]> granicus.if.org Git - nethack/log
nethack
6 years agoMerge branch 'NetHack-3.6'
nhmall [Thu, 30 May 2019 22:10:41 +0000 (18:10 -0400)]
Merge branch 'NetHack-3.6'

6 years agomnearto/mnexto/enexto
PatR [Thu, 30 May 2019 14:50:38 +0000 (07:50 -0700)]
mnearto/mnexto/enexto

This doesn't solve the <0,0> problem but it does prevent mnexto()
from using uninitialized coordinates if enexto() fails.  It also adds
several debugging messages.

enexto() was ignoring map row #0 (unlike column #0, row #0 contains
valid map locations).  Fixing that doesn't matter for Plane of Water
though since that row is stone there--that's probably a bug.  It was
also repeatedly re-testing the top+1 and bottom rows and left and
right columns after they had already failed to be acceptable.  It
still does some of that, but less.

6 years agosortloot memory leak
PatR [Wed, 29 May 2019 11:40:54 +0000 (04:40 -0700)]
sortloot memory leak

Unfreed memory noticed after interrupting the fuzzer and quitting.
query_objlist() has an early return--for touching a cockatrice
corpse--that was skipping release of sortloot info (an array with
one element per object from whichever object list was being used).

Some formatting that's been sitting around for a while got mixed in
and I decided not to take that back out.

6 years agoDeleting test file which was just checked-in.
Bart House [Wed, 29 May 2019 03:30:10 +0000 (20:30 -0700)]
Deleting test file which was just checked-in.

6 years agoMerge branch 'NetHack-3.6' of https://rodney.nethack.org:20040/git/NHsource into...
Bart House [Wed, 29 May 2019 03:19:33 +0000 (20:19 -0700)]
Merge branch 'NetHack-3.6' of https://rodney.nethack.org:20040/git/NHsource into NetHack-3.6

6 years agoTesting ability to cause chaos.
Bart House [Wed, 29 May 2019 03:18:27 +0000 (20:18 -0700)]
Testing ability to cause chaos.

6 years agoMerge branch 'NetHack-3.6'
nhmall [Wed, 29 May 2019 01:36:52 +0000 (21:36 -0400)]
Merge branch 'NetHack-3.6'

6 years agoMerge branch 'NetHack-3.6' of https://rodney.nethack.org:20040/git/NHsource into...
nhmall [Wed, 29 May 2019 01:24:56 +0000 (21:24 -0400)]
Merge branch 'NetHack-3.6' of https://rodney.nethack.org:20040/git/NHsource into NetHack-3.6

6 years agoclean up Windows panic results further
nhmall [Wed, 29 May 2019 01:24:04 +0000 (21:24 -0400)]
clean up Windows panic results further

6 years agowizard mode memory leak in endgame
PatR [Wed, 29 May 2019 00:08:48 +0000 (17:08 -0700)]
wizard mode memory leak in endgame

Leaving the Plane of Water to return to a previously visited endgame
level didn't free the air bubbles unless/until you visit a new level.
Returning to that level creates a new set of air bubbles, losing track
of the previous set.  Likewise with Plane of Air and its clouds.  (Not
an issue with actual save and restore when on those levels, or when
just moving forward to not-yet-visited levels.)

Not applicable to normal play where it isn't possible to return to a
previously visited endgame level.

For 3.7, bubble save/restore ought to become part of savlev() instead
of being handled by savegamestate().

6 years agoMerge branch 'NetHack-3.6'
nhmall [Tue, 28 May 2019 17:11:26 +0000 (13:11 -0400)]
Merge branch 'NetHack-3.6'

6 years agocurses getline()
PatR [Tue, 28 May 2019 09:27:40 +0000 (02:27 -0700)]
curses getline()

After going back and forth between prompts causing message lines
to be overwritten and to be skipped, this yoyo might have finally
run out of string.  Fingers crossed....

6 years agocurses message window
PatR [Tue, 28 May 2019 08:52:37 +0000 (01:52 -0700)]
curses message window

Fix a 'FIXME': don't follow a message with two spaces in anticipation
of combining with the next one, precede the next one with two spaces
when they're being combined.  Keeps nethack's message window <mx,my>
coordinates in sync with curses' internal coordinates.

