the default value of True then this eliminates most--but not
all--of the beeping that happens when it is running. I'm not sure
where the remaining beeps are coming from.
Modify the random keystroke selection to implement some bias towards
direction keys that I thought had already been there, plus a higher
chance to entering digits to initiate number responses.
The curses interface wouldn't build with HILITE_STATUS disabled. I
started adapting it to handle genl_status_update() but that was taking
too much effort with each niggling detail leading to another. This
goes the opposite direction: forcing the old STATUS_VIA_WINDOWPORT
behavior without having that #define available. That dragged along a
bunch of unexpected changes too.
I didn't noticed this because I've been building for tty+curses+X11
and either of the first two cause iflags.extmenu to exist. Make it
unconditional; there's not much benefit from trying to suppress it
for configurations that don't need it.
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.
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.
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.
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.)
This started out as an attempt to document the curses options in the
Guidebook, but I didn't actually get that far. Instead, integrate
the curses options better via more consistent WC/WC2 usage. This
prevents 'guicolor' from showing up as a boolean option for non-curses
interface in curses+other binary.
For curses itself, let 'petattr' be set/reset via 'O'. Also, accept
'Dim' as a possible pet highlight attribute since it already handles
all the other ordinary attributes. I'm not sure what leftline and
rightline highlighting are supposed to do. They were missing for
ncurses (or maybe they're misspelled for PDcurses?) but adding them
didn't produce any visible effect (using TERM=xterm-256color on OSX
with default font/character set).
Not addressed:
1) general confusion about compile-time vs run-time option filtering;
2) curses pet highlighting only works if 'color' option is enabled.
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.
This is branched from Alex's hallu-rng-stability branch,
with two build corrections (detect.c, zap.c), and merged
with the isaac64 branch that we have ready to go.
Alex's dual rng is supported by setting up the array
of multiple isaac64 contexts.
I stuck with Alex's approach of passing the rng function
name around as the parameter (rng or rn2_on_display_rng)
for the new additional parameter needed for
set_random(), init_random(), reseed_random(),
and init_isaac64().
Linux and BSD system have random number source devices that can be used
as source for a unguessable seed source.
Other platforms fall back to generate the seed with gettime().
This started out removing one tab and I got carried away. It moves
some labels to column 2, removes some parentheses where sizeof is
used on strings rather than types, adds or revises several comments,
replaces a couple of 'while' loops which can be simplified as 'for'
loops, and updates named fruit handling.
"glob of black pudding" became "candied glob of black pudding" if used
as a fruit name, but "small glob of black pudding" was used as-is and
became indistinguishable from an actual small glob. Unless you had
more than one; then you could try to check whether they merged into a
stack or coalesced into a bigger glob (but if neither of those changes
happened, you still couldn't tell which was the glob and which was the
named fruit).
The 'O' handling for bouldersym was updating the display value for
boulder even if the value had been rejected, and if it still had the
default of '\0', the map would end up with <NUL> characters. (When
examined via '//' or ';', those matched dummy monster class #0 and
led to the impossible "Alphabet soup: 'an("")'" that was suppressed
yesterday.)
Attempting to set bouldersym to ^@ or \0 would also be rejected as
duplicating a monster symbol. That is now accepted and used to reset
the boulder symbol to default. However, other control characters are
also accepted--not due to this patch, they already are, and from a
config file in addition to via 'O'--so bouldersym can still disrupt
the map. But that's no different from putting control characters
into a symbol set or setting them from config file via S_foo:^C.
I ran the fuzzer with MONITOR_HEAP enabled and heaputil found a dozen
or so un-free'd allocations, all made by the same dupstr() call in
special_handling() for "symset" and "roguesymset". (Reproducible with
a few tens of thousands of fuzzer moves, although you have to take
over from the fuzzer and make a clean exit rather than just interrupt
it or there'll be lots of other un-free'd memory.) I haven't actually
figured out how/why it was leaking, but reorganizing the code has made
the leak go away (according to a couple of even longer fuzzer runs) so
I'm settling for that.