6 years agoMerge branch 'NetHack-3.6'
nhmall [Mon, 27 May 2019 22:44:33 +0000 (18:44 -0400)]
Merge branch 'NetHack-3.6'

6 years agomake the rc file match the build target name for Windows
nhmall [Mon, 27 May 2019 22:42:25 +0000 (18:42 -0400)]
make the rc file match the build target name for Windows

6 years agomerge branch NetHack-3.6 and update version info
nhmall [Mon, 27 May 2019 22:23:50 +0000 (18:23 -0400)]
merge branch NetHack-3.6 and update version info

6 years agoMerge branch 'NetHack-3.6'
nhmall [Mon, 27 May 2019 22:21:34 +0000 (18:21 -0400)]
Merge branch 'NetHack-3.6'

6 years agoWindows .rc files
nhmall [Mon, 27 May 2019 21:38:54 +0000 (17:38 -0400)]
Windows .rc files

bump the version in the win/win32/*.rc files

side note: winhack.rc wasn't properly updated for the 3.6.2 release

6 years agoMerge branch 'NetHack-3.6'
nhmall [Mon, 27 May 2019 20:38:09 +0000 (16:38 -0400)]
Merge branch 'NetHack-3.6'

6 years agoend of game oddities: thrownobj, ball&chain
PatR [Mon, 27 May 2019 01:44:25 +0000 (18:44 -0700)]
end of game oddities: thrownobj, ball&chain

If you died while Punished but with attached ball and chain temporarily
off the map (changing levels and when swallowed are the cases I looked
at; there may be others), the ball and chain objects would not appear
in bones (for the falling-down-stairs case; bones are never saved if
hero dies while swallowed) and they weren't being freed.  Put them
back on the map so that they'll be included in bones and also freed as
part of normal map cleanup.

This caused a problem if the attached ball had state OBJ_FREE due to
being thrown rather than being temporarily off the map.  'thrownobj'
was being deallocated without first cancelling punishment, so uball
object was freed via thrownobj pointer but stale uball pointer still
referenced it.  Unpunishing would introduce sequencing issues because
that would need to be after attribute disclosure.  So instead of
deallocating thrown or kicked object, put it/them (can't actually have
both at the same time) back on the map.  This has a side-effect of
saving thrown Mjollnir in bones if it kills hero when failing to be
caught upon return.  (I thought that that had been fixed ages ago?)

6 years agoball.c formatting
PatR [Mon, 27 May 2019 01:23:05 +0000 (18:23 -0700)]
ball.c formatting

Mostly a couple of block comments.

6 years agoMerge branch 'NetHack-3.6' into NetHack-3.7
Patric Mueller [Sun, 26 May 2019 16:16:22 +0000 (18:16 +0200)]
Merge branch 'NetHack-3.6' into NetHack-3.7

6 years agosavelev() pasta
PatR [Sun, 26 May 2019 14:24:52 +0000 (07:24 -0700)]
savelev() pasta

Avoid some of the spaghetti [mostly the alternate call to
savecemetery()] in savelev().  There should be no change in behavior.

6 years agoMerge branch 'NetHack-3.6'
nhmall [Sun, 26 May 2019 12:18:03 +0000 (08:18 -0400)]
Merge branch 'NetHack-3.6'

6 years agofree level.bonesinfo
PatR [Sun, 26 May 2019 07:40:40 +0000 (00:40 -0700)]
free level.bonesinfo

Bones information for the current level was freed during save but not
at end of game.  Have freedynamicdata() call savelev(,,FREE_SAVE) to
throw away current level instead of trying to duplicate the actions
that performs.

6 years agosave.c cleanup
PatR [Sun, 26 May 2019 07:11:53 +0000 (00:11 -0700)]
save.c cleanup

Mostly 'sizeof' usage.  This has been sitting around for a while and
I wanted to get it out of the way before making some other save.c
changes.

6 years agofree overview->final_resting_place
PatR [Sun, 26 May 2019 06:43:42 +0000 (23:43 -0700)]
free overview->final_resting_place

The #overview command can provide some feedback about levels loaded
from bones files; that data wasn't being released at end of game.

(There are two copies of that data, one set always in memory with
the overview data [final_resting_place field in the 'mapseen' data],
and another set with portions attached to each relevant level [via
level.bonesinfo].  Neither set was being properly freed; this only
addresses one of them, so far. The per-level data can probably be
eliminated--for post-3.6--since DUNGEON_OVERVIEW isn't a conditional
feature as it was when that was implemented.)

6 years agorefer to the same level the same way in fixes file
nhmall [Sat, 25 May 2019 16:25:05 +0000 (12:25 -0400)]
refer to the same level the same way in fixes file

6 years agoMake sure the correct luckstone is the prize
Pasi Kallinen [Sat, 25 May 2019 16:15:53 +0000 (19:15 +0300)]
Make sure the correct luckstone is the prize

The first generated luckstone is marked as the prize in mines' end.
Make sure we generate the intended one first, before any other
(possibly randomly generated) luckstones.

6 years agoMerge branch 'NetHack-3.6' of https://rodney.nethack.org:20040/git/NHsource into...
nhmall [Sat, 25 May 2019 16:01:58 +0000 (12:01 -0400)]
Merge branch 'NetHack-3.6' of https://rodney.nethack.org:20040/git/NHsource into NetHack-3.6

6 years agodo not mark multiple luckstones as the prize in mines-end
nhmall [Sat, 25 May 2019 15:58:32 +0000 (11:58 -0400)]
do not mark multiple luckstones as the prize in mines-end

if one of the random objects happened to be a luckstone then
it and the explicit one got marked as a prize.

Following this change, only one will be marked as the prize,
but a follow-up on the order of things in mines.des may be
warranted to ensure it is the explicitly placed luckstone.

6 years agocurses memory leak
PatR [Sat, 25 May 2019 14:37:08 +0000 (07:37 -0700)]
curses memory leak

Back in February, my e991dd1b0c4192495c7e9092c10fa83bb6d81a24 added
ESC (when there's no input) as an early return for curses' getline,
but it neglected to clean up some allocated memory.

6 years agowhitespace bit
nhmall [Sat, 25 May 2019 04:42:29 +0000 (00:42 -0400)]
whitespace bit

6 years agoMerge branch 'NetHack-3.6'
nhmall [Sat, 25 May 2019 04:35:48 +0000 (00:35 -0400)]
Merge branch 'NetHack-3.6'

6 years agoswap places with pet onto boulder location
nhmall [Sat, 25 May 2019 04:33:47 +0000 (00:33 -0400)]
swap places with pet onto boulder location

when co-located with a boulder you could sometimes swap places
with pets of any size

bg72

6 years agovortex database update
nhmall [Sat, 25 May 2019 02:36:21 +0000 (22:36 -0400)]
vortex database update

Reported:
"The data.base entry for "vortex" is erronous, as fire and energy vortices have passive attacks (which count as "touching")"

6 years agoMerge branch 'NetHack-3.6'
nhmall [Fri, 24 May 2019 12:01:50 +0000 (08:01 -0400)]
Merge branch 'NetHack-3.6'

6 years agocurses message window vs prompting
PatR [Fri, 24 May 2019 08:33:45 +0000 (01:33 -0700)]
curses message window vs prompting

Fix a problem introduced by f218e3f15e2d5525e817fb07e6a8b4773f296073
and/or a19e64e470f21ef30b96122dd2d40c26ef9b1527.  Sometimes the line
after a prompt would be empty and the next message get shown on the
line after that.  a19e64e470 was intended to fix the opposite problem
so probably overshot the mark....

6 years agocurses message window refresh
PatR [Fri, 24 May 2019 01:56:20 +0000 (18:56 -0700)]
curses message window refresh

Sometimes curses tears down and recreates all its windows (when the
display is resized, for instance) and after doing that it repopulates
the message window with data saved for use by ^P.  But it was showing
the oldest messages available rather than the most recent ones.

There is still room for improvement.  That process combines short
messages but the refresh is based on the available number of lines;
combining messages can result in lines at the bottom of the message
window being left blank.  This could be fixed by reverse-scrolling the
window and inserting more messages at the top, or by combining short
messages in history data instead of at refresh time.  The second seems
easier but won't handle changing the message window's width sensibly,
and neither method handles wrapped, long lines well.  A More>> prompt
(possibly more than one) is issued if the refresh shows too many lines
(either because long messages already took multiple lines or because
the window has become narrower and ones which used to fit now need to
be wrapped).

6 years agocurses: remove duplicate wincap2 bit
PatR [Fri, 24 May 2019 00:54:37 +0000 (17:54 -0700)]
curses: remove duplicate wincap2 bit

WC2_HITPOINTBAR was OR'd into wincap2 bitmask twice.

6 years agoMerge branch 'NetHack-3.6'
nhmall [Thu, 23 May 2019 04:12:55 +0000 (00:12 -0400)]
Merge branch 'NetHack-3.6'

6 years agovs 2019 update bit
nhmall [Thu, 23 May 2019 04:11:18 +0000 (00:11 -0400)]
vs 2019 update bit

6 years agoxans fly, but could not reach your feet if you flew
nhmall [Wed, 22 May 2019 22:27:03 +0000 (18:27 -0400)]
xans fly, but could not reach your feet if you flew

6 years agofix #H8769 - steed drowns on Plane of Water
PatR [Wed, 22 May 2019 21:59:43 +0000 (14:59 -0700)]
fix #H8769 - steed drowns on Plane of Water

Air bubble movement on the Plane of Water manipulated <u.ux,u.uy>
directly when changing hero's coordinates, leaving steed with old
coordinates, resulting in dunking it when the old spot switched from
air to water.  Switch to u_on_newpos() which moves the steed with
the hero and also handles clipping when the screen is too small to
show the whole map at once.

6 years agoMerge branch 'NetHack-3.6'
nhmall [Wed, 22 May 2019 04:48:39 +0000 (00:48 -0400)]
Merge branch 'NetHack-3.6'

6 years agoball and chain sanity check revisited
PatR [Wed, 22 May 2019 00:40:57 +0000 (17:40 -0700)]
ball and chain sanity check revisited

Move some duplicated debugging code into its own routine.

6 years agocloses #79
nhmall [Tue, 21 May 2019 03:45:51 +0000 (23:45 -0400)]
closes #79

6 years agoMerge branch 'spixi-engravings' into NetHack-3.6
nhmall [Tue, 21 May 2019 03:43:18 +0000 (23:43 -0400)]
Merge branch 'spixi-engravings' into NetHack-3.6

6 years agoMerge branch 'NetHack-3.6.0' of https://github.com/spixi/NetHack into spixi-engravings
nhmall [Tue, 21 May 2019 03:41:40 +0000 (23:41 -0400)]
Merge branch 'NetHack-3.6.0' of https://github.com/spixi/NetHack into spixi-engravings

6 years agoMerge branch 'NetHack-3.6'
nhmall [Tue, 21 May 2019 02:10:01 +0000 (22:10 -0400)]
Merge branch 'NetHack-3.6'

6 years agowintty.c comments
PatR [Tue, 21 May 2019 01:08:32 +0000 (18:08 -0700)]
wintty.c comments

Some minor stuff I had pending that I stripped away for the status
conditions patch.

6 years ago3.7.0 WIP Makefile build required change
nhmall [Mon, 20 May 2019 20:00:11 +0000 (16:00 -0400)]
3.7.0 WIP Makefile build required change

6 years agoMerge branch 'NetHack-3.6'
nhmall [Mon, 20 May 2019 19:50:17 +0000 (15:50 -0400)]
Merge branch 'NetHack-3.6'

6 years agostatus line count (word two) bit
nhmall [Mon, 20 May 2019 19:16:57 +0000 (15:16 -0400)]
status line count (word two) bit

6 years agostatus_sanity_check is most useful in WIP and beta so restrict it
nhmall [Mon, 20 May 2019 16:08:51 +0000 (12:08 -0400)]
status_sanity_check is most useful in WIP and beta so restrict it

6 years agomore #H8609 - tty status condition rendering
PatR [Mon, 20 May 2019 12:26:19 +0000 (05:26 -0700)]
more #H8609 - tty status condition rendering

The earlier fix removed a valid optimization which happened to be
implemented incorrectly.  Put that back.  It also left an invalid
optimization when applied to conditions.  Remove that one.

I don't think either of these explains truncating 'y' off of "Hungry"
which was shown in one of the reports.

6 years agomake it tougher for incomplete render_status() to go unnoticed
nhmall [Mon, 20 May 2019 05:33:33 +0000 (01:33 -0400)]
make it tougher for incomplete render_status() to go unnoticed

Adds a sanity check that will write a paniclog
message if a code change prevents completion of
render_status() for each dirty (changed) field.

6 years agofix a botl status display issue
nhmall [Mon, 20 May 2019 03:35:21 +0000 (23:35 -0400)]
fix a botl status display issue

Reported as #H8609 (1679)

Some code recently added to render_status() for BL_CONDITION:
    if (!tty_condition_bits)
        continue;
was short-circuiting the required copy of NOW
values to BEFORE values for later comparison
further down in the for-loop.
    tty_status[BEFORE][idx] = tty_status[NOW][idx];

This caused some fields to be bypassed for rendering
once no more tty_condition_bits were set because the
length comparisons would match.

6 years agoremove a debug bit from Makefile.msc
nhmall [Sun, 19 May 2019 15:55:05 +0000 (11:55 -0400)]
remove a debug bit from Makefile.msc

6 years agoMakefile.msc bit for Windows
nhmall [Sun, 19 May 2019 15:48:20 +0000 (11:48 -0400)]
Makefile.msc bit for Windows

6 years agoMerge branch 'NetHack-3.6'
nhmall [Sun, 19 May 2019 14:12:39 +0000 (10:12 -0400)]
Merge branch 'NetHack-3.6'

6 years agofixes github issue #190 - EDIT_GETLIN for curses
PatR [Sun, 19 May 2019 06:52:04 +0000 (23:52 -0700)]
fixes github issue #190 - EDIT_GETLIN for curses

Fixes #190

Add EDIT_GETLIN support for curses.  It remains disabled by default.

6 years agodon't merge globs with differing BUC status
nhmall [Sun, 19 May 2019 02:56:27 +0000 (22:56 -0400)]
don't merge globs with differing BUC status

6 years agogrammar bit
nhmall [Sat, 18 May 2019 20:44:57 +0000 (16:44 -0400)]
grammar bit

6 years agoMerge branch 'NetHack-3.6'
nhmall [Sat, 18 May 2019 20:30:43 +0000 (16:30 -0400)]
Merge branch 'NetHack-3.6'

6 years agofurther improve additional glob interaction scenarios within a shop
nhmall [Sat, 18 May 2019 20:24:48 +0000 (16:24 -0400)]
further improve additional glob interaction scenarios within a shop

Scenarios:
1. shop_owned glob merging into shop_owned glob
2. player_owned glob merging into shop_owned glob
3. shop_owned glob merging into player_owned glob
4. player_owned glob merging into player_owned glob

6 years agocurses followup
PatR [Sat, 18 May 2019 15:12:43 +0000 (08:12 -0700)]
curses followup

Some prompts were being overwritten by the message that followed.

And clear_nhwindow(WIN_MESSAGE) gets called for just about every
keystroke so try to reduce the overhead I unwittingly added.  The
"scroll up one line earlier than the next message" mentioned in
the prior commit is much more obvious that I realized and prompt
erasure might need to be redone.

6 years agofix #H8753 - curses message window anomalies
PatR [Sat, 18 May 2019 09:25:48 +0000 (02:25 -0700)]
fix #H8753 - curses message window anomalies

Autodescribe feedback and multi-digit count prompts are always shown
on the last line of the message window and are suppressed from message
history (both ^P and DUMPLOG).  When the message window is using all
available lines, the last one was being overwritten (until the count
or the feedback was completed or dismissed, then last line returned).
Adopt the suggestion that it be scrolled up a line instead of being
overwritten.  [I haven't been able to reproduce the reported problem
where shorter overlaid text left some of longer underlying text visible
but that should now become moot.]

Bonus fix:  while testing, I noticed that if your screen only has room
for a one-line message window and you used ESC to cancel 'pick a spot
with cursor' prompting before moving the cursor, the prompt was left
intact on the message line.  tty erases it in that situation, but the
clear_nhwindow(WIN_MESSAGE) was a no-op for curses because it usually
doesn't erase old messages.  This changes the curses behavior when the
core asks it to erase the message window:  now it forces one blank line
of fake autodesribe feedback (causing the prompt or other most recent
message to scroll off top), then removes that fake feedback (leaving
a blank message line).  For multi-line message window, the old messages
scroll up by one line sooner than they would when waiting for the next
real message but are otherwise unaffected.

6 years agounpaid glob formatting
PatR [Fri, 17 May 2019 20:38:34 +0000 (13:38 -0700)]
unpaid glob formatting

Make unpaid (shop owned, that is) globs show same weight information
as for-sale globs.  And don't treat required arguments to globwt() as
if they were optional.

6 years agoMerge branch 'NetHack-3.6' of https://rodney.nethack.org:20040/git/NHsource into...
nhmall [Fri, 17 May 2019 20:34:07 +0000 (16:34 -0400)]
Merge branch 'NetHack-3.6' of https://rodney.nethack.org:20040/git/NHsource into NetHack-3.6

6 years agoMerge branch 'NetHack-3.6'
nhmall [Fri, 17 May 2019 20:31:26 +0000 (16:31 -0400)]
Merge branch 'NetHack-3.6'

6 years agowarning bit
nhmall [Fri, 17 May 2019 20:30:52 +0000 (16:30 -0400)]
warning bit

6 years agoobj sanity checking specific to globs
PatR [Fri, 17 May 2019 20:28:41 +0000 (13:28 -0700)]
obj sanity checking specific to globs

Verify that objects with the globby bit set are actually glob objects,
that their quantity is 1, and that their weight at least superficially
makes sense.

6 years agoMerge branch 'NetHack-3.6'
nhmall [Fri, 17 May 2019 20:19:17 +0000 (16:19 -0400)]
Merge branch 'NetHack-3.6'

6 years agocomment typo and an unused variable
nhmall [Fri, 17 May 2019 20:15:24 +0000 (16:15 -0400)]
comment typo and an unused variable

6 years agoMakefile.msc update for Windows
nhmall [Fri, 17 May 2019 19:57:04 +0000 (15:57 -0400)]
Makefile.msc update for Windows

6 years agoMerge branch 'NetHack-3.6'
nhmall [Fri, 17 May 2019 16:06:58 +0000 (12:06 -0400)]
Merge branch 'NetHack-3.6'

6 years agofix some billing and pricing issues when globs coalesce
nhmall [Fri, 17 May 2019 16:04:01 +0000 (12:04 -0400)]
fix some billing and pricing issues when globs coalesce

payment issue caused by glob coalescing
glob pricing did not consider coalesced weight

6 years agoglob floor merge message
nhmall [Fri, 17 May 2019 15:44:39 +0000 (11:44 -0400)]
glob floor merge message

Make it more obvious that globs that just merged on the
floor may be at an adjacent location.

6 years agofarlook feedback for overloaded symbols
PatR [Thu, 16 May 2019 22:26:35 +0000 (15:26 -0700)]
farlook feedback for overloaded symbols

When using '/' or ';' and picking--not just viewing the autodescribe
feedback for--a space or '#' on the map, the game would produce
 That can be many things (stone)
or
 That can be many things (corridor)
unlike the usual
 -     the interior of a monster or a wall or an open door (wall)
when the symbol matched more than 4 things.  I first changed it to
append the full sentence's missing period, but ultimately switched to
 #     can be many things (corridor)
so that the symbol that "many things" refers to isn't hidden.  This
works better for ^P where player isn't looking at the symbol anymore.

6 years agoMerge branch 'NetHack-3.7'
nhmall [Mon, 13 May 2019 00:33:41 +0000 (20:33 -0400)]
Merge branch 'NetHack-3.7'

6 years agoMerge branch 'NetHack-3.6'
nhmall [Mon, 13 May 2019 00:31:02 +0000 (20:31 -0400)]
Merge branch 'NetHack-3.6'

6 years agoversion number 3.7.0 for vms
PatR [Sun, 12 May 2019 22:57:05 +0000 (15:57 -0700)]
version number 3.7.0 for vms

6 years agoversion number 3.6.3 for vms
PatR [Sun, 12 May 2019 22:52:04 +0000 (15:52 -0700)]
version number 3.6.3 for vms

6 years agomatch the vs version info values with include/patchlevel.h in 3.7
nhmall [Sun, 12 May 2019 19:50:33 +0000 (15:50 -0400)]
match the vs version info values with include/patchlevel.h in 3.7

6 years agoMerge branch 'NetHack-3.6'
nhmall [Sun, 12 May 2019 19:31:31 +0000 (15:31 -0400)]
Merge branch 'NetHack-3.6'

6 years agoadd VERSION_IN_DLB_FILENAME to makedefs detection
nhmall [Sun, 12 May 2019 19:14:31 +0000 (15:14 -0400)]
add VERSION_IN_DLB_FILENAME to makedefs detection

6 years agosupport version-specific dlb file
nhmall [Sun, 12 May 2019 18:51:26 +0000 (14:51 -0400)]
support version-specific dlb file

There was a post-3.6.2 discussion on a forum where someone had
tried to copy the NetHack 3.6.2 exe file overtop of an
existing NetHack 3.6.0 playground, and then try to run it.

We have never suggested trying that, nor do we attempt to
provide any backward or forward compatibility between the
supporting files found in nhdat that would allow that. Any
particular version of NetHack expects to have matching
support files designed and matched to that version.

This adds optional support for helping to prevent the
opening of nhdat containing support files from an
unmatched version of NetHack.

If you #define VERSION_IN_DLB_FILENAME in your
platform's include/*conf.h file, it will use a
name such as nhdat362, instead of plain nhdat, and
will exit more gracefully than the fault/crash
mentioned in the discussion if it doesn't find the
file it is looking for.

Developers - please note that if you do
to cause NetHack to look for an nhdat* file with
the version info appended to the name, you will likely
have to modify your build/clean/spotless mechanics
beyond the C compile itself to properly deal with the
new generated file name.

6 years agofixes37.0 section header bit
nhmall [Sun, 12 May 2019 12:16:01 +0000 (08:16 -0400)]
fixes37.0 section header bit

6 years agocatch up on fixes37.0 entries
nhmall [Sun, 12 May 2019 12:10:21 +0000 (08:10 -0400)]
catch up on fixes37.0 entries

6 years agoMerge branch 'NetHack-3.6'
nhmall [Sun, 12 May 2019 02:05:32 +0000 (22:05 -0400)]
Merge branch 'NetHack-3.6'

6 years agoadd windows startup msg fix to fixes36.3
nhmall [Sun, 12 May 2019 02:01:23 +0000 (22:01 -0400)]
add windows startup msg fix to fixes36.3

6 years agoadd windows startup msg fix to fixes36.3
nhmall [Sun, 12 May 2019 02:01:23 +0000 (22:01 -0400)]
add windows startup msg fix to fixes36.3

6 years agoMerge branch 'NetHack-3.6'
nhmall [Sun, 12 May 2019 01:58:56 +0000 (21:58 -0400)]
Merge branch 'NetHack-3.6'

6 years agoReport of no error msg deliverd by NetHackw.exe for some startup issues
nhmall [Sun, 12 May 2019 01:55:27 +0000 (21:55 -0400)]
Report of no error msg deliverd by NetHackw.exe for some startup issues

6 years agoMerge branch 'master' into NetHack-3.7
nhmall [Sat, 11 May 2019 12:50:32 +0000 (08:50 -0400)]
Merge branch 'master' into NetHack-3.7

6 years agoMerge branch 'NetHack-3.6'
nhmall [Sat, 11 May 2019 12:50:05 +0000 (08:50 -0400)]
Merge branch 'NetHack-3.6'

6 years agofix #H8712 - curses menu selector overflow
PatR [Sat, 11 May 2019 10:04:53 +0000 (03:04 -0700)]
fix #H8712 - curses menu selector overflow

The curses interface would assign menu selector characters a-z, A-Z,
and then 0-9, but trying to type 0-9 would start a count rather than
select an entry, and if the display was tall enough for more than 62
entries, the ones after '9' were ASCII punctuation characters.
Limit the number of entries per page to 52 + number_of_'$'_entries
(which should be 0 or 1) so that it won't run out of normal letters.

The perm_invent window, if enabled, ought to allow more than that
because it isn't used to make selections and might have an arbtirary
number of '#' overflow entries.  But I'll leave that for somebody
else to tackle.

Tested by temporarily setting the limit to 26 instead of 52 since
I'm not able to display anything tall enough to exercise the latter.

6 years agoRemove built-in speaker support
Pasi Kallinen [Sat, 11 May 2019 07:30:41 +0000 (10:30 +0300)]
Remove built-in speaker support

The changes to amiga, mac and msdos builds are untested.

6 years agoMerge branch 'NetHack-3.6'
nhmall [Sat, 11 May 2019 02:54:38 +0000 (22:54 -0400)]
Merge branch 'NetHack-3.6'